CPM in

Construction
Management

James J. O’Brien, P.E., PMP

Fredric L. Plotnick, Esq., P.E.

Sixth Edition

McGraw-Hill
New York Chicago San Francisco Lisbon London Madrid
Mexico City Milan New Delhi San Juan Seoul
Singapore Sydney Toronto
Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc. All rights
reserved. Manufactured in the United States of America. Except as permitted under the United States
Copyright Act of 1976, no part of this publication may be reproduced or distributed in any form or by
any means, or stored in a database or retrieval system, without the prior written permission of the pub-
lisher.

0-07-158911-2

The material in this eBook also appears in the print version of this title: 0-07-145769-0.

All trademarks are trademarks of their respective owners. Rather than put a trademark symbol after
every occurrence of a trademarked name, we use names in an editorial fashion only, and to the bene-
fit of the trademark owner, with no intention of infringement of the trademark. Where such designa-
tions appear in this book, they have been printed with initial caps.

McGraw-Hill eBooks are available at special quantity discounts to use as premiums and sales promo-
tions, or for use in corporate training programs. For more information, please contact George Hoare,
Special Sales, at george_hoare@mcgraw-hill.com or (212) 904-4069.

TERMS OF USE

This is a copyrighted work and The McGraw-Hill Companies, Inc. (“McGraw-Hill”) and its licensors
reserve all rights in and to the work. Use of this work is subject to these terms. Except as permitted
under the Copyright Act of 1976 and the right to store and retrieve one copy of the work, you may not
decompile, disassemble, reverse engineer, reproduce, modify, create derivative works based upon,
transmit, distribute, disseminate, sell, publish or sublicense the work or any part of it without
McGraw-Hill’s prior consent. You may use the work for your own noncommercial and personal use;
any other use of the work is strictly prohibited. Your right to use the work may be terminated if you
fail to comply with these terms.

THE WORK IS PROVIDED “AS IS.” McGRAW-HILL AND ITS LICENSORS MAKE NO GUAR-
ANTEES OR WARRANTIES AS TO THE ACCURACY, ADEQUACY OR COMPLETENESS OF
OR RESULTS TO BE OBTAINED FROM USING THE WORK, INCLUDING ANY INFORMA-
TION THAT CAN BE ACCESSED THROUGH THE WORK VIA HYPERLINK OR OTHERWISE,
AND EXPRESSLY DISCLAIM ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING BUT
NOT LIMITED TO IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A
PARTICULAR PURPOSE. McGraw-Hill and its licensors do not warrant or guarantee that the func-
tions contained in the work will meet your requirements or that its operation will be uninterrupted or
error free. Neither McGraw-Hill nor its licensors shall be liable to you or anyone else for any inaccu-
racy, error or omission, regardless of cause, in the work or for any damages resulting therefrom.
McGraw-Hill has no responsibility for the content of any information accessed through the work.
Under no circumstances shall McGraw-Hill and/or its licensors be liable for any indirect, incidental,
special, punitive, consequential or similar damages that result from the use of or inability to use the
work, even if any of them has been advised of the possibility of such damages. This limitation of
liability shall apply to any claim or cause whatsoever whether such claim or cause arises in contract,
tort or otherwise.

DOI: 10.1036/0071457690
 Professional

Want to learn more?

We hope you enjoy this
McGraw-Hill eBook! If
you’d like more information about this book,
its author, or related books and websites,
please click here.
 For more information about this title, click here

Contents

Preface xv
Acknowledgments xix

Part 1. Introduction to CPM Planning and Scheduling

Chapter 1. Introduction to CPM Planning and Scheduling 3
1.1. Scheduling is for Everyone 3
1.2. We Teach Carpentry—Not “How to Use Your New Power Saw” 4
1.3. History of Scheduling Systems 5
1.4. The Ordered “ToDo” List 6
1.5. Gantt Charts and Bar Charts 6
1.6. Development of the Critical Path Method of Scheduling 9
1.7. Development of the PERT Method of Scheduling 11
1.8. Comparison of CPM and PERT 12

Chapter 2. Project Control Systems Evolve in Academia 15

2.1. 1960–1965: Logic Systems Gain Acceptance 15
2.2. 1966–1970: The Fight between CPM and PERT 17
2.3. PDM 18
2.4. SPERT and GERT 19
2.5. RDCPMTM 20

Chapter 3. Project Control Systems Evolve in the Marketplace 25

3.1. Commercialization 1965–1970 25
3.2. Commercialization 1970–1980 26
3.3. Early Legal Recognition 26
3.4. PCs 1980–1990 27
3.5. PCs 1990–2000 27
3.6. PCs 2000–2005 28
3.7. PCs 2005–2010 29
3.8. The Sixth Edition 30

iii
iv Contents

Part 2. The Theory of CPM Planning and Scheduling

Chapter 4. Your New Tool—Read Before Using 35
4.1. Primavera and Your Power Saw—Useful but Dangerous Tools 35
4.2. How Does Primavera (or Other Scheduling Software) Work? 37
4.3. What Goes In . . . 37
4.4. The Initial Logic Network—Input 37
4.5. The Logic of the Logic Network 38
4.6. Arrow Diagram 38
4.7. Logic Diagrams 41
4.8. Logical Loop 45
4.9. Non-construction Examples 46
4.10. Summary 48

Chapter 5. Network Construction 49

5.1. Form and Format 49
5.2. Events 51
5.3. Problems with Multisheet Networks 61
5.4. Summary 63

Chapter 6. The Durations of the Logic Network 65

6.1. Definition of an Activity 66
6.2. Setting a Minimum and Maximum Duration 67
6.3. Estimating versus Scheduling Durations 69
6.4. CPM versus PERT Durations 70
6.5. Summary 72

Chapter 7. What Comes Out . . . 73

7.1. Attributes of an Event 74
7.2. Attributes of an Activity 74
7.3. The Forward Pass—TE, ES, and EF 75
7.4. The Backward Pass—TL, LF, and LS 75
7.5. The Backward Pass—TF, FF, and IF 75
7.6. Calculating the Attributes of an Event or Activity 77
7.7. The Forward Pass—TE, ES, and EF 77
7.8. The Backward Pass—TL, LF, and LS 78
7.9. The Backward Pass—TF, FF, and IF 79
7.10. Summary 80

Chapter 8. Cranking the Engine 81

8.1. Manual and Computer Solutions for PERT and ADM—
The Matrix Method 81
8.2. Manual and Computer Solution for PERT and ADM—
The Intuitive Method 88
 Contents v

8.3. Activity Start and Finish Times 93

8.4. Critical Activities 95
8.5. Total Float 97
8.6. Free Float 100
8.7. Independent Float 102
8.8. Time Scale Network 104
8.9. Computation Time 105
8.10. Writing Your Own CPM Software 105
8.11. Manual and Computer Solution for PDM with Durations
Between Activities 108
8.12. Summary 110

Chapter 9. Adding Complexity 111

9.1. Enhancements to the Basic System 111
9.2. Original versus Remaining Durations 112
9.3. Percent Complete 112
9.4. Defined Subtasks and Check-off Updating 113
9.5. Calendar versus Work Period Conventions 115
9.6. Multiple Calendars 116
9.7. Multiple Starting and Ending Activities 118
9.8. Artificial Constraints to Dates 120
9.9. Artificial Constraints to Algorithms 123
9.10. Negative Float 123
9.11. Definition of Criticality 124
9.12. Continuous versus Interruptible Performance 129
9.13. Actual Start and Finish Dates 129
9.14. Retained Logic versus Progress Overrides 130
9.15. Events and Milestones 131
9.16. Hammocks and Summary Network Logic 132
9.17. Summary Activity Bars 133
9.18. User-defined Code Fields 136
9.19. Adding Resources to Activities 136
9.20. Adding Costs and Cost Codes to Activities 137
9.21. Resource Driven Scheduling 139
9.22. Master Schedules Local versus System-wide Updating 140
9.23. Activity Types 142
9.24. Hierarchical Codes 144
9.25. Summary 146

Chapter 10. PDM and Precedence Networks 147

10.1. Precedence Logic 148
10.2. Work Package Calculations 152
10.3. Computer Calculation 153
10.4. Project Example 156
10.5. Summary 159
vi Contents

Chapter 11. Respecting the Power of PDM 161

11.1. Durations between Activities: Percent Lead/Lag Relationships 161
11.2. Defining Overlapping Activities: Durations between Activities 163
11.3. Negative Durations between Activities 170
11.4. Remaining Durations between Activities 170
11.5. Impact of Percent Complete Upon Durations between Activities 171
11.6. PDM and Hammocks 172
11.7. Continuous versus Interruptible Progress 175
11.8. Undefined Subtasks and Relationships to Other Activities 177
11.9. Multiple Calendars 178
11.10. Retained Logic versus Progress Override 179
11.11. Total Float Calculation 180
11.12. Erroneous Loop Errors 181
11.13. Summary 182

Chapter 12. Enhancements and Extensions by Software Vendors 183

12.1. Overview of Primavera Project Planner P3 183
12.2. Overview of Suretrak Project Planner 184
12.3. Overview of P3e/c Primavera Program Manager 186
12.4. Overview of Primavera Contractor 188
12.5. Overview of Microsoft Project Professional 2003 188
12.6. Overview of Welcom Open Plan 189

Part 3. The Tools of CPM Planning and Scheduling

Chapter 13. Measure Twice—Cut Once 197
13.1. Preparing to Collect the Input 197
13.2. The Pure Logic Diagram 198
13.3. A Team Effort . . . on the Blackboard or Sketch Pad 198
13.4. Format for Ease of Data Collection versus for Ease of Data Entry
to Chosen Software 198
13.5. Bar Chart: May Be Based upon Logic, but Is not a Logic Network 199
13.6. Logic Restrained Bar Chart 200
13.7. Freehand 200
13.8. PERT 200
13.9. Summary 205

Chapter 14. Choosing Codes 207

14.1. Calendar 207
14.2. Deliverable and Responsible Entity: SHT1, SHT2, RESP, and SUBC 210
14.3. Key Resources: CRTY, CRSZ, MHRS, SUPV and EQUIP 211
14.4. O.T., Night Work, and Special Supervision or Inspection 211
14.5. Quantities and Rates of Productivity 211
14.6. Location, Location, Location 212
14.7. Budget Codes for Cost of Labor, Equipment, and Materials 213
 Contents vii

14.8. A Word About Codes Tied to the Activity ID 215

14.9. A Word About Aliases 216
14.10. Summary 216

Chapter 15. Acquiring Information to Initial Schedule 219

15.1. The Activity Description—a Gross Abbreviation 220
15.2. Activity ID, Activity Codes and Logs 221
15.3. The Activity Further Defined by Resources Assigned 221
15.4. The Activity Further Defined by Predecessors and Successors 221
15.5. The Checklist of Sub-tasks 222
15.6. The Checklist of Sub-deliverables (Events) 222
15.7. Summary 222

Chapter 16. Acquiring the Durations 223

16.1. Best Estimate with Utilization of Resources Envisioned 223
16.2. Compare with PERT Durations: Optimistic, Most Likely, Pessimistic 224
16.3. Schedule Durations versus Estimating Durations 224
16.4. Estimated Durations versus Calculated Durations 225
16.5. Do We Add Contingency Here? 225
16.6. Estimated Durations versus Expected Completion Dates:
“as Good as the Promise” 226
16.7. Productivity 226
16.8. Durations and the Project Calendar or Calendars 226
16.9. Summary 230

Chapter 17. Specifying the Relationships Between Activities 231

17.1. Mandatory and Discretionary Physical Restraints 231
17.2. Mandatory and Discretionary Resource Restraints 232
17.3. Mandatory and Discretionary Timing Constraints 233
17.4. The Misuse of Restraints and Constraints: “Nailing the Bar Down
Where It Belongs” 234
17.5. The Need to Document the Basis of Each Restraint and Constraint 234
17.6. Choosing the Type of Relationship between Activities 234
17.7. The Case for Restricting Relationships to Traditional “FS”
Without Lag 235
17.8. The Need for Non-Traditional Relationships 235
17.9. The Desire for Non-Traditional Relationship and Resulting Misuse 237
17.10. Non-Traditional Relationships Supported by Popular Software 238
17.11. Minimum Restrictions for Proper Usage of PDM 239
17.12. Review the Strengths of ADM: Expand the Definitions 239
17.13. Start of Each Activity Must Have Predecessor 239
17.14. Finish of Each Activity Must Have Successor 239
17.15. Real World Relationships between Activities 240
17.16. The Final Forward Pass 240
17.17. The Final Backward Pass 241
17.18. Choosing the Algorithm for the Initial Schedule 241
17.19. Summary 243
viii Contents

Chapter 18. Example Project: The John Doe Project 245

18.1. Acquiring Information to Initial Schedule 245
18.2. Choosing Appropriate Codes 246
18.3. Activity List 246
18.4. Could We Prepare a Bar Chart? 253
18.5. Network Logic in ADM 254
18.6. Logic Changes Examples 262
18.7. Network Logic in PDM 263
18.8. Populating the Codes 265
18.9. Checking the Output 265
18.10. Calendar Dates 274
18.11. Summary 276

Part 4. The Practice of CPM Planning

Chapter 19. Equipment and Workforce Planning 277
19.1. Workforce Leveling 279
19.2. Computerized Resources Planning 287
19.3. Multiproject Scheduling 292
19.4. Turnaround Application 296
19.5. Examples of Resource Loading on John Doe Project 300
19.6. Resource Leveling Significance 302
19.7. Summary 305

Chapter 20. Procurement 307

20.1. Scheduling Materials Procurement 307
20.2. John Doe Example 317
20.3. Summary 320

Chapter 21. Preconstruction 321

21.1. Predesign Phase 323
21.2. Design 325
21.3. Summary 333

Chapter 22. Evolution of the Project Schedule 335

22.1. Preliminary Schedule 335
22.2. Preconstruction Analysis 336
22.3. Contractor Preconstruction Analysis 336
22.4. Milestones 337
22.5. The John Doe Schedule 338
22.6. Resources 339
22.7. Fast Track 341
22.8. Responsibility 342
 Contents ix

22.9. Schedule versus Calendar 342

22.10. Contingency 344
22.11. Schedule Manipulation 349
22.12. Working Schedule 350
22.13. Summary 353

Chapter 23. CPM and Cost Control 355

23.1. CPM Cost Estimate 355
23.2. Progress Payments 358
23.3. Cost Forecasting 362
23.4. Network Time Expediting 369
23.5. Minimum Cost Expediting 375
23.6. Summary 379

Part 5. The Practice of CPM Scheduling

Chapter 24. Let’s Look Under the Hood at the Engine 383
24.1. Primavera Project Planner P3 383
24.2. Project and Activity Codes for Organizing Project Data 392
24.3. Entering Data 393
24.4. Debugging and Diagnostic Tools: Improving the Logic Network 395
24.5. Viewing Output 397
24.6. On-screen Formats 397
24.7. Viewing Output: Tabular Formats 400
24.8. Viewing Output: Graphical Formats 401
24.9. Entering Update Data 404
24.10. Primavera Project Manager: P3e/c 405
24.11. Summary 411

Chapter 25. Converting the Team Plan to the Calculated Schedule 413
25.1. Data Entry Made Easy 413
25.2. Check and Set Schedule Algorithm Options 414
25.3. First Run and De-bugging the Logic 417
25.4. Loop Detection and Correction 418
25.5. Technical Review: The Primavera Diagnostic Report 419
25.6. Beyond the Primavera Diagnostic 422
25.7. First Review of Calculated Output: Reality Check #1 427
25.8. Detail Views of Output of Schedule Calculations 427
25.9. Time Scaled Logic Diagram 434
25.10. Tailoring Initial Output to the Chosen Audiences 435
25.11. Whatever Owner Wants, Owner Gets 435
25.12. “You Can’t Always Get What You Want, But . . . You Get
What You Need” 436
25.13. Reports and Views for the Foreman Performing the Work 437
25.14. Reports and Views for the Contractor’s Superintendent 440
x Contents

25.15. Reports and Views for the Contractor’s Upper Management 442
25.16. The Narrative Report for Each Audience 443
25.17. Summary 444

Chapter 26. Engineer’s Review of the Submitted Initial CPM 445

26.1. Legal Aspects of a Review 447
26.2. Reviewing the Plan 448
26.3. Technical Review 449
26.4. Reschedule and Review the Diagnostic Report 449
26.5. But Is the Logic Realistic? The Smell Test 452
26.6. Project Calendar or Calendars 453
26.7. Summary 453

Chapter 27. Updating the Schedule 455

27.1. Why Update the Schedule? 456
27.2. Acquiring the Data for an Update 457
27.3. Distinguishing Updates from Revisions 459
27.4. Purpose of an Update 459
27.5. The Purpose of a Revision 460
27.6. Who Should Collect Data for an Update 460
27.7. Who Should Prepare Data for a Revision 460
27.8. Information Required for Schedule Control: AS, RD, AF 461
27.9. Determination of Actual Start and Actual Finish Dates 461
27.10. Determination of Remaining Duration of Activities:
Repeat the Steps of the Master 461
27.11. Expected Completion and Renewing Promises 462
27.12. Automatic Updates 463
27.13. The Forgotten Step: Determination of Remaining Duration
Between Activities 463
27.14. Save and Rename: Naming Strategies 465
27.15. Reports and Views to Assist Acquiring Data for an Update 468
27.16. Electronic Tools to Assist Acquiring Data for an Update 469
27.17. Choosing the Correct Algorithm for Updates 469
27.18. Scheduling the Update: Interpreting the Results 469
27.19. Technical Review: the Primavera Diagnostic Report 470
27.20. What to Look for when Reviewing the Update 471
27.21. Tailoring Update Output to the Chosen Audiences 472
27.22. The Narrative Report for Each Audience 473
27.23. Summary 474

Chapter 28. Engineer’s Review of the Submitted Update 475

28.1. Is This an Update or a Revision? 475
28.2. The Technical Review 476
28.3. The Critical Path 477
28.4. What to Accept and What to Reject 477
28.5. Summary 478
 Contents xi

Chapter 29. Revising the Logic Network 479

29.1. “What If” versus Committed Changes 479
29.2. Changes: Approved, Constructive and at Contractor’s Cost 481
29.3. Revised Baseline 482
29.4. Update then Revise 483
29.5. Summary 483

Chapter 30. Engineer’s Review of the Submitted Revision 485

30.1. Minor Revision 485
30.2. Major Revision 487
30.3. Recovery Schedule 487
30.4. Summary 487

Chapter 31. Case Histories 489

31.1. Chicago Courthouse 489
31.2. Times Tower 490
31.3. Airport Construction 490
31.4. High-Rise Construction 491
31.5. NASA 493
31.6. Housing 493
31.7. Manufacturing Facilities 494
31.8. SEPTA RailWorks 495
31.9. New Jersey Turnpike Authority 1990–1995 Widening Program 496
31.10. JFK Redevelopment 497
31.11. Toronto Transit’s “Let’s Move” Program 500
31.12. Phoenixville-Mont Claire Bridge 503
31.13. Graduate Hospital Diagnostic Building 506
31.14. Interstate 76, 202 & 422 Interchange 508
31.15. CPM Preparation Time 510
31.16. Summary 513

Chapter 32. Additional Exercises for Students of Project Controls 515

32.1. Sample Problems 515
32.2. Navigating the Enclosed CD-ROM 517
32.3. Answers to Sample Problems 520

Part 6. Advanced Topics

Chapter 33. Specifying CPM 525
33.1. Attorney’s Viewpoint to Writing a CPM Specification 525
33.2. Pure Logic Drawing 526
33.3. Content of the Logic Network 527
33.4. Updates and Revisions 529
xii Contents

33.5. Standard References 530

33.6. Sanctions 532
33.7. Sample CPM Specification 533
33.8. Summary 534

Chapter 34. CPM in Claims and Litigation 535

34.1. Introduction 535
34.2. Evidentiary Use of CPM 537

Chapter 35. Delay Analysis 545

35.1. Delay versus Disruption 545
35.2. Responsibility/Types/Force Majeure 546
35.3. As-Planned Logic Network 548
35.4. As-Should-Have-Been CPM Network 550
35.5. As-Planned Schedule 551
35.6. As-Built Schedule 552
35.7. As-Built Logic Network 553
35.8. Causative Factors 554
35.9. As-Impacted Logic Network 556
35.10. As-Impacted Schedule 557
35.11. Time Impact Evaluations 557
35.12. Zeroing Out 558
35.13. Zeroing to a Collapsed As-Impacted Logic Network 559
35.14. Limitations of the TIE Methodology 560
35.15. TIE Example of John Doe Project 561
35.16. Windows Analysis 567
35.17. Zeroing Out within the Windows Analysis 569
35.18. Windows Example of John Doe Project 569
35.19. Summary 570

Chapter 36. Disruption Analyses 571

36.1. Traditional Methodologies 571
36.2. CPM Out-of-Sequence Methodology 572
36.3. Summary 574

Chapter 37. Advanced Topics: Resource Leveling and Smoothing 575

37.1. Resource Leveling and Smoothing 575
37.2. Limitations of Algorithms 576
37.3. Driving Resources 576
37.4. Resource Calendars 577
37.5. Practical Solutions 578
37.6. Summary of Resource Leveling 579

Chapter 38. Advanced Topics: PERT, SPERT, and GERT 581

38.1. PERT 582
38.2. SPERT 582
 Contents xiii

38.3. GERT 586

38.4. Computers Add Power 587
38.5. Summary 588

Chapter 39. Conclusion 589

Appendix A. Sample CPM Specification as a Guideline for

Preparing Your Own Specification 591

Appendix B. Unified Facilities Guide Specification 607

Appendix C. Notation for RDCPM 631

Glossary 635
Acronyms and Symbols 641
Index 643
ABOUT THE AUTHORS
JAMES J. O’BRIEN, P.E., PMP, CVS, was vice chair of the board
of O’Brien-Kreitzberg & Associates, Inc., the construction
management firm that handled the renovation of San
Francisco’s cable car system. He was also the program
manager for the redevelopment of JFK International Airport.
Mr. O’Brien is the author or editor of many other books,
including Contractor’s Management Handbook, Second
Edition; Construction Management: A Professional Approach;
Value Analysis in Design and Construction; and Scheduling
Handbook, all published by McGraw-Hill. He is a registered
Professional Engineer in New York, New Jersey,
Pennsylvania, and many other states. Mr. O’Brien is a Fellow
of the American Society of Civil Engineers, the Project
Management Institute, the Construction Management
Association of America, and the Society of American Value
Engineers International. He is also an active member of the
Association for the Advancement of Cost Engineering
International. Mr. O’Brien is a charter member and a vice
president of the PMI College of Scheduling.

FREDRIC L. PLOTNICK, ESQ., P.E., is CEO and principal

consultant of Engineering & Property Management
Consultants, Inc. He has bachelor’s and master’s degrees in
civil engineering and is a registered Professional Engineer
in Pennsylvania, New Jersey, and Florida. He is also an
attorney and a member of the bars of Pennsylvania, New
Jersey, and Florida. Mr. Plotnick is an adjunct professor of
the departments of Civil Engineering, Engineering
Management, and Construction Management at Drexel
University, Philadelphia, Pennsylvania. He is a past
president of the Philadelphia chapter of the Pennsylvania
Society of Professional Engineers, past Construction Group
chair of the Philadelphia section of the American Society of
Civil Engineers, and member of the Association for the
Advancement of Cost Engineering International. He is a
charter member and director of Academic Liaison and the
Technical Track of the annual conference of the PMI College
of Scheduling.

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
 Preface

The original purpose of this book, in 1965, was to present and discuss the
critical path method (CPM) and its use in the construction industry. At
that time, CPM was a young but proven technique—usually considered
to be optional. When the second edition was published in 1971, the net-
work approach to scheduling was becoming a regular requirement in con-
struction contracts. The third edition, published after 25 years of
experience in the application of CPM, described highlights of that expe-
rience and its significance to the practical use of CPM.
The basic strength of CPM continues to be its ability to represent
logical planning factors in network form. One reviewer noted: “Perhaps
the most ironic aspect of the critical path method is that after you under-
stand it, it is self-evident. Just as an algebra student can apply the
rules without full appreciation of the power of the mathematical con-
cepts, so can the individual apply CPM or its equivalent without fully
appreciating the applicability of the method.”
The book first describes the development of CPM and its practical use
in the construction industry. The basic technique is described in suffi-
cient depth for the reader to apply it to practical construction situations.
The John Doe case study is used throughout the book to describe basic
CPM network techniques and then to illustrate such special functions
as updating, cost control, resource planning, and delay evaluation.
Optimum methods of specifying the use of CPM are described in suffi-
cient detail to be incorporated directly into construction specifications.
Since the second edition, CPM has become widely utilized as an ana-
lytical tool in the evaluation, negotiation, resolution, and/or litigation of
construction claims. This aspect is thoroughly explored in the current edi-
tion. Legal precedents for the use of CPM during litigation are provided.
In the 1980s, computer calculation shifted from mainframe programs
to personal computers (PCs). PCs were the wave of the past two decades.
The ubiquity in the 2000s of the internet and the wave of additional
interconnectivity linking individual PCs now has the appearance of
coming full circle and bringing back to CPM many of the strengths and

xv

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
xvi Preface

weaknesses of the era of the mainframe. However, the approaches and

procedures suggested in the first five editions are, almost without excep-
tion, still valid.
Network techniques are basic and logical, but assimilation of the net-
work concept does take time. Further, an effort is required to build an
experience level, which in turn builds confidence. This book aims to be
a useful element in the development of that conceptual experience and
confidence on the part of new users of CPM techniques.

James J. O’Brien, P.E., PMP

I was introduced to the concepts of CPM as a student in college for 2 weeks

in a course covering many aspects of construction management. It was a
revelation and led to additional independent study, including a grant of
computer time (on the giant mainframe) from Drexel University’s
Computer Center (Philadelphia, PA), on which my first CPM software
program was written. It was at this time that I realized the potential
value of CPM to resolve disputes involving delay that planted the seed for
my future legal education.
Several years past, during which I worked for several construction and
consulting firms, and a stint as assistant corporate counsel for a large firm
involved in international construction. In 1983, I formed EnProMaC
(Engineering & Property Management Consultants, Inc.) Interestingly,
in 1983 Joel Koppelman and Dick Faris formed Primavera Systems. One
of my first efforts was to rewrite my CPM software program to run on my
Osbourne I (a pre-IBM PC with 64 KB of RAM and 90 KB of floppy disk
storage) running as a routine under dBASE II (a database program by
Ashton Tate). At that time, I never dreamed that a market might exist
for such software—assuming such could be rewritten for user friendliness.
The success that Messrs. Koppelman and Faris achieved in launching
Primavera is largely based upon their attention to making their software
user friendly—and in giving their customers that which is asked for.
CPM theory has a number of limitations, as does any system that
attempts to model reality. Bending the rules of CPM analysis can, in some
instances, circumvent these limitations. In many cases special features
have been added to Primavera, which have legitimate uses in very lim-
ited situations, but which should be used with extreme care. The many
competitors of Primavera also have added features that extend and
modify the basic concepts of CPM—each in their own fashion—and each
that differ subtly from each other. One of my contributions to the 5th edi-
tion was to address these special features, and the proper use of them.
One of the factors in forming EnProMaC was that, in 1982, Drexel
University asked me to create a course on CPM. I have been teaching that
 Preface xvii

course, as well as courses on contracts and specifications, engineering law,

and project administration, ever since. The give and take of classroom
discussion with students ranging from candidates for a 30-hour certifi-
cate (through the local GBCA) to candidates for the Ph.D. degree over
these many years has further pushed me to more fully examine the
mathematics behind CPM and other scheduling systems. It is my hope
that my contributions to this edition will bring the confluence of the
basic theory of mathematics, the applied discipline of engineering, and
the framework for collaboration by adversarial parties provided by the
study of law, all to assist the practitioner of planning and scheduling.

Fredric L. Plotnick, Esq., P.E.

This page intentionally left blank
 Acknowledgments

The writing of this sixth edition has involved the assistance of numer-
ous individuals who have provided technical advice on the computer
software products discussed, feedback from the field for the case
studies discussed, editorial review to combine the writing styles of
the two authors into one more readable style and moral support
throughout the process of writing and rewriting and editing and
proofreading and publishing. Special recognition is accorded to the
following individuals.
Rita O’Brien and Kim Plotnick for strong silent (and not too silent)
support
Larry Hager of McGraw-Hill for coordination of efforts and setting
deadlines for the authors
Pamela Pelton of McGraw-Hill for coordination of publishing
Neha Rathor of International Typesetting and Composition for editorial
review
Ben Kolstad of International Typesetting and Composition for layout
of text and graphics
Kristy Tan and Jennifer Coyle of Primavera Systems for technical
questions regarding P3 and P3e/c
Nicole Styer of Primavera Systems for coordination of the CD-ROM
John Owen, Robert Blakeney and Andrew Ely of Welcom for techni-
cal questions regarding Open Plan
Dan Patterson of Pertmaster for technical questions regarding
Pertmaster and MCA
Jon Wickwire, Esq. of Wickwire Gavin, P.C. for reference to current
legal cases involving CPM
Dan Lynch, P.E. of SEPTA for the SEPTA Railworks case study
Tom Driscoll of URS for the NJ Turnpike case study

xix

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
xx Acknowledgments

Richard Smyth of Jet Blue for the JFK Airport Redevelopment case
study
Louis Tucciarone of URS for the Toronto Transit case study
Gordon Bryan and Jon Mountenay of American Infrastructure for
the Route 76/202/422 case study
Pradip Mehta of PA of NY&NJ for the Port Authority of NY and NJ
case study
Joseph Fisher of Hunt Construction Group for the Maricopa County
case study

and many others who have assisted in the Sixth and previous editions.
 Part

Introduction
1
to CPM Planning
and Scheduling

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
This page intentionally left blank
 Chapter

1
Introduction to CPM
Planning and Scheduling

This introduction discusses some factors that make the case for why
planning and scheduling is best performed by the Critical Path Method
(CPM). It covers some of the history behind its development and relays
some thoughts on where the process may be going in the future. The
interplay between the theory of mathematics that underlies the method-
ology and the modifications needed to make the methodology more prac-
tical are themes that continue throughout the text. It is hoped that the
conclusion drawn by the reader will be that it is the Scheduler who
must balance these two ideals, mathematics and engineering, to provide
a useful and user friendly tool to the users of CPM in construction man-
agement, manufacturing, software design and other users in the world
of projects that must be finished on time and within budget.

1.1. Scheduling is for Everyone

Scheduling is a discipline that is performed by every person, every day.
Should you first shave or brush your teeth in the morning? If you are
scheduling for one person only, the process is rather simple. You can pre-
pare a “ToDo” list and then choose in what order the items on it are to be
performed. However, the choice of what to do first is not completely
random. Perhaps there are physical restrictions, such as “shower before
dressing” or “cook breakfast before eating.” Perhaps there are logistical
restrictions, such as combining one trip to buy milk, pick up the dry clean-
ing, and refuel your car, rather than making three trips to accomplish
these three items on your “ToDo” list. Perhaps the order of performance
is pure personal choice, such as put on right shoe before left shoe.

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
4 Introduction to CPM Planning and Scheduling

Even at this simple level, not all is what it seems to be. If you are in a
hurry, you might begin eating a portion of your breakfast while still cook-
ing the rest. If your dry cleaner is open only from 10:00 AM to 6:00 PM
and if your car is very low on gas, you may have to refuel on your way to
work, pick up your dry cleaning at lunch, and buy milk on the way home
from work. If you have a foot or leg injury, you may need to put on your
left shoe first.
If you want to schedule the tasks of two or more persons or the work
flow of two or more machines (even if both are under the supervision of
one person), the process becomes much more complex.

1.2. We Teach Carpentry—Not “How to Use Your New Power Saw”

If the process of scheduling were a simple matter, requiring merely rote
actions without the need for thought, perhaps good schedules could be
created by a software Wizard. After clicking your way through a preset
series of screens, you would have your schedule. Perhaps, then, the
request of an old client for a device where building blueprints are fed in
one end and a schedule is printed out the other end would be feasible.
Alas, it is not so–scheduling is a complex process and the mathemati-
cal underpinning is at the level of rocket science.
Scheduling is an application of special knowledge or judgment of the
mathematical, physical, or engineering sciences to the conception or
implementation of creative work. Scheduling, formally or informally,
good or bad, is practiced in the planning, progress, and completion of
designs, analyses, or implementation performed in connection with util-
ities, structures, buildings, machines, equipment, processes, systems,
works, projects, and industrial or consumer products or equipment of a
mechanical, electrical, electronic, chemical, hydraulic, pneumatic, geot-
echnical, or thermal nature. The wording of the preceding sentence is
taken from the statute defining Engineering for a reason, that being,
that scheduling is a branch of engineering.
Implicit in the teaching of engineering, or of the supporting fields
of science and mathematics, is the need to understand the process and
not merely to blindly trust the black box desired by that old client. It
is important to understand the mathematical underpinning of modern
CPM software rather than to merely begin clicking away. Children
are still taught how to add and how to spell even though they have
access to calculators and computer aided spell-checks. One reason is
that even the best spell-check software can leave errors uncaught.
Another reason is to understand what the numbers on calculators
mean. Many of us may remember a freshman physics class where we
were taught that 2.5 × 3.01 is not equal to 7.525, but rather to 7.5, since
 Introduction to CPM Planning and Scheduling 5

the result will never be more accurate than the least accurate input.
(For those who have not taken Freshman Physics, 2.50 × 3.01 = 7.53
and 2.500 × 3.010 = 7.525.)
Even the terminology can be misleading. CPM was once noted as a
tool in the process of Planning and Scheduling. First we must plan,
then we can use the computer to perform the rote calculations (that we
understand and could perform given time) to generate the schedule,
and then we must read the output with a knowledge of the assumptions
and tolerances involved. Today, however, we can purchase software that
includes a Wizard to simplify or ignore the need for planning, perform
the calculations while allowing user overrides to generate the “correct”
or “desired” result, and provide killer report and graphics applications
to display the schedule results.
It is the purpose of this text to teach carpentry and not merely the fea-
tures and benefits and how to use your new power saw. It is the purpose
of this text to teach the process of planning and scheduling by means of
the Critical Path Method of Analysis. We can best start by reviewing how
this field of mathematics and engineering was developed.

1.3. History of Scheduling Systems

The Critical Path Method (CPM) was developed specifically for the plan-
ning of construction. The choice was fortuitous, since construction
accounts for more than 10 percent of the annual gross national product.
Almost every activity and every person is affected to some degree by new
construction or the need for it. Most projects are started well after the
need has been established, seeming to follow the whimsy, “If I’d wanted
it tomorrow, I’d have asked for it tomorrow.”
The construction industry is a heterogeneous mix of companies rang-
ing in size from the large operations to one-person operations. No matter
the size, construction companies face similar situations and, to some
degree, similar pressures. Many factors, such as weather, unions, acci-
dents, capital demands, and work loads, are either beyond individual
control or difficult to control. New problems in project approvals due to
increased public awareness include pollution and ecological controls.
CPM does not offer clairvoyance, but it does assemble all the informa-
tion to the project managing team.
Initially, CPM spotlighted construction and the contractor. The owner,
architect, engineer, and public agencies involved in a project are like the
backer, producer, and director of a Broadway show: Without them, the
show cannot go on, and any lack of competence, motivation, or interest
on the part of any one of the team members can delay a project. However,
the contractor is the performer who ultimately makes or breaks the
construction show.
6 Introduction to CPM Planning and Scheduling

The typical contractor is a planner who generally uses instinctive

methods rather than formal scheduling. Prior to 1957, contractors had
little choice than to operate this way because no comprehensive, disci-
plined procedures for planning and scheduling construction projects
existed. And prior to the mid 1980s, contractors desiring to utilize the
benefits of the newer methods had to rely upon outside consultants,
who in turn had to rely upon computer service bureaus and their large
mainframe computers.
One of the keys to the success of CPM is that it utilizes the planner’s
knowledge, experience, and instincts in a logical way first to plan and
then to schedule. CPM can save time through better planning, and in
construction, time is money.
The Egyptians and Romans worked construction miracles in their
day, and surviving ruins attest to the brilliance of their architecture, but
little is known of their construction planning and scheduling. Other
historical project managers included Noah, Solomon, and the unknown
architect who designed the tower of Babel. Again, history records much
about the construction details but little about the methods of control.

1.4. The Ordered “ToDo” List

Many of us make lists of things to do (i.e., a ToDo list.) Those who are
well organized may make the list in a logical order–for example, a shop-
ping list based upon the layout of a store or supermarket. Perhaps a
fanatic to organization may first make a list of activities (or, from our
example, items to be purchased) and then copy it a second time to the
preferred order that it is to be performed. The use of word processing
or organizing software adds a modern wrinkle to this age-old method of
planning and scheduling. However, there are no rules widely published
to guide the development of “ToDo” lists.

1.5. Gantt Charts and Bar Charts

In the mid-nineteenth century, at least one writer discussed a work
versus time graphical representation very similar to today’s bar charts,
but it remained for Henry L. Gantt and Frederick W. Taylor to popularize
their graphical representations of work versus time in the early l910s.
Their Gantt charts were the basis for today’s bar graphs, or bar charts.
Taylor and Gantt’s work was the first scientific consideration of work
scheduling. Although their work was originally aimed at production
scheduling, it was readily accepted for planning and recording the
progress of construction. Today, the bar graph remains an excellent
graphical representation of activity because it is easy to read and under-
stood by all levels of management and supervision.
 Introduction to CPM Planning and Scheduling 7

If the bar graph is so well suited to construction activity, why look for
another planning aid? Because the bar graph is limited in what infor-
mation it can retain. In preparing a bar chart, the Scheduler is influ-
enced almost necessarily by the desired completion dates, often working
backward from the completion dates. The resultant mixture of planning
and scheduling is, unfortunately, no better than wishful thinking.
When a bar graph is carefully prepared, the Scheduler goes through
the same thinking process as the CPM planner. However, the bar graph
cannot show (or record) the interrelations and interdependencies that
control the progress of the project. And, at a later date, even the origi-
nator is often hard pressed to explain the plan by using the bar graph.
Figure 1.5.1 is a simplified bar chart of the construction of a one-
story office building. Suppose that, after this 10-month schedule has
been prepared, the owner asks for a 6-month schedule. By using the
same time for each activity, the bar chart can be changed as shown in
Figure 1.5.2. Although the chart looks fine, it is not based on logical plan-
ning; it is merely a juggling of the original bar graph.
The general contractor usually prepares the overall construction plan,
which is sensible because the schedules of the other major contractors
depend on the general contractor’s schedule.
Note that in Figures 1.5.1 and 1.5.2, the general contractor’s work is
broken down in some detail, with both the mechanical and electrical work
shown as continuous lines that start early and end late. In conformance
with the bar graph “schedule,” the general contractor will then often
push the subcontractors to staff the project as early as possible with as

Figure 1.5.1 Bar chart for a one-story building.

8 Introduction to CPM Planning and Scheduling

Figure 1.5.2 Revised bar chart for a one-story building.

many mechanics as possible. Conversely, the subcontractors want to

come on the project as late as possible with as few mechanics as possi-
ble. The result is that the general contractor will often complain that
the subcontractors are delaying the project through lack of interest. At
the same time, the subcontractors will often complain that the general
contractor is not turning work areas over to them, forcing them to pull
out all of the stops to save the schedule.
As in most things, the truth lies somewhere between the extremes.
CPM offers the means to resolve these differences with specific infor-
mation rather than generalities.
The bar chart often suffers from a morning glory complex: It blooms
early in the project but is nowhere to be found later on. We can suppose
some general reasons for this disappearing act. Prior to the construc-
tion phase, the architect, the engineer, the owner, or all three are trying
to visualize the project schedule in order to set realistic completion
dates. Most contracts will require the submission of a schedule in bar
graph form soon after a contract is awarded. Once the project begins to
take shape, however, this early bar chart becomes as useful as last
year’s calendar because it does not lend itself to planning revisions.
Although progress can be plotted directly on the schedule bar chart,
the S curve has become popular for measuring progress. The usual S
curve consists of two plots (Figure 1.5.3): the scheduled dollar expen-
ditures versus time and actual expenditures versus time. Similar S
curves can be prepared for labor hours, equipment and material acqui-
sitions, concrete yardage, and so on. Though this presentation can be
interesting, it does not provide a true indication of project completion.
 Introduction to CPM Planning and Scheduling 9

Figure 1.5.3 Typical S curve.

For instance, a low-value critical activity could delay the project com-
pletion far out of proportion to its value.
Misuse of bar charts does not prove that they should be discarded. To
throw out bar charts is like throwing out the baby with the bath water.

1.6. Development of the Critical Path Method of Scheduling

In 1956, the E. I. DuPont de Nemours Company set up a group at its
Newark, Delaware, facility to study the possible application of new man-
agement techniques to the company’s engineering functions.1 The plan-
ning and scheduling of construction projects was one of the first areas
studied. The group had a UNIVAC I computer (the third unit built) at
its disposal, and they decided to evaluate the potential of computers in
scheduling construction work. Mathematicians worked out a general
approach; they theorized that if the computer was fed information on the
sequence of work and the length of each activity, it could generate a
schedule of work.2
In early 1957, the Univac Applications Research Center, under the
direction of Dr. John W. Mauchly, joined the effort with James E.

1
Hayward and Robinson, Preliminary Analysis of the Construction Scheduling Problem,
internal paper, Engineering Department, DuPont Company, December 1956.
2
James E. Kelley and Morgan R. Walker, “Critical Path Planning and Scheduling,”
Proceedings of the Eastern Joint Computer Conference, pp. 160–173, Dec. 1–3, 1959; see
also James E. Kelley, “Critical-Path Planning and Scheduling: Mathematical Basis,”
Operations Research, vol. 9, no. 3, pp. 296–320. 1961.
10 Introduction to CPM Planning and Scheduling

Kelley, Jr., of Remington Rand (UNIVAC) and Morgan Walker of DuPont

in direct charge at Newark, Delaware. The original conceptual work was
revised, and the resulting routines became the basic CPM. It is inter-
esting that no fundamental changes in this first work have been made.3
In December 1957, a test group was set up to apply the new technique
(then called the Kelley-Walker methods). The test team (made up of six
engineers, two area engineers, a process engineer, and an estimator) and
a normal scheduling group was assigned to plan the construction of a
$10 million chemical plant in Louisville, Kentucky.
As a control, the new scheduling team worked independently of the
normal scheduling group. This is the only documented case of a com-
prehensive comparative CPM application. The test group had not been
part of the development of the CPM method, but members were given
a 40-hour course on the technique before starting the test.
The network diagram for the project was restricted to include only the
construction steps. The project was analyzed beginning with the com-
pletion of its preliminary design. The entire project was subdivided into
major areas of scope, and each of the areas was analyzed and broken
down into the individual work activities. These activities were dia-
grammed into a network of more than 800 activities, 400 of which rep-
resented construction activities and 150 design or material deliveries.
The ability of the first team was such that a larger-capacity computer
program had to be developed for support. By March 1958, the first part
of the network scheduling was complete. At that time, a change in cor-
porate outlook, plus certain design changes, caused a 40 percent change
in the plan of the project. Both planning groups were authorized to modify
the plan and recompute schedules. The revisions, which took place during
April 1958, required only about 10 percent of the original effort by the
CPM test team, substantially better than the normal scheduling group.
One significant factor involved the determination of critical delivery
items. The normal scheduling group arbitrarily assigned critical cate-
gories, which the CPM group determined from its network analysis.
From the analysis, it was determined that only seven items were criti-
cal, and three of these were not included in the normal scheduling
group’s list.
The initial test scheduling was considered successful in all respects.
In July 1958, a second project, valued at $20 million, was selected for
test scheduling. It also was successfully scheduled. Since the first two
projects were of such duration that the complete validity of the system

3
James E. Kelley, “Computers and Operations Research in Road Building,” Operations
Research, Computers and Management Decisions, Symposium Proceedings, Case Institute
of Technology, January 31–February 1 and 2, 1957.
 Introduction to CPM Planning and Scheduling 11

could not be established, a shorter project, also at DuPont in Louisville,

was selected for scheduling.
The third project was a shutdown and overhaul operation involving
neoprene, and one of the materials in the process was self-detonating,
so little or no maintenance was possible during downtime. Although
the particular maintenance effort had been done many times, it was con-
sidered to be a difficult test of the CPM approach.
In the first CPM plan, the average shutdown time for the turnaround
was cut from 125 to 93 hours, and in later CPM applications, it was fur-
ther cut to 78 hours. The resultant time reduction of almost 40 percent
far exceeded any expectations.4

1.7. Development of the PERT Method of Scheduling

The development of CPM was enhanced when the U.S. Navy Polaris pro-
gram became interested in it. The Polaris program staff had developed
its own network system known as performance evaluation and review
technique (PERT). The DuPont work is considered antecedent material
for the development of PERT.
The Polaris fleet ballistic missile (FBM) system was initiated in early
1957. To manage the program, a Special Projects Office (SPO) was estab-
lished under the direction of Admiral Raborn. The Office is generally
credited with having developed the PERT system.
One of the key people involved in the development of PERT was
Willard Fazar, who noted that the various management tools available
for managing the Polaris program did not provide certain information
essential to effective program evaluation. In particular, they did not
furnish the following:

1. Appraisal of the validity of existing plans in terms of meeting program

objectives
2. Measurement of progress achieved against program objectives
3. Measurement of potential for meeting program objectives

The search for a better management system continued throughout

the fall of 1957. At that time, the Navy was cognizant of the develop-
ment of CPM at DuPont. In January 1958, the SPO initiated a special
study to determine whether computers could be used in planning and

4
Hayward and Robinson, Preliminary Analysis of the Construction Scheduling Problem,
Engineering Department, DuPont Company, December 1956.
12 Introduction to CPM Planning and Scheduling

controlling the Polaris program, and on January 27, 1958, the SPO directed
a group to undertake the task of formulating the PERT technique.5
The goal of the group was to determine whether improved planning
and evaluating research and development work methods could be
devised to apply to the Polaris program, which involved 250 prime con-
tractors and more than 9,000 subcontractors.
The PERT program evolved, and included the development of detailed
procedures and mechanics phases, which were reported in formal doc-
uments. The PERT method, as described in the phase II report, was
designed to provide the following:

1. Increased orderliness and consistency in planning and evaluating

2. An automatic mechanism for identifying potential trouble spots
3. Operational flexibility for a program by allowing for a simulation of
schedules
4. Rapid handling and analysis of integrated data to permit expedi-
tious corrections

The PERT system, programmed at the Naval Ordinance Research

Calculator, was implemented in the propulsion component, which was
followed by an extension to the flight control and ballistic shell compo-
nents, and finally, to the re-entry body and guidance component.
About a year after the start of the PERT research, the system was
operational. This was outstanding considering the typical 36 percent
time overrun for developing other weapons systems.
Following its success in the Polaris program, PERT was incorporated
voluntarily in many aerospace proposals in 1960 and 1961. In some
proposals, PERT was added principally as window dressing to make the
proposal more attractive to the government. But thanks to its basic
soundness and the acumen of the engineering staff members involved,
PERT often stayed on as a useful planning tool even though it had
entered some companies through the backdoor.

1.8. Comparison of CPM and PERT

The key difference between CPM and PERT is that one identifies activi-
ties of finite and reasonably estimated duration while the other identifies

5
D. G. Malcolm et al., A Network Flow Computation for Project Cost Curves, Rand Paper
P-1947, Rand
Corporation, March 1960; D. G. Malcolm, J. H. Roseboom, C. E. Clark, and W. Fazar,
“Applications of a Technique for Research and Development Program Evaluation,”
Operations Research, vol. 7, no. 5, pp. 646–699, 1959; and W. Fazar, “The Origin of PERT,”
The Controller, vol. 30, pp. 598 ff., December 1962.
 Introduction to CPM Planning and Scheduling 13

events of zero duration separated by “some form of activity” only loosely

understood to be performed within a range of possible durations. This
range of durations varies from an optimistic estimate (or shortest time until
the next event will occur) to a “most likely” estimate to a pessimistic esti-
mate. This dichotomy was understandable since the duration of an activ-
ity, relating to a known quantity of work, was fairly capable of estimation;
the duration between events, based upon a scope only vaguely under-
stood, was much more a “guestimate.”
The theory behind the PERT method was based upon the interplay
between these estimates of duration and the statistical likelihood of a
project outcome as the actual duration experienced may fluctuate among
the three. However, the early computers of the 1950s and even the 1960s
did not have the necessary speed or memory to fully utilize the theory
and the three estimates were usually combined into one (often by sep-
arate calculation by hand alongside the computer) using the formula

(O + 4M + P)
DUR =
6

where O = Optimistic,
M = Most Likely and
P = Pessimistic.
The important distinction to remember, before considering the newer
offshoots of CPM, is that CPM measures performance of defined activ-
ities and the durations of defined activities, while PERT measures the
reaching of defined events and the passage of time between these events.
Another important difference is that CPM durations are of defined
events, while PERT durations are of undefined activity between events.
This page intentionally left blank
 Chapter

2
Project Control Systems
Evolve in Academia

By 1960, John Mauchly had left Sperry Rand and formed Mauchly
Associates in Fort Washington, PA. Jim Kelley rejoined Mauchly
Associates as a principal. Mauchly Associates had a consulting group
who both taught CPM principles to in-house industry groups, such as
petrochemical and pharmaceutical, and applied CPM to projects, in par-
ticular, in the construction industry.

2.1. 1960–1965: Logic Systems Gain Acceptance

True conceptual design and testing of CPM was accomplished from 1955
to 1960. In the 5 years that followed, an almost evangelical enthusiasm
spurred the conversion of the conceptual into the practical utilization.
Many public seminars were given and great project engineer exposure
to the techniques was achieved.
Development was spurred especially by three factors: First, the orig-
inating DuPont group disseminated information on the planning tech-
nique to DuPont customers as part of an overall service policy. Second,
the Remington Rand Company, in further computer applications,
assisted many of its computer clients in the application of CPM to plan-
ning problems. Third, the originating team went into private practice
and actively developed the concept and the techniques of applying CPM
to a broad range of projects and problems.
The construction industry in general (and the petrochemical indus-
try in particular) became the greatest single area of CPM application.
This was fortunate, because CPM had no sponsorship by a particular
agency or group; it had to develop and grow on its own merits.

15

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
16 Introduction to CPM Planning and Scheduling

A 1965 survey revealed that only 3 percent of the nation’s contrac-

tors actively used CPM, but since most of the users were larger con-
tractors, about 20 percent of the nation’s major construction companies
were actively scheduling with CPM. Of the contractors using CPM, 90
percent were satisfied with the investment in time and effort it
required. Actual dollars-and-cents savings in scheduling time and costs
were hard to identify, but CPM users believed that savings often
exceeded 10 percent.
PERT owed much to the earlier work by Kelley and Walker. Ironically,
after a courtesy review of their own work as converted into PERT, Kelley
and Walker were astute enough to use the term “critical path” as the new
caption of their Kelley-Walker (“main chain”) technique.
CPM enthusiasts saw PERT as a competitor and as a factor frag-
menting the enthusiastic, but limited, market for network techniques.
This feeling was intensified in 1962 when Secretary of Defense
MacNamara drafted an executive regulation stating, in effect, that the
existence of two different network-based scheduling systems was con-
fusing and that, henceforth, all Department of Defense organizations
would use PERT. At the time, this appeared to enhance the development
of PERT as a system at the expense of CPM.
PERT was applied to part of the Atlas E and to all the Atlas F site acti-
vation programs. It was also used in the Titan I, Titan II, and Minuteman
site activation programs. Although the application varied from site to site
and program to program, the approach used in Titan I is representative.
A site activation PERT network was developed for each site that was
limited to the events that would occur at that site. Within that site net-
work, individual networks were developed. The networks were so
arranged that they were compatible with networks prepared by Corps
of Engineers contractors as well as planned delivery schedules.
The Corps of Engineers Ballistic Missile Coordinating Office
(CEBMCO) used a network monitoring system to monitor the current
status of the Titan complexes. The cost of the monitoring system was
about 0.5 percent of the site construction cost.
Although large weapons systems and space systems accounted for
the largest number of PERT networks and the greatest expenditures on
PERT, a number of other agencies picked up the new technique. The
Atomic Energy Commission (AEC) used PERT to plan and control the
development of new components for atomic weapons. The National
Aeronautics and Space Administration (NASA) made broad use of PERT
and a form of PERT termed “NASA-PERT” (actually an activity-on-
arrow CPM-type network) in its space program planning. Also, such
firms as RCA and General Electric, which had recognized the potential
of networking in the late 1950s, applied network techniques to their
space projects.
 Project Control Systems Evolve in Academia 17

2.2. 1966–1970: The Fight between CPM and PERT

The concept period of the 1950s and the training and development of the
1960 to 1965 period, continued in the latter part of the 1960s. Although
not apparent at the time, acceptance of network techniques broadened
as the result of a number of independent factors:

1. The size of programs, such as Apollo, demanded an integrated proj-

ect control system, and NASA-PERT (or CPM) offered the best vehi-
cle for this type of system.
2. The evolution of network scheduling as a device for controlling a
single project was extrapolated into a program control system in
which a number of projects could be simultaneously integrated and
controlled.
3. The logical basis of the network approach, irrespective of its computer-
oriented identity, resulted in an increasing acceptance of its usefulness.
4. Academicians, particularly in civil engineering curricula, recognized
the validity of network scheduling as a project control approach, and
they were incorporating it into the undergraduate curriculum.
Graduating engineers were predisposed to use networks.

The Corps of Engineers, the Navy, and NASA were already utilizing
network systems. Other agencies, such as the AEC, the Veterans
Administration, and the General Services Administration, followed in
their footsteps.
The initial development of CPM included a sophisticated cost opti-
mization approach developed by Kelley and Walker that was included
as part of the basic CPM algorithm. This algorithm combined infor-
mation on crash and normal costs for each activity and estimated an
optimal completion time for the overall project. From a theoretical
viewpoint, the system is most interesting, but difficulties in collect-
ing the supporting cost and time information have precluded its wide
use.
The Kelley-Walker group (Mauchly Associates) also developed a com-
puterized approach to using CPM networks for scheduling labor, which
was called the resource planning and scheduling method (RPSM).
Concurrently, the CEIR computer consulting organization worked in
collaboration with DuPont to develop the resource allocation and labor
planning system (RAMPS). Although used on a very limited basis, the
extensions were well tested in field applications.
Current computer capabilities have resulted in a number of approaches
and proprietary systems. Although today’s computer technology greatly
facilitated the efficiencies of the computer program systems, the basic
principles have not changed.
18 Introduction to CPM Planning and Scheduling

By 1962, the PERT team had released PERT/Cost, which combined

cost reporting with the PERT network and came to be required in many
aerospace and defense contracts. The system is technically correct,
although it is based on a rather simple premise that the combined cost
of the various components completed in a project, when extended, will
provide a meaningful prediction of the completion date of the overall
project.
Most of the difficulties encountered in using the system have occurred
in collecting costs that can be meaningfully combined with the network.
The difficulties in reconciling an internal accounting system with the
special PERT/Cost breakdown lead the government to the approach
designated cost/schedule control systems criteria (CSCS).
International Business Machines Corporation (IBM), NASA, the Navy,
and others prepared their own versions of PERT and PERT/ Cost. IBM
and McDonnell Automation combined forces to prepare a coordinated
version of PERT and PERT/Cost, designated project management sys-
tems, or PMS. Although substantial technology was applied in the pro-
gramming and testing of computer systems for PERT and PERT/Cost,
applications tended to simplify theoretical approaches.
Variations of both CPM and PERT were developed by many organ-
izations, usually to get special systems to respond to special require-
ments. Variations of PERT included SPERT, GERT, MERT, and other
systems with acronyms that designated the changes entailed. CPM
was recast into precedence networks (PDM), which were substan-
tially different in approach but provided essentially the same calcu-
lated result.

2.3. PDM
Professor John W. Fondahl, of Stanford University, the early 1960 expert
on noncomputerized solutions to CPM and PERT networks, was one of
the early supporters of the precedence method, or PDM. He called it the
circle and connecting arrow technique. His study for the Navy’s Bureau
of Yards and Docks included descriptive material and gave the technique
early impetus, particularly on Navy projects.
An IBM brochure credited the H. B. Zachry Company of San Antonio
with the development of the precedence form of CPM. In cooperation
with IBM, Zachry developed computer programs that could handle
precedence network computations on the IBM 1130 and IBM 360. This
was particularly significant because in 1964, C. R. Phillips and J. J.
Moder indicated the availability of only 1 computerized approach to
precedence networks versus 60 for CPM and PERT.
Creation of an alternate format for preparing CPM networks required
new naming conventions to distinguish between the two. The form for
 Project Control Systems Evolve in Academia 19

traditional CPM networks was originally termed the AOA or “activity

on arrow” variant of CPM. The form for new style precedence networks
was originally termed the AON or “activity on node” variant of CPM. In
the AON variant, the activity description is shown in a box (or node) with
the sequence, or flow, shown by interconnecting lines. In most cases,
arrowheads are not used, although this leaves more opportunity for
ambiguous network situations.
Because the terms AOA and AON are similar, and possibly because a
box only represents a node to a mathematician, AOA became known as
ADM or “arrow diagramming method” and AON became known as PDM
or “precedence diagramming method.” Often, specifications copied from
older specifications may refer to the CPM being prepared in the AOA or
AON method. A sad reflection upon the care in which such engineering
documents are written is that it is not unusual for a specification to
require the CPM to be prepared using the AOA methodology and to run
the schedule upon the latest version of Primavera software (which only
supports PDM.)
Computer users have always preferred the PDM format because it
readily lends itself to graphical output. Another advantage claimed for
PDM is that the diagram is “cleaner” and, therefore, easier to follow. The
simplifying factor results from the fact that “redundant” restraints are
not required in PDM (as they are in CPM) to create unique activity
numbers (i.e., when activities span between the same two events).
Until recently, schedulers could request that their network computer
calculations be performed in either ADM (activity-on-arrow) or PDM
(precedence diagramming method). Primavera’s scheduling software
had been typical of this two-way option (i.e., ADM or PDM). However,
when Primavera software writers created a Windows version, they opted
to use PDM as the platform for the flagship program. The impact on
scheduling in the construction industry is substantial and is addressed
in this book.

2.4. SPERT and GERT

CPM and PERT are based upon mathematics and professors of mathe-
matics were quick to note many of the new insights opened by this new
branch of mathematics. If an estimate of duration is merely an estimate
and subject to a level of uncertainty, what might happen if randomly
some of the durations were raised and others lowered? If two or more
paths of the logic network were fairly close, this modification may well
shift the critical path and overall duration of the project. CPM provides
a set date upon which a project is expected to be complete. What is the
probability of the project finishing on that date, on an earlier date or on
a later date? If each of the durations of activities (in CPM) or between
20 Introduction to CPM Planning and Scheduling

events (in PERT) were randomly chosen between the Optimistic, Most
Likely, and Pessimistic Durations, we would get one value for an end
date. Repeat this process 100 or 1000 times and we get probabilities of
completion over a range of dates.
Unfortunately, computers in the 1950s were not powerful enough to
perform such analyses for more than demonstration logic networks of
few activities. Today, there are software programs (and supporting hard-
ware) that can perform 1000 iterations for logic networks of several
thousand activities in under a minute. Thus, not only can the Scheduler
determine the date on which the project can be expected to be complete
but also the probability of that expectation.
The mathematics behind neither CPM nor PERT permits Boolean
·OR· logic. If an activity in the logic network “A” is followed by two
other activities “B” and “C,” it is assumed that both can start upon the
completion of “A.” It is also assumed that “B” can start independently
of “C” and vise versa, either starting before the other or both at once.
In the real world this is not always true; sometimes you can start “B”
or “C” but only one at a time (the Boolean ·OR·.) Sometimes you can
only perform “B” and “C” if both are performed concurrently.
Sometimes the choice of which can be performed first is subject to the
status (started or completed) of a fourth activity “D.” And sometimes
the choice of successor is based upon a test—pass and go down one
path—fail and go down the other path. In the case of a failed test, the
logic network can even loop around to retake the test after corrective
measures have been taken. None of these possibilities are supported
by the mathematics of CPM or PERT. However, many of these possi-
bilities are supported by mathematical models envisioned in the 1950s
through today and more recently supported (at least in part) by
modern software programs generically noted as GERT programs. As
PERT was the acronym for Performance Evaluation and Review
Technique, GERT became the acronym for Generalized Evaluation
and Review Technique.

2.5. RDCPMTM
“The more things is different, the more they’s the same.” At of the turn
of the millennium, PDM supplanted ADM in the majority of the sched-
uling world. And yet, numerous serious practitioners noted flaws in the
implementation of PDM and many bemoan the loss of rigor of the ADM
system. Recent developments highlighted in the professional and tech-
nical societies and in academia have drawn attention to the focus upon
information relating to individual and groups of activities and the lack
of focus upon the relationships between these activities that was the
hallmark of the original ADM and PERT methodologies.
 Project Control Systems Evolve in Academia 21

Key among these concerns was a lack of a specific definition of rela-

tionships between activities other than the traditional “Finish 100% to
Start Next” relationship, the description of the reason for a specific rela-
tionship between activities and whether the relationship is mandatory
or optional, the ability to footnote the duration between activities (lag)
to the same degree as that afforded durations of activities, and the fail-
ure to calculate or provide access to the calculation of the attributes of
these relationships. Research publications submitted in various tech-
nical venues, including the Project Management Institute’s College of
Scheduling, the Association for Advancement of Cost Engineering
International, and the American Society of Civil Engineers, all raise
these issues and suggest various means to work around the problems.
Invariably, the proposed solutions suggest something akin to the need
for “dummy activities” that carry logic between “real activities” (this
being the hallmark of the original ADM system).
Posited by the authors of this text is a system that adds the follow-
ing functionality to a scheduling system. Since the focus of the major-
ity of these embellishments relates to the relationships between
activities, a suggested label for this system is RDM or Relationship
Diagramming Method. A fully integrated system is also under develop-
ment by the authors for general dissemination but subject to certifica-
tion under the trademark of RDCPMTM.
The key aspects of a RDCPMTM or RDM system include the ability to:
■ Identify nodes representing events or points of time at each point
where restraints converge. Such nodes will be similar to the i-node of
ADM but are established not for the purpose of data entry (as in
ADM), but rather for identifying points of merge bias (where several
restraints or logic lines come together) and the “mini-milestones” that
these points represent.
■ Identify the rationale or reason for each restraint, both by a code and
description thereof. A physical restraint (erect the walls before the
roof) is the most obvious example. Other types of restraint are
resources including crew, equipment, reusable forms and others all
already part of the thinking of the team preparing the CPM. However,
by expanding the recording of assumptions behind the plan used to
prepare the schedule, additional power may be gained such as (1)
automated guarantees that each activity in the network is preceded
by a physical restraint, (2) permitting “what-if ” analyses of the impact
of limiting or not limiting crews by various craft, reusable forms, or
other specific resources, (3) sorting and selecting by reason for rela-
tionship, and (4) providing an automated guide to areas of possible cor-
rective action when various events threaten to delay or disrupt timely
completion of a project.
22 Introduction to CPM Planning and Scheduling

■ Expand sort and select capabilities to the text of the activity descrip-
tion and various activity codes of the predecessors and successors of
an activity. For example, a selection may highlight each instance
where work by the mechanical subcontractor is immediately followed
by work by the electrical subcontractor.
■ Expand the types of relationships between activities to account for how
people actually plan their actions, rather than to match the options
set by software designers. For example, few people would say “Bob is
starting a 30-day activity next week and Mary will start her activity
15 days after Bob has started without regard to how much progress
Bob has made.” Rather, more people will say, “Bob is starting a 30-day
activity next week and Mary will start her activity when Bob is 50 per-
cent complete.” Thus, if the scope of Bob’s activity changes or if his
productivity is other than expected, there will be an automatic change
to the lag between the start of work for Bob and Mary.
■ Provide the same level of control over lag durations (between activi-
ties) as is (1) provided for activities, such as choice of calendar and (2)
range of duration for those systems that support PERT and SPERT
style calculations.
■ Expand the types of duration to include a Trend Duration (“TD”)
based upon an adjustably damped comparison between original and
actual durations classified by similar work scopes. A separate SPERT
style calculation could then be run based upon both the original and
trend durations.
■ Wrap in the power of GERT types of relationships including (1) B or
C to follow A but not both at once, (2) logical loops to cover test fail-
ure, corrective action, and retesting, and (3) choice of action based
upon progress or status of other activities within the logic network.
■ Expand the algorithms used to handle situations where actual per-
formance bypasses the planned logic and where work is performed out-
of-sequence. In addition to the choices of “retained logic” that assumes
an activity started out-of-sequence to be suspended until its prede-
cessors are 100 percent complete, or “progress override” that assumes
that once an activity is started out-of-sequence that the violated pred-
ecessor logic is no longer important, there may be a “modified progress
override” that assumes that the activity started out-of-sequence may
continue to zero remaining duration, but that successors thereof may
not start until its predecessors are complete. The choice of algorithm
should further be expanded from a project-wide decision to one which
may be set by type of restraint (“retained logic” for “physical”
restraints, “modified progress override” for “resource” restraints) or
even on a restraint by restraint setting by the Scheduler.
 Project Control Systems Evolve in Academia 23

■ Expand the types of float to include (1) a multi-calendar float (“MF”)

attribute to uniformly report float on a chain of activities having dis-
similar calendars and (2) a secondary or junior float (“JF”) attribute
to indicate the float of support activities leading to an i-node or merge
point that is selected to be driven by another chain of activities. (This
is similar to free float but is attributed to all preceding activities
rather than only the immediate preceding activity.)

Many other academic users and practitioners of CPM have suggested

similar and other extensions to CPM. Many variants of CPM and exten-
sions thereto exist in university computers. Eventually one of the estab-
lished software vendors or a new entrant will make such extensions
commercially available, and if they are successful, all other vendors
will rush to copy the new algorithms. Thus, just as PDM replaced ADM
and has become the primary method used in the construction industry
today, so too will RDCPMTM or some other diagramming method become
the standard of tomorrow. Notwithstanding, the basic rules of planning
and scheduling are immutable and it is the hope of the authors that all
users of CPM will understand and appreciate the basics, whichever
conventions and software is used.
This page intentionally left blank
 Chapter

3
Project Control Systems
Evolve in the Marketplace

By the period of 1965 to 1970, CPM theory had been disseminated

through academia and the value of computerization had been proven.
It was now up to the commercial venture groups to write and market soft-
ware services to bring this new technique to the marketplace. Again, that
this initial development was specifically for construction was fortuitous,
since it was in this field that there was a general understanding that
the manager of a construction project was to be given unrestricted con-
trol over that project. The evolution of CPM was to support this person
and not to report the performance of this individual to an “upper man-
agement.”

3.1. Commercialization 1965–1970

From 1965 to 1970, networking tools evolved into project control systems
(PCS), usually for the purpose of managing large programs or multi-
project programs. PCS approaches were developed for many projects,
including the World’s Fair in New York City, Expo ’67 in Montreal, con-
struction for the State University of New York, the Apollo launch com-
plex at Cape Canaveral, and the Bay Area rapid transit system (BART),
but the availability of tremendous amounts of project information, how-
ever important and meaningful, presented a new problem. Previously,
although decisions had been based on sparse and limited information,
the executive mind was essentially uncluttered by facts. Now, with proj-
ect and resource information flowing in, managers had to determine
which data were important and which could be disregarded to reach or
establish alternatives for decisions.

25

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
26 Introduction to CPM Planning and Scheduling

3.2. Commercialization 1970–1980

The 1970s were highlighted by several diverse new influences that encour-
aged the acceptance and utilization of the PCS approach. First, engineer-
ing school curricula added both network techniques and computer
applications to their undergraduate curricula, which resulted in a more
natural utilization by recent engineering graduates. Second, the evolution
of construction management and better management control became an
important corollary by the utilization of construction management. Also
during this time, a dramatic increase in construction litigation citing delay
as a reason for damages made schedules and their utilization more impor-
tant to both plaintiff and defendant. The existence and proper utilization
of a CPM plan was a significant factor in either supporting a contractor’s
claim or defending the role of the owner-construction manager in coordi-
nating a project. Finally, the dramatic evolution in computer compatibil-
ity not only made basic network systems more available, but also provided
an economical support for the implementation of network systems that
both tracked a schedule and correlated it with costs and resources.

3.3. Early Legal Recognition

The courts gave early recognition to the validity of CPM. In 1972 (Appeal
of Minmar Builders, Inc. GSBCA No. 3430, 72-2 BOA), the court rejected
a claim based on bar graph schedules, stating: “The schedules were not
prepared by the Critical Path Method (CPM) and, hence, are not pro-
bative as to whether any particular activity or group of activities was
on the critical path or constituted the pacing element for the project.”
Also in 1972, a Missouri Court (Natkin & Co. v. Fuller. 347 F Supp 17)
stated that bar charts did not “afford an overall coordinated schedule
of the total work covered by the contract.” An Illinois court (Pathman
Construction Co. v. Hi-Way Electric Co. 65 Ill. App. ad 480, 382 N.E. 2d
453,460) in 1978 noted that “technological advances and the use of com-
puters to devise work schedules and chart progress on a particular proj-
ect have facilitated the court’s ability to allocate damages.”
Early courts stressed the transparency of the original CPM presen-
tations. This may be compared to the court’s reaction to the modern vari-
ant of PDM as cited in Donahoe Constr Co. ASBCA #47,310 et al. 98-2
BCA ¶30.076 (1998.) This case, as discussed in Construction Scheduling,
Preparation, Liability and Claims, 2nd edition, by Jon Wickwire,
Thomas Driscoll, Stephen Hurlbert, and Scott Hillman (Aspen,) notes
that the court found “the utility of the baseline CPM schedule as a
benchmark for measuring delays in a window analysis was rendered
largely ineffective due to improper use of leads and lags.” Perhaps the
most succinct comment by the court in this 1992 case was that the court
found incredible the contractor’s expert analysis that “only the first five
 Project Control Systems Evolve in the Marketplace 27

days of each activity [footings and slab on grade] were on the critical
path.” Perhaps only a portion of the footing and slab were critical, but
since there was only one activity each without detail, the court was not
going to take the “say so,” by even a well-respected expert.
Thus the shift from more difficult to code to a computer but trans-
parent ADM to the more easy to enter to a computer but opaque PDM
could not come at a more problematic time than as the courts trans-
formed from the Frye, or “follow the expert you feel more credible,”
approach to the Daubert, or “show me, Mr. Expert, what you did,” stan-
dard now used in federal and many state courts. The key to the early
legal recognition of CPM was its total simplicity once it was explained.
But as computers got more powerful, software incorporated new features
and extensions that might not be deemed so simple.

3.4. PCs 1980–1990

The 1980s saw a shift from mainframe software (MSCS/Project2/Artemis)
to personal computer-oriented programs (Primavera, Aldergraf, MicroPert).
This shift brought Schedulers face-to-face with the computer screen.
Because many engineering undergraduates became personal computer
(PC) users in college and scheduling software became so affordable, many
smaller organizations began applying scheduling in-house.
In 1982, a review of 40 CPM/PDM programs showed:
Number of programs Percentage of total
Arrow diagram 35 87.5
Precedence 32 80.0
Both 26 65.0

Of the 40 programs, 30 required expensive mainframe hardware. Of the

10 mini-computer programs, the purchase price for 9 averaged $35,500. The
tenth sold for $1.1 million. Most of the programs could be leased for $1200
to $3500 per month (with lease payments credited to purchase). Thus, the
high cost of software made service bureaus a practical way to process net-
works. At least 5 of the 40 programs were offered only through service
bureaus. In the early 1980s, 8 of the 40 programs in the 1982 survey had
been converted into a PC version. The conversions included PROJECT/2,
by Project Software Development, Inc. and MSCS, by McAuto. The third
edition of this text (1984) listed 68 sources for CPM/PDM software.

3.5. PCs 1990–2000

By 1992, 32 of the 40 programs available in 1982 had disappeared, and
so had most, if not all, of the service bureaus. The 68 sources for
CPM/PDM software listed in 1984 showed only 10 “survivors” by 1992.
28 Introduction to CPM Planning and Scheduling

Primavera Systems was on both the 1984 and 1992 lists of CPM/PDM
software firms. Having crossed the millennium, Primavera Systems
has become THE software for the construction industry, with more than
35,000 members of the construction industry holding over 350,000
licensed copies of P3 Primavera Project Planner and Suretrak. This
well exceeds 95 percent of the market.
The widespread use of PCs by the 1990s transformed the use of CPM.
Previously, each “run” of the software had a significant expense and often
was available only once per day (running overnight from a service bureau)
unless the user was willing to pay an even higher fee for “premium time”
use. Project personnel often spent many hours carefully reviewing the
input to assure that a “run” would not be wasted by either an error mes-
sage or a bad result. All at once users could make multiple runs, allow-
ing the computer to be used to “locate” errors of the input. Additional runs
could be made for various “what if ” scenarios and further refinement of
claims analysis allowed the use of the Windows (or multiple time period)
methodology first discussed in the 4th edition (previously known but
deemed too expensive to pursue other than as an academic exercise).
Coincident with the more popular use of the Windows methodology
for analysis of delay, was the introduction of the similarly named
Windows operating system by Microsoft. And with the more powerful
computer architectures now possible with this more powerful operating
system, many software companies, including those providing CPM soft-
ware, were required to rewrite their software from scratch. In addition,
a new and potentially “larger than the construction world” customer base
for CPM software, that of the IT or information technology world, began
to develop a desire for CPM. This customer base, having a less detail-
oriented need for the rigor of classical CPM, usually desired the PDM
variant that allowed a good deal of “fudge” in defining the relationship
of one “activity” to another. This general shift from ADM to PDM through
the 1990s led to further discussion of this variant in the 5th edition.

3.6. PCs 2000–2005

The use of Primavera’s newer flagship software, P3e Primavera Program
Manager, by the intellectual technology, aerospace, manufacturing, and
other industries now exceeds 1,000,000 licensed copies, although this rep-
resents only a fraction of the larger market, which is dominated by
MicroSoft Project. (The IT and other industries account for some portion
of the P3 and Suretrak licenses, and a growing number of the construc-
tion industry companies are converting to Primavera’s P3e/c for con-
struction as well as Primavera’s latest offering Primavera Construction.)
The large number of non-construction users and their differing needs
now drives the development of software upgrades. This group typically
sees a smaller distinction between tasks, activities, and projects—one
 Project Control Systems Evolve in the Marketplace 29

individual’s “project” may be to personally prepare the code for one module
of a larger piece of software, all of which are part of a “program.” Keeping
track of the progress of these highly trained individuals (or resources) who
may be working at various locales around the world, is more of a matter
of coordination than the choreography necessary for running a once-off
construction project.
The newest features of the software, therefore, are focused upon collab-
oration and reporting of various levels of detail (from “task” to “activity” to
“summary activity” to “project”) to upper levels of management, prefer-
ably on a real-time basis. It is expected that any individual or “resource”
performing “work” will do so on a computer. The concept that productivity
can be measured by each swing of a hammer recorded on a real-time basis
much like a click on the keyboard is obviously not correspondent with the
real world of construction. However, the software continues to provide the
basic calculations necessary to schedule a logic network if a proper logic net-
work is prepared by the project team.
The latest “improvements” to the software provided by various vendors
also assist project managers (and the Scheduler serving the project man-
ager) to provide better communication to upper management and other
interested parties. The tools are still good and new features add value; how-
ever, the default is that the features are tweaked to the larger marketplace
of IT and not construction. It is a paradox that this new power made avail-
able to the casual user makes it ever more important that the user be
knowledgeable in the underlying theory of CPM. Proper understanding
and use of these new tools that are not necessarily tweaked to construc-
tion and related fields is required and is addressed in this edition.

3.7. PCs 2005–2010

The mighty ship USS Scheduling is coming full circle. While once-upon-
a-time, project managers could not understand “these new-fangled” logic
networks that were run by “computer specialists” from headquarters and
preferred the use of “old fashioned” bar charts, the advent of untethered
PCs in the 1980s led to the general acceptance of CPM by project man-
agers and their staff members. The consolidation of PCs to a worldwide
network (and the attendant need for network administrators setting stan-
dards above the level of a project, even to the suggestion that there should
be a standard method or “module” to install pipeline or erect a wall) has
led many project managers to leave the preparation of professional sched-
ules to the professionals and to plan their jobs with the module barchart
blocks, rather than at the level of detail grudgingly accepted in the 1970s
to 1990s. It is all fine and well that the accounting department need not
expend resources to measure project productivity in the field, but if the
cost is to add an additional burden to the already overworked project
manager and reduce the usefulness of the schedule to this project, the
30 Introduction to CPM Planning and Scheduling

field may fall back upon the bar chart and even the old ToDo list to
manage the project.
A successful project trumps all, however, and the successes of better
project managers who properly use the new tools will again lead others
to relearn the lessons of the 1970s and 1980s. Perhaps we will have
stealth project managers running bootleg CPM schedules under the
radar. Perhaps these most successful project managers will demand
and get the autonomy necessary to base their project plan upon the
needs of the project and not other more global concerns. Perhaps the
mighty ship USS Scheduling will continue another 180 degrees and
usher in a new era of personal computing where a wireless connection
to the web is used to provide and not drain resources from the project
managers’ credo, that is, the project comes first.

3.8. The Sixth Edition

The construction industry has long recognized this text as the leading
scheduling authority. Many specifications prepared by the leading engi-
neering design firms throughout the United States specify that the proj-
ect CPM is to be prepared and administered in accordance with the
principles stated in the text. In June 2004, the Planning & Scheduling
Committee of AACE International recommended the 5th edition as the
primary reference to be used to prepare for their new Planning and
Scheduling (PSP) Professional Certification examination.
In 2002, a charter group (including co-authors O’Brien and Plotnick)
formed the Project Management Institute (PMI) College of Scheduling
(COS). One of the several reasons for forming the COS is the concern
of many experienced schedulers that many bad construction schedules
are being generated by “screen jockeys” using powerful software, such
as Primavera P3, without a sound underlying understanding of sched-
uling and/or construction. A number of the issues raised with misuse
of the additional power of such powerful software are addressed in
Chapter 9 (Adding Complexity.)
All prior editions emphasized the development of a logical plan as
the base of a meaningful schedule. In PDM output, the logic of the plan
may not be clear or evident. This edition demonstrates the proper
use of popular scheduling software products (such as Primavera’s P3
and P3e/c). It also demonstrates how the software can be misused
(intentionally or unintentionally) to produce or force incorrect results.
A comparison is made between the P3 and P3e/c programs where appli-
cable and the features of these products are compared with other pop-
ular software products.
There are options and practices described in the 6th edition that can
result in output in which the logic plan can be more readily apparent.
 Project Control Systems Evolve in the Marketplace 31

Two published articles in the Technical Track at the COS first annual
conference in 2004 addressed this problem with a new mathematical
approach, which was highlighted in the ENR coverage of the conference.
These insights are added to the 6th edition.
Case studies of the successful use of CPM scheduling on major proj-
ects in the past 5 years are updated. The chapter on the use of CPM in
claims and litigation is expanded to include references to relevant liti-
gation in the past 5 years. Use of a Windows approach in Time Impact
Analysis is discussed in more detail. This consideration is important in
determining whether delays are concurrent (or non-concurrent), which
is important in allocating responsibility for delay. A chapter on the use
of CPM in determining and proving disruption has also been added.
In the 5th edition, Chapter 26 closed the book with a brief reference
to advanced topics such as PERT, SPERT, and GERT. These topics and
risk analysis in scheduling are expanded with examples and illustrations.
The 6th edition expands the hands-on coverage of microcomputer
software systems first covered in the 5th edition. A significant part of
this coverage in the previous edition consisted of 43 screen inputs and/or
outputs of the Primavera P3 System. Comments from users of the text
noted that these screen shots were shrunk to 6'' × 9'' page size, making
them hard to read. This problem was exacerbated by printing the screens
in black and grey tones and equated to watching color TV on a black and
white set. This edition necessarily has similar screens: however, the
attached CD has the same screens in color and larger size (i.e., the size
of the computer screen, which is almost three times the size of the
screens in the book). We also added new P3 and P3e/c screen examples
to this edition.
Over time, a dichotomy developed between ADM and PDM users.
Primavera cut the Gordian knot by selecting PDM exclusively. On the
other hand, to toss out all ADM experience would be an unconscionable
waste. Two construction management professors, Richard Smyth at New
York University and Fredric Plotnick at Drexel University, both said that
ADM was the only way to teach scheduling theory, and that PDM had to
be given its due as THE way to calculate and present schedules today.
Each said that they do that by separating theory and computer practice.
As a result, this book is organized into the following sections:

I. INTRODUCTION TO CPM PLANNING AND SCHEDULING

II. THE THEORY OF CPM PLANNING AND SCHEDULING
III. THE TOOLS OF CPM PLANNING AND SCHEDULING
IV. THE PRACTICE OF CPM PLANNING
V. THE PRACTICE OF CPM SCHEDULING
VI. ADVANCED TOPICS
This page intentionally left blank
 Part

The Theory of CPM

2
Planning and Scheduling

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
This page intentionally left blank
 Chapter

4
Your New Tool—
Read Before Using

Attached to the rear cover of this text is a CD-ROM disk including a fully
functional demonstration copy of the Primavera P3 software product.
This demonstration copy is limited only in the number of activities that
can be included in schedules calculated with the software. Also pro-
vided is a fully functional, 90-day trial copy of Primavera Construction.
(This may be extended by purchase of a license from Primavera).
However, like the instructions provided when you purchase a new power
saw, it is highly recommended that you fully read how these products
work before you attempt to use them.

4.1. Primavera and Your Power Saw—Useful but Dangerous Tools

The defacto standard for scheduling software in the construction indus-
try has been Primavera Project Planner, also known as P3. In 2000,
Primavera Software Systems improved upon its P3 software with the
introduction of Primavera Program Manager P3e for use by the manu-
facturing and IT industries, and subsequent release of Primavera
Program Manager for Construction P3e/c. However, due to the addi-
tional power of P3e/c, requiring additional computer resources and train-
ing (both at the user and at the administrator level) in return for limited
additional benefits for most small- to medium-sized contractors, engi-
neers, and owners involved in the construction industry, P3e/c has not
quite yet taken off in the construction venue. This shift from “the super-
intendent does it all on his laptop” to “the organization’s administrator
sets company-wide coding standards” created some resistance to con-
version. This resistance may change as the additional features of ver-
sion 5.0 are recognized or yet other features are added that are of benefit

35

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
36 The Theory of CPM Planning and Scheduling

to this large base of smaller users. At some point, there may be a mas-
sive shift to the newer system. Primavera recognized this in 2004 to
some extent with the release of a “P3e/c Lite,” marketed as Primavera
Construction and aimed at small- to medium-sized construction firms.
Then again, as databases become larger, tie-in applications more
desired, and wireless technologies become ubiquitous, P3e/c may be
leapfrogged by an even more useful platform by Primavera. Similarly,
the competitors of Primavera are not sleeping. However, for purposes
of demonstration of the working of a standard CPM software tool, this
text features Primavera Project Planner P3 with references to the other
software provided by Primavera and its competitors where appropriate.
The major differences that distinguish one software system from another
are the ease of input, the ease of and variety of means of reporting output,
and the features that permit the user to get around some of the limitations
of the mathematics upon which CPM is founded to more perfectly model
the real world. As noted in the introduction to this text, in the real world,
you may begin eating a portion of your breakfast while still cooking the
rest. But would you ideally plan to do it on a regular basis? Even here, you
are planning and scheduling your activities. You may begin drinking your
coffee while cooking an egg, but you will want to heat the pan prior to pour-
ing the egg in. If you desire a breakfast sandwich, you will hopefully start
the toast going before the egg is fully cooked. Primavera (and other soft-
ware systems) permits you to show cooking and eating the breakfast as
overlapping activities without the need to show the detail of how and how
much overlap is to be accomplished. The bar chart provided by Primavera
(see Figure 4.1.1) may well show the two activities as starting at the same
time. Worse, if not careful, the software may show you completing the
eating before completing the cooking. To reiterate, it is the intention of this
text to teach the proper means of planning and scheduling as well as the
shortcut tools available so that your schedules will not make this mistake.

Figure 4.1.1Original plan is start cooking (10 min) and begin eating while finishing cook-
ing. Actual was started cooking breakfast, dropped eggs, cleaned up, continue cooking.
CPM scheduled at start and rescheduled at 5-min point.
 Your New Tool—Read Before Using 37

4.2. How Does Primavera (or Other Scheduling Software) Work?

Quite a bit of discussion of the mathematical origins of CPM analysis and
the migration of pure mathematics to an engineered system for the plan-
ning, scheduling, and control of a project, has been presented. At first, the
process of CPM may seem trivial or it may seem strange and rather coun-
terintuitive, but like any system based in mathematics, it is grounded in
basic axioms. Modern software systems are forgiving and may allow the
user to ignore some of these axioms, but only at the risk that if not truly
understood, the calculated output may be less than accurate.

4.3. What Goes In . . .

Any project can be subdivided into a list of activities or tasks. For pur-
poses of precise terminology of the Critical Path Method of Planning and
Scheduling, tasks are scopes of work that are components of activities.
An activity can be composed of one or several tasks. The order in which
the tasks of an activity are performed may be irrelevant or may be so
obvious that formal instructions would not be given to those perform-
ing such tasks. Thus, when changing a tire on a vehicle, the activity
“remove lug nuts” is composed of tasks “remove first lug nut,” “remove
second lug nut,” etc. and need not be separately specified, while the
tasks of the activity “replace and tighten lug nuts” may either be stated
as separate activities or not, depending upon the experience of the
person placed in charge of this scope of work.
The level of detail of specific activities may vary, but it is an axiom of
all Critical Path Methods of scheduling that each activity (or specified
portion thereof ) can start only upon 100 percent completion of some
other activity (or specified portion thereof), except for a first or starting
activity that does not have a stated predecessor. With an event-based
system, such as PERT, each specified event can occur only after some
other event has occurred, other than a first or starting event that does
not have a stated predecessor. The methodology supports only “hard”
relationships, even if the relationship is loosely defined, and does not
support “fuzzy” relationships, of what may occur. Even the probabilis-
tic methodologies of SPERT and GERT are based upon “hard” estimates
of probabilities, and not merely upon random choice. CPM is a discipline
of engineering and not merely an application of intuition. This is the
basis of its many strengths, as well as its weaknesses.

4.4. The Initial Logic Network—Input

Often forgotten in classes on use of scheduling software is the idea that the
software is only a tool and that the output will only be as good as the input.
38 The Theory of CPM Planning and Scheduling

This is the GIGO or “Garbage-In, Garbage-Out” rule. The purpose of the

software tool is to assist with some of the laborious calculations that are
required by this mathematical analysis but still requires care with the
input, much like a power saw cuts through wood faster and cleaner than
a hand saw, but still requires care in measuring where to cut.

4.5. The Logic of the Logic Network

The backbone of the traditional implementation of the Critical Path
Method is a graphical model of a project. The basic component of the
model is the arrow. Each arrow represents one activity in the project.
The tail of the arrow represents the starting point of the activity, and
the head represents the completion. The arrow is not a vector and it is
not drawn to scale. It can be curved or bent as required; however, it
cannot be interrupted because it is a separate entity.

4.6. Arrow Diagram

The arrows are arranged to show the plan, or logical sequence, in which
the activities of a project are to be accomplished. This is done by answer-
ing two questions with each arrow:

1. Which arrows (activities) must precede this one?

2. Which arrows (activities) must follow this one?

The resulting logical flowchart is a network of arrows, usually referred

to as either the arrow diagram or the network. For example, consider a
routine checkup of your car as a project. Assume that you want the fol-
lowing work done:

■ Rotate tires
■ Lubricate
■ Change oil
■ Wax and polish
■ Drain antifreeze

CPM is often referred to as a “decision maker.” This is a misnomer

because CPM, being inanimate, cannot make decisions. However, the use
of CPM encourages the user to make decisions to draw the arrow diagram.
 Your New Tool—Read Before Using 39

In this example, a decision is required before any arrows can be drawn.

The mechanic must decide whether to do the hoist work first or last.
Assume that the mechanic decides to do the hoist work first. Accordingly,
the first arrow will be

Following this are all of the arrows that could logically follow hoist-
ing the car. From the work list, they are rotate tires, lubricate, and
change oil.

When the activity, lower car, is added, note that the general work list
is not broken down into enough detail to show the mechanic’s work
plan. Adding this activity after the hoist work:

What does this really say? It says that the activities cannot start
until the hoist is raised and must finish before the hoist is lowered.
Something is missing, however. The activity, rotate tires, indicates
that the mechanic must get the spare tire out while the car is on the
hoist. That is not logical, and it certainly is not what the mechanic
might be expected to do. Also, it is usual practice for the mechanic to
loosen the tire lug nuts before raising the wheels clear of the ground.
Change
40 The Theory of CPM Planning and Scheduling

This part of the network then becomes

For lubricate, the first network indicates that oiling and checking
items under the hood (battery, alternator, radiator, brake fluid, etc.)
must be done while the car is up on the hoist. To do this, the mechanic
would need stilts or a ladder.

Similarly,

This part of the network then becomes

Figure 4.6.1 Arrow diagram, car checkup.

 Your New Tool—Read Before Using 41

Combining the portions of the network and adding the two activities
not previously shown, drain antifreeze and polish and wax car, the
arrow diagram representing this everyday operation is as shown in
Figure 4.6.1.
Preparing the arrow diagram focuses on one activity or a group of
related activities at a time. The reason is obvious: Only one arrow is
drawn at a time. The very simplicity of the reasoning gives strength to
the technique. No one can thoughtfully consider all details of a multi-
million dollar project simultaneously, but using the arrow diagram to
record thoughts spotlights and plans one area at a time. As each area
is completed, thoughts and plans are recorded by the arrow diagram.

4.7. Logic Diagrams

The logic diagram is the most important single feature of the CPM
method. Logic diagrams have long been used by mathematicians, and it
was assumed by many that mathematician Kelley used the logic dia-
gram to convey the basic plan sequence to the computer. In a 1983 meet-
ing, Kelley stated that the entire algorithm was envisioned
mathematically. He used the logic diagram, initially, to explain the
approach to DuPont management. Introducing the logic diagram to reflect
the intended sequence of a plan had a dramatic impact on the planning
process. A number of abstract logical rules are useful in the preparation
of a network. If activities A, B, and C occur in series, their network rep-
resentation is

If the statement is that B and C follow A, this is one solution. A more

correct one would be

Examine the latter solution. Unlike the first one, it shows B and C as
independent activities. When drawing network sequences, it is not
proper to add logical connections that are not stated. Perhaps this is an
obvious caution, but you must constantly guard against subtle, unin-
tentional logical interconnections.
42 The Theory of CPM Planning and Scheduling

If activity C follows B and activity D follows A, what is the network

expression? It may seem to be

However, no connection between C and A or B and D was stated.

Therefore, the proper relation is

Now, if both A and B precede both C and D, the network expression is

However, this is not correct if A and B precede C but only B precedes D.

Stating this diagram as

means that B is not shown as a precedent to C. And stating the diagram

as

means that B is not shown as a precedent to D. The problem is that the

arrow B cannot be broken into two parts; the arrow diagram is not per-
mitted to “speak with a forked tongue.” The dilemma is solved by intro-
ducing the logical connection, an arrow that represents logic flow but
no work. To differentiate from regular arrows, the no-work connections
 Your New Tool—Read Before Using 43

are dashed-line arrows. In this example, the logical connection (or log-
ical restraint) is

The network now shows that C follows A and B but D follows only B.
The concept of the logical connection is common sense, but it is indis-
pensable in CPM.
Now consider a network example with two parallel chains of activi-
ties. One of these chains is made up of activities A, B, and C in series.
The other is made up of X, Y, and Z in series. A and X are the starting
activities; C and Z are the terminal activities. This gives

Now add an activity M originating at the project start. If activity M

must precede C and Y, the result is

The point is that any number of logical restraints can originate from the
finish of an activity. Similarly, any number can lead into the start of an
activity. In the network

adding terminal activity E, which follows A but is independent of C, is

not accomplished by
44 The Theory of CPM Planning and Scheduling

Figure 4.7.1 Logic network examples. (a) Activities C and D

follow A; activity E follows C. Activity F follows D; and E and
F precede B. (b) G follows F but precedes H; K follows A but
precedes L; F follows A; A and D start at the same time; J
and L precede completion of the logic network; G follows D;
J follows G; H follows G but precedes M; and M precedes L.
(c) A and D start at the origin; J follows F but precedes K; C
follows A but precedes G; H follows D but precedes L; B fol-
lows A but precedes E and F; E follows B and C; K follows G
and H; and E, K and L precede completion of the logic net-
work.
 Your New Tool—Read Before Using 45

This is typical when unintentional logical connections are made. To

keep E independent of C, add another logical restraint after A:

This might be termed a logic splitter or logic spreader. Logic cannot

back up from B against the arrowhead, which functions as a check valve.
Figure 4.7.1 offers more examples.

4.8. Logical Loop

If activities A, B, C, and D are in series and activity E following C pre-
cedes B,

The portion B, C, and E is a logical loop. It is a question of “Which

comes first, the chicken or the egg?” Since a loop is illogical, it has no
place in a logical network. It might seem unlikely that anyone would
draw a loop. In large complex networks, however, it is quite common for
loops to be inadvertently inserted.
Figure 4.8.1 shows the site layout for a hospital project. Because the
existing hospital was in a prime location, the new building was to be con-
structed immediately behind it. However, an annex building had to be
demolished before new construction started. Since the service annex
included the kitchen-cafeteria area, a temporary kitchen-cafeteria had
to be established in the existing building until a new kitchen-cafeteria
could be constructed and the new building was ready for occupancy. At
that time, the temporary kitchen-cafeteria was to be vacated. This is
easily shown in arrow diagram form:
46 The Theory of CPM Planning and Scheduling

Figure 4.8.1 Hospital site layout.

The same information can also be shown in PDM (shown here for
comparison only.)

However, a factor not noticed until the preparation of the arrow dia-
gram was the location of the electric power distribution vault for the new
building. The vault was to be the site in the old building occupied by the
temporary kitchen. Adding that information to the network resulted in
the following loop.

The situation was pointed out to the owner and the architect. Since
the power vault was not needed until a year later, a new vault location
was designed and constructed. Through the use of the CPM plan, a
costly and inconvenient time loss was foreseen and avoided.

4.9. Non-construction Examples

Any number of non-construction projects can be planned using CPM.
Some actual projects include:
 Your New Tool—Read Before Using 47

■ Shipbuilding
■ City planning
■ Refinery maintenance
■ Architectural design
■ Staffing a new plant
■ Researching a project
■ Embarkation of a construction battalion
■ Cooking a meal
■ Creating procedures for state approval of a new school
■ Bringing a show to Broadway
■ Preparing a corporate budget
■ Preparing a city budget
■ City approval of plans
■ Purchasing a new house
■ Purchasing a car
■ Manufacturing one car
■ Creating a family camping trip activity list

Although there is no one correct activity list for a family camping

trip, this example assumes that a family consists of a father, a mother,
and two children. A typical list might be the following.

■ Prepare budget
■ Pack car
■ Collect site information
■ Select site
■ Purchase equipment
■ Make equipment list
■ Prepare food list
■ Make camp site reservations
■ Schedule vacation
■ Plan clothing list

Figure 4.9.1 presents one plan that could be used to coordinate these
activities.
48 The Theory of CPM Planning and Scheduling

Figure 4.9.1 Family camping trip.

4.10. Summary
This chapter discussed the concept of the network, as well the premise
that CPM can encourage decision making but cannot make decisions
itself. Preparing arrow and logic diagrams helps the planner to under-
stand a project by clearly defining the activities required to complete it.
CPM is particularly applicable to construction work, but its usefulness
is by no means limited to the construction field.
 Chapter

5
Network Construction

Chapter 4 discussed the concept and the fundamentals of construction of

the CPM network. This chapter covers the practical mechanics of network
construction. Since CPM is a logical and organized planning system, it is
important that the physical layout of the network reflect the same logi-
cal organization. The thought required to separate the network’s parts into
practical subdivisions contributes to the overall plan. The network is
often used to present the plan to strangers to the project. If the physical
layout is clear, concise, and well arranged, first impressions will be good.
However, CPM can also expose poor planning.
Figure 5.0.1 shows two networks with the same information. Both are
logically correct, but the top network was drawn directly from a problem
description without careful attention to physical layout. The bottom net-
work is a rearrangement of the top one. It has only 12 activities. In a proj-
ect network, the differences between network layouts and the possible
resulting confusion would be multiplied a hundredfold.

5.1. Form and Format

The network is traditionally drawn on reproducible paper or Mylar. In
preparing it, trial layouts should be sketched out before drawing it in fin-
ished form. The sketches are usually done on a blackboard, nonrepro-
ducible paper, vellum, or grid paper. Grid paper with non-reproducing
squares is especially helpful in laying out a network. It can be used in the
freehand sketch phase or for the finished network.
These freehand sketches should be saved and otherwise treated as the
calculations of the design engineer’s job book. After all, a specific size for
a steel beam is not chosen merely because it “looks right” but only after
preparation of a dimensionless rough draft, followed by selection of appro-
priate equations and careful calculation. Traditionally, the page upon

49

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
50 The Theory of CPM Planning and Scheduling

Figure 5.0.1 Rough and rearranged networks.

which these calculations are performed is stamped with a form noting the
original designer with boxes for initials by the reviewer and often by a
second reviewer. A check mark in the same color ink as the reviewer’s sig-
nature is placed next to each number or drawing element to show that each
was checked. This calculation sheet is compiled into the job book that
then goes to archived storage.
While today’s design engineers struggle with how to duplicate this
process in a digital world, some variation of this method is still the norm.
The job book may still be the base document reviewed if there is litigation
relating to when design decisions were made. Similarly, the initial sketches
of the CPM logic may be the base document reviewed if there is litigation
relating to the contractor’s initial “plan of execution” of the work.
Because early networks were modest in size, drawing size was not a
problem. As networks became larger, the size of the drawings increased.
A huge network is unwieldy and difficult to handle, however. Although
there may be times when long, rollout drawings are practical, for most
work, it is better to break down larger networks onto a number of sheets.
The selection of the scope of each sheet is important. The sheet should
not be crowded, but it should be well used. In subdividing the project so that
it can be presented on a number of sheets, keep the practical use of the net-
work in mind. For instance, if all the foundation work for a building appears
 Network Construction 51

Figure 5.1.1 Random line example.

on one sheet, the field office will find the network easier to use, since the
current field status can be located on one network sheet or two at a time.
There is no fixed rule for optimum sheet size. The Army Corps of
Engineers uses a 34 × 44-inch sheet. A larger size can be used for draw-
ing and then reduced for better handling. Many of the early diagrams
were drawn with random direction lines (Figure 5.1.1) or wide-sweep-
ing curves (Figure 5.1.2.) The clarity of hindsight obscures whatever rea-
sons there might have been for originally using this method, which is
mentioned here because people continually rediscover abandoned tech-
niques and try to use them.

5.2. Events
The intersection of two or more activity arrows is termed an event. An event
has a zero time dimension. However, all activities leading into an event
must be completed before any of the activities leading out of the event can
be started. This is just a restatement of the rules of network logic.

Figure 5.1.2 Sweeping curve example.

52 The Theory of CPM Planning and Scheduling

Figure 5.2.1 Example of milestone event: ready to advertise

for bids.

Certain key events are called milestones; they represent important

intermediate goals within the network. For instance, “ready to adver-
tise for bids” (Figure 5.2.1), is an important event. It represents an
instant in time but has no time dimension of its own. To reach this par-
ticular event, all activities pertaining to the design and specifications
for the project must be first completed. No action toward getting a con-
tract can be taken until the logic flow has passed through the event.
On the CPM diagram, important events can be identified by name.
Event titles are not emphasized; instead, events are assigned numbers.
Because each activity is bounded by a starting and completion event, the
event can be identified by the number.

The number assigned to the starting event is referred to as the i; the

number assigned to the completion event is the j. (These designations
were used by the founders of CPM and have remained in general use,
probably because of their brevity.) Thus, the typical activity looks like:

The i-j number for an activity can be used as an abbreviated name for
the activity. A number of rules must be followed in assigning event num-
bers to a network.

Rule 1. Each activity must have a unique i-j description, but often two
or more activities span the same events. For instance, between events
1 and 4 could be the following:
 Network Construction 53

TABLE 5.2.1 Activity List,

Common Activity

i–j Description

1–4 Plumbing lines

1–4 Power conduits
1–4 Steam piping

A list of these activities would read as shown in Table 5.2.1. This con-
fusing situation is corrected by adding logical restraints originally called
dummies. The term “dummy” was used because the connections say
nothing new; it was added only so that unique event numbers could be
introduced. The more proper term “restraint” is used now. The activity
list now reads as shown in Table 5.2.2.

Rule 2. When event numbers are assigned, the number at the head
(or j end) of the arrow should be greater than the event number at the
tail (or i end). That is, j > i. In early computer programs, the ability of
the computer to calculate the network often depended on this rule, as
well as on the consecutive numbering of events. All computer pro-
grams handle nonconsecutive event numbers and random numbering

TABLE 5.2.2 Activity List,

Unique Numbers

i–j Description

1–2 Plumbing lines

1–3 Steam piping
1–4 Power conduits
2–4 Restraint
3–4 Restraint
54 The Theory of CPM Planning and Scheduling

(random numbering can be j > i, j < i, or both in a network). Not only

is random numbering a convenience, it is often a necessity. For
instance, consider the partial network

Assume that the network continues for perhaps 50 more event numbers.
Now, suppose it is discovered that the activity clear and grade, which
should follow activity 1–2, survey and lay out site, and precede both 2–3,
install septic tank, and 2–4, underslab plumbing, was forgotten. Without
random numbering, the network would have to be renumbered as follows.

Since there would now be 51 event numbers, 50 of them would have to

be changed (all except event 1). With random numbering, the revised
network could be

No event numbers would have to be changed and only one would have
to be added. Since many of today’s networks have in excess of 1000
events, random numbering is very important when activities must be
added to the network.
Since random numbering is available, why even try to follow rule 2,
which might be called the traditional rule for event numbering? First,
numbering in the j > i manner makes it easier to locate events on the
diagram. Second, logical loops are more easily identified. Using the
example of a loop and numbering the events,
 Network Construction 55

Note that 4 > 2, or i > j, for activity E indicates a loop. Reverse the posi-
tions of 2 and 4:

Then j > i for activity E but not for activities B and C.

Event numbers should not be added until a network is completed and
is ready for the first computation and numbers should be assigned in a reg-
ular fashion. This can be done horizontally (Figure 5.2.2a) or vertically
(Figure 5.2.2b). Either one is acceptable. In the horizontal method, event
numbers are assigned along a chain of activities until a junction event
(a meeting of more than one activity) is reached. The routine is repeated
until all chains into the junction event are numbered. In vertical event
numbering, the numbers are assigned up and down vertically but still
observe the j > i rule.
The vertical numbering system localizes numbers in areas of the dia-
gram, which makes it easier to locate a particular activity on the network.
The horizontal numbering system results in logical groupings of activities,
so the i-j list (or printout) has groupings of activities that are logically
related. However, horizontal numbering can produce networks that make
it difficult to locate an event number. Similarly, random numbers can
make it difficult to locate a particular event on the network.
The number of digits in an event is limited by the computer program
used. Older programs are often 3-digit-oriented. Since the average ratio
of activities to events is about 1:5, the 3-digit concept limits the network
size to about 1500 activities.
Today’s major programs can accept 5 digits, which permits a net-
work of 150,000 activities. Many programs can also accept alphabetics,
so that the maximum network size is essentially unlimited. Increased
capacity allows many events, which can then be assigned digits by area
or function, such as purchase material or equipment or drawing review.
In drafting the network, it is optional whether the event is circled or not.

There is no significance to the event numbers except their value in iden-

tifying the activities. To the one who assigns the numbers, the logic is
obvious. However, people not familiar with CPM often try to read unin-
tended significance into the event numbers.
56 The Theory of CPM Planning and Scheduling

Figure 5.2.2 (a) Horizontal numbering, and (b) Vertical numbering.

 Network Construction 57

Figure 5.2.3 Horizontal format.

Activity descriptions should be written horizontally. To do this, a part

of each arrow (except restraints) must be drawn on the horizontal
(Figure 5.2.3). A comparison of the three cases shown in Figures 5.1.1,
5.1.2, and 5.2.3 illustrates the advantages of horizontal activity titles.
Another temptation for the drafter is to code activities rather than use
full titles. The example network shown in Figure 5.2.4 is coded; compare

Figure 5.2.4 Untitled (coded) network.

58 The Theory of CPM Planning and Scheduling

Figure 5.2.5 Significant primary activities.

it with Figure 5.2.3. A coded network is easier to prepare than a titled

one, but is almost useless because not even the person who prepared the
diagram can read it directly.
When arranging a network, center the significant activities on the sheet
so that they function as the backbone of the network. These main activi-
ties put visual emphasis on the important areas and minimize crossover
of arrows. Figure 5.2.5 illustrates the technique. In the past, some speci-
fications have required that the critical path be the network backbone. This
is not a valid requirement though, because the critical activities have not
yet been identified when the network is being prepared. However, activi-
ties that are usually critical can be identified from experience; these are
referred to as significant.
The arrow size and spacing are quite important. If the arrows are too
long and are widely spread, the diagram will become too large and
unwieldy. On the other hand, if the arrow arrangement is too tight, the
network will be difficult to read. Also, a crowded network cannot be read-
ily revised or amended. The usual arrow length is 2 to 3 inches but is not
mandatory. In the example activity 1–4 must be the sum of the lengths of
1–2, 2–3, and 3–4. A minimum vertical distance of 2 to 3 inches between
arrows leaves room for revisions.
 Network Construction 59

Backward arrows should be avoided because they are confusing and

are drawn against the time flow of the network. They also increase the
possibility of introducing unintended logical loops. In the example shown
in Figure 5.2.6, the first network did not show a requirement that the
hydro testing and insulation must precede the start of lath because
part of the piping is enclosed by lath and plaster. The restraint arrow
added to show this logic is a backward arrow.
Crossovers are a problem. It is inevitable that some lines of logic must
cross others, but many crossovers can be eliminated by careful layout
(see Figure 5.2.4.) There is no one method for showing crossovers.
However, it is important that the lines do not intersect. In the example
shown in Figure 5.2.7, the intersection of activities 12–14 and 9–16, illus-
trated in the lower left-hand corner, is not proper because it implies a
logical crossroad that does not exist.
One solution to this problem is to use a pipeline technique (see Figure
5.2.7, upper left-hand corner). The crossover is shown in the same way
that a pipe crossing is shown on piping drawings. Another solution is to
show a broken arrow (see Figure 5.2.7, upper right-hand corner). Any
good crossover technique can be used, but the same technique should
be used consistently so that the network user can become accustomed
to it. A second version of the broken arrow is shown in the lower right-
hand corner of Figure 5.2.7. In this case, the parts of the broken arrow
are in line. On large networks, it may not be practical to maintain the
straight-line relation.
A broken arrow can also connect events on different sheets of a mul-
tisheet network, which is necessary when preparing large networks.
However, coupled with backward arrows, broken arrows can lead to
unintended loops. The best guard against loops is to use traditional
event numbering, j > i. To make this effective, the events should not be
numbered until the network is completed.

Figure 5.2.6 Section of network with backward arrow.

Figure 5.2.7 Arrow crossover techniques.

Figure 5.2.8 Standard versus bus bar

technique.

60
 Network Construction 61

At the project start, a number of activities usually originate. The

result often looks like a traffic jam (Figure 5.2.8). The bus bar tech-
nique can reduce unproductive congestion of a network. Some net-
work purists object to this technique because it violates the rule of
intersecting arrows at points that are not events. Thus, the criterion
the diagrammer uses in deciding whether to use the bus bar tech-
nique should be the clarity of the resulting network. If the technique
is clever but confuses the user, it is a case of “the operation was a suc-
cess but the patient died.”

5.3. Problems with Multisheet Networks

A difficult factor in multisheet networks is where to cut off the arrows
on one sheet to start the next. For ease in drawing and to facilitate draw-
ing use, the network should be interrupted at the point where the least
number of arrows must be cut. Assume that the portion of the network
shown in Figure 5.3.1 is to be on the end of one sheet and the start of the
next. If the network is split as shown in Figure 5.3.2, it is more difficult

Figure 5.3.1 Multisheet network example: continuous

network.
62 The Theory of CPM Planning and Scheduling

Figure 5.3.2 Multisheet example: poorly split network.

for the drafter to draw. More significantly, it does not present a clear pic-
ture to field workers or other users of the diagram. In Figure 5.3.3, the
network is split at the end of the foundation work and prior to steel erec-
tion. Splitting the network at an important event meets the needs of both
the diagrammer and the user of the diagram.
Figure 5.3.3 illustrates another useful technique when connecting
events from sheet to sheet. The connecting event is highlighted with a

Figure 5.3.3 Multisheet network example: network split at logical place.

 Network Construction 63

00 02 04 06 08 10 12 14 16 18 20 22 24 26 28 30

Act#1 Act#2 Act#4 Act#8 Act#10 A

A 1000 1005 1015 1035 1045 1F02
6 6 10 4 6
SS0
Act#5
B 1020 B
SS5 3
Act#6 FF2
C 1025 C
1 FS2
D 1010 Act#3 1030 Act#7 1040 Act#9 D
2 4 5
E E

Act#11 1055 Act#12 1060 Act#13 1070 Act#15 1075

F 1A26 1050 F
8 4 10 4
Act#14 FF0
G 1065 G
12

H PL## H
Page–Line-Offset

00 02 04 06 08 10 12 14 16 18 20 22 24 26 28 30

Figure 5.3.4 Page and grid location method.

hexagon. The number of the sheet to which the event connects is writ-
ten outside the hexagon.
Yet another method used by EnProMaC is to provide coding symbols and
an identification system for use in navigating the pure logic diagram but
not regularly incorporated in the computer model and being strictly for
locating activity or logic lines that traverse one or more sheets of the logic
diagram. In this instance, a typical sheet is marked with grid of lines “A”
through “T” for say, 20 lines, and “01” through “20” across the top of the
sheet. When an activity arrow or logic line is to run off the end of the sheet,
or needs to be connected to another activity at the far end of the same
sheet or different sheet, the arrow is drawn to end at a diamond. The page
and grid location of the diamond is noted and notated to another diamond
where the arrow or line continues on a separate sheet, and the page and
grid location of the second diamond is notated to the first diamond as in
Figure 5.3.4 as shown. (Also see Figure 11.6.1c, the P3 for DOS pure logic
graphic format.)

5.4. Summary
This chapter discussed the practical mechanics of network construc-
tion. Primarily, the network layout must be logical and organized. A con-
fused diagram exposes confused planning. The drawing size should be
reasonable, and multiple sheets should be used if necessary.
Activity descriptions should be on horizontal lines. Avoid wide-sweep-
ing lines or random lines. Center significant chains of activities to form
64 The Theory of CPM Planning and Scheduling

a network backbone. Space the arrows so that additions may be made.

Crossovers of logic lines can take a number of forms, but the form used
should be consistent.
i-j event numbers are abbreviated activity designations and must be
unique for each activity. The careful assignment of event numbers makes
the network easier to use and avoids unintended logical loops. Random
is used here in its literal sense: “without direction, rule, or method.”
 Chapter

6
The Durations of the
Logic Network

The preparation of the arrow diagram furnishes a number of advan-

tages, including a

■ disciplined method of preparing a plan,

■ method of considering the project in detail, and
■ graphic record of the plan, which can be useful in exchanging opin-
ions and constructive criticism about the plan.

One thing that the arrow diagram lacks thus far is the dimension of
time. It might be said that the portion of CPM described thus far has
been qualitative but not quantitative.
The logic network may determine the order in which activities must be
performed, but not when. To determine when the earliest time each activ-
ity first may be performed and the latest time when each activity must
be performed requires not only logic but durations. Similarly, setting
minimum and maximum limits upon the duration of an activity will have
an impact upon the level of detail and the definition of specific activities.
But keep in mind the order in which this material is presented—the accu-
racy of the pure logic network is paramount in a CPM analysis.
Thus an inspector, reading a specification limiting durations to 20
days or less, may reject a submission where certain activities have
greater durations. But keep in mind that every specification has the
added provision (either explicitly in writing or implicitly at law) that
such limitations are “subject to the sound discretion of the Engineer.”
An engineer is expected to understand the reason for this limitation and
relax the limitation appropriately. Strict reliance upon the specification

65

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
66 The Theory of CPM Planning and Scheduling

without understanding the rationale of it can be treated in this, or any

other review situation, as “practice outside the engineer’s field of
expertise,” contrary to the obligations of licensure and placing the
engineer at risk for discipline.

6.1. Definition of an Activity

The first step in determining activity duration is made when defining
the activity. In the construction industry, a proper activity is a set of
instructions, given to a competent foreman, who is then expected to
complete such without further supervision or interaction with other
than his/her own subordinates. In the manufacturing or IT industry, an
activity is a process in a black box that once started does not require fur-
ther interaction or supervision. Returning to the construction industry,
an activity is further defined as being under the control of one, and only
one, responsible individual. Thus, notwithstanding a specification indi-
cating separate activities for each trade, if a crew of electricians is stand-
ing by to assist in a massive concrete pour that includes embedded
conduit, the activity POUR CONCRETE includes the efforts of this sep-
arate trade and it would be improper to list such scope of work as a sep-
arate activity. Similarly, the installation of a MSE wall involves some
backfill, often by a separate subcontractor or crew. But this portion of
the backfill is dictated by the foreman of the MSE crew and thus this
scope of work is included in the ERECT MSE WALL activity.
The definition of an activity is also controlled by its predecessors and
successors. Even in PDM, where overlaps between activities are sup-
ported, at some level (even if not stated) each activity may start only
when its predecessors are 100 percent complete and each activity is fin-
ished when any of its successors may start.
As an example, see Figure 6.1.1, which includes an activity “Form/
Rebar/ Pour Walls.” This is preceded in the logic network by “Form/
Rebar/ Pour Footing” and succeeded by “Rig Joists On To Wall.” Presume
that it is intended to prefabricate a portion of the rebar. This subset of
scope does not require the footing to be poured and, therefore, it should
not be included as part of the wall activity. The prefab scope should be
listed as a separate activity with predecessors back to delivery of rebar

Figure 6.1.1 Activity description further defined by predecessors and successors.

 The Durations of the Logic Network 67

and access to site. Even if this activity is not included in the logic net-
work, the duration of the wall activity should not include such scope.
Note how the duration of the activity would then differ from the bid
estimate.
Similarly, once joists are rigged upon the wall, the wall activity is fin-
ished as far as the CPM analysis is concerned. From a cost perspective,
an inspector may wish to reserve some retainage to provide assurance
that the non-structural honeycombing is corrected, but the CPM activ-
ity should still be noted as 100 percent complete for schedule purposes.
It is usually unrealistic to include an additional activity after each pour
to provide for such scope and such scope may be included in the general
punchlist activity or not at all. Nevertheless, the activity duration should
not include time for this contingency. Again, this means that the dura-
tion may differ from the number of days reserved for this activity in the
bid estimate.

6.2. Setting a Minimum and Maximum Duration

Setting a minimum duration is a matter of usage of the software sys-
tems available. In the construction industry, the typical minimum is 1
day. In the maintenance and turnaround industry, the minimum dura-
tion typically can be as low as 15 minutes. In city planning, in which
activity descriptions are fairly broad, weeks can be used. In the manu-
facturing industry, it can range from as high as one shift to as low as
one cycle of a machine. In the IT industry, minimum duration can be as
low as a clock tick. However, full time units are usually used in CPM.
If an activity is expected to take 3 days and 6 hours, the Scheduler will
use 4 days.
What if the logic network provides for four activities of 2 hours each
and by separate subcontractors in 1 day? The accuracy of the pure logic
network is paramount in a CPM analysis, but the tool to be used may
be limited to only one minimum unit; either all activities are measured
in hours or none. We may further note that somebody better be provid-
ing careful supervision if four independent foremen are going to mesh
their work in 1 day. The answer may be to combine all four to one activ-
ity or to improperly show all four as concurrent with a side comment to
the users of the output or provide some other fudge. What is important
is to understand and convey to all users of the schedule is that there is
a problem with network logic at this point.
P3e/c provides a work-around for this problem by use of “Activity
Steps.” Using this feature, each of the four activities may be listed as
being components of a master activity that may be automatically statused
as the components are completed. However, the internal logic between
steps is not recorded or used as part of the calculations performed.
68 The Theory of CPM Planning and Scheduling

Setting a maximum duration is another matter, based upon the usage

by the people using the schedule. The primary reason for setting max-
imum activity duration is to improve the quality and ease of updates to
the schedule during the course of the project. A second reason, often
cited, is to assist the reviewer in verifying the reasonableness of activ-
ity durations. A seasoned superintendent is a genius in the field of esti-
mating the most likely duration for an activity. A myriad of factors
influence the duration specified. It is important for the Scheduler to elicit
these factors and record them along with the number provided to permit
the superintendent to later verify his/her work or to allow less-knowl-
edgeable individuals to understand the basis for the duration.
However, in practice, updates are not handled by the superintendent
but by a less experienced individual. The choice of maximum duration
and frequency of update is, therefore, set by the level of error that can
be tolerated in the update. A seasoned superintendent may look at a two
specified scopes of work and estimate their durations as 15 days and 35
days, respectively. After 2 weeks have passed, the superintendent may
look at the work-in-progress and note remaining durations of 7 days and
18 days, respectively, noting less than anticipated performance in the
first and better than anticipated performance in the second case.
However, it is the most recently hired junior engineer or clerk on the site
who may likely be assigned to collect update data (after taking job
trailer lunch orders). Looking over the 15 day activity, the junior engi-
neer may note that it looks “about” 60 percent complete, but after con-
sulting the last update choose to enter 67 percent complete. Looking over
the 35-day activity, the junior engineer probably will have no opinion at
all and simply record 35 – 10 = 25 days remaining. In a project of 3-years’
duration being updated every 2 weeks does an error of 2 days in the
remaining duration of one activity make any difference? Even if wrong,
it will be corrected in the very next update. The 7-day error is more prob-
lematic and may well lead to the mis-scheduling of a follow-on subcon-
tractor. If the possible error is greater than one update period, then the
project team has an even greater problem. Thus many specifications
base the setting of a maximum duration as not greater than twice the
frequency of updates. However, as will be discussed in chapter 27, the
contractor may well desire to update more frequently than required by
the specification.
Returning to the question of the “sound discretion of the Engineer,”
consider the situation where the duration is based strictly upon a meas-
ured quantity that will also be tracked for payment purposes. Thus if
the contractor estimates 200,000 CY of soil to be moved over a duration
of 70 days based upon an average production rate of 2850 CY/day, and
the engineer is already tracking the actual quantities being moved, it
is probably more accurate and certainly easier for the project team to
 The Durations of the Logic Network 69

base remaining duration as a percent of 70 days rather than artificial

segmentation of this scope of work into “acceptable” activities. However,
the Engineer may require the contractor to check his estimate and
desire to check the contractor’s estimate of this quantity against the
Engineer’s estimate.

6.3. Estimating versus Scheduling Durations

One way to estimate the duration of an activity is to estimate labor
hour requirements for the activity and divide that figure by the
assumed size of the work crew. However, labor hour requirements
are usually not available because almost all construction estimates are
prepared by subtracting the work quantities by the physical cate-
gories. An activity often includes more than one work category, but it
rarely includes all major categories.
Using basic CPM, it is not possible to make an accurate time esti-
mate for an entire project on an off-the-cuff basis. If an estimator is
experienced, however, it is possible to make very accurate time esti-
mates once a project is properly broken down into discrete activities.
The project can be compared to a steer: The meat cannot be consumed
on the hoof, but by breaking the steer into hamburgers, it can be easily
consumed.
There are situations in which it is not practical to forecast a time
requirement, but the estimator makes the best judgment of the proba-
ble time factor. In subgrade work, for instance, unusual situations can
develop or weather conditions might be a big factor. In such a situation,
it is proper to add some contingency time. The more uncertain the con-
ditions, the greater the contingency time that should be included.
Breaking down the overall project into well-defined activities helps to
reduce the contingency time required.
When a unique new structural or architectural system is planned, the
architect-engineer is usually reluctant to place a time estimate on activ-
ities. In this case, a bracket approach is useful. The first tack is to ask
how long the activity might take, starting with a high figure, such as
10 months, and working down. Then start with a low figure and work
up from the minimum time the activity could take. The result is almost
always a reasonable time range in which the activity could be accom-
plished. Within that range, a specific time estimate can then be selected.
While on this topic, it is important that the determination of duration
be prepared independent of the contractor’s bid estimate. The bid esti-
mate, prepared in a short timeframe, may be subject to errors and the
preparation of the CPM is often the first post-award opportunity to
cross-check the estimate. In addition, a seasoned superintendent may
deploy resources in a manner differing from those assumed during the
70 The Theory of CPM Planning and Scheduling

Figure 6.3.1 Use of custom data items and global change to calculate productivity rates.

rushed bid process, using larger or smaller crews or equipment. And, as

previously noted, the scope of an activity for schedule purposes may vary
from that for cost estimation. Such differences will, of course, have an
impact upon the number of working days to perform an activity.
Two points made in the preceding paragraphs should be further dis-
cussed. The duration determined by the superintendent or project man-
ager will be based upon certain assumptions and it is an estimate subject
to some level of risk. The estimate of duration will be based upon the
resources assigned thereto. This, in turn, will be based upon the quantity
of work to be performed. Finally, whether verbalized or not, there will be
a range of durations considered before choosing the most likely dura-
tion. As a Scheduler, it is important to record all of this background infor-
mation and not simply the estimate of duration provided.
The recording of quantities permits the calculation of a rate of produc-
tivity that can be used to verify estimates of duration. Merely by glancing
down the column of productivity rates for similar activities, the Scheduler
(and entire project team) can detect durations that do not seem consistent
and review each accordingly. (Figure 6.3.1)
In Primavera P3 and P3e/c software, use one custom data code for
quantity and a second for productivity rate. A standard activity code may
also be reserved to note the units of quantity recorded. Enter the quan-
tity (numeric) and create a Global Change to calculate the productivity
rate for all non-zero duration activities (equal to quantity divided by orig-
inal duration.)

6.4. CPM versus PERT Durations

Note how the various precursors to modern CPM converge at this point.
PERT requires input of the Optimistic, Most Likely, and Pessimistic esti-
mates of duration between each event. Many modern software systems,
 The Durations of the Logic Network 71

Figure 6.4.1 MicroSoft Project permits the user to provide the earliest dates that the user
hopes to start and finish (or optimistic duration), the latest dates (or pessimistic dura-
tion) anticipated, and the expected (or most likely) dates or durations anticipated.

such as MicroSoft Project (Figure 6.4.1), provide as a default that the dura-
tions will be calculated from the inputs of quantities of work and resources
assigned. SPERT analysis, initially limited to an academic exercise, is now
supported by software such as Monte Carlo and Pertmaster and provides
default values for Optimistic and Pessimistic estimates of duration, sub-
ject to direct entry of these values by the Scheduler.
The PERT methodology of asking for Pessimistic, Optimistic, and
Most Likely estimates of duration has a psychological as well as math-
ematical use. Project personnel first asked “In the near worst circum-
stances (or 95 percent of the time), what is the maximum duration that
this may take?,” “In the near best circumstances (or best 5 percent of
the time), what is the minimum duration that this may take?” and then
“What do we expect is the most likely duration that this will take?,” will
likely give a more accurate and less padded estimate than if asked
straight out, “What duration should we assign to this activity?” Even if
the answers to the first two questions are not recorded and are thrown
out, the more accurate answer to the third question may justify the
effort to ask three questions. On the other hand, if the information is
given, why not record it? Even some basic CPM software programs,
such as Microsoft Project, support recording and actually can use all
three data points. Even if the software used does not support recording
72 The Theory of CPM Planning and Scheduling

all this information, the Scheduler may still record such (in raw notes
or in a code field) for potential use.
Microsoft Project also provides for variations from the original
PERT algorithm, allowing the user to vary the standard weighting from
(O + 4M + P)/6 and to compare the results to networks using only the
Optimistic, Most Likely, or Pessimistic durations. (For reasons deal-
ing with merge bias, or the mathematical issues raised when two or
more paths of activities merge, this comparison can provide only a
rough approximation of the probable optimistic, most likely and pes-
simistic duration for the project as a whole. This is further discussed
in Chapter 38).

6.5. Summary
Activity durations are based upon the project manager’s estimate of
the scope of work, resources to be assigned, and other factors and
assumptions, all which should be recorded by the Scheduler. A proper
activity is a set of instructions, given to a competent foreman, who is
then expected to complete them without further supervision or inter-
action with other than his/her own subordinates. The maximum scope
and duration for an activity should also be based on the ability of junior
personnel to assess partial completion. The duration estimates of the
project manager should be fresh, based upon the resources that the
project manager intends to assign to the activity and should not influ-
enced by the estimates based upon quantity takeoff or other method pre-
pared by an estimating department during the bid process. Asking for
a Pessimistic, Optimistic, and Most Likely estimate of duration may be
of use in obtaining unpadded and more accurate estimates, even if the
extra information is not recorded.
 Chapter

7
What Comes Out . . .

As may be noted from the preceding pages, a good deal of effort is

required to prepare the logic network and make it ready for the software
to do its job. Additional practical detail on acquiring the information for
the logic network is provided in Part III of this text. But what do we
expect the software to accomplish that makes the output more useful
than a bar-chart and merits all this effort? A bar-chart indicates when
the preparer intends or hopes to perform each activity. Thus, the prepa-
ration of a bar-chart calculates two attributes for each activity on the
chart—the scheduled start and scheduled finish dates.
While this information is an improvement over that of a numbered
to-do list, it is still somewhat limited. Other scheduling information that
may be useful for management of a project includes whether and which
of these activities may be capable of starting earlier than scheduled if
additional resources become available or productivity exceeds expecta-
tions. A project manager may wish to know which of these activities are
critical to the completion of the project and which may slip without such
an impact, and to know the latest dates upon which an activity must
start and be completed for the project to be completed by a specified date.
Even if the completion date of the project is not imperiled, the slippage
of one activity may have an impact upon the scheduling of other work,
be this work by the same foreman or by another crew or subcontractor.
Therefore, a project manager may wish to know which activities may slip
without impact to a successor and by how much.
If a project manager is concerned with limited crews or other resources,
he/she may even plan to allow some activities to slip. Some such shifts
could have an impact to project completion. Some may only impact the
start of another activity. And some may merely reduce the ability to
allow other activities preceding the activity to slip. But others can be

73

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
74 The Theory of CPM Planning and Scheduling

shifted without any impact to a successor or reduction of options of a pred-

ecessor. As the project is being planned and scheduled at the outset, a
project manager may wish to know these attributes for each activity.
The mathematics of CPM systems expand this level of information to
indicate when an activity first may start and may finish based upon a
stipulated start date for a project, when an activity must start and
must finish if the project is to be completed by a stipulated date or by
the earliest time possible, the number of days (or other units of time)
between the time when an activity may and must so start or may and
must so finish, and other attributes relating to the timing of perform-
ance of the activity, and those points in time immediately preceding
and following an activity.
The last comment is important because the initial development of
CPM was as a mathematical exercise with a primary emphasis upon
events, or points in time, and only a secondary emphasis upon the activ-
ities between these events. While this is downplayed by some manuals
teaching PDM and software tutorials, the mathematics of these systems
continue to be based upon these concepts.

7.1. Attributes of an Event

One of the advertising campaigns of Primavera Systems was “This is an
Activity Based World.” However, in many applications, precise knowledge
of an activity or what must be done is rather sketchy. This would be
especially true in an R&D (research and development) project, such as
the Polaris Missile System for which PERT was developed in 1958. What
are known are the various events or milestones that mark the path from
project inception to project completion. Moreover, the mathematics
of CPM that require “each activity to be 100 percent complete before
the next activity may start” was initially accomplished by requiring
that each activity (an arrow) double as the logic connecting two events
(or nodes) representing points in time. Whether in CPM or PERT, the ini-
tial calculation was to determine the earliest and latest time that such
events may occur based upon the logic and durations of activities (CPM)
or between events (PERT). These attributes are expressed as TE and TL.
The difference between TE and TL is the attribute known as total float
in the United States, or slack in the UK, and is expressed as TF.

7.2. Attributes of an Activity

An activity, unlike an event, has duration. An activity, unlike an event,
has a distinct start and finish and some period of time between the two.
An event, being a single point in time, either has or has not occurred.
However, an activity, having duration, may yet to have started, have
 What Comes Out . . . 75

started but not yet finished, or have started and finished. Thus there
exist a greater number of attributes affiliated with an activity than
with an event. The mathematics of CPM places these activities between
events to carry both the logic of the order of events and the durations
between such events.

7.3. The Forward Pass—TE , ES, and EF

Each activity will, therefore, have an attribute for the earliest time that
the activity may start, expressed as ES and equal to the TE of the event
from which the activity springs, and for the earliest time the activity may
finish, expressed as EF.

7.4. The Backward Pass—TL, LF, and LS

Each activity will also have an attribute for the latest time that the activ-
ity must finish (if the project is to be complete on time or at the earliest
possible time), expressed as LF and equal to the TL of the event to which
the activity goes, and for the latest time the activity must start,
expressed as LS. Note that the attributes EF and LS are new and are
not necessarily equal to one of the event attributes TE or TL.

7.5. The Backward Pass—TF, FF, and IF

For an activity, the attribute measuring the difference between when an
activity may start and must start, the number of time units that the
activity may slip without impact to timely completion of the project, is
known as total float (or slack) and is expressed as TF. Similarly, the
attribute measuring the difference between when an activity may finish
and must finish is also known as total float (or slack) and is also
expressed as TF. In the original and traditional format of CPM, now
known as arrow diagramming method (ADM) or activity-on-arrow
method (AOA), where the calculation of activity attributes require the
intermediate step of calculation of event attributes (and thus guaran-
teeing that each activity be 100 percent complete before the next starts),
the two attributes are indeed the same since LS − ES = TF = LF − EF.
(The derivation of this equality is demonstrated in the next section.)
The attribute measuring the number of time units that an activity
may slip without impact to another activity that may follow (or succes-
sor activities) is known as free float and is expressed as FF. The attrib-
ute measuring the number of time units that an activity may be
deliberately deferred without reducing the ability to defer any other
activity of the logic network is known as independent float and is
expressed as IF.
76 The Theory of CPM Planning and Scheduling

There are several additional attributes for each activity that are not truly
calculated by the mathematics of CPM but rather by various add-ons.
One of the benefits of CPM over the use of a bar-chart is that it empow-
ers the project superintendent (and perhaps his/her subordinates) to intel-
ligently choose which activities to pursue most vigorously and which may
be allowed to slip to some extent without adverse effect. Thus we speak of
early start and early finish dates, upon which the superintendent may first
expect to be able to perform an activity, and late finish and late start
dates, upon which the superintendent must perform on or before, but we
have not explicitly specified the dates between these two extremes, upon
which the superintendent would like to perform the scope of the activity.
There are possibly three steps from the attributes calculated by the math-
ematics of CPM to those calculated or assigned by other methodologies.
The first is the correction of the assumption of CPM of unlimited
resources for multiple operations. While the resources assigned to any one
activity may be carefully and properly chosen for optimal performance,
pure CPM does not have the means to limit the number of concurrent
operations. The results of the CPM calculation may at any one time
require an unreasonable or unattainable number of craftsmen, pieces of
support equipment or quantity of materials. Thus there may be the need
for an additional calculation to determine the attributes of leveled start
and leveled finish such that the usage of resources does not exceed the
limits set. This additional analysis, if performed, must be reviewed care-
fully. If the analysis calculates leveled starts and finishes between the
early and late starts and finishes calculated by the CPM algorithm, all
is well. If, on the other hand, the leveling calculation pushes completion
beyond the desired end date, the project team must revisit the original
plan to determine if an alternate means exists to perform the project in
the stipulated timeframe with the limited resources provided.
Often, a project manager will be satisfied that the schedule can be lev-
eled, then throw away the printout and run the project using only the
CPM algorithm generated attributes. This is because once the CPM is
subject to leveling or further restriction, it returns to being a bar-chart.
Rather than being a flexible guide to decision making, it can become a
static picture of precisely how the day-to-day detail of the project should
be run, which will quickly become out of date as the real world diverges
from the model. In addition, a project manager will often include ele-
ments of “soft logic,” including the sequencing of crews, formwork, and
other resources into the original logic such that the calculated early
start is, in fact, a leveled start. (In chapter 29 on revisions, we discuss
that this “soft logic” is the first area subject to adjustment if a need to
recover from delay occurs.) This informal means of “leveling” is thus
acceptable if not optimal on most small- to medium-sized projects. On
the other hand, as projects become larger and involve a greater variety
 What Comes Out . . . 77

of resources, a leveled schedule may become useful if not too restrictive

of the initiatives of the project manager and the team.
Going beyond leveling, the degree of flexibility can be further nar-
rowed by a process called smoothing. This algorithm will not lengthen
the project but rather will first look at the possible choices for leveling
the project and then choose the one most likely to result in the fewest
instances of layoff and rehiring or fluctuation in resource usage. Thus
two additional attributes, smoothed start and finish, are assigned.
Finally, a project manager may, with or without the use of leveling and
smoothing calculations, simply incorporate an additional layer of con-
trol as a command decision to stipulate a scheduled start and finish for
each activity, preferably somewhere between the early start and late
finish. This last step may be performed at the start of the project (and
thus almost certain to be subject to later revision) or at the last moment
during the standard 1, 2, or 3 week look-ahead field planning and sched-
uling sessions that have always occurred with or without a CPM.

7.6. Calculating the Attributes of an Event or Activity

A program manager, in setting a series of milestones towards the com-
pletion of a project, and a project manager, in ordering a number of
activity bars (of length equivalent to duration) to create a bar-chart,
both are manually performing the first half of the calculations required
to ascertain the CPM attributes of an event or activity. Starting from
the beginning of the project, each determines what is the first event or
activity and then determines those events and activities that follow. The
process continues until project completion is accomplished. Calculation
of the earliest time that an event may occur or the earliest time that
an activity may start and finish is performed in the same manner,
starting from the beginning and running to the end—a forward pass
through all of the events and activities of the project.

7.7. The Forward Pass—TE , ES, and EF

By definition, the earliest time of the first event TE or activity start ES
is traditionally set as zero. (Primavera Software Systems sets such time
as one.) The early finish of any activity EF is defined as equal to ES plus
the duration of the activity. Thus far, the calculation is simple. However,
other than the first event or activity, each other event or activity is pre-
ceded by one or more other events or activities. So the question becomes,
“what is the TE or ES of the next event or activity?”
If for example, an event is preceded by three other events, then that
event’s TE cannot occur until all three are completed and specifically
until the latest TE of the three predecessor events has occurred. If, for
78 The Theory of CPM Planning and Scheduling

a a
b
b
c
c
d
d

Figure 7.7.1 The forward pass – TE, ES & EF.

example, an activity is preceded by three other activities, then that

activity’s ES cannot start until all three are completed and specifically
until the latest EF of the three predecessors. (Figure 7.7.1.)
Thus the attributes of TE, ES, and EF may be defined by the equations:

TE0 = 0 ES0 = 0
TE = Latest (TE PRED + DPRED) EF = ES + D
ES = LEFPRED

7.8. The Backward Pass—TL, LF, and LS

The key improvement of a CPM over the bar-chart methodology is to cal-
culate and report not only the earliest or scheduled dates on which an
event may occur or an activity may start and finish, but also the latest
dates. There are some planning and scheduling problems that call for this
as the primary information required, such as planning for a wedding or
a company or industry conference. (What is the last date we can make
changes to the invitation or brochure?) However, it is the combination of
early dates and late dates that provides true empowerment to the proj-
ect manager. Knowing that he/she may start five activities but only has
the resources to perform four at any time is now replaced by knowing the
impact to the project if any one or more of the activities are deferred.
The means of determining the latest dates for each event or activity
is to go from the end of the project working backward—a backward pass
through all of the events and activities to the start of the project. By def-
inition, the latest time of the last event TL or activity finish LF is tra-
ditionally set as equal to the earliest time of the last event TE or activity
finish EF. This is because, all other things being equal, it is more eco-
nomical to complete a project as early as reasonably possible. Later
additions and expansions to the original CPM algorithm permit the
user to set a stipulated or mandated FNET finish-not-later-than date,
but the original system did not have this option. Next, the late start of
any activity LS is defined as equal to LF minus the duration of the
activity. And, working backwards, this event or activity is preceded by
another until we reach the start of the logic network. So the question
becomes, “what is the TL or LF of the preceding event or activity?”
 What Comes Out . . . 79

a a
b b
c c
d d

Figure 7.8.1 The backward pass – TL, LF & LS.

If, for example, an event is succeeded by three other events, then that
event’s TL must occur prior to the earliest TL of the three successor
events. And if, for example, an activity is succeeded by three other activ-
ities, then that activity’s LF must occur before the earliest LS of the
three successors. (Figure 7.8.1.)
Thus the attributes of TL, LF, and LS can be defined by the equations:

TL end = 0 LFend = EFend

TL = Earliest (TL SUCC − DSUCC) LS = LF − D
LF = ELSSUCC

7.9. The Backward Pass—TF, FF, and IF

At this point, arithmetic calculations can determine the number of time
units between the time when an event may start and when it must
finish (if the project is to be completed by the earliest possible time),
between when each activity may start and must start, and between
when each activity may finish and must finish. The attribute of TF may
be defined by the equations:

TF = TL − TE TF = LF − LS
TF = EF − ES

In the case of traditional CPM, these three equalities, all equal to TF,
are thus all equal to each other. The activity attribute of free float is
defined as the difference between the earliest of the early starts of all
successors to an activity and the calculated early finish of that activity.
(Figure 7.9.1)

from other predecessors

FF = EESSUCC – EF
a
b
c
FF
d

Figure 7.9.1 Free Float.

80 The Theory of CPM Planning and Scheduling

a IF = EESSUCC – LEFPRED – D
b
c
d
e IF
f

Figure 7.9.2 Independent Float.

The activity attribute of independent float is defined as the difference

between the earliest of the early starts of all successors to an activity
and the latest of the early finishes of all predecessors of that activity.
(Figure 7.9.2.)
Calculation of the attributes of leveled start and finish (as well as
smoothed start and finish) is not so easily performed and cannot be solved
by a single equation. This will be more fully discussed in Chapter 37.

7.10. Summary
By merely recording the relationships and probable durations between
events, or the relationships and estimated durations of activities, the
mathematics of PERT and CPM will calculate a number of attributes
about the event or activity that would otherwise not be readily deter-
mined. These include, for an event, not only the earliest time that the
event is expected to occur, but the latest, and the difference between
these two times or dates. These include for an activity, not only the
earliest time that the activity may be expected to start and finish, but
the latest time that the activity must finish if the project is to be com-
pleted in the earliest possible time. Also computed are the attributes
of total float, free float, and independent float, concepts that are dis-
cussed in the Chapter 8.
 Chapter

8
Cranking the Engine

The first CPM and PERT logic networks calculated schedules by the new
invention, the electronic computer. Manual simulation of the steps taken
by the computer program, to prove that the model worked and to allow
individuals to calculate smaller schedules, involved the use of a matrix.
This was a natural step because mathematicians often used a graphic
grid to solve problems. Figure 8.0.1 shows a portion of the logic network
for the John Doe Project that will be developed in Chapter 18 with
assigned time estimates.

8.1. Manual and Computer Solutions for PERT and ADM—

The Matrix Method
Figure 8.1.1 (a) through (e) shows the development of the matrix (grid)
for this small portion. The i-nodes are listed on the X axis down the side
of the matrix, while the j-nodes are listed across the top of the matrix.
Two columns are reserved for recording the earliest time the node can
occur (TEi) and calculation thereof; two rows are reserved for recording
the latest time by which the node must occur (TLj) and calculation
thereof. A tabular listing of the CPM activities (i – j,) duration and cal-
culated attributes of ES, EF, LS, LF, and TF are at the right and this
list may be longer or shorter than the number of rows assigned to the
matrix. A true PERT system (unlike the NASA CPM/PERT) will report
only the TE and TL for each event.
We previously noted the need in the original computer programs for
sequential numbering of the nodes such that 1–2–3 or 3–5–12 are accept-
able but 1–3–2 or 3–12–5 are not acceptable. This was explained as a
means of reducing the amount of computer memory required to solve the
CPM algorithm and is the basis for such restrictions and extensions

81

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
82 The Theory of CPM Planning and Scheduling

Figure 8.0.1 Activity time assignment: site preparation.

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0
1
2
3
9
10
11
12
T Lj
MIN

Figure 8.1.1 (a)

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 —
1 — —
2 — — —
3 — — — —
9 — — — — —
10 — — — — — —
11 — — — — — — —
12 — — — — — — — —
T Lj
MIN

Figure 8.1.1 (b)

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 1 2 3 4 5 6 7
1 — — 8 9 10 11 12 13
2 — — — 14 15 16 17 18
3 — — — — 19 20 21 22
9 — — — — — 23 24 25 28 = 8 ⫻ 7/2 = n(n − 1)/2
10 — — — — — — 26 27
11 — — — — — — — 28
12 — — — — — — — —
T Lj
MIN

Figure 8.1.1 (c)

 Cranking the Engine 83

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0–1 3
1 — —
2 — — —
3 — — — —
9 — — — — —
10 — — — — — —
11 — — — — — — —
12 — — — — — — — —
T Lj
MIN

Figure 8.1.1 (d)

(such as skip numbering) found in older specifications. The reason this

numbering restriction reduces use of memory is that it drastically reduces
the number of cells of the matrix. Thus for our example of a nine activ-
ity network using eight nodes, the number of cells possibly required can
be reduced from 56 to 28, or from n*(n–1) to n*(n–1)/2. (If calculation of
attributes for the milestones or events of the network, such as “activity”
3–3 are desired, an additional n cells will be required.) (Figure 8.1.1(c).)
Once we have prepared the matrix, we may do the “take-off” from the
logic network. We copy activity i – j of 0 – 1 and duration of 3 to the tab-
ular listing at the right, then transfer the duration of 3 to the intersec-
tion on the matrix of i = 0 and j = 1. (Figure 8.1.1(d).)
Continuing, we duplicate this effort for the other activities of the logic
network. A similar effort may be used to copy only the logic and dura-
tions between PERT event nodes to the matrix. (Figure 8.1.1(e).)
At this point we begin the solution. By definition, TE0, the earliest time
that the first event node may occur, is set to zero. Next, to calculate the
TEi for node 1, we go to the column for j = 1 and choose the maximum of
duration plus previously calculated TEi for each row where a duration
is entered, in this case 3. For a CPM, the TEi is copied to all activities
having that i node.
The process is repeated until the forward pass is completed, as shown
in Figures 8.1.2 (a) through (h.)

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0–1 3
1 — — 2 1–2 2
2 — — — 2 2–3 2
3 — — — — 10 1 6 3–9 10
9 — — — — — 5 3 – 10 1
10 — — — — — — 5 3 – 12 6
11 — — — — — — — 3 9 – 11 5
12 — — — — — — — — 10 – 11 5
T Lj 11 – 12 3
MIN

Figure 8.1.1 (e)

84 The Theory of CPM Planning and Scheduling

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 1–2 2
2 — — — 2 2–3 2
3 — — — — 10 1 6 3–9 10
9 — — — — — 5 3 – 10 1
10 — — — — — — 5 3 – 12 6
11 — — — — — — — 3 9 – 11 5
12 — — — — — — — — 10 – 11 5
T Lj 11 – 12 3
MIN T Ei

Figure 8.1.2 (a) Determine MAX (TEi + Dij) = MAX (earliest time for node i plus duration
from i to j).

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 3 3+0 1–2 2 3
2 — — — 2 2–3 2
3 — — — — 10 1 6 3–9 10
9 — — — — — 5 3 – 10 1
10 — — — — — — 5 3 – 12 6
11 — — — — — — — 3 9 – 11 5
12 — — — — — — — — 10 – 11 5
T Lj 11 – 12 3
MIN T Ei

Figure 8.1.2 (b) Determine MAX (TEi + Dij) = MAX (earliest time for node i plus duration
from i to j).

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 3 1–2 2 3
2 — — — 2 5 2+3 2–3 2 5
3 — — — — 10 1 6 3–9 10
9 — — — — — 5 3 – 10 1
10 — — — — — — 5 3 – 12 6
11 — — — — — — — 3 9 – 11 5
12 — — — — — — — — 10 – 11 5
T Lj 11 – 12 3
MIN T Ei

Figure 8.1.2 (c) Determine MAX (TEi + Dij) = Earliest time for node i plus duration from
i to j.

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 3 1–2 2 3
2 — — — 2 5 2–3 2 5
3 — — — — 10 1 6 7 2+5 3–9 10 7
9 — — — — — 5 3 – 10 1 7
10 — — — — — — 5 3 – 12 6 7
11 — — — — — — — 3 9 – 11 5
12 — — — — — — — — 10 – 11 5
T Lj 11 – 12 3
MIN T Ei

Figure 8.1.2 (d) Determine MAX (TEi + Dij) = Earliest time for node i plus duration from
i to j.
 Cranking the Engine 85

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 3 1–2 2 3
2 — — — 2 5 2–3 2 5
3 — — — — 10 1 6 7 3–9 10 7
9 — — — — — 5 17 10 + 7 3 – 10 1 7
10 — — — — — — 5 3 – 12 6 7
11 — — — — — — — 3 9 – 11 5 17
12 — — — — — — — — 10 – 11 5
T Lj 11 – 12 3
MIN T Ei

Figure 8.1.2 (e) Determine MAX (TEi + Dij) = Earliest time for node i plus duration from
i to j.

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0=0 0–1 3 0
1 — — 2 3 1–2 2 3
2 — — — 2 5 2–3 2 5
3 — — — — 10 1 6 7 3–9 10 7
9 — — — — — 5 17 3 – 10 1 7
10 — — — — — — 5 8 1+7 3 – 12 6 7
11 — — — — — — — 3 9 – 11 5 17
12 — — — — — — — — 10 – 11 5 8
T Lj 11 – 12 3
MIN T Ei

Figure 8.1.2 (f) Determine MAX (TEi + Dij) = Earliest time for node i plus duration from
i to j.

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 3 1–2 2 3
2 — — — 2 5 2–3 2 5
3 — — — — 10 1 6 7 3–9 10 7
9 — — — — — 5 17 3 – 10 1 7
10 — — — — — — 5 8 3 – 12 6 7
11 — — — — — — — 3 22 5 + 17 9 – 11 5 17
5+8
12 — — — — — — — — 10 – 11 5 8
T Lj 11 – 12 3 22
MIN T Ei

Figure 8.1.2 (g) Determine MAX (TEi + Dij) = Earliest time for node i plus duration from
i to j.

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 3 1–2 2 3
2 — — — 2 5 2–3 2 5
3 — — — — 10 1 6 7 3–9 10 7
9 — — — — — 5 17 3 – 10 1 7
10 — — — — — — 5 8 3 – 12 6 7 25
11 — — — — — — — 3 22 9 – 11 5 17
12 — — — — — — — — 25 6+7 10 – 11 5 8
3 + 22
T Lj 25 11 – 12 3 22 25
MIN T Ei T Lj

Figure 8.1.2 (h) Determine MAX (TEi + Dij) = Earliest time for node i plus duration from
i to j.
86 The Theory of CPM Planning and Scheduling

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 3 1–2 2 3
2 — — — 2 5 2–3 2 5
3 — — — — 10 1 6 7 3–9 10 7
9 — — — — — 5 17 3 – 10 1 7
10 — — — — — — 5 8 3 – 12 6 7 25
11 — — — — — — — 3 22 9 – 11 5 17 22
12 — — — — — — — — 25 10 – 11 5 8 22
T Lj 22 25 11 – 12 3 22 25
MIN 25 – 3 T Ei T Ei T Lj

Figure 8.1.3 (a) Determine MIN (TLj – Dij) = Latest time for node j minus duration from
i to j.

At this point, we copy the TEi for the last node to the TLj for the last
node since it is assumed that the project should finish as soon as pos-
sible. For the CPM tabulation, this late time is copied to the late finish
attribute of those activities having the same j node.
Now we begin the backward pass. To calculate the TLj for node 11, we
look to the row for i = 11, locate the durations listed on that row, and
subtract them from the TLj previously calculated for that column. Thus,
25 – 3 = 22. This process is repeated until we get back to j = 0 as shown
in Figures 8.1.3 (a) through (g).
Finally, if we are preparing a CPM, we calculate the secondary attrib-
utes of EF = ES + D, LS = LF – D and TF = LF – EF = LS – ES. This
step is not required if we are working with an event-based system, such
as PERT. (Figure 8.1.4)
Although the matrix served its purpose in early work, there is an
easier and more direct solution. When James Kelley, a member of the
original CPM group, was asked why his group had not immediately
seen an easier solution, he explained it this way: If both the mathe-
matician and the engineer are confronted with the problem of how to
move a pan of water from the kitchen table to the stove, both will solve
it by lifting the pan from the table directly to the stove. The next day,

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 3 1–2 2 3
2 — — — 2 5 2–3 2 5
3 — — — — 10 1 6 7 3–9 10 7
9 — — — — — 5 17 3 – 10 1 7 17
10 — — — — — — 5 8 3 – 12 6 7 25
11 — — — — — — — 3 22 9 – 11 5 17 22
12 — — — — — — — — 25 10 – 11 5 8 22
T Lj 17 22 25 11 – 12 3 22 25
MIN 22 – 5 T Ei T Ei T Lj

Figure 8.1.3 (b) Determine MIN (TLj – Dij) = Latest time for node j minus duration from
i to j.
 Cranking the Engine 87

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 3 1–2 2 3
2 — — — 2 5 2–3 2 5
3 — — — — 10 1 6 7 3–9 10 7 17
9 — — — — — 5 17 3 – 10 1 7 17
10 — — — — — — 5 8 3 – 12 6 7 25
11 — — — — — — — 3 22 9 – 11 5 17 22
12 — — — — — — — — 25 10 – 11 5 8 22
T Lj 17 17 22 25 11 – 12 3 22 25
MIN 22 – 5 T Ei T Ei T Lj

Figure 8.1.3 (c) Determine MIN (TLj – Dij) = Latest time for node j minus duration from
i to j.

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 3 1–2 2 3
2 — — — 2 5 2–3 2 5 7
3 — — — — 10 1 6 7 3–9 10 7 17
9 — — — — — 5 17 3 – 10 1 7 17
10 — — — — — — 5 8 3 – 12 6 7 25
11 — — — — — — — 3 22 9 – 11 5 17 22
12 — — — — — — — — 25 10 – 11 5 8 22
T Lj 7 17 17 22 25 11 – 12 3 22 25
MIN 25 – 6 T Ei T Ei T Lj
17 – 1
17 – 10

Figure 8.1.3 (d) Determine MIN (TLj – Dij) = Latest time for node j minus duration from
i to j.

i\ j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0
1 — — 2 3 1–2 2 3 5
2 — — — 2 5 2–3 2 5 7
3 — — — — 10 1 6 7 3–9 10 7 17
9 — — — — — 5 17 3 – 10 1 7 17
10 — — — — — — 5 8 3 – 12 6 7 25
11 — — — — — — — 3 22 9 – 11 5 17 22
12 — — — — — — — — 25 10 – 11 5 8 22
T Lj 5 7 17 17 22 25 11 – 12 3 22 25
MIN 7–2 T Ei T Ei T Lj

Figure 8.1.3 (e) Determine MIN (TLj – Dij) = Latest time for node j minus duration from
i to j.

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0 3
1 — — 2 3 1–2 2 3 5
2 — — — 2 5 2–3 2 5 7
3 — — — — 10 1 6 7 3–9 10 7 17
9 — — — — — 5 17 3 – 10 1 7 17
10 — — — — — — 5 8 3 – 12 6 7 25
11 — — — — — — — 3 22 9 – 11 5 17 22
12 — — — — — — — — 25 10 – 11 5 8 22
T Lj 3 5 7 17 17 22 25 11 – 12 3 22 25
MIN 5–2 T Ei T Ei T Lj

Figure 8.1.3 (f) Determine MIN (TLj – Dij) = Latest time for node j minus duration from
i to j.
88 The Theory of CPM Planning and Scheduling

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0 3
1 — — 2 3 1–2 2 3 5
2 — — — 2 5 2–3 2 5 7
3 — — — — 10 1 6 7 3–9 10 7 17
9 — — — — — 5 17 3 – 10 1 7 17
10 — — — — — — 5 8 3 – 12 6 7 25
11 — — — — — — — 3 22 9 – 11 5 17 22
12 — — — — — — — — 25 10 – 11 5 8 22
T Lj 0 3 5 7 17 17 22 25 11 – 12 3 22 25
MIN 3–3 T Ei T Ei

Figure 8.1.3 (g) Determine MIN (TLj – Dij) = Latest time for node j minus duration from
i to j.

the engineer, on finding the pan of water on the floor, will again move
it directly to the stove. Under the same circumstances, the mathemati-
cian would first move the pan from the floor to the table and then from
the table to the stove. Why? Because the mathematician has already
solved the table-to-stove problem.
Similarly, having used the matrix approach before, it was natural for
the CPM mathematicians to use it in solving the network manually.

8.2. Manual and Computer Solution for PERT and ADM—

The Intuitive Method

Intuitive manual computation

The manual CPM computation now in use was probably developed con-
currently by several persons. There is a famous phrase used by almost
all college professors at some time or other in explaining a mathemat-
ical solution: “Intuitively we can understand this next step . . .” In this
case, however, the computation is based upon common sense and is
intuitively obvious. Since the matrix was still in use by the CPM origi-
nating team in late 1960, the intuitive solution probably originated in
1961. The mental block that probably deterred the mathematicians

i\j 0 1 2 3 9 10 11 12 T Ei MAX i−j dur ES EF LS LF TF

0 — 3 0 T E0 = 0 0–1 3 0 3 0 3 0
1 — — 2 3 1–2 2 3 5 3 5 0
2 — — — 2 5 2–3 2 5 7 5 7 0
3 — — — — 10 1 6 7 3–9 10 7 17 7 17 0
9 — — — — — 5 17 3 – 10 1 7 8 16 17 1
10 — — — — — — 5 8 3 – 12 6 7 13 19 25 12
11 — — — — — — — 3 22 9 – 11 5 17 22 17 22 0
12 — — — — — — — — 25 10 – 11 5 8 13 17 22 9
T Lj 0 3 5 7 17 17 22 25 11 – 12 3 22 25 22 25 0
MIN T Ei T Ei T Lj

Figure 8.1.4 Determine EF = ES + D; LS = LF – D; TF = LF – EF = LS – ES.

 Cranking the Engine 89

from arriving at it is that the intuitive solution is logical rather than

mathematical.

Early event times TE

Look at the first activity in Figure 8.0.1,

If the project is started at event 0, what is the earliest time for reach-
ing event 1? According to estimate, 3 days would finish clearing the
site. The early time TE for event 1 is then 3 days. How early could event
2 be reached? The answer is, of course, 3 + 2, or at the end of the fifth
project day. To keep track of those results, show them in a box just over
the event:

The earliest schedule for reaching event 3 is the sum of the times
required to accomplish the first three activities, 3 + 2 + 2, or 7. Now look
at event 9. Do not go back to the originating event to determine the TE
(early event time) for this event. Add the duration to the TE for event 3,
and the result is a TE of 17 for event 9. To go on to event 11, two logic
paths lead into this event:

The earliest time for reaching event 11 is along path 3–9–11. This is TE
for event 9 plus the duration, or 17 + 5, or 22. Note this without enclos-
ing it in a box, and then investigate the path through events 3–10–11:
90 The Theory of CPM Planning and Scheduling

For event 10, the TE is 7 + 1, or 8. For event 11 along path 3–10–11, the
early event time is 8 + 5, or 13. The activities along path 0–1–2–3–10–11
can be accomplished in as early a time as 13 days; along path
0–1–2–3–9–11 they would take 22 days. What is TE for event 11? The ear-
liest time for reaching event 11 is the end of the twenty-second project day.
Accordingly, discard the 13-day solution and select the longer 22-day
answer as TE for event 11:

TE is always the larger value when there is a choice between two or

more values.
A caution is in order here. Remember that the event numbers have
no significance other than identification. Unfortunately, it is easy to
add them in accidentally as durations or to use the event number rather
than the TE. This is particularly the case with one- or two-digit event
numbers. To avoid the error, circle the event numbers, use three-digit
event numbers, or do both.
Figure 8.2.1 is the entire site preparation network with times assigned
and early event times noted. The TE at event 12 is the choice of the time

Figure 8.2.1 Early event times: site preparation.

 Cranking the Engine 91

along path 11–12 (22 + 3 = 25) or along path 3–12 (7 + 6 = 13). The TE
at event 12 is the longer time, or 25. The early event time at event 13
along this lower path is 25 + 5 = 30.
Now observe the two upper paths. The path through events 3–4–5–8
totals 25 days. That, added to the TE at event 3, gives an early time
along the path to event 8 of 7 + 25, or 32. Along the path through
events 3–6–7–8, the activities total 24 days. This 24 + 7 is 31 days,
which is less than 32. Thus, the TE at event 8 is 32. The early time to
event 13 along the upper path is 34 days. Since this is larger than 30
days, the TE for this network is 34.
The result is 34 days, but what is the significance? Based on our log-
ical sequence and time estimates, the shortest time in which this work
could be completed is 34 working days, or about 7 weeks.

Late Event Time TL

The late event time TL for an event is defined as the latest time at
which an event can be reached without delaying the computed project
duration. Keep in mind that “late” in this context is late in terms of this
computed completion time rather than a desired or prescribed comple-
tion time. To determine late event times, work backward through the
network. From Figure 8.2.1, the final event 13 has two activities (8–13
and 12–13) leading into it:

By definition, the late event time at event 13 is 34 days, since the late
event time for the terminal event equals the early event time for that
event. If event 13 is to be reached by time 34, event 8 must start no later
than 34 less the duration of activity 8 – 13 (34 – 2). Thus, the late event
time for event 8 is 32. The late event time for event 12 is 34 – 5, or 29.
In showing the late event times TL, on the diagram, put them in cir-
cles to differentiate them from the TE values. Figure 8.2.2 shows the late
event times for this network. In determining To values, there is a choice
between values when two or more arrow tails converge. On Figure 8.2.2
that occurs only at event 3, where the tails for five arrows converge.
Figure 8.2.3 is an enlargement of the network at event 3.
From Table 8.2.1, the path backward from event 4 results in the
“earlier” late event time at event 3. TL is always the earlier value when-
ever there is a convergence of two or more arrow tails. Accordingly, TL
92 The Theory of CPM Planning and Scheduling

Figure 8.2.2 Late event times: site preparation.

Figure 8.2.3 Network at event 3.

 Cranking the Engine 93

TABLE 8.2.1 Late Event Times at Event 3

Late event time

along this path
Activity Late event time Duration, days from event 3

3–4 22 15 7
3–6 12 4 8
3–9 21 10 11
3–10 21 1 20
3–12 29 6 23

at event 3 is time 7. As a check, the late event time for the originating
event should always be zero.
Computed event times, both early and late, is fundamental informa-
tion. Nonetheless, network events are not very descriptive. For instance,
how would you describe event 3 in Figure 8.2.3? You would probably term
it “completion of grading.” But how would you indicate that it marks the
logical starting point for five other activities? Certain key events, or mile-
stones, are easily identified and are of interest. Among them are com-
plete foundations, start steel erection, start studs, complete drywall, and
start piping.
Because construction is work-oriented, activity descriptions better
define the CPM plan. Accordingly, activity time information is the most
useful format.

8.3. Activity Start and Finish Times

The source of activity start and finish times is event time calculation.
Look at the typical activity:

i j
i j
Activity
i j
D

Each activity must be bound by two events. The earliest time that an
activity can start is when the TE for its starting (or i) event has been
reached. That is,

i
Early start = ES = TE (event i) =
94 The Theory of CPM Planning and Scheduling

TABLE 8.3.1 Activity Information First 9 Activities

Activity Duration, days Description

0–1 3 Clear site

1–2 2 Survey and layout
2–3 2 Rough grade
3–4 15 Drill well
3–6 4 Water tank foundations
3–9 10 Excavate sewer
3–10 1 Excavate electrical manholes
3–12 6 Pole line
4–5 2 Well pump

If the early start (ES) is known, the earliest time the activity can be
completed is the start time plus the job duration (D):

Early finish = EF = ES + duration = ES + D

After determining the early time for an activity, the late time is the
TL for the finishing (or j) event; that is,

j
Late finish = LF = TL (event j) =

After late finish, the late start is obviously

Late start = LS = LF – D

Certain information about activities can be summarized before any

calculations are made. For instance, from Figure 8.2.1, the first nine
activities offer the information shown in Table 8.3.1. After the event
times are computed, the additional information shown in Table 8.3.2

TABLE 8.3.2 Event Time Calculations First 9 Activities

Duration,
Activity days Description ES LF

0–1 3 Clear site 0 3

1–2 2 Survey and layout 3 5
2–3 2 Rough grade 5 7
3–4 15 Drill well 7 22
3–6 4 Water tank foundations 7 12
3–9 10 Excavate sewer 7 21
3–10 1 Excavate electrical manholes 7 21
3–12 6 Pole line 7 29
4–5 2 Well pump 22 24
 Cranking the Engine 95

TABLE 8.3.3 Activity Time Calculations First 9 Activities

Duration,
Activity days Description ES EF LS LF

0–1 3 Clear site 0 3 0 3

1–2 2 Survey and layout 3 5 3 5
2–3 2 Rough grade 5 7 5 7
3–4 15 Drill well 7 22 7 22
3–6 4 Water tank foundations 7 11 8 12
3–9 10 Excavate sewer 7 17 11 21
3–10 1 Excavate electrical manholes 7 8 20 21
3–12 6 Pole line 7 13 23 29
4–5 2 Well pump 22 24 22 24

from Figure 8.2.2 can be listed. Adding duration to the ES column and
subtracting it from the LF gives what is shown in Table 8.3.3.

8.4. Critical Activities

The early CPM team referred to the critical path as the “main chain.”
The term was dropped in favor of “critical path,” which was used by the
early PERT group. The critical path determines the length of the proj-
ect. It is the longest part into the last event, since it establishes the latest
TE for the last event. Accordingly, the longest chain or path of activi-
ties through the network is the critical path.
The critical path is not always obvious. Look at the network for the
interior work for the John Doe plant (Figure 18.5.4.) You might guess
at the critical path based upon experience, but without a project time
estimate for each activity, you cannot identify it. Figure 8.4.1 is a plot
of activity information on a time scale. Note that the activities 0–1, 1–2,
2–3, 3–4, and 4–5 show a solid connection; they are on the path of crit-
ical events (0–1–2–3–4–5, etc.). Look at the activity times for activity
4–5. The ES is 22, and the LF is 24. The time span between them is
24–22, or 2. Since the time span available equals the duration for activ-
ity 4–5, the activity must start on its ES and finish on its EF if the proj-
ect is to finish by time 34. Note that for these critical activities, early
start equals late start and early finish equals late finish.
In Figure 8.2.2, the critical path goes through events 0–1–2–3–4–5–8–13.
Three conditions that each critical activity must meet are:

1. The early and late event times at the activity start must be equal:

i = i
96 The Theory of CPM Planning and Scheduling

Figure 8.4.1 Plot of activity information on a time scale.

2. The early and late event times at the activity completion must be
equal:

j = j

3. The difference between the ES and LF must equal the duration.

The first two conditions are easy to recognize when the network is
manually computed with TE and TL on the diagram. People often forget
to test for the third rule, however. Add an activity 3–5 to the network
and call it deliver pipe. The delivery cannot start until the site is rough-
graded (event 3), and it is needed before piping installation starts
(event 5). If the delivery takes a week (duration = 5),

Deliver pipe

7 5 24

7 22 22 24
Drill well Well pump
3 4 5
15 2

Activity 3–5 meets the first two conditions, but 24 – 7, or 17, is greater
than an activity duration of 5; accordingly, activity 3–5 is not critical
even though it spans two critical events.
Note that there can be any number of critical paths through the net-
work. One path can spread out into a number of paths, and a number
 Cranking the Engine 97

of critical paths can converge into one. However, the critical path(s)
must be a continuous chain of activities; it cannot be intermittent. Also,
there must be at least one critical path from the first to the last event
of the project.

8.5. Total Float

In preparing the CPM diagram for a channel improvement project,
Corps of Engineers planners were certain that the critical path would
be through the pile-driving activities, because pile driving had always
been critical in the past. However, the Corps had reckoned without its
own foresight.
Based on past experience, the Corps construction group had devised
a scheme that enabled them to utilize two pile-driving rigs instead of
one in the limited space available. That cut pile driving off the critical
path. It was replaced by a land acquisition handled by the Corps real
estate group. Their time estimate also was based on experience. In this
case, the diagram served as a communication medium to advise all cog-
nizant Corps groups of new planning factors. If activity 3–5, deliver
pipe, is not critical, what differentiates it from a critical activity? Since
it has an available working time span of 17 (24 – 7) and a duration of
5, there is a latitude in scheduling it equal to 17 – 5, or 12. We call this
characteristic float:

Float = F = (LF – ES) – D

Since EF = ES + D,

Float = (LF – ES) – D = LF – (ES + D)

= LF – EF

Also, since (LF = LS + D) and (EF = ES + D),

Float = LF – EF = (LS + D) – (ES + D)

= LS – ES

Getting away from formulas, it is reasonable for the difference

between the early and late starts to equal the scheduling flexibility, or
float. Also, the difference between the late and early finishes furnishes
the same values.
In the network shown in Figure 8.2.2, the total float for all activities, by
using each of the previously mentioned formulas, is shown in Table 8.5.1.
Case 1, shown in Figure 8.5.1, is a time scale plot of activities 3–9, 9–11,
11–12, and 12–13. The total float for each of the activities is 4. Does this
mean that each of the activities has 4 days of float to use? The answer
98 The Theory of CPM Planning and Scheduling

TABLE 8.5.1 Float Calculations

Formula: F = LF – ES – D

Activity LF –ES –Duration = Float

0–1 3 0 3 0
1–2 5 3 2 0
2–3 7 5 2 0
3–4 22 7 15 0
3–6 12 7 4 1

Activity LF –ES –Duration = Float

3–9 21 7 10 4
3–10 21 7 1 13
3–12 29 7 6 16
4–5 24 22 2 0

Formula: F = LF – EF

Activity LF –EF = Float

5–8 32 32 0
6–7 22 21 1
7–8 32 31 1
8–13 34 34 0

Formula: F = LS – ES

Activity LS (LF – D) –ES = Float

9–11 21 17 4
10–11 21 8 13
11–12 26 22 4
12–13 29 25 4

is a qualified yes. If none of the prior activities in this same chain has
used the float, the answer is yes (see Table 8.5.2).
In case 2, shown in Figure 8.5.1, assume that activity 3–9 used the 4
days of float. That is, it started at time 11 instead of the ES of 7. The
result is a solid link of activities following 3–9. When total float is used
up by any one activity or a series of activities, all succeeding activities
become critical.
Case 3, shown in Figure 8.5.1, illustrates the use of total float by dif-
ferent activities in the chain. Activity 3–9 starts two days after its early
start, which reduces the float to two days. Activity 11–12 delays its start
until the late start, and no float remains. Look at the float picture in the
broader view. The TE for event 3 is 7; the TL for event 13 is 34. The dif-
ference, or 27 days, is the time span within which the four activities must
be accomplished. Adding the durations of these four activities results in
what is shown in Table 8.5.3.
 Cranking the Engine 99

Figure 8.5.1 Time scale plot of activities.

TABLE 8.5.2 Total Float Path

Activity ES EF LS LF Float, days

3–9 7 17 11 21 4
9–11 17 22 21 26 4
11–12 22 25 26 29 4
12–13 25 30 29 34 4

TABLE 8.5.3 Total Float

Path Duration

Activity Duration, days

3–9 10
9–11 5
11–12 3
12–13 5
Total 23
100 The Theory of CPM Planning and Scheduling

The available time span from event 3 to event 13 (27 days) less the
total time for the activities in this chain (23 days) is 4 days float. This
is another illustration of the shared aspect of float.

8.6. Free Float

The originators of the critical path method defined a variety of floats,
including total float, free float, and independent float. The measure of
float previously described is known as “total float.” It is both the most
widely used version and the most practical one. Of the three types orig-
inally defined, only two appeared to have any practical use: total float
and free float.
Free float is defined as that which, if used, will not delay the early
start of a succeeding activity. The definition appears to offer a very
useful identification. The formula, compared with the total float formula,
is as follows:

j – i – D = total float = LTj – ETi – D = LF – EF = LS – ES

j – i – D = free float = ETj – ETi – D = EES(succ) – ES – D

= EES(succ) – EF

Looking past the formula, though, free float loses its luster. As an
example, take Figure 8.6.1, which is part of the initial John Doe network
between event 3 and event 13. All these activities have total float, how-
ever, as a string of activities emerges from a junction event, such as event
3, the early start for all activities has been controlled by the selection
of the longest of all paths leading into that junction event.
In this example, the critical path from event 0 to event 3 determined
that the early start time is 7. For a string of more than one activities,
such as 3–9 or 3–10, in which the early finish for the j event is deter-
mined only by the early start figure coming out of the junction point, the
formula necessarily produces a free float of 0. It is only when the string
of activities joins another junction event, at which a new early start
figure is determined by the longest path leading into the new juncture,
that the free float formula produces a non-zero number. This number is
produced because one or more other paths coming into the junction
point establish an early start for that key junction, which is greater than
the early finish time of the series of activities under a study.
Free float is really a comparative value of floats in parallel paths. All
the activities shown in Figure 8.6.1 have float, and the lowest float
value is 4. Thus the free float values are 0 for the lowest relative float
path (3–9–11–12–13). However, the free float is also 0 on the activity
 Cranking the Engine 101

Figure 8.6.1 Free float compared with total float: John Doe project.
102 The Theory of CPM Planning and Scheduling

3–10, which initiates the path 3–10–11, but it is 9 on the second activ-
ity because that is the last activity before a junction point.
The free float for activity 3–12, which has only a single activity in the
string, is dependent on the early event time at event 12, which is estab-
lished by the longer path 3–9–11–12 and, therefore, has a non-zero free
float value. Free float is, therefore, deceptive because it shows a zero
value for the parallel path with the lowest total float and also for any
series of initial activities with early finishes that are not dependent on
another chain. In some cases, the free float will equal the total float value
where a path of non-critical activities re-enter a critical path string of
activities. It may be less than total float, but it will never be more.
Many programs still print out free float even though it is virtually never
used. Other times, the program may continue to generate free float but
the printout is blanked off by request.
When a column listing, Free Float, is included in a report, it is usu-
ally to note the amount of slippage permitted for delivery of fabricated
materials that will not delay the early start of a subsequent erection or
installation activity. But as noted previously, the calculated attribute is
misleading as project personnel would desire similar information relat-
ing to the submittal, approval, and fabrication activities preceding the
delivery activity. Instead, each of these preceding activities has a cal-
culated free float of zero because their successors each have but one pred-
ecessor. Non-zero free float can only exist where an activity has more
than one predecessor. It is the consequence of the merger of multiple
paths of logic.
Two solutions to this problem are theoretically possible. Each involves
assignment of a new attribute “path free float” to calculate, record, and
report the latest dates on which an activity must start or finish if not
to delay the earliest start of its successor(s.) The first is to have a code
field reserved for designation of an activity as being the last of a string
or sub-path leading to the “not to be delayed” activity. The second
requires coding of the restraints (such as in a RDM system) leading to
a merger of paths with calculation of the “path free float,” (or “junior
float” as suggested in Chapter 2), whenever the restraint is of a speci-
fied (such as “deliver material”) type. Each method has its benefits and
limitations, but the programming for each is relatively simple once
there is sufficient demand in the marketplace.

8.7. Independent Float

Independent float represents the attribute that an activity start or
finish may be deferred without reducing the ability or float of any other
activity’s start or finish to be deferred. To some extent, it is a more reli-
able indicator of when an activity is “needed” than free float. However,
 Cranking the Engine 103

as noted previously, when independent float was first defined, there did
not appear a practical use for its calculation. Thus, as early computers
were rather limited in power, it was not calculated in most situations
and became a largely ignored artifact.
As computers became more powerful, the ability to level (stay below
a set level) and smooth (minimize cycles of increasing and decreasing—
or hiring and firing) use of resources became part of commercially avail-
able software. It became clear that in deciding which activity may be
deferred, the already calculated free float attribute was useful. However,
it appears that since independent float was not already calculated, the
programmers adding the leveling and smoothing modules to the CPM
software did not see the benefit of modifying the basic calculation mod-
ules to provide this attribute.
The formula for independent float also is a bit more complex and this
may have put off the programming team. The formula is expressed as:

IF = independent float = EES(succ) – LLF(pred) – Dur

or the earliest of the early starts of all successors to an activity, minus

the latest late finish of all predecessors to an activity, minus the dura-
tion of the activity. In Figure 8.7.1, the only activity to have independ-
ent float would be Activity “H.” The early start of “I” is 40. The late finish
of “A” is 10. The duration of “H” is 10. The independent float = IF =
40 – 10 − 10 = 20. Compare this to Activity “G.” ES(I) – LF(F) – DUR(G)
= 40 – 30 – 10 = 0.

A B C D I

0 10 10 10 10 20 20 10 30 30 10 40 40 10 50
0 10 10 20 20 30 30 40 40 50
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

F G

ES DUR EF 10 10 20 20 10 30
LS LF 20 30 30 40
TF FF IF 10 0 0 10 10 0

10 10 20
30 40
20 20 20

Figure 8.7.1 Comparison of total float, free float, and independent float.
104 The Theory of CPM Planning and Scheduling

8.8. Time Scale Network

Figure 8.4.1, which demonstrates the critical activities, is the front end
of a plot of the site work activities according to a time scale. If all the
activities are plotted according to a time scale, the result is a graphical
calculation of the network. (See Figure 8.8.1 for a time scale network of
the John Doe project.) The activities are plotted in solid line to scale, with
dashed-line connections to the event connection point. The dotted sec-
tion is equal to the float in the chain of activities.
In plotting a network in which a computer or manual calculation
has not been made, all activities are plotted by early start. Float will
appear as dotted lines following the last activity in a series. If the net-
work has been calculated, either manually or by computer, the pre-
ferred plot is by late start. The early start plot gives the CPM
calculation, but experience confirms that activities do not start at the
earliest point. Accordingly, an early start plot will be patently incor-
rect at each update. And if the network is to be updated correctly, each
review will require a time-consuming redraft. On the other hand, if
the graphical plot is to a late start, redrafting will not be required
unless a major change in approach is decided on. In fact, if the
sequence and durations go unchanged, the graphical network (late
start plot) can be made to remain correct by a simple shift of the hor-
izontal time scale.

Figure 8.8.1 Time scale network: John Doe project. (Plotted to early times).
 Cranking the Engine 105

8.9. Computation Time

How long it takes to compute a network manually and how large a net-
work can be hand-computed cannot be specifically answered because net-
work characteristics vary. Why would anyone care to do these calculations
by hand rather than by computer? Beyond the simple answer that a
Scheduler or engineer should understand what is going on inside that
black box, there are often situations in the field where a small network is
developed for immediate use and there is not time to go back to the office.
This provides the overall time frame, and often picks up obvious errors.
The John Doe networks have about 130 activities. (A rule of thumb:
The number of activities in a network is about equal to 1.6 times the
number of events.) You can hand-compute the John Doe networks faster
than you could input data to a computer for one run. However, if you
expect several runs, the computer is much faster. If a computer is avail-
able, you should probably use it for networks above 100 to 200 activi-
ties if you expect reruns. If you have a complex, tightly interconnected
network, a network of 100 or 200 activities can be tedious to compute.
Thus, there is no specific limit to hand computation. You will have to set
your own limits based upon your own situation and experience.

8.10. Writing Your Own CPM Software

The basic rules for activity time computations are relatively simple, so
simple that they are intuitively obvious. To reiterate the rules:

1. The early start (ES) of the first activity is defined as zero.

2. The early finish (EF) of any activity is the ES + duration (D).
3. The ES of any other activity is the latest of the EFs of all predeces-
sors to that activity.
4. The late finish (LF) of the last activity is defined as equal to the EF.
5. The late start (LS) of any activity is the LF – D.
6. The LF of any other activity is the earliest of the LSs of all succes-
sors to that activity.
7. The total float (TF) of any activity is equal to the LS – ES, which is
also equal to the LF – EF.

As an aid, refer to the following simple diagram of a CPM activity:

Activity Description
(i)------------------------------------------------( j)
ES Duration EF
LS LF
TF
106 The Theory of CPM Planning and Scheduling

Armed with these equations and some common sense, you can write
a fairly sophisticated software program in whatever language you prefer
for solving activity time computations.
First, identify the first activity in the network. Intuitively, you can
do this by looking at the left-hand side of the pure logic diagram.
However, if our diagram was our first draft, it may look more like
Figures 5.1.1 or 5.1.2 or perhaps an even rougher diagram where the
first activity is not clearly at the left. (It is interesting to note that
Figures 5.0.1, 5.1.1, and 5.1.2 could not be solved using the matrix
method discussed in this chapter or by early computer systems that
required a single starting activity.) So how do we know which is (or are)
the first activity in a network?
Look at the preceding diagram. Note that the (j) node for each activ-
ity will be the (i) node for the next activity. Similarly, the (i) node of each
activity is the (j) node of another activity—except for those activities that
do not have a predecessor—first activities. So for the first module of your
program, you can assign an ES of zero to all first activities of a logic net-
work. After assigning an ES of zero, compute the EF as the ES + dura-
tion (D). You can then compute the ESs of other activities.
Look at the next activity in the list of activities (or the next record in
a database) without concern for the order in which the activities are
listed. Note the (i) node and search for the activities having the same
node number in their (j) column (see Figure 8.10.1.). Note its EF. If it
has been previously calculated, store this number and look for others.
You can then assign the latest EF as the ES of your target activity, and
compute the EF as the ES + D of that activity.
If the EF has not yet been defined, then ask, “Which is larger, any
known number or undefined?” The answer is always “undefined,” which
is the entry you assign to the ES of your target activity. Complete each
activity in your list until you reach the end of your list, then return to
the top of your list and repeat the process for all activities with an
“undefined” ES. Eventually, you will have determined an ES and cal-
culated an EF for each activity in your list. This concludes the forward
pass of your intuitive program.
The first step in the backward pass procedure is to determine the
last activity (or multiple last activities, which is discussed in later chap-
ters). Simply, the last activity is that in which the (j) node does not
appear as an (i) node in a list of activities. The remainder of the program
is left as an exercise for the student.

CLASS EXERCISE: Write, compile, and execute a CPM program

for the first 17 activities of the John Doe project as depicted in
Figure 18.5.1
 Cranking the Engine 107

Figure 8.10.1 Manual calculation of ED, EF, LS, LF, and TF by simulated com-
puter method.

You can expand your program to include features of modern propri-

etary software. For example, you can assign a title or description to
each activity based on its unique i–j designation. Similarly, you can
assign a date to each time designation, even addressing weekends and
holidays by skipping them in your conversion list.
108 The Theory of CPM Planning and Scheduling

The basic system is indeed very simple and can be easily improved
upon. You can improve the ease of use, include additional features, or
add the capability to select and sort activities for more informative
reports and graphics.

8.11. Manual and Computer Solution for PDM

with Durations Between Activities
The addition of non-traditional types of restraints (leads) and durations
between activities (lags) can make the calculation more tedious, but it
is still understandable. Figure 8.11.1 illustrates the additional inputs
to calculation of the early finish during the forward pass, and the late
start during the backward pass. Notice that since it is no longer guar-
anteed that EF = ES + D, nor that LS = LF – D, the calculation of total
float changes and that it is possible that LS – ES = TF(Start) is not equal
to LF − EF = TF(Finish).
Figure 8.11.2 illustrates the new calculations required to determine
the ES and EF attributes of the forward pass. The initial ES is still
defined as zero (or DataDate during an update) but the EF is now the
greater of ES + D or EF(pred) + lag or ES(pred) + lag. The ES of subsequent
activities is now calculated as the greater of the latest early finish of all
predecessors plus lags or preceding early starts plus lags.
Figure 8.11.3 illustrates the new calculations required to determine
the LF and LS attributes of the backward pass. The final LF is still
defined as equal to the EF but the LS is now the lesser of LF – D or
LS(succ) – lag or LF(succ) – lag. The LF of preceding activities is now cal-
culated as the lesser of the earliest late starts of all predecessors minus
lags or succeeding late finishes minus lags.

SF lag duration
Other activity
FF lag duration
Other activity

Activity
ES duration EF
LS LF
TFstart TFfinish

SS lag duration
Other activity

SF lag duration
Other activity
Figure 8.11.1 Additional Inputs to Early Start in PDM calculation.
 Cranking the Engine 109

Other SF lag duration Other FF lag duration • ESstart = 0

activity activity • EF = later of ES + Dur
or latest EFpredecessors + finish-finish
Activity lag
duration or latest ESpredecessors + start-finish
lag
ES EF
LS LF
TFstart TFfinish

Other SS lag duration • ESstart = 0

activity • ES = later of latest EFpredecessors
or latest EFpredecessors + finish-start
Activity lag
duration or latest ESpredecessors + start-start
lag
ES ES
LS LS
FS lag
TFstar TFfinis
duration

Figure 8.11.2 Forward pass.

Activity
duration
• LFend = EFend or as set by contract
ES EF • LS = earlier of LF − Dur
LS LF or earliest LSsuccessors + start-start lag
TFstart TFfinish or earliest LFsuccessors + start-finish lag

Other
SF lag duration activity
Other
SS lag duration activity

Activity
duration
• LFend = EFend or as set by contract
ES EF • LF = earlier of earliest
LS LF LS
FS lag or successors
earliest LS + finish-start lag
TFstar TFfinis duration or earliest LFsuccessors + finish-finish lag
successors
FF lag duration Other
activity

Figure 8.11.3 Backward pass.

110 The Theory of CPM Planning and Scheduling

Other FS lag duration SF lag duration

Other
activity
activity
Other SS lag duration FF lag duration
Other Forward
activity activity pass
FNET constraint
Other SNET constraint Other subject to calculation
event subject to calculation event
Mandatory start
overrides calculation Mandatory finish Other
Other ES EF overrides calculation event
event LS Activity LF
duration
TFfinis
Other Start-on constraint TFstar
event subject to calculation
FS lag duration Other FS lag duration Other
activity activity
FS lag duration
FF lag duration Other Other
Backward
pass activity activity
FNLT constraint
Other subject to calculation Other
event event

Figure 8.11.4 Additional calculations required to account for non-traditional restraints

and constraints.

The problem only gets worse as we add in the additional calculations

necessary to incorporate constraints that may override the logic of the
CPM network as illustrated in Figure 8.11.4.
Solution of Figure 8.11.5 is left as an exercise for the student.

8.12. Summary
This chapter discussed the use of event times to compute activity times,
specifically early start, early finish, late start, and late finish. The three
rules for identifying a critical activity were started, and float time was
defined.

1015
Activity #4
1005 10 days 1035
Activity #2 Activity #8
6 days 4 days
1020
Activity #5
1000 SS0 3 days 1045
Activity #1 FF2 Activity #10
SS5
6 days 6 days
1025
Activity
1010 1040 FS2
#6
Activity #3 Activity #9
2 days 5 days
1030
Activity #7
4 days

Figure 8.11.5 Sample network for student solution.

 Chapter

9
Adding Complexity

The basic ADM model requires only three data fields: an i node, a j
node, and a duration. As we saw in previous chapters, preparation of a
computer program to perform the calculations of activity attributes for
such a simple model is an easy exercise. To appreciate the multitude of
possible misunderstandings that can be created, we examine some of the
enhancements to the basic model.

9.1. Enhancements to the Basic System

Many features have been added to the basic concept of CPM. Some of
these enhancements include the following.
■ Separate tracking of original duration versus remaining duration
■ Input or calculation of percent complete
■ Defined subtasks and checkoff updating
■ Reporting early starts/late starts/finishes with calendar dates
■ Use of multiple calendars
■ Multiple starting and ending activities
■ Restraints and constraints to activities extraneous to the pure logic
network
■ Negative float and modifying the definition of criticality
■ Continuous and interruptible performance
■ Assigning actual start and finish dates to activities
■ Choice of algorithm for work performed out-of-sequence of retained
logic versus progress override

111

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
112 The Theory of CPM Planning and Scheduling

■ Events and milestones

■ Hammocks and summary network logic
■ Association of user-defined code fields to activities
■ Association of resources to activities
■ Association of costs to activities
■ Driving resources
■ Multi-project scheduling
■ Hierarchical coding structures
■ Using the precedence diagramming method (PDM)
■ Resource leveling and smoothing
■ PERT, SPERT, and GERT

9.2. Original versus Remaining Durations

Creation of separate data fields for original and remaining duration
may seem trivial. However, in performing updates to the network, it is
important to remember that you should only update the remaining
duration of activities that have actually started.
If new information leads you to desire to change the duration of an
activity not yet started, such a change is a revision to the network rather
than an update of the existing schedule. Thus, since no work has yet
begun, the duration to be changed would be the original duration. As we
have discussed, the mixing of information based upon observations
(updates) and hopes and expectations (revisions) can dilute the value of
the resulting calculations as a tool of analysis for the project. Another
problem that occurs when changing the remaining duration of an activ-
ity not yet started is an erroneous report of progress.
But what remaining duration should we use if the activity is started
and then work is to be suspended for a period of time? One school of
thought suggests increasing the remaining duration to cover both the
anticipated period of inactivity plus the remaining duration of actual
work. Notice that the definition of remaining duration now is remaining
duration plus-something-else, linguistically a poor definition. A better
method to handle such a situation is to report a remaining duration of
anticipated work days only, and subject the remaining portion of the
activity to a constraint. This method recognizes that your statement that
work remaining will be deferred until a future date is a revision to the logic.

9.3. Percent Complete

If a new field is added for percent complete, the first issue is to deter-
mine percent complete of what. To illustrate, different personnel may
 Adding Complexity 113

report differing percents complete for the same activity. The project cost
accountant may be interested in the percent of budget expended or per-
cent of earned value for the activity. For installation of a pump, an activ-
ity which can take 5 days, 90 percent of the cost is both expended and
earned when the pump is rough rigged and set in the first day. Final posi-
tioning, milling, and connections may take another 4 days, and 90 per-
cent of the labor, so that from a foreman’s viewpoint, only 10 percent of
the activity is complete. From the Scheduler’s viewpoint, 4 of 5 days
remain, so 20 percent of the activity is complete.
From the owner’s viewpoint, the installation will not be 100 percent
complete until the pump is successfully tested. Then again, from the
Scheduler’s viewpoint, the activity will be 100 percent complete when
its successors are capable of starting.
If after the pump has been rough rigged a problem is encountered,
requiring reporting the remaining duration as 7 days, will we report neg-
ative percent complete, 20 percent complete, or 90 percent complete?
Most software programs will report 0 percent complete.
If an actual start date field has been added and an actual start
reported for this activity (see more on problems on Actuals later in this
chapter), and for an activity of original duration of 10 days, 10 days have
passed to reach the 50 percent complete point, is the remaining dura-
tion to be calculated as the remaining 50 percent of the original dura-
tion (that is, 5 days remaining) or based upon the performance to date
(that is, another 10 days)?

9.4. Defined Subtasks and Check-off Updating

Part of the definition of an activity is that it is a set of instructions given
to an entity that may perform it without further intervention. However,
the performance of the activity may involve several discrete tasks. These
tasks may be performed in a specified order, or one of several specified
orders or in any order. For example, in rotating the tires of an automobile,
removal of the nuts holding each wheel in place may be performed in any
order but must be replaced in one of several specified orders. Rarely need
the master mechanic specify an exact order to an assistant. A pre-flight
checklist is another example. However, although a specified order of these
tasks may not be required, some means to check off and record that each
step has been taken is desired.
Several methods can be used to implement such additional functional-
ity. The description of the activity may refer the user of the CPM printout
to a separate check-off list. Or the activity description may be annotated
on the printout by means of logs or notes, as depicted in Figures 9.4.1, 2
and 3. Or a true, interactive check-off system may be implemented, as by
Primavera’s P3e/c’s Step function as depicted in Figure 9.4.4.
Figure 9.4.1 Primavera P3 Logs note detail of utility installation activity under
direction of one foreman.

Figure 9.4.2 Primavera P3 Logs note individual bents of a drainage pipe. Each line is
checked off as performed.

Figure 9.4.3 Primavera P3 Logs note individual bents of a drainage pipe. Each bent is
checked off as performed.

Figure 9.4.4 Primavera P3e/c Steps allows user to specify tasks within the “Place 2 SUR-
FACE COURSE” activity. User may check-off as work performed – 19mm 1st Course is
67 percent, 9.5mm 2nd Course is 33 percent of this 4-day activity.

114
 Adding Complexity 115

The sole caveat is that defined subtasks or Steps cannot take the place
of activities. Since there is no logic between the subtasks, much less logic
that may transcend the group of subtasks within an activity (for example,
deliver rebar to the rebar subtask of “Form/Rebar/Pour Wall” activity), this
feature regresses scheduling back to a ToDo list. The sole danger created
by the addition of “Steps” is not technical but psychological, that is, the user
must use these to supplement the description of an activity and not to
replace the proper use of activities and fall back to the use of a ToDo list.

9.5. Calendar versus Work Period Conventions

The original implementation of ADM, including i node, j node, and dura-
tion, works solely with numbers and not dates. Thus an activity may be
reported to have an early start of Day 5 and early finish of Day 12. If for
each of the calculated fields ES, EF, LS and LF we add fields to report such
day numbers in date format, our output will be much more useful to the
user. However, the use of dates does create new opportunities for misun-
derstanding.
Let us assume a 5-work day per week calendar with day ZERO being
01FEB99 (Figure 9.5.1):
Our first option is to assign a date to each day number (Figure 9.5.2).
This assumes that each day entails 24 hours. Activity A would finish at
7:59 AM on 08FEB and Activity B would start at 8:00 AM on 08FEB. This
can be misleading since, in the real world, we would probably finish Activity
A at 4:00 PM on 05FEB, and a foreman reading the schedule might think
that he or she had until 08FEB to complete Activity A.
Our second option is to assign two dates to each day number, one if the
day number is an early (or late) start and one if the day number is an early
(or late) finish. Here, we explicitly understand that the “day” ends at 4:00
PM, and that even with overtime, we will certainly finish before midnight.

FEBRUARY 1999 WORK DAYS

MON TUE WED THU FRI SAT SUN MON TUE WED THU FRI SAT SUN
1 2 3 4 5 6 7 0 1 2 3 4 – –
8 9 10 11 12 13 14 5 6 7 8 9 – –
15 16 17 18 19 20 21 10 11 12 13 14 – –
22 23 24 25 26 27 28 15 16 17 18 19 – –

Act A OD=5 Act B OD=7 Act C OD=1 Act D OD=2

1-------------2--------------3---------------4------------5
ES/EF 0 5 5 12 12 13 13 15
option 1 01FEB 08FEB 08FEB 17FEB 17FEB 18FEB 18FEB 22FEB
option 2 01FEB 05FEB 06FEB 16FEB 17FEB 17FEB 18FEB 19FEB

Figure 9.5.1 Calendar days versus project days.

116 The Theory of CPM Planning and Scheduling

Act A OD=5 Dummy OD=0 Act B OD=7 Act C OD=1 Act D OD=2
1-------------2--------------3--------------4---------------4------------5
ES/EF 0 5 5 5 5 12 12 13 13 15
option 1 01FEB 08FEB 08FEB 08FEB 08FEB 17FEB 17FEB 18FEB 18FEB 22FEB
option 2 01FEB 05FEB 08FEB 05FEB 08FEB 16FEB 17FEB 17FEB 18FEB 19FEB

Figure 9.5.2 Problems with calendars. Option #1 work “ends” at 7:59 AM . Option #2 work
“ends” at 4:00 PM.

This second option looks less likely to be misunderstood at first glance.

The 1-day Activity C both starts and finishes on 17FEB. But what if
Activity B is a logic restraint or milestone having zero duration?
Option 2 now causes confusion by listing the late start for Activity B as
occuring before its early start. If the logic restraint (or “dummy”) spans a
weekend, the late start may be reported several days earlier than the
early start. Schedulers and users of the software know what is meant, but
third parties think the software is flawed or worse. A third option was cre-
ated by several software vendors, which reported the early (and late) fin-
ishes for logic restraints or milestones as equal to the early (or late) starts
(Figure 9.5.3). This caused even more confusion for some users. At least
one software vendor gave individual users the choice of options 1, 2 or 3
in the setup or configuration screen for the software. (Imagine the confu-
sion when a data disk prepared by a contractor is run by the engineer using
a different configuration choice.) At least one software vendor solved this
dilemma by declaring that logic restraints or milestones, having no dura-
tion and thus being a point in time, will not report any value for an early
(or late) finish. However, such logic restraints or milestones must be so
declared, and an activity having zero duration, but not declared as a mile-
stone, will default to option 1.

9.6. Multiple Calendars

In the real world, there are some activities which may only be performed
during the work week, and if not finished on Friday, will continue on the
following Monday. There are also some activities, such as the curing of
concrete, which proceed equally well on weekends as on weekdays. In the
original implementations this was addressed by accepting that the specific
dates for any activity, being merely an estimate, may be off by several days.

Act A OD=5 Dummy OD=0 Act B OD=7 Act C OD=1 Act D OD=2
1------------2------------3-------------4--------------4-------------5
ES/EF 0 5 5 5 5 12 12 13 13 15
option 3 01FEB 05FEB 08FEB 08FEB 08FEB 16FEB 17FEB 17FEB 18FEB 19FEB
option 4 01FEB 05FEB 08FEB --- 08FEB 16FEB 17FEB 17FEB 18FEB 19FEB

Figure 9.5.3 Problems with calendars. Option # 2 has different rules for duration > zero
and duration = zero. Option #4 does not print finish dates for “dummy” activity logic
restraints and milestones with duration = zero.
 Adding Complexity 117

But as computers became more powerful and software more complex,

the capability of multiple calendars was introduced, first into special high
end software, then into the basic products marketed by the software com-
panies. As expected, the use of multiple calendars creates several poten-
tial misunderstandings.
Our first problem is that multiple calendars create a dilemma in
defining and calculating float. As we have learned, TF (total float) is
equal to the LS (late start) of an activity minus its ES (early start). When
we calculate Day 10 minus Day 5 we always get 5 days of float. But
exactly how much is 10FEB99 minus 01FEB99?
Typically, the total float is reported in units from the same calendar
as the original duration. Thus, if an activity performed on a 5 day per
week calendar has an ES = 01FEB99 and an LS = 08FEB99, the soft-
ware will calculate TF = LS – ES = 5 days. But if the same activity were
performed on a 7 day per week calendar, the software will calculate TF =
LS – ES = 7 days. If requesting a report sorted by criticality, that is, by
total float, the activity on the 7 day per week calendar will not be located
in the proper position.
Even more disconcerting, in changing from a 5 day per week calen-
dar to a 7 day per week calendar and back, especially if weekends are
spanned, is that the software may calculate a critical path with vary-
ing amounts of float on the path (Figure 9.6.1.).
Since, mathematically, the use of multiple calendars degrades the value
of calculations of the total float attribute, the use of them should be lim-
ited unless the progress of work during any one timeframe will vary sig-
nificantly depending upon the calendar used. For example, if cure time for
concrete is 1 week, the choice of 5 work days or 7 calendar days is irrele-
vant. However, if the cure time is 3 days, it will make a difference if the
pour is made on a Monday or Friday. However, when one considers the tol-
erance or chance of error of the activities preceding and succeeding the cure
activity, the error raised by leaving the cure on a 5 day or work calendar
is probably insignificant compared to the loss of accuracy of the total float
attribute. Thus the decision of which calendar to use for the cure activity
is left to the sound discretion of the Scheduler or knowledgeable Engineer.
Yet another dimension of multiple calendar issues is when one activ-
ity may have multiple calendars. This issue occurs when separate cal-
endars are assigned for activities and for individual resources. As an
example, suppose an activity requires two limited resources, special
equipment and an inspector. The activity may only be worked only on
weekdays. The equipment is only available on the 1st through 10th of
each month. The last day of the activity requires an inspector, who is
never available on Fridays.
Different software vendors treat the use resource calendars in dif-
ferent fashions. Microsoft Project resource calendars work in conjunc-
tion with Activity calendars, thus a day off in either means no work.
118 The Theory of CPM Planning and Scheduling

Figure 9.6.1 Problems with multiple calendars. Confused reporting of total float.

Primavera resource calendars override the activity calendars, thus, an

activity non-work day designated as a resource available day, is worked.
Obviously, the user must read and understand the rules relating to cal-
endar priority prior to use of this feature.

9.7. Multiple Starting and Ending Activities

The original CPM model, based upon the matrix mathematical approach
and the limitations of the limited memory of 1950s computers, required
that every network start with only one activity and end with only one activ-
ity. Several of the lower end software programs today still have this limi-
tation. In addition to being a software limitation, this is usually good
practice, as it precludes “dangling” activities. However, in many instances.
there is a legitimate reason for having multiple starts and completions.
An example of multiple starts is when two (or more) projects, with differ-
ing notice-to-proceed dates, are combined into one larger network to
account for the interrelationships between the two projects. Obviously,
this could be handled without special software by having a common start-
ing activity named “START OF NETWORK” followed by the two specified
 Adding Complexity 119

notice-to-proceed activities. A more difficult problem is when there are

two (or more) end products to the network. An example of this situation
is a building with commercial and residential rental space, each of which
may be rented and occupied prior to completion of the other section, as
shown in Figure 9.7.1. In this situation, it is advantageous to have two
critical paths, one to completion of the commercial section as-soon-as-pos-
sible, and one to the residential section, as-soon-as-possible.

Figure 9.7.1 Multiple completions of project, both being calculated critical.

120 The Theory of CPM Planning and Scheduling

The original CPM model, and many programs even today, cannot
handle this type of problem. (As noted in Chapter 8, a simple computer
program that you can write can handle this problem.) A schedule pre-
pared using a software package that can handle this problem, and sub-
sequently loaded to a software package which cannot handle this
problem, may either fail, yielding only an error message, or create a
hidden internal logic restraint to the latest finish, resulting in one “true”
critical path and the mistaken impression that work on the other sec-
tion may be deferred without economic consequence.

9.8. Artificial Constraints to Dates

The ability to add artificial constraints (not based upon explicitly stated
logic) is of great benefit to the user of CPM software. Such constraints
should be divided into two classes, those that could be handled by the
original CPM model with the addition of hidden internal logic restraints,
and those that may require overriding of the basic precept of CPM, that
each activity must be finished before its successor may begin.
If we want to say that an activity can not start until at least a speci-
fied date has been reached, we may provide that activity with a start-
not-earlier-than (SNET) constraint. This could be stated in the traditional
model by creating a logic restraint, from the starting activity to the activ-
ity in question, having a duration sufficient to delay the activity until (at
least) that date. Of course, each update will require laborious recalcula-
tion of the remaining duration required to push to (at least) that date.
Similarly, if we state a specific activity may finish-not-later-than (FNLT)
a specified date, we could add a logic restraint to the ending activity with
a sufficient duration to assure that a deadline is included in the net-
work. (Note the problem of multiple ending activities, and that this solu-
tion, used by some software programs, creates the one “true” critical path
problem there stated) (Figure 9.8.1).
On the other hand, constraints such as SNLT, FNET, mandatory-
start-on, and mandatory-finish-on will override the basic premises of
CPM and must be used with extreme caution.
First, we must agree on what the terms mean. The SNLT constraint may
be interpreted as saying that an activity may start on the specified date,
notwithstanding predecessor logic or unanticipated delays to other activ-
ities. Or, the SNLT constraint may be interpreted as saying that an activ-
ity must start on or before a specified date. The impact of the constraint
is to the late start of the activity. In this case, the forward pass of the CPM
calculations will not be impacted by this constraint, and the project will
still show completion based upon the logic-based calculation. This second
definition is used by Primavera Systems software.
Comparing, in Figure 9.8.2, Example #1 to Example #2, note that the
SNLT constraint is highlighted only for the late start of Activity #3.
However, both the late start and late finish of Activity #3 and all of its
 Adding Complexity 121

2000 2001
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC JAN
3 10 17 24 31 7 14 21 28 6 13 20 27 3 10 17 24 1 8 15 22 29 5 12 19 26 3 10 17 24 31 7 14 21 28 4 11 18 25 2 9 16 23 30 6 13 20 27 4 11 18 25 1 8 15 22 29
0 10 0,47 40 50 15,0 120 130 15,0140 150 0,0
NOTICE TO SLAB ON 1ST FLOOR 1ST FLOOR TURNOVER FOR
PROCEED GRADE FINISHES COMMERCIAL RENTAL INCOME
10 20 5,47 30 40 15,0 60 70 15,0 80 90 15,0 100 108 10,0 130 140 10,0
MOBILIZE STRUCTURAL STEEL SLAB ON ROOF DECK 1ST FLOOR 1ST FLOOR
FLOORS 1 TO 5 DECK FLOOR 3 & ROOFING MECHANICAL PUNCHLIST
0 20 52,0 20 30 20,0 50 60 15,0 70 80 15,0 90 100 20,0 110 120 10,0 150 450 30,0
INTERNAL ACTIVITY TO FOUNDATIONS - START NOT SLAB ON SLAB ON CURTAIN WALL 1ST FLOOR INTERNAL ACTIVITY TO
SUPPORT SNET 15MAR00 EARLIER THAN 15MAR00 DECK FLOOR 2 DECK FLOOR 4 INSTALLATION STUD & DRYWALL SUPPORT FNLT 14OCT00
105 109 7,3 225 430 0,25 430 440 10,0
1ST FLOOR PUNCHLIST FOR
ELECTRICAL ROUGH-IN FLOORS 2 THROUGH 4
INTERNAL ACTIVITY 0–20 CREATED 109 110 0,3 420 430 15,0
4TH FLOOR
TO SUPPORT SNET 15MAR00 CONSTRAINT 109 205 0,12
FINISHES
410 420 10,0
4TH FLOOR
STUD & DRYWALL
200 208 10,0 440 450 0,0
2ND FLOOR FLOORS 2 - 4 TURNOVER
MECHANICAL FOR RESIDENTIAL RENTAL

INTERNAL ACTIVITY 150–450 CREATED 300 308 10,0 425 430 0,5
3RD FLOOR
MECHANICAL
TO SUPPORT FNLT 14NOV00 CONSTRAINT 400 408 10,0
4TH FLOOR
MECHANICAL
205 209 7,9 420 425 10,5
2ND FLOOR 4TH FLOOR
ELECTRICAL ROUGH-IN ELECTRICAL FINISHES
209 305 0,9

209 210 0,23 325 430 0,15

120 125 10,5

1ST FLOOR
ELECTRICAL FINISHES
125 130 0,5

210 220 10,20

2ND FLOOR
STUD & DRYWALL
220 430 15,20
2ND FLOOR
FINISHES
305 309 7,6
3RD FLOOR
ELECTRICAL ROUGH-IN
309 405 0,6

309 310 0,13

220 225 10,25

2ND FLOOR
ELECTRICAL FINISHES
310 320 10,10
3RD FLOOR
STUD & DRYWALL
320 430 15,10
3RD FLOOR
FINISHES
405 409 7,3
4TH FLOOR
ELECTRICAL ROUGH-IN
409 410 0,3

320 325 10,15

3RD FLOOR
ELECTRICAL FINISHES

act id rd tf 1002 Sheet 1 of 1

Activity Bar/Early Dates CPM IN CONSTRUCTION MANAGEMENT - 5HT EDITION
Data Date 3JAN00 Critical Activity
Project Start 3JAN00 Milestone/Flag Activity Date Revision Checked Approved
Project Finish 26DEC00

SNET / FNLT EXAMPLE

(c) Primavera Systems, Inc.

Figure 9.8.1 Internal logic for supporting SNET and FNLT constraints.
122 The Theory of CPM Planning and Scheduling

PRIMAVERA PROJECT PLANNER

START DATE 1FEB99 FIN DATE 25FEB99
DATA DATE 1FEB99 PAGE NO. 1
----- ----- ---- ---- - --- ---------- --------------------- -------- -------- -------- -------- -----
ORIG REM ACTIVITY DESCRIPTION EARLY EARLY LATE LATE TOTAL
PRED SUCC DUR DUR % START FINISH START FINISH FLOAT
----- ----- ---- ---- - --- ---------- --------------------- -------- -------- -------- -------- -----
EXAMPLE #1 -- NO CONSTRAINTS
1010 1015 5 5 0 ACTIVITY #1 1FEB99 5FEB99 1FEB99 5FEB99 0
1015 1020 5 5 0 ACTIVITY #2 6FEB99 10FEB99 6FEB99 10FEB99 0
1020 1025 5 5 0 ACTIVITY #3 11FEB99 15FEB99 11FEB99 15FEB99 0
1025 1030 5 5 0 ACTIVITY #4 16FEB99 20FEB99 16FEB99 20FEB99 0
1030 1035 5 5 0 ACTIVITY #5 21FEB99 25FEB99 21FEB99 25FEB99 0
EXAMPLE #2 -- SNLT CONSTRAINT OF 8FEB99 TO ACTIVITY 1020
1010 1015 5 5 0 ACTIVITY #1 1FEB99 5FEB99 29JAN99 2FEB99 -3
1015 1020 5 5 0 ACTIVITY #2 6FEB99 10FEB99 3FEB99 7FEB99 -3
1020 1025 5 5 0 ACTIVITY #3 11FEB99 15FEB99 8FEB99* 12FEB99 -3
1025 1030 5 5 0 ACTIVITY #4 16FEB99 20FEB99 16FEB99 20FEB99 0
1030 1035 5 5 0 ACTIVITY #5 21FEB99 25FEB99 21FEB99 25FEB99 0
EXAMPLE #3 -- MANDATORY START CONSTRAINT OF 8FEB99 TO ACTIVITY 1030
1010 1015 5 5 0 ACTIVITY #1 1FEB99 5FEB99 29JAN99 2FEB99 -3
1015 1020 5 5 0 ACTIVITY #2 6FEB99 10FEB99 3FEB99 7FEB99 -3
1020 1025 5 5 0 ACTIVITY #3 8FEB99* 12FEB99 8FEB99* 12FEB99 0
1025 1030 5 5 0 ACTIVITY #4 13FEB99 17FEB99 13FEB99 17FEB99 0
1030 1035 5 5 0 ACTIVITY #5 18FEB99 22FEB99 18FEB99 22FEB99 0
EXAMPLE #4 -- START ON CONSTRAINT OF 8FEB99 TO ACTIVITY 1020
1010 1015 5 5 0 ACTIVITY #1 1FEB99 5FEB99 29JAN99 2FEB99 -3
1015 1020 5 5 0 ACTIVITY #2 6FEB99 10FEB99 3FEB99 7FEB99 -3
1020 1025 5 5 0 ACTIVITY #3 11FEB99* 15FEB99 8FEB99* 12FEB99 -3
1025 1030 5 5 0 ACTIVITY #4 16FEB99 20FEB99 16FEB99 20FEB99 0
1030 1035 5 5 0 ACTIVITY #5 21FEB99 25FEB99 21FEB99 25FEB99 0

Figure 9.8.2 Compare effect of constraints.

predecessors are impacted by this constraint. Note also the gap of 3 days
between the late finish of Activity #3 and late start of Activity #4, in viola-
tion of the basic algorithm of CPM. Since the forward pass is not impacted
by this constraint, project completion is calculated as the same time as if no
constraint were used. However, an independent critical path is charted to
this activity. In summary, the use of the SNLT constraint is treated exactly
as the FNLT constraint, creating an independent completion deadline (com-
pletion of all activities required for the start of this activity), but not impact-
ing the mandated completion date for other activities or for the project.
Primavera recognizes the other definition noted previously under the
designation of a mandatory start. Comparing, in Figure 9.8.2, Example #1
to Example #3, note that here both the early start and late start of
Activity #3 are highlighted as set to the constrained date of 8FEB99.
Here, the CPM calculated completion of 25FEB99 has been overridden,
and a newly calculated completion date of 22FEB99 was calculated
based upon Activity #3 starting on 08FEB99. Although Activity #3’s
early and late starts are highlighted in the tabular report, and activi-
ties precedent to Activity #3 are noted as having negative float, the
assumption stated, that Activity #3 will start on 08FEB99, is accepted
and used in all other calculations.
Analogous definitions and modification to basic CPM theory applies to
the use of FNET and mandatory finish constraints. Here, the impact of the
FNET constraint is to the early finish, isolating such activity as an inde-
pendently starting activity for purposes of float calculation, but not alter-
ing the project length. Similarly, the mandatory finish constraint will
impact all successors to the constrained activity and push the project com-
pletion date back as if a SNET constraint had been used.
 Adding Complexity 123

CLASS EXERCISE: Modify your CPM computer program pre-

pared in Chapter 8 to permit the use of SNET and FNLT constraints.
What additional modifications are required for SNLT, FNET, and
mandatory start and finish activities?

9.9. Artificial Constraints to Algorithms

Although the constraints noted previously plug in a specific date to over-
ride what would otherwise be calculated by the standard CPM algorithm,
other constraints accomplish the same end by substituting a date calcu-
lated during the backward pass with one calculated during the forward
pass. Two such constraints are the zero total float (ZTF) and zero free float
(ZFF.) The ZTF constraint does exactly what it sounds like; it substitutes
the LF date calculated during the backward pass with the EF date cal-
culated during the forward pass. Since the LF now equals the EF, LF –
EF = TF = 0. Thus a new additional critical path will be calculated from
this point back to the start of the network. This feature is often useful
when there is a commercial reason for completing an interim milestone
as soon as possible (such as the commercial space in a mixed residential
and commercial building) without setting a specific contractual deadline.
The ZFF constraint may be used to plan to delay an activity until the
succeeding activity (which may have more than one predecessor) is
ready to start. An example would be for delivery of equipment that is
to be placed upon a new foundation. It may be desirable to not plan the
delivery until the date that the completion of the foundation is expected.
Note that if the foundation is completed early, some advantage may be
lost by the use of ZFF. The revised algorithm for the zero free float con-
straint substitutes the EF and ES calculated during the forward pass
with the LF and LS calculated during the backward pass. Since the for-
ward pass has been completed by the time the backward pass is being
made, the impact of these substitutions will affect only the one activ-
ity having the ZFF constraint, unlike the ZTF constraint, which also
impacts all activities preceding the constrained activity.

9.10. Negative Float

Once we permit an activity, or even a project, to have a constraint to
its completion date, we alter one of the basic theory rules of CPM,
namely that the late finish of the last activity is equal to the early
finish of the last activity (reflecting the desire to complete as early as
possible.) If the FNLT constraint is earlier than the calculated LF of
an activity, then the activity must be completed earlier than it may be
completed and the calculation of TF will be a negative number.
124 The Theory of CPM Planning and Scheduling

There are two ways of looking at the question of criticality for a

schedule that has negative float. The first is that all activities having
negative float must be expedited to bring the project back on schedule.
The second is that only the most negative activities now constitute the
critical path. This question leads to the more general problem stated
in the next section.

9.11. Definition of Criticality

The classic definition of critical, as on the critical path, is that total float
equals zero. Two caveats to this definition are required due to extensions
of traditional CPM. The first is when a designated completion date is spec-
ified (for example, by a FNLT) and such a constraint creates negative
float, a question arises as to whether all activities having total float less
than or equal to zero are critical, or only those having the maximum neg-
ative float.
A subspecies of this problem may occur when a specified FNLT com-
pletion date is beyond the calculated completion date. In this case, vari-
ous computer programs may either use the earlier of the calculated or
FNLT date, or treat the FNLT date as a mandated completion date. In the
first case, the total float along the critical path will be calculated as zero.
In the second case, the total float along the critical path will be a positive
number.
As an example, Primavera Project Planner software allows two means
to designate a FNLT deadline for a project. In the opening or OVERVIEW
screen, a field exists for noting the FNLT deadline (Figure 9.11.1). In
addition, the specified activities at the end of the network may be con-
strained by a FNLT deadline (Figure 9.11.2). In the first case, if the FNLT
field is used on the OVERVIEW screen (whether or not such information
is duplicated for the ending activity on the network), the software will cal-
culate a positive total float for activities on the critical path. If the FNLT
field in the OVERVIEW screen is left blank, but a FNLT constraint is
entered for the ending activity in the network, the software will calculate
a total float of zero for activities on the critical path.
The second caveat to the traditional definition of criticality is based
upon experience. Considering that the original duration of each of the
activities in a network is merely an educated estimate, and considering
that a project may last several months or years, many practitioners
believe that it is misleading to designate the activities having zero float
as being “critical” for the purposes of highlighting, but ignoring those
activities having one or 2, 5, or even 10 days of float.
In a tabular printout, this problem may be solved by appropriate use
of filters (selections) and sorts. For example, preparation of a critical
activity report may involve a filter permitting only activities with a
Figure 9.11.1 FNLT box, if used, will set this date as LF of project.

Figure 9.11.2 FNLT box, if used, will set earlier of this date
or calculated LF as the LF of this activity.

125
126 The Theory of CPM Planning and Scheduling

Figure 9.11.3 Filter defining “critical” as all activities with less than 11 days total
float.

total float less than 11 days, then, sorted by early start (Figures 9.11.3,
4 and 5).
In a graphical representation, where the critical path may be high-
lighted (for example in another color or solid versus hollow bar), a spe-
cial software switch or dialog box is required to designate criticality
(Figures 9.11.6, 7 and 8).

Figure 9.11.4 Sort instruction to list by early start, then by most critical for each date.
Figure 9.11.5 Graphic created by use of filter and sort instruction to list only “near-crit-
ical” activities.

Figure 9.11.6 Critical Activities to be designated as all those with under 11 days of total
float.

127
Figure 9.11.7 Critical Activities now designated as those with under 11 days of total float.

Figure 9.11.8 Critical Activities now designated as those with under 11 days of total float.

128
 Adding Complexity 129

9.12. Continuous versus Interruptible Performance

With the ability to constrain the finish of an activity occurs the ques-
tion of whether this will also constrain the start of the activity or may
the activity start as planned and then, after a period of non-work, the
remainder of the activity scope may be completed. Another way to pose
this issue is to ask if work on a specific activity must be performed con-
tinuously or may work on the activity be interrupted (Figure 9.12.1). If
the activity may be interrupted, an associated issue is to determine
where or when. Computer software that converts the CPM output to
a bar chart graphical format has a particularly difficult problem with
this and tends to solve it by showing a bar spanning from start to finish,
without regard for the stated work day duration. Obviously, at this
point, the float calculation of EF – ES is no longer equal to that of LF –
LS. Although it may seem obvious that the determination of inter-
pretability may differ from activity to activity, software systems that
support this ability do it on a project basis. As this issue is raised more
often in the PDM variant of CPM by the use of FF or finish-to-finish rela-
tionships, it is further discussed in Chapter 11.

9.13. Actual Start and Finish Dates

The assigning of actual start and finish dates to activities can cause addi-
tional confusion and, depending upon the software algorithm, create addi-
tional misunderstandings. As noted previously, nominal preparatory work
or material deliveries can result in incorrect reporting of actual start
dates (for scheduling purposes), while nominal or schedule unrelated
work remaining can result in incorrect reporting of actual finish dates. A
typical problem caused is the reporting of work out-of-sequence, with
activities reported started (or even complete) prior to completion of their
predecessors. Several software systems, recognizing such reporting as
antithetical to proper scheduling rules, refuse to accept such data, report
an error, and stop processing. Other programs accept the data, generat-
ing output of questionable validity. Still other programs accept the data,
but print an exception report highlighting the potential problem.

FF 5 FF 5
A 20 A 20

SS 5 SS 5
B 10 B 10

Figure 9.12.1 Continuous vs. Interruptible Activities.

130 The Theory of CPM Planning and Scheduling

Reiterating the issues raised in the previous section, an actual start

date should not normally be reported until an activity’s predecessors
are complete. Thus nominal preparatory work should not generate an
actual start date. An indication this has occurred is when an actual
start is reported, but no reduction in remaining duration (RD) or pos-
itive percent complete is reported. Similarly, remaining minor or cos-
metic repair work, typically considered punchlist work, should not
delay reporting of an actual finish date. The actual start of incurring
costs or earning revenue should not trip the actual start; holdbacks for
costs unrelated to the start of successors should not delay reporting of
actual completion.

9.14. Retained Logic versus Progress Overrides

Although the basic algorithm of CPM calls for performing the forward
pass from the beginning of the network, using zero durations for those
activities that are complete, a special problem can occur when work is
performed out-of-sequence. In such cases, when an activity has started
even though its predecessors may not be complete, a question arises: can

Figure 9.14.1 Choice of Retained Logic vs. Progress Override.

 Adding Complexity 131

the remainder of this activity be completed or must work on this activ-

ity stop until all predecessors are complete?
There are logically three possible answers to this question under the tra-
ditional method of CPM calculation. (There is also a fourth possible answer
using the PDM implementation of CPM, discussed in Chapter 11.) The first
is that the work performed is incidental to the main thrust of this activ-
ity and that all further work must await completion of previously stated
predecessors, the traditional answer. Second, that this activity may con-
tinue, but its successors will be delayed until this activity’s predecessors
are complete. This is an implied FF relationship permitted in PDM but not
in ADM without very special computer software. And third, that having
shown that we can break the logic relationship, it is construed to be broken,
and that further work on this activity and all successors may continue
without regard to the uncompleted predecessor work.
In Primavera software, the first option is called RETAINED LOGIC
and the third is called PROGRESS OVERRIDE. Option 2 is not sup-
ported. The user selects which of these two algorithms are to be used
for the project in a project configuration screen (Figure 9.14.1).

9.15. Events and Milestones

Noted previously were the calculation algorithms for both the original
ADM variant of CPM and PERT and also some of the issues related to con-
verting the calculated numerical output of the CPM algorithm to dates.
It is important to remember that the mathematical basis for these algo-
rithms is to calculate the early and late times of events, or points in time
TE and TL, and only as a secondary calculation to determine the attributes
of the activities for early and late start and finish, ES, EF, LS, and LF. It
is unfortunate that the concept, as well as the proper reporting of event,s
has often been ignored by software vendors focusing upon converting the
CPM output to a bar chart style graphical user interface (GUI.) What is
missed is that events, including important events or milestones, do not
have a start and finish but are in fact one point in time. The confusion
engendered over whether the ES – EF dates for a milestone should be
reported as 08FEB – 08FEB or as 08FEB – 05FEB and the requirement
of software systems to designate milestones as “START” or “FINISH” are
only symptoms of the problem.
A recognition of this problem was made by some of the early vendors of
the ADM variant of CPM by allowing milestones to be designated by having
identical i and j nodes, such as “45 – 45.” These early programs had lim-
ited GUI capabilities and were limited to reporting the milestone date as
either an “early” or “late” date in appropriate tabular columns. Recognition
that milestones and other zero duration events are, in fact, points in time
132 The Theory of CPM Planning and Scheduling

Figure 9.15.1 Possible depiction of a milestone or zero duration activity.

and neither “start” nor “finish” dates is now possible to properly depict, as
in Figure 9.15.1, but not currently supported by software vendors.
From a conceptual point of view, the erosion of the concept of events,
or points in time, is also troubling. As repeated throughout this text, the
basic concept of CPM is that each activity may start only at a point in
time after the finish, or 100 percent completion, of all predecessors to
that activity. The backward pass likewise requires that the activity
must finish prior to that point in time prior to the earliest of late start
dates of all successors to that activity. Where these rules are relaxed,
as will be discussed in Chapter 11 on the PDM variant of CPM, there is
a significant danger of loss of accuracy in the initial CPM and a further
loss of accuracy in updates during the course of the project.

9.16. Hammocks and Summary Network Logic

The original ADM methodology is based upon points in time, or nodes,
separated by activities or additional “dummy activity” logic restraints.
Measurement from the first node to the last node of the logic network,
via the critical path, constitutes the total project duration. Often, as a
means of reviewing the forest of an overview of the project rather than
the trees of individual activities, it is desired to know the total duration
between two points or nodes of the project, whether these have been offi-
cially designated milestones or not. For this purpose, a special type of
activity was developed called a “hammock.”
The hammock activity is neither a true activity nor is it a logic
restraint, but merely designates the start and completion of a subset of
the total network. If CPM calculations are performed from the start of
this sub-network to its end, the longest, or “critical” path, will calculate
the total duration of this sub-network. This duration will then be
recorded and reported as the calculated duration of the hammock.
 Adding Complexity 133

Implementation of this concept by various software vendors may be

done correctly, that is, by requiring a true logic path from start to end, or
literally ignore the logic and merely subtract the TE of the i node from that
of the j node of the hammock activity to calculate the hammock duration.

9.17. Summary Activity Bars

The hammock activity summarizes portions of the logic, from one point
in time to another. There is an alternate means to summarize the many
activities of the project, the summary bar by activity code. While the
hammock approach may be used to condense as one bar all work from
the completion of a building foundation until completion of the roof, the
summary activity bar may condense all work by a specific subcontrac-
tor, such as the plumber. The first step, of course, is to have a code field
on which to summarize and to provide descriptions for the various values
that may populate the field. The second step is to organize by that code
field (Figure 9.17.1.)
A solid bar from the first to the last activity to be performed by this
subcontractor may be used, or the bar may be “necked” to distinguish
periods of activity and inactivity. Finally, it may be desirable to color bars
based upon some code field and this can be accomplished in the pattern
sub-dialog box. (See Figure 9.17.2 with “Neck” box checked and patterns
chosen by contractor type.)
Shown in Figure 9.17.3 is a portion of the detail of the project organ-
ized by contractor type. Figure 9.17.4 provides one view of the same
information in a summarized format, and necked to distinguish work
and non-work periods. A careful look at the black and white graphic in

Figure 9.17.1 Activity code dictionary and organize dialog boxes.

134 The Theory of CPM Planning and Scheduling

Figure 9.17.2 Bar dialog box. Pattern sub-dialog box.

the printed text, or full color illustration on the CD, may indicate the
value of setting pattern colors. Critical activities (per the Scheduler’s def-
inition of criticality), such as Activity 38, are indicated by a red line down
the middle of the bar in the detailed view. Obviously, this type of infor-
mation is not available in the summarized view.
If pattern colors are chosen, the summary bar may be one bar or it
may be separated into constituent activities. In Figure 9.17.5, the activ-
ities of the project are first reorganized by work area, then the summary
box dialog box is used to choose individual bars rather than the one sum-
marized bar of the underlying figure. Figure 9.17.6 displays the result,

Figure 9.17.3 Detail of project, organized by subcontractor.

 Adding Complexity 135

Figure 9.17.4 Summary of project, organized by subcontractor.

Figure 9.17.5 Summary Bar Dialog Box.

Figure 9.17.6 Summary Bar showing individual activities.

136 The Theory of CPM Planning and Scheduling

indicating the total number of crews performing work in each area at

any time and to which subcontractors these crews report.
Yes, this information is being presented in bar chart format. The point
is that the effort to record the logic behind the CPM allows the software
to then produce such sophisticated bar charts. And these bar charts
may be updated on a monthly, weekly, or even daily basis by merely
adding current update data and recalculation.

9.18. User-defined Code Fields

The ability to attach codes to various activities is generally a boon to users
of a schedule. Such user-defined codes, including perhaps a code for respon-
sible party, or area of the project, permits software filters and sorts to pre-
pare custom reports of say, activities to be performed by one responsible
party, sorted by area of the project. A hidden benefit of such code usage is
promoting the discipline of having only one party responsible for any one
activity. But in the real world some exceptions may occur, such as (on a
union project) having electricians as observers on a large concrete pour that
includes numerous embedded conduits. How can this activity be coded to
assure it being listed on the electrical responsibility report as well as the
concrete crew report? More importantly, in an effort to achieve perfect
coding, will adding phantom activities or other manipulations compro-
mise the network? The key is to remember that the coding structure is to
augment and not control the network preparation.

9.19. Adding Resources to Activities

Similarly, the addition of assigning of resources to an activity is a ben-
efit to the Scheduler. Assignment of resources helps clarify the limited
description assigned to the activity. Assignment of resources helps define
the duration assigned to the activity. For example, a 2-craftspeople crew
may take 10 days to perform a task, while a 10-craftspeople crew might
take only 2 days. A statement of the crew size helps validate the dura-
tion assigned by defining how the Scheduler chose such duration.
However, such listing of resources as an explanation of the scope and
duration of the activity does not imply that no other resources may be
required to accomplish the activity, nor that the resources listed will be
used exclusively for such activity.
Let us assume that a crew of three boilermakers will rig and set two
pumps in two adjacent but distinctly separate structures on the same
day. Each activity will be given the minimum duration used on this
project, that is, 1 day. Each activity will be assigned three boilermak-
ers. The total manpower to be used that day, however, will be three and
not six. Might we assign 1.5 craftspeople to each activity? Yes, if we wish
to count beans rather than produce a schedule.
 Adding Complexity 137

If measurement of total manhours or mandays is desired, then pos-

sibly two separate data fields may be assigned to the activity. The
first to state the resources required to perform the activity (three
craftsmen, one backhoe and driver, etc.) and another to state the quan-
tity of resources to be used (three craftspeople × 4 hours each or 12
manhours, 1 hour of the backhoe). Again, the key is to remember that
the assignment of resources is to augment and not control the network
preparation.

9.20. Adding Costs and Cost Codes to Activities

Assignment of costs to activities creates even a greater risk of com-
promising the scheduling benefits of the CPM network. For although
a cost-loaded network has many benefits, the key is to remember that
the assignment of costs is to augment and not control the network
preparation.
If, however, the CPM is to be used for accounting purposes, and to the
tolerances of accounting (to the penny!), then the viability of the CPM
as a scheduling tool may be gravely compromised. The chance of its
meeting the needs of an accounting department are also low. The term
tolerances used here is instructive. Costs to an accountant have a low
tolerance. If payroll is to be generated from any system, it had better
be correct, to the penny, and correct to the penny relating to benefits and
taxes. A looser tolerance is required for estimating purposes and for
cost engineering or productivity studies. An even rougher estimate or
looser tolerance is required for scheduling purposes. If such a looser tol-
erance is permitted, if costs are added to augment the network rather
than control the network, then several benefits may accrue to the users
of the network.
These include (1) additional clarity to the definition of an activity, (2)
a means to roughly forecast cash flow, and (3) a means to compare the
validity of the network versus the bid estimate. As noted previously, the
resources attached to an activity are approximate. Extrapolating such
resources (labor, equipment, and material) with average wages, rental
costs, and purchase prices, therefore, will be expected to create an
approximate cost.
Such approximations are acceptable for payment purposes also since
(1) even the most detail-oriented project engineer would measure and
pay for concrete only to the nearest cubic yard and (2) even if one activ-
ity is overvalued by several hundreds of dollars, the total for the proj-
ect will be correct, to the penny.
The use of a cost-loaded CPM for payment purposes raises another seri-
ous issue for the scheduling professional. An implied definition of any
activity is a scope which requires completion of its predecessors and is
138 The Theory of CPM Planning and Scheduling

required for the start of its successors. Thus, in the real world, some por-
tion of any named activity may be performed and payment earned before
its predecessors are complete and the truly defined activity begins; and
some portion of any named activity may remain incomplete, although its
successors can begin, and the truly defined activity can be deemed com-
plete for scheduling purposes. As an example, consider the erection
sequence of form-rebar-pour slab-on-grade, walls, and elevated slab. It
is likely that the delivery of all rebar for the three activities would be
accomplished at once, thus, depending upon the wording of the con-
struction contract, justifying payment for delivered materials for all
three activities. It is also possible that some cosmetic flaws may exist in
the concrete wall, although it is capable of supporting the elevated slab.
In the first case, it would be inappropriate to indicate an actual start
or schedule percent complete for the walls and elevated slab rebar prior
to the pouring of the slab-on-grade. In the second case, it would be inap-
propriate to fail to provide an actual finish date or grant less than 100
percent for work on the wall. Many software systems provide a means
to either unlink percent complete from schedule progress or to report
two percent completes for each activity, a schedule percent complete
and a cost percent complete. Primavera provides both of these options
as well as per-project configuration switches (Figure 9.20.1).

Figure 9.20.1 Configuration screen to choose linkage of percent complete schedule versus cost.
 Adding Complexity 139

Another problem with assignment of costs to activities is that any one

activity may have a number of different “costs” associated therewith. The
first is the budgeted cost. The next is the actual cost experienced in the
field. Then, there is the earned value, including unassigned overheads
and profits. If an activity has been altered by a change order, these
three “costs” will be duplicated. If multiple change orders impact the
activity, multiple duplications may be called for. Software products may
track one or several of these multiple fields and the subsidiary fields
associated with them. For example, in Primavera, the user may report
an activity 50 percent complete for schedule, have spent 90 percent of
its budget, and yet have earned only 30 percent of its specified value.

9.21. Resource Driven Scheduling

Throughout our discussions of various scheduling algorithm alterna-
tives, we have so far used a model that requires the Scheduler to deter-
mine an original duration for each activity. As a large project is broken
down into small definable tasks, the estimation of duration for such
tasks is made easier. In most cases, a Scheduler or project manager
may estimate the duration of such tasks with a reasonable degree of
accuracy. Factors that influence the estimate of duration include, but are
not limited to, the number of manhours estimated for the task, the
chosen utilization of resources, and an understanding that both the
crew size may vary and even the number of hours may vary during a
day, depending upon the progress of the work.
Sometimes, however, estimation of duration may be both more and
less difficult. Less difficult because performance is based strictly upon
the usage of key resources (people, computer access time, etc.); more dif-
ficult because the availability of key resources may run according to their
own calendars rather than the common calendar of the project. On the
other hand, the effort of determining an estimated duration now becomes
a purely mechanical task that is best done by the computer. The
Scheduler need only enter the estimated number of manhours (or other
units of resource usage), the number of craftspeople (or other units of
resource), and a calendar of resource availability to permit the software
to determine the duration of the activity.
Again, seemingly simple, but if a room of project managers were given
the same information, it is likely that more than one estimate of indi-
vidual task duration would be the end result. One area of divergence
may be the assumption of linear or constant usage of a resource versus
an expectation of ramp-up, production, and taper-off usage. Here, some
project managers may assume in their duration estimates that although
a nominal crew, of say 10 masons, will build a wall, for the first 2 days
only two key craftspeople will begin at the corners, and at the end of the
140 The Theory of CPM Planning and Scheduling

activity, only two craftspeople will top off the parapet. In a large wall
forming activity, although the number of craftspeople remains constant,
productivity will be lower for the first 2 days as a learning curve devel-
ops. Note that in both of these instances, finer detail could better alle-
viate the estimating difficulty, but would make reporting and updating
more difficult.
Software solutions may also result in divergent results. While one
software product may assume a linear assignment of resources, another
may assume or permit a nonlinear assignment, such as the bell-shaped
curve or slow-at-first then full-production triangular curve suggested
previously.
If more than one resource is assigned to an activity, which one or ones
will be used to determine the activity duration? Some (usually limited)
resources may be designated as driving the scheduling of an activity,
while others (openly available) may passively be called for as required.
If two or more resources are designated as driving, the activity may be
constrained to production only when both are present or different por-
tions may proceed independent of the resource needs of the others. Note
that the latter case may be represented by two (or more) activities
having common predecessors and successors.
Both P3 and the newer P3e/c support the type of activity where the cal-
culated duration increases incrementally as one of several resources become
available until all work is accomplished. The type of activity when the cal-
culated duration increases incrementally only during periods of time when
two or more resources are jointly available until all work is accomplished
is supported by P3 but not by P3e/c. However, since this type of activity is
used rarely and the same calculation can be obtained via a work-around
(creating a special “activity calendar” limited to dates when both resources
are available), this should not be seen as a significant problem.
Finally, if a limited quantity of some resource is available for the proj-
ect, to be divided among several activities, which one should go first?
This question, relating to resource leveling and smoothing, and dealing
with the various software algorithms for adding such additional sequenc-
ing restraints, is discussed in Chapter 37.

9.22. Master Schedules Local versus System-wide Updating

Managing two or more projects that are somehow related poses a new
set of requirements upon the process of planning and scheduling.
However, it is again important to remember that the primary purpose
of the CPM is for planning and scheduling for the benefit of the project
(usually matching the needs of those performing the work) and not for
reporting to upper management, the cost department, regional resource
directors, or third-party dignitaries. Any additional benefit that may be
 Adding Complexity 141

obtained from a good CPM is welcome, so long as obtaining it does not

reduce the quality or usefulness of the CPM to the project.
One implementation of a master schedule involves designating several
independently controlled projects as being part of a “group.” Perhaps the
group shares common resources, such as heavy construction equipment
or key personnel, or a common program, such as multiple projects occur-
ring at an airport site or modules to a computer program, or a common
supervisor. By taking away power from the project manager for an indi-
vidual project and requiring all project managers to adopt common codes
and procedures, the CPM may provide benefits to managers who are not
directly responsible for the implementation of this project. There may
even be additional benefits for the project manager who has sublimated
control over the CPM due to the reciprocal information from other proj-
ects subjected to this process.
What may be these potential benefits? In a large construction com-
pany, the heavy equipment manager will be able to see not only the
short-term needs of various projects, but the mid- and long-term needs.
Allocation of equipment could be planned upon strategic reasons rather
than “who asked first.” Decisions could even be centralized as to whether
an individual project manager has the right to rent equipment outside
of the company to keep a project “on schedule.”
What are the potential costs? Obviously, the process of planning must
be somewhat institutionalized so that each activity requiring heavy equip-
ment is so coded, and that common company-wide codes are used for such
equipment. If using software such as Primavera P3, the order of the code
fields must also be consistent as roll-ups among projects within a group
are by code field #1, #2, #3, etc., rather than by the code field name. In any
case, it is likely that all project managers will be required to use the same
software system. If the heavy equipment manager desires reliable infor-
mation that is good for other than long-range purposes, what will be the
impact if project One is updated each Monday, project Two each Friday,
project Three on the first Monday of the month, and project Four on the
1st and 15th of each month? And yet, the timing of updates may be more
dependent upon the owner’s specification or the availability of the com-
pany’s in-house Scheduler to perform only so many updates on any one day
than upon the needs of the heavy equipment manager. Finally, if the proj-
ect manager has a weekly meeting of foremen on Monday mornings and
updates are set by company needs to be on Monday afternoons (with
results available Wednesday,), the team will likely schedule its work based
upon the reliable, old–fashioned, hand-drawn, 3-week schedules and the
entire CPM effort will be converted from a tool to help the field to just an
extra paperwork burden.
Another problem with the concept of master schedules is determining
the proper hierarchy of projects. The program manager of the airport
142 The Theory of CPM Planning and Scheduling

authority wants to know how many large cranes are on the site and tells
all contractors to use a specific code for a crane in code field #5. This may
well conflict with the construction company master schedule system.
What to do? Have a “company-wide” program with one or several major
projects excluded or have duplicate CPMs for the airport project, that
is, one coded to the company master schedule and one coded to the air-
port authority master schedule?
Other rollup or summarization issues may arise from differing sys-
tems of measurement from one project to the next. It is all well and good
that productivity by quantities installed be reported for company-wide
accounting, but if some projects use metric and others English meas-
urements, the specified coding field is not going to work.

9.23. Activity Types

The original activity of the 1956 era CPM was generally described as
the effort between two points in time. Alas, that simple definition is no
longer so straightforward. The bridging of many simple activities from
the point in time before the first to the point in time following the last
has been described previously as a hammock. As noted, a hammock
“activity” is not assigned a duration, rather the duration between the
two aforementioned points in time is calculated. In the Primavera for
DOS software product, the means to designate such an activity was on
the constraint screen, listing such as “hammock constraint.” Probably
more accurately, the Primavera for Windows product creates a new
input attribute for each activity, the activity type, and designates such
summary activities by coding such as a Hammock type activity (Figure
9.23.1 through Figure 9.23.3).
Primavera does not check the logic validity of designated hammocks,
instead only calculating the EF – ES = Duration between the start of
its earliest predecessor and finish of its latest successor. It is up to the
individual Scheduler to validate the logic actually runs from start to end.
Primavera also supports another faux hammock type designated as a
WBS activity. Here, the logic is determined automatically from the first

Figure 9.23.1 Activity types.

 Adding Complexity 143

Figure 9.23.2 P3 for DOS designation of Hammock as a constraint type.

activity of a specific WBS code to the last. As the schedule changes

during the course of the project, the logic also will shift accordingly.
Other options on how an activity is to be treated have also been added.
The traditional activity is designated as a task activity, composed of one
or more discrete tasks to be performed. In attempting to simulate a
zero duration event (or node in ADM), Primavera required the desig-
nation of a zero duration activity (or node in PDM) as a milestone. Since
even a zero duration activity has a start and finish, it became impor-
tant to distinguish whether the simulated event was before the start or
after the finish of the zero duration activity, thus being designated as a
start milestone or finish milestone. Since Primavera has always sup-
ported PDM, even its treatment of ADM milestones (or events) was in
this fashion, rather than the more traditional method of designating an
event as activity #12 – 12 (with the j node equal to the i node.)
Another activity type is the flag. This special activity type may be
placed following one or more other activities and calculates the earliest
early start date of its predecessors if designated a start flag, or the
latest early finish of its predecessors if designated a finish flag.

Figure 9.23.3 P3 for Windows designation of Hammock as an activity type.

144 The Theory of CPM Planning and Scheduling

Primavera P3 also supports two types of activities when the duration

is driven by resource calendars rather than by the activity calendar. The
first is the independent type activity. If such an activity is driven by one
resource, it progresses according to the resource calendar of that
resource. If such an activity is driven by two or more resources, it pro-
gresses whenever any one of the resources is available.
The second of the resource activity types is the meeting activity. As
the name implies, for this activity to progress, requires that all driving
resources are available. Thus a meeting activity requires (as driving
resources) both Steve and Mary, with Steve being available only the first
2 weeks each month and Mary being available only on Thursdays and
Fridays, the activity can only experience progress for 4 days each month.
P3e/c no longer supports the meeting activity. A workaround sug-
gested to Primavera by the authors is to code such activities as task
activities assigned to a “Steve and Mary” calendar. The only stipulation
to this workaround is that the “Steve and Mary” calendar will have to
be manually adjusted rather than automatically following the individ-
ual calendars of Steve and Mary as they schedule their individual vaca-
tions and other appointments.

9.24. Hierarchical Codes

Professional Schedulers know that it is the plan and the process of plan-
ning that brings the greatest return in the use of CPM. The reporting of
progress and its impact upon the remainder of the plan is an important
secondary aspect of CPM. The reporting of progress for the purpose of
measurement of past performance is by far third priority. However, once
progress may be measured, whether against a good plan, a poor plan, or
no plan, some members of upper management want that measurement
summarized according to hierarchies that often have little meaning at
the level of the project and in assisting the project team to achieve good
performance and a project completed in a timely manner.
It is the job of the Scheduler to understand these forms of hierarchy and
to implement each in a manner that does not unduly conflict with the pri-
mary purposes of CPM. The literature on these topics, from such sources
as the Project Management Institute (PMI), Association for Advancement
of Cost Engineering International (AACEi), and American Society of Civil
Engineers (ASCE) each deal with reporting within these hierarchies.
Three such hierarchies discussed in Primavera’s P3e/c manual “Moving
from P3 to P3e/c” include EPS (Enterprise Project Structure), WBS (Work
Breakdown Structure), and OBS (Organizational Breakdown Structure),
Although the definitions used by Primavera, MicroSoft, other software ven-
dors, and in various scholarly articles may vary, a general definition of each
may be as follows.
 Adding Complexity 145

■ EPS, Enterprise Project Structure: “The EPS is a hierarchy that rep-

resents the breakdown of projects in a company.” This may include
“phases of projects or other major groupings . . . while projects always
represent the lowest level of the hierarchy.” Higher levels of the hier-
archy may group projects by project manager, client, office, fiscal year,
or other level of summary. Note that the choice of group upon which
to summarize is entirely subjective. Whether a summary can be made
of all client projects serviced by each project manager for the vice
president of projects or of all project managers servicing each client
for the vice president of sales is a decision that will be made above the
vice presidential level.
■ WBS, Work Breakdown Structure: “The WBS is a hierarchical
arrangement of the products and services produced during and by a
project.” In many industries other than construction, hierarchies exist
where a product is constructed of modules or components provided by
other entities. These components may, in turn, be constructed of sub-
components provided by others. However, at each level of component
assembly, there may be little interaction among individuals.
In the construction industry, while some work is of a modular nature,
it is more likely that the entire team of diverse individuals will weave
in and out during the course of the entire project; thus making the
design of the hierarchy a much more academic exercise. Thus, in the con-
struction industry the WBS is often tied to the summarization format
of an estimating or accounting department. Examples may include:
■ Location, e.g., structure #1, #2, floor #1, #2, quadrant NE, SW, etc.

■ system, e.g., potable water, utility water, etc.

■ CSI division, materials, e.g., rebar, concrete, pipe, conduit

■ work type, e.g., formwork, rebar, concrete, cure, strip

However, the decision of whether to summarize by structure then

system or by system then structure is entirely subjective. More seri-
ously, activities spanning the WBS breakdown may cause problems.
Thus, while large foundations may involve separate activities for forms,
rebar, and pouring, small foundations may require only one activity of
1-day duration. The temptation of upper management to demand that
this activity be broken into three (to match three WBS codes) often
will prevail. Reporting by the field against three “accounting” codes
rather than one activity can be expected to be less than complete. Field
crews attempting to utilize a 3-week look-ahead cluttered with dupli-
cate codes for the same “activity” will abandon it for more simple field
scheduling tools.
■ OBS, Organizational Breakdown Structure: “The OBS is an outline
of managers responsible for the projects in your company.” Whether
this hierarchy is of named individuals (John, Mary, and Steve) or
146 The Theory of CPM Planning and Scheduling

roles (Superintendent, Foreman, and Craftsman), the applicability

of this structure is more useful in an organization where personnel
have fixed responsibilities. The OBS also presumes a simple hierar-
chy of reporting and authorizations and may be difficult to imple-
ment in an organization where matrix management is practiced.
In non-construction industries such a code is often useful where a
named individual (or resource) is required for two or more separate
products or services that may occur on different branches of the WBS
tree. When the work product requires several layers of review and
approvals, a WBS may be very useful. In the construction industry,
even when projects get large enough to call for area superintendents,
there is rarely enough of an OBS below the project manager to war-
rant separate coding.

The trick with hierarchical codes, like other codes, is to have them add
value to the reporting system of the CPM schedule without distracting
from the basic strengths (the choice of activities and logic relationships)
of the CPM logic network.

9.25. Summary
As we have seen, the basic Critical Path Methodology, while bringing logic
to the planning and scheduling process and being a vast improvement over
simple bar charts, has limitations inherent in any model of the real world.
The good news is that the methodology is flexible enough to permit numer-
ous enhancements while still maintaining the basic concept, that each
activity must await completion of its predecessors before starting and in
turn, must be complete before its successors may start. Each of the
enhancements noted brings additional usefulness to the users of the
Critical Path Methodology, but at the cost of requiring both the CPM pre-
parer and CPM reviewer to address the ambiguities of non-standard ter-
minology and algorithms, and requiring both to verify that the
enhancements have not been used to accidentally or purposefully obfus-
cate this model of reality.
 Chapter

PDM and
10
Precedence Networks

In the early 1960s, Professor John W. Fondahl of Stanford University,

an established expert on noncomputerized solutions to CPM and PERT
networks, was one of the early supporters of the precedence method, or
PDM. He called it the circle-and-connecting-arrow technique. His study
for the Navy’s Bureau of Yards and Docks included descriptive materi-
als and gave the technique early impetus, particularly in Navy projects.
An IBM brochure credited the H. B. Zachry Company of San Antonio
with the development of the precedence form of CPM. In cooperation
with IBM, Zachry developed computer programs that could handle
precedence network computations on the IBM 1130 and IBM 360. This
was particularly significant because, in 1964, C. R. Phillips and J. J.
Moder indicated the availability of only one computerized approach to
precedence networks vs. 60 for CPM and PERT.1
The form for precedence networks was originally termed “activity on
node.” The activity description is shown in a box or oval, with the sequence
or flow shown with interconnecting lines. In some cases, arrowheads are
not used, although this leaves more opportunity for ambiguous network
situations.
Figure 10.0.1 shows the John Doe network in precedence form.
Seventeen precedence activities are shown, the same number as the
regular activity-oriented CPM network. Simplicity of form is pur-
ported to be one of the advantages of precedence networks. When

1
Joseph J. Moder and Cecil R. Phillips, Project Management with CPM and PERT,
Reinhold New York, 1964.

147

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
148 The Theory of CPM Planning and Scheduling

Figure 10.0.1 John Doe network in precedence form.

activities have to be subdivided to show phased progress, the prece-

dence network can result in a lower number of notations, in some
cases, more than a 50 percent reduction. Consequently, the prece-
dence network has the advantage of a simple appearance and, to those
who use them continually, interpreting them can be straightforward.
Unfortunately, the ability to interpret them is not as easily acquired
by someone accustomed to CPM.

10.1. Precedence Logic

One reason for the apparent simplicity of precedence networks is that
a work item can be connected from either its start or its finish. This
allows a start-finish logic presentation with no need to break the work
item down. The translation of the John Doe network into precedence
form shown in Figure 10.0.1 consists of only one type of connection: end
to start. Figure 10.1.1 illustrates the three basic precedence relations:
start to start, end to end, and end to start. Although precedence networks
are simpler in appearance than regular CPM diagrams, greater thought
must be given to reading and interpreting them.
 PDM and Precedence Networks 149

Figure 10.1.1 Typical precedence relations.

Another characteristic of PDM diagrams is the use of lead and lag fac-
tors. In CPM, lead activities that logically delay the start of a particu-
lar activity or group of activities can be introduced (Figure 10.1.2).
Assigning a duration to the lead activity imposes a delaying factor in
the CPM calculation. (The effect can be achieved in many CPM computer
programs by locking in an event date to occur “not earlier than.”)

Figure 10.1.2 Lead factors.

150 The Theory of CPM Planning and Scheduling

Figure 10.1.3 Shows use of lead and lag factors.

Similarly, a large activity can be imposed to direct the completion of an

activity to occur some period of time after either the start or the com-
pletion of another activity (Figure 10.1.3).
The lead/lag factors assigned to PDM work packages can replace the
multiple activities required in CPM to reflect start-complete or start-con-
tinue-complete; that is, they can replace the multiple activities required
in CPM to create an interim event or events at which other activities
start or conclude (Figure 10.1.4 and 10.1.5).

Figure 10.1.4 ADM version of network for 1-mile highway

 Figure 10.1.5 PDM version of network for 1-mile highway
151
152 The Theory of CPM Planning and Scheduling

The result can be a network diagram that is apparently simpler than

a regular CPM network because it takes fewer work package “boxes” to
describe the same set of circumstances. Although the depiction appears
simpler, PDM diagram users have to think harder to understand the
logic depicted. Perhaps the greatest strength of the CPM network dia-
gram is its ability to first record the logical sequence of a plan and then
to communicate that logic. PDM, in its sophistication, takes a step back-
ward in communications capability.
Experienced schedulers using PDM on a regular basis have stated that
they can fine-tune and change schedules more readily with computerized
PDM. At the same time, the leads and lags make hand calculation of PDM
less practical, if not impractical. Further, time scaling of PDM is more dif-
ficult than time scaling of CPM. Since time scaling is, in itself, a calcula-
tion, the difficulty in doing it confirms two things: (1) that manual
calculation of PDM is impractical, and (2) that PDM obfuscates the use of
a network as a means of communicating information.
That is a very significant loss. From the earliest period of using net-
work methods, it was clear that communicating the results is vital to
the effective implementation of a network-generated schedule. The
network schedule itself becomes moot and meaningless if project man-
agers are unable or unwilling to understand the output. Early CPM pro-
grams suffered from overenthusiasm and overwhelming pages of
computer printouts. Since then, sensitivity to the communications
aspects of CPM has become a vital part of ensuring the effective uti-
lization of CPM results.
PDM has the paradoxical characteristics of apparent simplicity
and built-in sophistication. The result is that the PDM scheduler
becomes the project guru rather than a participating project team
participator.

10.2. Work Package Calculations

In theory, work package time calculations are quite similar to CPM
event calculations. The first stage is the establishment of a work item
and duration chart. A table of relations is then constructed on the
basis of the typical relations shown in Figure 10.1.1. The early start
time for the first work item is zero, although a calendar start date can
be inserted later. The early start time at the beginning, or each of the
other work items, is the greatest of the paths entering the beginning
of the work item. The value of the paths is computed by the following
methods:

1. Start to start: The early start time for the preceding work item is the
early start time for the work item.
 PDM and Precedence Networks 153

2. End to start: The early finish time for the preceding work item is the
early start time for the work item.
3. End to end: The early finish time for the preceding work item less the
duration of the work item itself is the early start time for the work
item.

The longest path to the beginning of a work item determines the

item’s early start time. The early finish time for a work item is the
item’s early start time plus the duration. By definition, the late finish
time for the last work item is set equal to the early finish time for that
item, which establishes a critical path. The late finish times for other
work items are determined by subtracting or a backward pass from the
late, or the finish, time for the terminal event. The late finish times for
other work items are the least of the paths leading into completion of
the work item, as follows:

1. End to start: The late finish time is the latest start time for the fol-
lowing work items.
2. End to end: The late finish time for the work item is equal to the late
finish time for the following work item.
3. Start to start: The late start time for the following work item plus the
duration of the work item itself determines the late finish time for
the work item.

Late start time equals late finish time less duration. Float for a work
item can be calculated by the same formulas utilized in the CPM
approach. Similarly, the critical path can be identified using the stan-
dard rules. As noted, the introduction of lead and lag factors (easily
handled by computer) makes manual calculations difficult, if not
impractical.

10.3. Computer Calculation

For a golden period in the 1980s and early 1990s, computer programs
could handle either PDM or ADM. Ironically, the initial programs uti-
lized an internal translation of PDM into an ADM format, calculation
by ADM, and retranslation back into the PDM format.
One problem in inputting the PDM diagram is the lack of event num-
bers. If all of the activities were end-to-start, the work package numbers
could be used similarly to i-j. However, the complexity introduced by
start-to-start, start-to-end, and end-to-end relations required a cum-
bersome cataloging of predecessor and successor work items.
Figure 10.3.1 presents a simplified ADM printout. Figure 10.3.2 pres-
ents the simplified PDM printout for the John Doe example. It is very
154 The Theory of CPM Planning and Scheduling

Figure 10.3.1 i–j sort of initial baseline of John Doe project.

similar to the CPM printout with restraints deleted. (In fact, CPM out-
puts can have any activities, including restraints, suppressed to simplify
the volume of output.) Although the output is simple in appearance, it
cannot be used in this form to track a path through the diagram.
In a field situation, in which the master network and printouts are in
PDM format, a CPM-oriented contractor scheduler complained that he
could not match the PDM output with the diagram. The PDM project
manager scheduler retorted, “Of course not, only I can do that.” What the
 PDM and Precedence Networks 155

Figure 10.3.2 John Doe project PDM printout.

project manager really meant was that the contractor scheduler had not
been given sufficient output to understand the PDM. In effect, the basic
output is a scheduling directive, not a scheduling tool for mutual use.
Figure 10.3.3 presents the John Doe PDM master output with pred-
ecessors. This output demonstrates that the purportedly simple PDM
can become cumbersome when presented in usable form.
156 The Theory of CPM Planning and Scheduling

Figure 10.3.3 John Doe project PDM output with all precedence activities.

10.4. Project Example

Figure 10.4.1 is a sample project network consisting of 34 work items.
The work item identification numbers identify work items by functions.
For instance, concrete items are grouped in the 300 series, and the elec-
trical items are in the 700 series.
The networks indicate interrelations between work items and also
options having to do with lag time factors. Three of the four lag time
157

Figure 10.4.1 Precedence example.

158

Figure 10.4.2 Precedence output, early start sort (partial).

 PDM and Precedence Networks 159

options have been included in the network. Duration for work items is
shown in the small boxes under the work items.
The network indicates that the drilling of piers (work item 110) may
begin after 50 yd3 of excavation has been completed in work item 100.
This is represented by the line leading from 100 to 110, which shows
that part of the duration of work item 110 may be concurrent with
work item 100. The estimated quantity of work item 100 is 150 yd3, so
work item 100 may start after approximately 33% of the excavation
operation has been done or, in direct proportion, after half a day has
elapsed. The top line leading from work item 100 to work item 110
indicates that the second work item cannot be completed until at least
half a day after the completion of the first work item 100.
The rebar of pier 200 is shown to begin at least 1 day following the
start of drilling the piers. This is shown by a lag time of 1 day on the
connecting line. The relation of work items 205 and 400 to item 300 and
following work item 310 is very similar to the other diagramming rela-
tion showing concurrent activity.
Between work items 310 and 410, a delay of 1 day is shown. The lag
permits 1 day of curing before the form stripping is started, and it could
have been included by adding one more day to work item 310 or by
introducing a work item 311 called initial cure. A CPM network can
duplicate the delay and lag options in the precedence network, but addi-
tional arrows or activities are required.
The work item report is a listing printed in early start sequence
(Figure 10.4.2). In addition to the obvious descriptive material, the PC
column in Figure 10.4.2 contains the amount of the operation completed
in shift, shifts per day, and days per week of the calendar factor.
Precedence programs will accept schedule dates and, therefore, can pro-
duce negative slack.

10.5. Summary
PERT has virtually disappeared from the construction scheduling scene,
but PDM use has grown dramatically. A more recent comer than CPM,
it offers the appeal of newness. Susceptible to ready adjustment and fine-
tuning, it can be readily utilized by a sophisticated scheduler.
Much has been claimed for the simplicity PDM offers in regard to both
network diagrams and printouts, but as in tip-of-the-iceberg cases, more
is hidden than seen in many PDM schedules.
This page intentionally left blank
 Chapter

Respecting the Power of PDM

11
We have all heard stories of how the modern computer is like the myth-
ical genie, that is, it does exactly what you tell it rather than what you
actually want it to do. One of the advantages of the ADM (Arrow
Diagramming Method) methodology is that in its stark simplicity, it
forces the user to say exactly what he or she wants. The simple rule of
ADM, that each activity may start only after its predecessors are fin-
ished, is easily understood. Moreover, although differing practitioners
and writers of computer software may try to “add” features to ADM, the
basic precepts of the methodology are difficult to abuse.
PDM (Precedence Diagramming Method) is the much more powerful
system; so powerful that its inner workings are rarely understood by the
user. One of the most cited advantages of PDM is its use of lead and
lag factors, or more succinctly, duration between activities or portions
thereof to supplement the information given by the duration of an activ-
ity. Unfortunately, there does not exist a universally agreed set of defi-
nitions relating to what is meant by lead and lag factors. One result of
this lack is that various software vendors may each use a differing def-
inition without even realizing the problem.

11.1. Durations between Activities: Percent Lead/Lag Relationships

As an example, let us look at the simplest lead/lag relationship: Activity
B is to start after 50 percent of A is complete. This is often the exact lan-
guage of the project manager or superintendent and what was said is what
was meant. However, popular software tools available cannot accept that
relationship as stated. To say that activity B is to start 5 days after the
start of a 10-day activity A is not the same as saying that activity B is to

161

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
162 The Theory of CPM Planning and Scheduling

Start B after 50% of A is complete

A 10 B 10 A

B
SS 5 12 3456789012345
123456789012345
Figure 11.1.1 Ambiguous language of PDM.

start 5 days before activity A is calculated (expected) to finish. Neither is

it the same as saying that activity B is to start when 5 days of activity A
are performed. None of these approximations and “fudges” is the same as
use of the percentile provided by the project manager, and as noted in
Chapter 9.3, the definition of percent complete is also somewhat vague.
Assume that both A and B have original durations of 10. To enter this
information into commercially available software systems, you might state
that A and B are connected by a start-to-start relationship with a lag factor
of 5, as shown in Figure 11.1.1. The computer relentlessly will accept what
we just said and perform its calculations accordingly, but it is not correct.
What we have just stated is not that Activity B may start after 50 per-
cent of activity A is complete (or having achieved 5 days of progress), but
rather that Activity B may start 5 days after Activity A has started
(or having achieved 5 days of passage). This misunderstanding can have
several unintended consequences.
What if actual progress on activity A is better than anticipated? If on
day 4 we report that activity A is 50 percent done (5 days progress and
5 days remaining duration), we would then need to adjust the start and
completion dates of all successors of A. In fact, such information will
impact only those successors of A following A’ s completion. Although we
thought we told the computer to start after A is 50 percent done, which
now is day 5, the computer will blindly schedule activity B to be incapable
of starting until day 6, as shown in Figure 11.1.2. Similarly, if on day 5
we report that activity A is only 20 percent complete (2 days progress and

Start B after 50% of A is complete Day 5 data day

A 10 B 10 A

B What we meant
SS 5
B Computed result
12 3456789012345
123456789012345

Figure 11.1.2 Ambiguous language of PDM.

 Respecting the Power of PDM 163

8 days remaining duration), the computer will blindly schedule activity

B to start on day 6 rather than properly reporting that the start of activ-
ity B will be delayed until day 8. This illustrates only one of the many
“sources of misunderstanding” common in PDM. And as different prac-
titioners and computer software writers add features to the basic system,
additional interpretations and misunderstandings will occur.

11.2. Defining Overlapping Activities: Durations between Activities

As discussed in Chapter 10, PDM permits logic relationships other than
FS. Some of the relationships available in theory and supported to vary-
ing degrees by software vendors are shown in Table 11.2.1. If lag factors
are included, or the number of time units between, say the finish of A and
start of B, the possible number of relationships expands, as illustrated
in Table 11.2.2. The distinctions between some of these types of rela-
tionships may appear obscure, but as shown in Table 11.2.3, the same
problem can have differing results based upon which type of start-to-start
or finish-to-finish relationship is used.
The MSCS program (Management Scheduling and Control System by
McDonnell-Douglas Aircraft Co.), written in the 1960s to run on a main-
frame computer, understood some of these nuances and had separate
codes for two types of lag, passage and progress. However, the differences
between their designations of the start-to-start and the begin-to-begin
codes and the finish-to-finish and end-to-end codes were poorly docu-
mented and subject to misinterpretation.
The list of Table 11.2.2 is not complete. Referring back to Figure 10.4.1
suggests a restraint based upon measured units performed or installed,
such as between Activity 100 and 110 or 520 and 530, which call for the
second activity to start after 0.50 cubic yards of excavation and 0.10 square
feet of clay tile are respectively complete. This is the information that
would be provided by the project superintendent performing the work. In
ADM, this information may be retained. In PDM, as supported by com-
mercial software, the conversion of 0.50 CY to a start-plus-lag-to-start
format will not convey this information to the project team.
The migration of software to microcomputers in the 1980s entailed,
at that time, severe memory limitations and resulted in the progress def-
initions being dropped from many software programs. Lag measuring

TABLE 11.2.1 Types of PDM Relationships

FS Finish to Start Activity B may Start after Activity A is Finished

SS Start to Finish Activity B may Start after Activity A is Started
FF Finish to Finish Activity B may Finish after Activity A is Finished
SF Start to Finish Activity B may Finish after Activity A is Started
 TABLE 11.2.2 Explanded Set of PDM Relationships
164

RDM P3 MSCS Formula Description

FS FS C ESb = EFa + Lagab FINISH to START

Progress FS “n” Finish to Start
B may start “n” units after duration of A is
reduced to zero
ES FS∗ ESb = AFa + Lagab END to START
Passage FS “n” Finish to Start
B may start “n” units after A reported finished
SS SS B ESb = ESa + Lagab START to START
Passage SS “n” Start to Start
B may start “n” units after the Early Start of
remainder of A, if A reported started
BS SS∗ ESb = ASa + Lagab BEGINNING to START
Passage SS “n” Start to Start
B may start “n” units after A reported started
PS — S for Lagab ≥ ODa; PROGRESSED to START
ESb = ESa – ODa + RDa + Lagab Progress SS “n” Start to Start
B may start after duration of A is reduced
by “n” units
RS — — for Lagab ≥ ODa; REMAINING to START
ESb = ESa + RDa – Lagab Progress SS “n” Start to Start
B may start after duration of A is reduced
to “n” units
FF FF E EFb = EFa + Lagab FINISH to FINISH
Passage FF “n” Finish to Finish
B may finish “n” units after A is reported finished
EF FF∗ EFb = AFa + Lagab END to FINISH
Passage FF “n” Finish to Finish
B may finish “n” units after A is reported finished
 FR — F for Lagab ≥ ODa; FINISH to REMAINDER
EFb = EFa + min(RDb, Lagab) Progress FF “n” Finish to Finish
B may finish “n” units after duration of A is
reduced to zero
FP — — for Lagab ≥ ODa; FINISH to PROGRESSED
EFb = EFa + min(RDb, (ODb – Lagab)) Progress FF “n” Finish to Finish
B may continue beyond “n” units to finish after
duration of A is reduced to zero
SF SF W EFb = ESa + Lagab START to FINISH
Passage SF “n” Start to Finish
B may finish “n” units after Early Start of
remainder of A, if A reported started
BF SF∗ EFb = ASa + Lagab Passage SF “n” Start to Finish
B may finish “n” units after A reported started
PR — — for Lagab ≥ ODa and Lagab ≥ ODb; PROGRESSED to REMAINDER
EFb = ESa – ODa + RDa + Lagab1 + min(RDb, Lagab2) Progress SF “n” “m” Start to Finish
NOTE TWO LAGS ARE REQUIRED B may finish “m” units after duration of A is
reduced by “n” units
RR — for Lagab ≥ ODa and Lagab ≥ ODb; REMAINDER to REMAINDER
EFb = ESa + RDa – Lagab1 + min(RDa, Lagab2) Progress SF “n” “m” Start to Finish
NOTE TWO LAGS ARE REQUIRED B may finish “m” units after duration of
A is reduced to “n” units
RP — for Lagab ≥ ODa and Lagab ≥ ODb; REMAINDER to PROGRESSED
EFb = ESa + RDa – Lagab1 + min(RDb, (ODb – Lagab2)) Progress SF “n” “m” Start to Finish
NOTE TWO LAGS ARE REQUIRED B may continue beyond “m” units after duration
of A is reduced to “n” units
PP — for Lagab ≥ ODa and Lagab ≥ ODb; PROGRESSED to PROGRESSED
EFb = ESa – ODa + RDa + Lagab1 + min(RDb, (ODb – Lagab2)) Progress SF “n” “m” Start to Finish
NOTE TWO LAGS ARE REQUIRED B may continue beyond “m” units after duration
of A is reduced by “n” units
165
166 The Theory of CPM Planning and Scheduling

TABLE 11.2.3 Differing Results Based upon Choice of Relationship Code.

Graphic Code Formula Calculation

PS ESb = ESa – ODa 10 − 10 + 5 + 7 = 12

A
10 5 + RDa + Lagab
B
SS 7
8
SS ES + Lag 10 + 7 = 17
AS = 6 DD = 10
BS AS + Lag 6 + 7 = 13
A
10 0
FF 7 FR EFb + EFa + min(RDb, Lagab) 20 + min(6, 7) = 26
B

AS = 6
8 6
FF EF + Lag 20 + 7 = 27
AF = 18
AS = 15 DD = 20 EF AF + Lag 18 + 7 = 25

Example for Start-to-Start: Activity A has 10 day original duration, started on day 6,
has 5 days remaining duration. It is linked to Activity B by a Start-to-Start restraint
with a 7 day lag. Activity B has a 8 day original duration. The update is on day 10.
■ PS algorithm starts at early start (data date) and reduces lag by 5 days to match the 5 day
reduction in activity duration, with resulting calculation of early start of B being day 12.
■ SS algorithm starts at early start (data date) and adds 7 days lag (unless lag is reduced
by update input,) with resulting calculation of early start of B being day 17.
■ BS algorithm starts at actual start date recorded as day 6 and adds 7 days lag, with
resulting calculation of early start of B being day 13
Example for Finish-to-Finish: Activity A was started on day 6 and finished on day 18.
It is linked to Activity B by a Finish-to-Finish restraint with a 7 day lag. Activity B has
a 8 day original duration, started on day 15 and still has 6 days remaining duration.
The update is on day 20.
■ FR algorithm starts at early start (data date) plus zero remaining duration of Activity
A and then reduces lag by 2 days to match the 2 day reduction in Activity B duration,
with resulting calculation of early finish of B being day 26.
■ FF algorithm starts at early start (data date) plus zero remaining duration of Activity
A and adds 7 days lag (unless lag is reduced by update input,) with resulting calcula-
tion of early finish of B being day 27.
■ EF algorithm starts at actual finish date recorded as day 18 and adds 7 days lag, with
resulting calculation of early finish of B being day 25.

days passage rather than days progress continues to mystify and plague
users of scheduling software today. One example often encountered is
when project personnel updating a schedule enter a percent complete
or remaining duration for an activity, but do not have the time or infor-
mation to note the start and completion dates for that activity.
This can create a problem if the network includes nontraditional rela-
tionships. Assume Activity A had a 10-day duration and a SS relationship
with a 5-day lag to Activity B. Assume Activity A now is 99.9 percent com-
plete and has a remaining duration of 0, but that an actual start date has
not been entered. Now we update and reschedule. The computer does
exactly what it is told. Five days after the reported start date, or since there
is none, 5 days after the data date, Activity B can start (Figure 11.2.1).
These special, nontraditional relationships require that start and
finish dates are entered to properly calculate the schedule. So although
 Respecting the Power of PDM 167

Start B after
A
A 10 0 B 10 50% of A is data
complete date B
123456789012345
SS 5

Figure 11.2.1 Updating the PDM network.

the software will allow the user to enter progress without reporting
actual start and finish dates, if the network has relationships other
than the traditional FSs, incorrect results can be calculated.
Even without the additional problems caused by including reported
actual start and finish dates in the calculation algorithms, additional effort
is required to update networks that include nontraditional relationships
with lags. In fact, many software products do not include the measurement
of lag from a reported actual start date but instead measure from the latest
data date. Some programs, such as Primavera, permit the user to choose
whether to measure from the actual start or early start (being the data date
for the first activity not already completed; Figure 11.2.2).
If actual dates are not used, then the lag must be manually updated
whenever the duration of an activity is updated. Here, even if an activity

Figure 11.2.2 Actual start versus early start options.

168 The Theory of CPM Planning and Scheduling

A need never finish

A 10
B
FF 5
B 10
SS 5 may start anytime C

C 10 C may start before NTP

A is not required for project completion
Figure 11.2.3 Orphaned relationships.

is started, completed, has reported actual start and finish dates, a remain-
ing duration of 0 and percent complete of 100 percent, if the lag is not man-
ually reduced to 0, successor activities will be scheduled based on the data
date plus the lag. The computer will accept what we just said and perform
its calculations accordingly.
Another common problem with nontraditional logic relationships is
their failure to ensure that each activity has a predecessor before its
start and a successor after its finish. This problem is exacerbated with
advanced software that plots the activity as a bar on the computer
screen when data relating to the activity is entered.
Reviewing Figure 11.2.3, if Activity A is connected to successors only
by means of start-to-start relationships, then its finish is not required
according to the logic of the network. Similarly, if Activity C is con-
nected to predecessors only by means of finish-to-finish relationships,
then it may start at any time, even before the notice to proceed, begin-
ning with the first day in the project calendar.
Some software programs address this problem. For example, the MSCS
program, with a combination or joint relationship code, used a “Z” code to
combine a start-to-start and a finish-to-finish relationship with similar lag.
Since this is a popular use of nontraditional relationships, it allows this
combination relationship to be designated with one entry and reduces the
chance of creating orphan activity starts or finishes.
Note that the example given in Figure 11.2.4 will only properly work
if the duration of A and B are the same, keeping the two activities in
lockstep. If the two durations are not the same, one of the two relation-
ships will be overridden by the other. For example, assume that two

Figure 11.2.4 “SS5” + “FF5” code equals MSCS

“Z5” code.
 Respecting the Power of PDM 169

A 10 C 12
FF 6 FF 5

SS 5 SS 6
B 12 D 10
Figure 11.2.5 Passage type of lags. Note only one lag will drive second activity if durations
are not equal.

activities of unequal duration are linked with the instruction that the
second may start when the first is 50 percent complete and the first
must finish before the final 50 percent of the second can be accomplished.
In Figure 11.2.5, the total duration of A plus B is the greater of 10 + 6 = 16
or 5 + 12 = 17. Thus, the FF6 lag may be ignored in the initial schedule.
However, if A takes longer than expected and has an actual duration of
15, B may still be expected to start 5 days after the start of A, but will
not be able to finish until 15 + 6 = 21 days after the start of A.
Similarly, the total duration of C plus D is the greater of 12 + 5 = 17
or 6 + 10 = 16. Thus the SS6 lag may be ignored in the initial schedule.
If the actual duration of C is better than expected, the SS6 lag may
become the driving relationship.
The two examples also highlight an important point that may be
missed if the lags measure passage of time rather than progress of
reducing the original duration of the activities. The expressed wording
of the project manager is that the SS lag relates to the first activity and
the FF lag relates to the second activity. This is more clearly depicted
in Figure 11.2.6, which displays the lags as progress in reducing the orig-
inal duration. Unlike the example in the prior figure, here the start of
B will be earlier if A is performed faster than anticipated. Similarly, for
a progress style lag, should activity D start and progress partially out
of sequence such that it has only a remaining duration of 4 by the time
C is complete, the lag will likewise be reduced from 5 to 4 days.
Because the SS lag relates to the predecessor activity while the FF
lag relates to the successor activity, a special problem exists for the SF
relationship. If we assign to a SF relationship a passage type of lag, the
lag merely represents how the two bars line up on the bar chart. On the
other hand, if we assign to a SF relationship a progress type of lag, we
need two lags, one to measure from the start of A and the second to meas-
ure to the end of B, as shown in Figure 11.2.7.

SS 5 A 10 SS 6 C 12

B 12 FF 6 D 10 FF 5
Figure 11.2.6 Progress type of lags.
170 The Theory of CPM Planning and Scheduling

A 10 A 10
SF 11 SF 5 6

B 12 B 12
Figure 11.2.7 Examples of passage versus progress type of start-to-finish lag.

11.3. Negative Durations between Activities

In addition to durations of activities and durations between activities
(or lags), there exists yet another possible type of duration in PDM con-
sisting of a negative duration. This may be used to convert the stated,
“Activity B may start 3 days before the Activity A is finished” to the lead
lag expression “FS –3.” However, even though such a statement may
commonly be made, and such a lead/lag may be used to “move the bars
on the screen” for the initial schedule printout, the comment and expres-
sion are inherently logically flawed.
The actual comment should be heard as “Activity B may start 3 days
before the expected finish of Activity A” and the same logical informa-
tion is better provided by a start-to-start restraint with a lag equal to
the original duration minus 3 days. In either case, the actual comple-
tion or finish of Activity A is not required by Activity B (nor any other
activity unless there are other restraints). Thus, although there appears
to be a finish-to-something relationship required by PDM, the reality is
that the finish of Activity A is orphaned, as shown in Figure 11.3.1.
Obviously, if the finish-to-start negative lag is greater than the dura-
tion of the activity, all that is being accomplished is the movement of bars
on the bar chart. Similarly, a start-to-start, or finish-to-finish or start-
to-finish relationship with negative lag does not have a physical refer-
ent and is “just moving bars.”

11.4. Remaining Durations between Activities

The very first enhancement to the Kelley-Walker basic CPM algorithm,
discussed in Chapter 9, is the distinction between original duration and
remaining duration. This added feature allows the Scheduler or other
individual reviewing an update to the CPM to note the status of work in

SS 7 A 10
FS − 3

SS 7
B 10
Figure 11.3.1 Finish-to-start negative 3
equals start-to-start plus 7.
 Respecting the Power of PDM 171

A 10 C 10

SS 7
B 10

Figure 11.4.1 The problem of 2 days of progress

over 5 days of time—when can B start?

progress, and further to compare the reduction in duration versus the

passage of time since the start of the activity. It would seem obvious
that similar tracking may be desired for the lags or durations between
activities.
As an example, assuming 2 days have been performed of activity A,
as in Figure 11.4.1, the Scheduler could note from the tabular printout
that 8 days of work remain for Activity A, after which its successor C
may start. However, the Scheduler could not determine the number of
days remaining until Activity B may start without solving an arith-
metic problem mentally or scribbled in the margin of the report. To
some extent this may be an artifact of the dichotomy between the two
definitions of the lag duration, that is, deducting the number of work
days of Activity A performed or counting the number of days from the
update data date to the early start of B. Either way, this is important
information if the lags are provided for reasons other than simply
moving bars on the Gantt chart view.

11.5. Impact of Percent Complete Upon Durations between Activities

If the lag in Figure 11.4.1 represents 70 percent of the performance of
Activity A, then when performance of Activity A reaches 50 percent,
this will also represent that 71.4 percent of the portion of A required to
start B is complete and only 2 days or 28.6 percent of that portion of A
remains incomplete. Note how this example explicitly portrays the basis
of the theory of PDM. There is a definable 70 percent portion of Activity
A preceding B rather than just any 70 percent of Activity A. Even if the
passage definition is used, stating that Activity B may start 7 days after
the start of A, there is an implicit understanding that the component
subtasks of A must be performed in a certain sequence, and that 5 days
after the start of A, 71.4 percent of these subtasks will be complete and
only 28.6 percent remain to be performed.
But what if the reality on the jobsite is that progress is being made
on Activity A, but not on the portions required to start Activity B? If the
lags are anything other than simply to move bars on the Gantt chart,
172 The Theory of CPM Planning and Scheduling

there should be some means to decouple the reporting of progress of A,

or passage of time since A started, from the countdown to B.

11.6. PDM and Hammocks

The use of hammocks in PDM creates a special problem of semantics.
A hammock is a summary activity from one event or point in time to
another, such events being the i node and j node of an activity in the
ADM system. The PDM system does not have such nodes. It truly does
not support point-in-time events. Even a milestone of zero duration is
calculated as having a separate start and finish, or must be designated
as a start milestone or finish milestone.
As a result, the machinations required to effectuate a hammock make
the proper usage of “dummies” in ADM look like “kid’s play.” Assume
from the John Doe project that we wish to create a hammock to sum-
marize the work between the completion of rough grading and the start
of building layout. In ADM this is rather simple. A hammock is created
from the j node of 2–3 to the i node of 13–14. The software then com-
putes the duration to be 27 days between these two activities (Figures
11.6.1a, b, and c).
However, in PDM the process is not quite so simple. Since there is
no event (node) following Activity #2, it is necessary to go to the suc-
cessors of Activity #2 and link to the new hammock by means of a start-
to-start relationship, as shown in Figure 11.6.2. Ideally, there should be
such a link from each and every successor the Activity #2, although the
figure shows only two successors so linked. The finish of the hammock

Figure 11.6.1(a) Hammock in ADM. Tabular view, P35.0 for DOS.

 Respecting the Power of PDM 173

1 2 3 4 5 6 7 8 9 10 11 12 13

PULL IN FEEDER:
OVERHEAD POLE LINE:6: 12 5:

14 INSTALL 14
EXCAVATE FOR INSTALL SEWER ELECTRICAL
SEWER:10: 9 AND BACKFILL:5: DUCT BANK:3:
11

13 EXCAVATE INSTALL 13
ELECTRIC ELECTRICAL
MANHOLES:1: 10 MANHOLES:5:

12 12
BUILDING
ROUGH GRADE:2: Hammock #1:27*: LAYOUT:1:
2 3 13 14

11 11
WATER TANK ERECT WATER TANK PIPING &
FOUNDATIONS:4: TANK:10: VALVES:10:
6 7

10 10
INSTALL WELL UNDERGROUND CONNECT WATER
DRILL WELL:15: PUMP:2: WATER PIPING:8: PIPING:2:
4 5 8

9 9

Figure 11.6.1(b) Hammock in ADM. Graphic view, P3 5.0 for DOS

is similarly linked to each of the predecessors of Activity #13 by means

of a finish-to-finish relationship. The hammock duration is calculated
from the start of the earliest successor to Activity #2 to the finish of the
latest predecessor of Activity #13, rather like reversing the calculation
rules for other activities.
This will move the bars on the initial Gantt chart, but will it continue
to be accurate as the project is updated? Since the calculation is from the

Figure 11.6.1(c) Hammock in ADM. Graphic view, P3 3.1 for Windows.

174 The Theory of CPM Planning and Scheduling

Figure 11.6.2 Hammock in PDM.

start of the successors of Activity #2 rather than its finish, should the suc-
cessors not start immediately after completion of the rough grading, the
calculated actual duration of the hammock will be incorrectly low.
A more serious problem occurs if we desire to start a hammock at the
end of an activity with successors that are not driven by that activity
but each has more critical predecessors. Assume that we desire a ham-
mock from end of the electrical work of “pull in feeder” to the start of
“set electrical load center.” This hammock, and its calculated duration,
may be very useful in that it shows the time period when the electrical

Figure 11.6.3 Cannot create a Hammock from Activity #12 to #300 as start of #13 is
driven by #8.
 Respecting the Power of PDM 175

Cannot create a Hammock from Activity #12 to #300 in PDM except with
Figure 11.6.4
“dummy” activity.

subcontractor will not be needed on the job site. But the only succes-
sor to Activity #12 is Activity #13, and that is driven by Activity #8. As
shown in Figure 11.6.3, it is not possible to create a hammock from #12
to #300. The workaround is to create a “dummy” activity of zero dura-
tion and meaning to carry the completion of #12, untainted by #8, to the
hammock as shown in Figure 11.6.4. (Shades of ADM!) The “fix” requires
continued maintenance through the duration of the project. When updat-
ing the project, it will now be necessary to remember to enter the actual
finish of #12 as the actual start and actual finish of #12-J (so named
because it represents the missing j node of activity #12).
It may also be noted that Activity #12J does not have a proper succes-
sor (or a finish-to-something relationship) since its only purpose is to carry
the start-to-start logic to the hammock. Thus, since the calculation mode
is to set to display the “finish float,” or LF – EF, the displayed float is mean-
ingless. The problem also carries through to the hammock that incorrectly
calculates 88 as the float, rather than the 4 days of Activity #12.

11.7. Continuous versus Interruptible Progress

Another means to alleviate this problem is to add the assumption that
CPM activities are of fine enough detail to be performed on a continuous
basis without interruptions. Thus, in Figure 11.7.1, because the last 2 days
of Activity C are restrained until day 13, the software will delay the start
until day 10.
Some programs give the user the option of utilizing this extension to
theory or not. Primavera provides this option, defaulting to continuous
176 The Theory of CPM Planning and Scheduling

AAAAA AAAAA
A5 C5
BBBBBBBBBB BBBBBBBBBB
CCC - - - - CC CCCCC
B 10
SS 5 FF 2 by theory by software requiring
continuous performance

day 1 2 3 4 5678 90123 4 5 1 2 3 4 5678 90123 4 5

Figure 11.7.1 Interruptible versus continuous progress.

but permitting the user to specify activities as interruptible. The option

is available on a per project basis and all activities within the project
are affected by the choice (Figure 11.7.2).
On the other hand, standard reports do not specify which option was
chosen (although it is specified in Primavera’s excellent diagnostic
report). Thus, reviewing only a tabular report or graphic bar chart or
time-scaled logic network will not reveal which algorithm was used.
Finally, on the topic of continuous versus interruptible durations,
note that the progress type of lags model the real world such that, in

Figure 11.7.2 Choice of contiguous versus interruptible options.

 Respecting the Power of PDM 177

Figure 11.7.3 Continuous versus interruptible duration for progress versus passage
type of lag.

Figure 11.7.3, Activity D will be performed for 5 days, rest 1 day, and con-
tinue for another 5 days for a total of 10 days worked over an 11-day
period. Use of the passage type of lag leaves totally open the question of
what portion of D cannot be performed and the timing of the rest period.

11.8. Undefined Subtasks and Relationships to Other Activities

The additional features of PDM may appear revolutionary, however,
most can be duplicated in ADM by splitting activities into greater
detail while utilizing only the traditional FS relationship. This is illus-
trated in Figure 11.8.1. Thus the 10-day duration Activity A followed
by a 5-day duration Activity B via a 3-day SS lag, and by a 6-day dura-
tion Activity C via a 2-day FF lag may be rewritten into ADM: A 3-day
duration Activity A1 followed by a 7-day duration Activity A2 and a 5-
day duration Activity B (presumably followed by some other activity),
and a 2-day duration Activity C2 preceded by a 4-day duration Activity
C1 (presumably following some other activity) and Activity A1. Finally,
if we desire that C1 and C2 be performed continuously and without
interruption, we would apply a ZFF (zero free float) constraint to
Activity C1.
The use of PDM to accomplish all of this is certainly easier. However,
the precise scope of A1 versus A2 and C1 versus C2 are unknown.

Figure 11.8.1 Conversion of PDM to ADM.

178 The Theory of CPM Planning and Scheduling

If Activity A is erecting a brick wall, is the distinction between A1 and

A2 the corners versus the middle (so that adjoining walls may be started)
or right side versus left side (so that a window frame may be installed)
or lower versus upper courses (where a scaffold is required)? (We have
ignored the other option of upper course before the lower course as we
know this to be nonsensical, but would we know this if the generalized
activity was “install instrumentation,” some of which interacted with
installations by other trades?) Also, in this example, the SS and FF
relationships are progress based rather than passage based, as we
assume that Activity B may start only when Activity A1 (or the first 3 days
of scope of Activity A) is completed, rather than merely 3 days after
Activity A has started.

11.9. Multiple Calendars

Although we discussed some of the problems associated with multiple
calendars used in ADM in Chapter 9, in PDM, the problems of multiple
calendars are raised to a whole new level. A typical application for using
a lag factor for the traditional FS (finish-to-start) relationship is form,

Figure 11.9.1 Problems with multiple calendars.

 Respecting the Power of PDM 179

pour, and cure concrete, with a duration of 3 work days and 7 calendar
days for curing. Typically, software designates one calendar associated
with the activity, which is also used to define all lag factors associated
with its successors (Figure 11.9.1).
In this case, because the cure time is 7 calendar days, we can over-
come the problem by stating the lag as 5 work days. However, if the cure
time is 2 calendar days, then it makes a difference if the pour is com-
pleted on Monday or Friday. Considering the variety of possible lag fac-
tors used with SS (start-to-start) and FF (finish-to-finish) relationships,
you can see how easily multiple calendars create multiple interpreta-
tions and misunderstandings.

Class Exercise: Discuss Preparation of a Network in ADM and PDM

Our concrete crew only works Monday through Friday. We are required to
form and pour a concrete slab, which we will pour in two segments and
which will take 20 days. We may begin forming a wall on a poured slab 48
hours after the concrete is poured. Therefore, 2 days after 50 percent of the
concrete slab is poured, we can begin forming the walls. The forming and
pouring of walls will also take 20 days.

11.10. Retained Logic versus Progress Override

In addition to the three possible answers to the question noted in
Chapter 9 on ADM networks, in theory, a fourth possible answer exists
for PDM networks. This is that some additional portion of an activity
can continue to be performed but that the activity cannot finish until
all predecessors are complete—an implied FF relationship (possibly
with some lag) included in all preceding FS relationships. In fact, this
option would alleviate many of the problems of how to address work per-
formed out-of-sequence.
An exciting extension of this concept would be for the software to
explicitly insert the appropriate FF relationships whenever out-of-
sequence work is reported, highlighted for notice to the scheduling pro-
fessional, and editable for modifying the amount of FF lag or deletion,
if appropriate. To our knowledge, neither of these two options have
been included in any of the commercial software programs available
(Figure 11.10.1).
180 The Theory of CPM Planning and Scheduling

Figure 11.10.1 Retained logic versus progress override options.

11.11. Total Float Calculation

We earlier learned that the value of total float (TF) is calculated as the
late start (LS) minus the early start (ES), which is also equal to the late
finish (LF) minus the early finish (EF ). This is because both:

ES + duration = EF and LS + duration = LF

If the finish of an activity is controlled not by the duration of the activ-

ity but by the FF relationship from another activity, you can add the
assumption of continuous, uninterruptible activities as noted previously
to determine when the desired, rather than earliest start can occur, but
the need for a true predecessor of that activity must still be addressed
in many situations.
In Figure 11.11.1, the critical path runs from A through the start of
B to C to D. B must be started on day 6 if the project is to be completed
at its earliest possible time. However, once Activity B has started, it has
2 days float. So how is the float defined for Activity B? You could choose
a start float as being the TF or a finish float as being the TF, or the more
critical of the two, a most critical float as being the critical float. Some
software explicitly states which calculation is used for determining TF;
 Respecting the Power of PDM 181

AAAAA
A5 B5 D5 BBBBB- -
CCCCC
C5 DDDDD
SS 2 1 2 3 4 5 6 78 9 01 2 3 4 5 6 78 9 0
Figure 11.11.1 Total float in PDM.

other programs require that you reference a diagnostic printout or the

reference manual for the software. Of course, when this problem is
combined with the problems of multiple calendars, an activity’s float
is closer to an opinion than a calculated number.

11.12. Erroneous Loop Errors

A final problem is erroneous reporting of loop errors for logic that would
not be correct in ADM, but is acceptable under PDM theory. Looking at
the right diagram of Figure 11.12.1, if we were to use the power of PDM
to combine the two drywall activities into one activity with a duration
of 7 days, a start-to-start relation to rough-in electrical of 2 days lag, and
a finish-to-finish relationship back to drywall of 2 days lag, we would
be saying the same as the diagram on the left. Logically, the diagram
makes sense. However, all but one of the software products we reviewed
declared a logical loop because Activity A was listed before Activity B,
which was listed before Activity A. The loop detection subroutine cre-
ated for the old ADM networks failed to accommodate the new possi-
bilities of PDM.
Interestingly, because hammock activities are treated differently
for calculation purposes in Primavera software, a hammock starting
from and going to the same activity will not trigger this error message
in P3 or SureTrak, but will do so in P3e/c and Primavera Construction
(Figure 11.12.2).

Drywall Drywall
side 1 side 2 Drywall
A5 C2 both sides
5 days R/I 2 days A+C7
electric R/I
7 days
B5 electric
SS 2 FF 2
5 days SS 2 B5 FF 2
5 days
Figure 11.12.1 Erroneous report of loop error.
182 The Theory of CPM Planning and Scheduling

Figure 11.12.2 Hammock as included activity accepted by P3 and SureTrak, but not by P3e/c.

11.13. Summary
As we have seen, the Precedence Diagramming Method variant of the
Critical Path Method of schedule analysis brings a great deal of addi-
tional power to the project control team in creating a model of the real
world of scheduling. However, it also brings the capability of ignoring
the basic regimen required of the planning professional in preparing a
proper logic network, and depicting possible schedules based upon guess-
work rather than logic. This new power is there for experienced sched-
ulers to be properly used and not abused.
 Chapter

Enhancements and Extensions

12
by Software Vendors

Numerous software vendors have written programs to solve the CPM

algorithm. Many of these vendors added additional features not found
in the basic theory. In addition, many of these vendors have added fea-
tures to better utilize the calculated results, including various filtering
and sorting routines for tabular reports and for graphical representa-
tion of the plan and the schedule for the target project.
It would be impossible in a, hopefully, enduring text to discuss all or
even many of the excellent software products that are available to the
project management team for schedule preparation and monitoring.
Most major vendors develop a “new and improved” version of their soft-
ware each year. Whereas keystroke instructions tend to remain con-
stant from year to year in an effort to maintain customer loyalty, bold
changes in approach (such as Windows 95 from the Windows 3.x family)
are not unknown. Changes in the operating system (such as Windows
3.x to 95 to 98 to 2000 to XP) will also play a role in the changing fea-
tures and benefits of a software product, as will the changes in com-
puting architecture and power (such as going from 16 to 32 to 64 bit
technology). Therefore, this text discusses several of the many soft-
ware products available, with keystroke details of one system to illus-
trate how any of these (or other) systems “add value” to the basic CPM
algorithm.

12.1. Overview of Primavera Project Planner P3

The personal computer (as opposed to mainframe computer) software with
which the authors have the most experience is Primavera Project Planner
by Primavera Systems, Inc. Primavera, like many other software vendors

183

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
184 The Theory of CPM Planning and Scheduling

over the past several years, migrated to the graphical user interface (GUI)
standards set by Microsoft Windows. Thus, although earlier versions may
use a <ALT-F><ALT-Q> sequence to Quit certain functions, current ver-
sions use Microsoft Windows’ <ALT-F><ALT-X> to eXit such functions as
well as the familiar mouse point and click exit commands.
Primavera Project Planner, also known as P3, is a high-end schedul-
ing software product. To the ability to solve the CPM algorithm, P3
adds the ability to examine the CPM calculations in a myriad of views,
both tabular and graphical. The high-end software package also permits
users to view summaries of numerous projects with data that may be
located on one computer, on other computers via a network, or even on
machines connected via the web.
Primavera’s P3 product provides a very large number of configuration
options that alter the fundamental purpose of the software, the algo-
rithm for solving the CPM analysis. Users may choose to have out-of-
sequence work override the dictated logic, or to retain the original logic.
Users may choose from several option for the definition of total float.
Users may choose to have lag durations counted from the early or actual
start of the activity. In effect, the users write the rules on how the soft-
ware will solve the CPM analysis, and individuals who review printed
output must be vigilant in determining what set of rules have been
used. With the significantly increased power created by these options,
comes the need for them to be used knowledgeably and responsibly.

12.2. Overview of Suretrak Project Planner

Primavera Systems also provides and supports Suretrak software.
Suretrak, sometimes referred to as “Primavera light,” was actually writ-
ten by a different group of software programmers and the similarities
to Primavera Project Planner (or P3) are by design rather than default.
As a result of this divergent background, several small but fundamen-
tal differences exist between the two software products.
Perhaps the most significant of these is that the algorithm used to cal-
culate the CPM attributes of early start, late start, early finish and late
finish differ. Because Primavera chose to have both programs use a
common data format, it is possible for a project prepared using one soft-
ware program to be read and even updated in the other. But, depend-
ing upon the specific CPM network and specific update, it is possible that
such an update of a network in Primavera Project Planner (P3) will
yield different results than in Suretrak. These types of problems, how-
ever, will only occur if non-traditional lead/lag relationships are used in
the network. (See Figures 12.2.1 for common Base Network in P3 and
Suretrak, and Figure 12.2.2 for divergent Updated Network in P3 and
Suretrak.)
 Enhancements and Extensions by Software Vendors 185

Figure 12.2.1 Comparison of P3 and Suretrak schedule option screens.

Figure 12.2.2 Comparison of schedule output for various algorithm options.

186 The Theory of CPM Planning and Scheduling

As of the printing of this 6th edition, Primavera is continuing to pro-

vide support for, but is no longer selling, either its classic P3 Project
Planner or SureTrak software. Its sales emphasis is now upon its new flag-
ship of P3e/c Primavera for Construction and Engineering (version 5.0)
and a reduced-cost (and feature) version for smaller construction project
use, Primavera Contractor.

12.3. Overview of P3e/c Primavera Program Manager

Primavera Engineering & Construction, formerly P3e/c for Construction,
is the latest extension of its Primavera Project Planner for the
Enterprise or P3e software launched in 1999. Version 5.0 rectifies a
number of the features lost from its P3 software, including reintroduc-
tion of recording of suspend and resume dates for calculation of actual
durations, and adds new features and benefits of particular use to con-
struction industry users, such as incorporating a new means of track-
ing critical and near-critical paths and an UNDO function.
The basic premise of Enterprise software is that a company, agency
or other organization has limits to the resources that it can field and that
the intelligent allocation of these resources requires a common database
that provides the status of all projects or programs (groups of projects)
under the common umbrella. In some industries, such as IT and phar-
maceutical, the ability to be “first-to-market” trumps all other factors
in determining the success of a project. In such a case, the allocation of
scarce resources to one or two champions is preferable to each project
taking its “fair share.”
While P3e/c has added a great deal of power to group projects and to
perform rollups on resource usage by one project or by the entire group,
expanding this to allow tracking of up to several hundred projects
required a more robust database engine. Primavera’s scaling up of its
software initially to handle enterprise issues inadvertently caused some
difficulties to users of P3 software due to problems similar to those
encountered between users of P3 and SureTrak. For example, the con-
vention for determining which calendar applies to “lag” duration
between activities differs between P3 and SureTrak, which assign the
calendar of the predecessor activity, and Microsoft Project (or “MSP”),
which assigns these to the calendar of the successor activity. (The old
MSCS mainframe system allowed the user to assign any defined cal-
endar to each lag. This is another example of “lost art.”) The new P3e
system, designed to accept input from many sources (including MSP),
chose the calendar of the successor activity. Introduction of P3e/c
solved this problem by allowing the user to set a system-wide choice
 Enhancements and Extensions by Software Vendors 187

of modes. This has been supplemented in more recent upgrades to also

include a system-wide choice of the project default calendar or a 24/7/365
calendar.
The greater capabilities of an enterprise system also comes at another
cost—that of increasing the complexities of operating the system. A
crash can no longer take down only one project for a few minutes until
the file copy of the last update (on a floppy diskette) is located; rather
the entire enterprise can be taken down. Thus such larger systems tend
to require a hierarchy of computer architecture, as noted in the P3e/c
Administrator’s Guide (see Figure 12.3.1.).
The manual to explain the concepts of project management and all the
functionality of the software requires 444 pages. The manual for the
administrator runs to 469 pages. While this may seem daunting for
most construction project managers, project engineers, and casual users,
in most cases, the larger systems will be set up in a company’s head-
quarters while smaller setups, even for a single laptop, are possible via
a user-friendly wizard. The administrator’s guide is so large because it

Clients Servers Database

Primavision
application
server

Primavision
HTTPHTTPS
Progress
Projects and
reporter
resources
HTTPHTTPS
Methodologies
Browser clients
Web
server

Progress reporter group

server & job services

Desktop clients

Figure 12.3.1 Architecture of fully installed P3e/c.

From P3e/c for Construction Administrator’s Guide p 7, with permission.
188 The Theory of CPM Planning and Scheduling

has options for handling companies of the size of Bechtel, all the way
down to a small Mom and Pop subcontractor.
Some of the other features made available in the P3e/c series include
the following.
■ linking of drawings, specifications, documents, photographs, and other
“objects” to specific activities,
■ assigning of “steps” or tasks as a subset or checklist for an activity,
■ adding the functionality of the previously separate software known as
Claims Digger,
■ setting thresholds to alert a manager (who may be reviewing hundreds
of projects) of a potential problem, and
■ providing superior graphics, especially for “management by exception.”

12.4. Overview of Primavera Contractor

Primavera also launched a low-end and lower cost product to work seam-
lessly with its P3e/c software and to replace its existing entry level
SureTrak software. While SureTrak may have been called “Primavera
Lite” due to lesser functionalities, Primavera Contractor has all the func-
tionality of P3e/c but is limited to 750 activities. This limit severely
reduces its usefulness for construction projects as a typical recommen-
dation for the number of activities is one per $10,000 of value, thus lim-
iting its use to projects under $8 million. (Other formulas may increase
this to projects of double this size depending upon the typical crew size.)
While there may be workarounds to the limits set, each has serious
drawbacks. Larger or longer duration “summary” activities can be sup-
plemented by an unlimited number of “steps” or tasks, but only at the
expense of reducing the rigor of the CPM logic network. Assuming that
a subcontractor could have only his/her work on a separate CPM will
only work if the work of the subcontractor is totally independent of
work by others on the project. It is predicted that other means of limi-
tation will be developed by Primavera, such as a maximum number of
projects or a lockout, which would permit a subcontractor to load an
entire (larger than 750 activity) network but be able to update progress
and view resource and cost data on its own activities.

12.5. Overview of Microsoft Project Professional 2003

Any discussion of microcomputer software that does not discuss the
software sold by Microsoft would be incomplete. Microsoft Project for
Windows brings to scheduling all of the strong points of other Microsoft
 Enhancements and Extensions by Software Vendors 189

products, namely a standard graphical user interface (GUI), a standard

set of instructions for navigating through the various menus, and a
high level of connectivity with other Microsoft products.
Microsoft Project is a relatively low-level scheduling product. This does
not imply that it is inferior, but merely that it is attuned to the needs of
“occasional” schedulers and managers, who rarely must utilize the prod-
uct. Its learning curve is low. Few options exist relating to fundamental
issues such as the solution algorithm. The downside of a low-learning
curve and an aversion to complexity, however, is an inability (or at least
difficulty) in handling various non-standard scheduling situations.
Microsoft Project is well suited for the small or non-complex project.
In writing our own scheduling software earlier in the text, we deter-
mined that the basic algorithm is quite simple. Microsoft Project may
be used to provide the high quality of graphics noted in other Microsoft
products. And in some areas, such as extremely short-term scheduling
(to the minute), it exceeds the abilities of higher end products. Also, its
inclusion of an UNDO feature is highly desirable, if difficult to imple-
ment in larger and more complex systems.
Several of the shortcomings of Microsoft Project are founded in
Microsoft’s trait of assuming users’ desires from limited user input and
acting thereon without overrides. A prime example of this is the
assumption that the schedule will be created and edited solely upon the
computer. As a result of this assumption, Microsoft Project assigns
activity numbers as activities are entered. If an activity is later added,
it will be assigned the next sequential number, or if inserted, the task
will be assigned the activity number of the task it was inserted before,
and that and all subsequent tasks will be automatically renumbered.
Thus, the ability to refer to a task by an abbreviation, its task number,
on printed output is severely limited. If a task description should
require modification (including correcting misspellings, etc.), there is
no means to compare a recent update to a prior update or the initial
schedule. In essence, the task number is useless as a reference, other
than for the temporary program usage, in establishing additional logic
restraints.
Despite these drawbacks, for individuals who want to schedule their
own time or schedule for a limited number of subordinates on a relatively
small project, Microsoft Project is an effective tool.

12.6. Overview of Welcom Open Plan

Another of the various scheduling software products that has gained a
recognized market share is Welcom Software’s Open Plan, currently in
Version 3.1 as of the publication of this text. Joining the trend of
Primavera, Version 3.x has migrated from a project centric product to
190 The Theory of CPM Planning and Scheduling

Overview
Welcome to Open Plan 3.1. This version of Open Plan contains exciting new features
and major enhancements to features in previous versions. Several enhancements
have been made to the Welcom Security Administrator as well.
To view information about each feature, click the appropriate link in the
navigation pane at the left.
Many enhancements to this version of Open Plan are in response to Software
Fault Reports (SFRs). Where appropriate, the relevant SFR numbers are identified.
Future releases of Open Plan will not be tested for compatibility with MS
Windows 95, 98, or 98SE.

Figure 12.6.1 Open Plan opening screen.

Figure 12.6.2 Open Plan Project Properties – User may set default calendar for lags, but
not to predecessor.

Figure 12.6.3 Open Plan Activity Details dialog box at Relationships tab.
 Enhancements and Extensions by Software Vendors 191

an enterprise wide product. Thus, Open Plan 3.x provides a single data-
base for all projects versus Version 2.6 and prior versions that allowed
multiple databases. This is noted in the Open Plan 3.0 Data Migration
manual (page 3) which states, “Since Open Plan 3.0 stores its data in only
one location, duplication of file names is not allowed. There can be only
one file of a given type with a given name. Previous versions of Open Plan
allowed you to store your data in almost any drive/folder. As a result you
may currently have more than one copy of a file in different folders.”
A prime concern with the software is that the product appears to be
going through growing pains. A great deal of the material included in
the manual and on the ReadMe files of the CD on which the software
is provided deals with “minor” changes, in response to “Software Fault
Reports” reported by current users (Figure 12.6.1). However, a strong
commitment to quality may be derived from an opening screen that
points out these faults and caveats of future limitations before even
going into what the product does (for new, first time customers).
Some of the “faults” described are not so much faults as much as
questions of style that have plagued all CPM software vendors. One
example is default rule for the calendar to be used for lag durations.
While most of the construction industry has used a standard of calcu-
lating the duration based upon the calendar of the preceding activity,
Open Plan has generally set its default to the calendar of the succeed-
ing activity. This difference may be more than compensated for by Open
Plan’s ability to allow the user to set the calendar for each individual
lag; however, individually resetting 16,000 lag calendars in a 10,000
activity network may be rather time consuming (Figure 12.6.2). Also, if
using this feature, the user must remember if changing the calendar of
an activity to change the calendar of each lag following the activity. One
may next expect that Welcom would follow the example of Primavera (in
P3e/c and Primavera Construction) to allow the user to choose which pro-
tocol will be used as a default, while Primavera and Microsoft may well
consider allowing users to individually set lag calendars.
The big change has been to accomplish the migration from project cen-
tric scheduling to enterprise wide scheduling, requiring new project
naming procedures and means to migrate information from the older
system. The newer database system has other unintended consequences,
leading to various special issues for those users migrating from prior ver-
sions of the software, much of which have been corrected by Welcom’s
SFR system and notification of appropriate patches to registered users.
Going beyond these issues of administration and default standards,
the Open Plan product appears robust and relatively user friendly. The
layout of the dialog box for input and review of relationships is highly
intuitive as shown in Figure 12.6.3.
The layout of the logic network may also be displayed graphically as
depicted in Figure 12.6.4. Obviously, as the number of activities
Figure 12.6.4 Open Plan Network View

Figure 12.6.5 Open Plan bar chart showing Monte Carlo risk analysis data.
Bar chart highlights criticality index and mean dates
Green = Never Critical Yellow = 0–50% chance of being critical
Red = 51–100% chance of being critical Gray bar represents Mean Early Start – Mean
Early Finish bar

192
 Enhancements and Extensions by Software Vendors 193

increases, the Open Plan Network View will become more difficult to read,
but the implementation with curved logic restraint arrows is well done.
One feature of particular use is the automated ability to trace the crit-
ical path of a project even where multiple calendars and non-traditional
relationships cause the activities of such to have positive float, as pre-
viously illustrated in Figures 9.6.1 and 11.9.1. Finally, since Open Plan
incorporates a SPERT style Monte Carlo analysis as part of its basic
package, it creates the ability to highlight not only the most probable
critical path, but also activities that have a lesser but still high proba-
bility of becoming critical as depicted in Figure 12.6.5. (See Chapters
22.10 and 38.2 for further detail on this type of analysis.)
This page intentionally left blank
 Part

The Tools of CPM

3
Planning and Scheduling

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
This page intentionally left blank
 Chapter

Measure Twice—Cut Once

13
The mathematical basis of CPM theory has been discussed. And the fea-
tures and benefits of various software tools have been discussed. So now
we can discuss the practice of CPM planning and scheduling. As noted,
these tools should assist the entire project team to better plan and
manage a project based upon the myriad of information known at the
outset of the project and as may be further known during the course of
the project. But these tools alone, like a power saw, will only lead you
into trouble faster if you do not take the effort to implement a system
of data acquisition and a system of data verification. The old carpen-
ter’s saying, “measure twice—cut once,” is also applicable to a CPM tool.

13.1. Preparing to Collect the Input

A CPM is not a mere checklist of tasks to be performed. CPM also is not
a mere ordering of items included in the estimate. Preparing the CPM
requires a fresh look at the detail of the subject project. The level of this
review requires making decisions and it is imperative that the person
who will be primarily responsible for the execution of the work is
involved during the entire data collection phase. In construction, this
person is usually the superintendent chosen to run the project. It is the
team of this superintendent, who will provide the intelligence for the
logic network, and the Scheduler, who will draw out this information and
code it for calculation by the computer, that makes CPM the tool that
it is. Read the help wanted advertisements in the newspapers or on the
internet. It is no coincidence that the highest paid construction position
is that of the superintendent and the second highest is that of the sched-
uler. The salary for construction estimators, who provide the life-blood
for the company, may be close, but are definitely in third place.

197

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
198 The Tools of CPM Planning and Scheduling

13.2. The Pure Logic Diagram

The mathematical basis of CPM is the logic network. If you are going to
use the mathematics of the CPM logic network to schedule the project,
rather than merely display your preconceived guesses or desires of how
the project should be scheduled, it is important to focus upon the input
to the logic network and not to the output of the dates to be calculated.
Remember, it is understood that you intuitively know how to schedule
the many tasks in your project and in your life. The benefit of CPM is to
do so via an analytical methodology as a crosscheck and possible improve-
ment over the intuitive process. More importantly, by stressing the plan-
ning function to be performed before scheduling, it allows the input of
multiple parties to be factored into the final product.

13.3. A Team Effort. . . on the Blackboard or Sketch Pad

The essence of the planning function is that the planning process should
be performed without the need for the calculations to be performed by
the computer. Also of primary importance is that the professional
Scheduler cannot perform this task alone, in a vacuum. The prepara-
tion of the plan must be a team effort. The role of the Scheduler is to
elicit the information known by the various members of the team that
will be responsible for the performance of the project. This role includes
the need to nudge, cajole, or otherwise obtain information from each of
these individuals that may not seem important for this individual to per-
form his or her scope of work, but may be necessary for other members
who must interface with the performance of such scope. The role of the
other members of the team is to provide an understanding of what each
must do, what resources are required to perform such scope, and what
is or may be required from other members of the team or other third par-
ties to perform such scope. It is important for the Scheduler to draw out
all assumptions of the necessary support for the performance of each
activity from the members of the team, and it is to be expected that the
process of recording all of this information will be sloppy, involving
repeated erasures and insertions. Thus, it is suggested that the entire
process of preparing the pure logic network is best performed on a black-
board (or whiteboard with copy-to-paper capability) or sketchpad.

13.4. Format for Ease of Data Collection versus for Ease of

Data Entry to Chosen Software
The absolute need for the Scheduler to determine the variety of infor-
mation to be collected for each activity included in the pure logic net-
work was previously discussed. Although obtaining the information
 Measure Twice—Cut Once 199

ACT STGE STGT

DES SHT1

SHT2

SHT3 Lead/Lag

OD SHFT CRSZ CRTY

RESP SUBC LOC1 LOC2

Figure 13.4.1 Large node format: activity in box with codes.

needed for “other team member’s use” during the planning session may
be deemed akin to “pulling teeth,” attempts to collect such information
at a later time are usually useless. Thus, the first session of the team
should be devoted strictly to what information each team member wants
from the final product for his or her constituencies (clients, customers,
immediate and senior managements, immediate and tiered subordi-
nates, regulatory agencies, and other “interested” parties.) A format to
record all of this information should then be adopted by the Scheduler,
both for ease of recording the information and as a reminder to extract
the information before moving onto the next activity.
Some schemes for such a format include either the creation of a table,
either within a printed (or drafted) box, or placement of the same infor-
mation in a regular manner without the need for a bordered box
(Figures 13.4.1 and 13.4.2).

13.5. Bar Chart: May Be Based upon Logic,

but Is not a Logic Network
Preparation of a bar chart or charts by the various team members of
their selected portions of the works (to be combined by the Scheduler) is
not only not the best practice, but may be counter-productive. Since each
of the team members, in preparing their bar charts, has (presumably)
spent some level of time and effort in the proper placement of the bars,
there will be a natural reluctance to repeating all of the thought processes

Title Excavate
ACT# 1010
OD cal shft crsz crty 10 1 1 4 EX
resp subc locl loc2 SS ... etc. ...
stage sht1 sht2 sht3 5
Title
lead & lag 1020 3

Figure 13.4.2 Small node format: activity on arrow with codes.

200 The Tools of CPM Planning and Scheduling

just expended. Moreover, many construction professionals view schedul-

ing solely as the scheduling of their crews, and only afford a cursory cross-
check that the physical infrastructure (upon which he/she may direct the
crew) will be ready and available to work upon. Thus, the concept of plan-
ning and scheduling is reduced to only that of scheduling.

13.6. Logic Restrained Bar Chart

A logic-restrained bar chart is a slightly better precursor to preparing
a CPM in that some of the thought behind the placement of bars is
recorded, but it is still not the best practice and may continue to be
counter-productive.

13.7. Freehand
As suggested previously, perhaps the most efficacious method for record-
ing the information acquired in this collaborative effort is a freehand dia-
gram on the blackboard or sketchpad. Activities can be placed in boxes with
activity data in preset positions within the box or represented by lines
(arrows) with the activity data arranged around the arrow, as depicted in
Figures 13.4.1 and 13.4.2. One advantage of the arrow method is a greater
ability to sneak an extra activity between two previously drawn items. If
non-traditional relationships (other than finish-to-start) or lags (dura-
tions between activities) are used, some care must be used to distinguish
a line representing an activity from that representing additional logic.
Remember that you, as the Scheduler, are in charge of this phase
of data collection and that any internally consistent means of recording
the data is acceptable so long as it will permit you to later transcribe the
data to a computer format for calculation. Thus, the freehand drawing
need not strictly comply with the rules relating to drafting according to
the standards of PERT, the ADM or PDM variants of CPM, the special
notations of GERT networks, or custom variations.

13.8. PERT
Several examples are provided in this section, ranging from the uncer-
tainties best recorded by PERT, to traditional ADM, to the more pow-
erful PDM, to a GERT network, to a format that may record all

Notice to 3 5 10 Foundation 25 30 35 Foundations 123 Foundations

proceed excavated poured backfilled

Figure 13.8.1 PERT—Events (deliverables) known, scope of activities (to achieve

deliverables) uncertain.
 Notice to Foundation North FDN North FDN North FDN North FDN North FDN
proceed excavation form rebar pour cure strip
0 5 6 2 1 2 1
East FDN East FDN East FDN East FDN East FDN
form rebar pour cure strip
6 2 1 2 1

South FDN South FDN South FDN South FDN South FDN
form rebar pour cure strip
6 2 1 2 1

West FDN West FDN West FDN West FDN West FDN Foundation
form rebar pour cure strip backfill
6 2 1 2 1 2

Figure 13.8.2 ADM—Events and activities known, relationships simple.

201
202

Notice to Foundation North FDN North FDN North FDN FS 2 North FDN
proceed excavation form rebar pour strip
0 5 6 2 1 1

East FDN East FDN East FDN East FDN

form rebar pour strip
6 2 1 FS 2 1

South FDN South FDN South FDN South FDN

form rebar pour strip
6 2 1 FS 2 1

West FDN West FDN West FDN West FDN Foundation

form rebar pour strip backfill
6 2 1 FS 2 1 2

Figure 13.8.3 PDM—Activities known, relationships complex. (Sole methodology currently supported by Primavera software.)
 Notice to FDN over- Select fill & Pass North FDN North FDN North FDN FS 2 North FDN
proceed excavation compacted Compaction form rebar pour strip
000 3 5 10 124 test 568 123 112 112
Fail

East FDN East FDN East FDN East FDN

form rebar pour strip
568 123 112 111

South FDN South FDN South FDN South FDN

form rebar pour strip
568 2 112 1
FS 2
West FDN West FDN West FDN West FDN Foundation
form rebar pour strip backfill
568 123 112 FS 2 111 123

Figure 13.8.4 GERT—Choice of path to be taken.

203
204

Notice to Foundation North FDN North FDN North FDN North FDN North FDN
proceed excavation form rebar pour cure strip Physical
000 3510 568 123 112 222 111

CP crew Rod crew Concrete crew Labor crew

East FDN East FDN East FDN East FDN East FDN
Physical form rebar pour cure strip Physical
568 123 112 222 111

CP crew Rod crew Concrete crew Labor crew

Forms #1
South FDN South FDN South FDN South FDN South FDN
Physical form rebar pour cure strip Physical
568 123 112 222 111

CP crew Rod crew Concrete crew Labor crew

Forms #2
West FDN West FDN West FDN West FDN West FDN Foundation
Physical form rebar pour cure strip backfill
568 123 112 222 111 123

Figure 13.8.5 RDCPMTM—Recording reasons for relationships.

 Measure Twice—Cut Once 205

necessary data for an implementation of RDCPMTM . A small foundation

is to be excavated, possibly requiring over-excavation and compaction
(and testing of the compaction,) and four foundation walls are to be
formed, set with reinforcing steel, and poured. A maximum of two sets
of forms is desired by the contractor. Upon completion, the foundation
may be backfilled (Figures 13.8.1 through 13.8.5).

13.9. Summary
The collection of data for the CPM logic network must be a team effort
led by the experienced Scheduler. Advance preparation of lists of activ-
ities or bar chart schedules may be counter-productive as the team
member performing such work may then be reluctant to repeat such
drudgery as part of the team effort. The exact format used by the
Scheduler to record the information acquired is not so important as the
need for consistency and ability for this (or another) Scheduler to tran-
scribe these notes to the selected software product.
This page intentionally left blank
 Chapter

Choosing Codes
14
The usefulness of the schedule produced by the CPM software from the
logic network will be determined to a large degree by two distinct fac-
tors: the validity of the input and the ability of the Scheduler to fash-
ion the output in a format that will be understood by management and
that will, in fact, assist those individuals charged with performance. Your
choice of appropriate codes at the outset of data acquisition is the key
to both of these endeavors. As the team develops the logic diagram and
determines appropriate estimates of duration for each activity, a myriad
of detail will be examined. If the Scheduler does not note these details
at the time they are first being determined, it is unlikely that any
member of the team will be inclined to go back and retrace the steps
taken. Thus the first step to be taken, before asking “what comes after
the notice to proceed,” is to set aside an hour to discuss who will be using
the CPM output and what these individuals would like to see.

14.1. Calendar
Many, if not most, projects are developed using only one calendar, being
either a 7-day per week calendar or a 5-day per week calendar or a 5-day
per week with major holidays excluded calendar. Durations of activities
that do not match this one calendar are modified accordingly. The incon-
sistencies that occur using this approximation are usually less than the
tolerance of error of the estimates of duration. This merely repeats
the comment that the printed dates calculated by the CPM software are
mere approximations and should not be taken literally.
For example, if a duration (or lag) is included for the curing of con-
crete for 7 calendar days, such may be noted as requiring 5 work days.
If those 5 work days extend over Thanksgiving weekend, which holiday

207

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
208 The Tools of CPM Planning and Scheduling

is often given 2 days, the total duration may equal 9 calendar days and
print the early start of the successor accordingly. This is not to suggest
that the contractor would wait until this printed date to perform the next
step of work. An inspector citing the printed CPM date as a reason to
deny the contractor permission or support for continued work would be
as wrong as one who rejected a 2 × 4 stud for being only 11/2'' × 31/2'' in
dimension.
If multiple calendars are used, there should be some careful thought
provided prior to beginning the process of developing the CPM. After all,
the durations of the activities, and possibly the logic, will be determined
by these decisions. The first comment is that if there is the possibility that
this project will be “rolled up” or merged with other projects for an enter-
prise view, you may wish to create several company-wide standard cal-
endars. Some of the software systems available will understand only one
Calendar #1 and will assume that the Calendar #1 of multiple projects
being merged (or “rolled up” for senior corporate review) is the same.
(P3e/c appears to have corrected this error by giving the user the choice
of enterprise-wide global calendars or project-specific calendars.) However,
beyond this is the use of the CPM by members of the project team and
their understanding of multiple calendars. It is recommended that
Calendar #1 (both globally and for each project) be the standard 5-day per
week with major holidays excluded calendar and Calendar #2 be a 7-day
per week with no holiday calendar.
In P3e/c there is the option to choose which calendar will be assigned
as a default. The drop-down box choice of calendar and long-name
description of calendar reduces the potential confusion of designation
of Calendar #1 or #2 or #3, but also increases the width of the column
required to display this information and may increase the effort required
to enter the information.
This too can be ameliorated by renaming your standard calendars as
“5d,” “6D” and “7D,” renaming the column from “Calendar” to “Cal” and
adjusting the column width appropriately (Figure 14.1.1).
In defining and choosing special calendars or weather calendars it is
important to differentiate between anticipated conditions and unantic-
ipated conditions. For example, in the northern portion of the United
States, there are contractual restrictions to performing certain types of
work during the winter months. Other work can be performed during this
period but may be subject to a loss of productivity. Yet other work can be
performed during this period without hindrance by the weather. If, for
example, certain work is prohibited or not capable of being performed
during the month of February, such as placement of an asphalt surface
coat (either due to temperature restrictions or a general closure of asphalt
plants), it is proper to have an “Asphalt Winter Calendar” with zero work
days in February. But if contractors in a specific locale have experienced
 Choosing Codes 209

Figure 14.1.1 Multiple calendars in P3e/c.

between 7 and 13 days over the past 5 years during which concrete
cannot be poured, the correct number of “weather days” to incorporate
in a “Concrete Winter Calendar” is 10 and not 13. While the contract doc-
uments (or common law) may provide that the contractor receive an
extension of time if the number of “weather days” exceed 13, it is not
anticipated that the contractor will experience 13 days of “weather” in
an average year.
The important element is to not add contingency to contingency. If the
durations provided by the superintendent or the team provide for the
contingency of an occasional “weather day” in June, the standard “Outdoor
Calendar” should not duplicate this contingency. Since one of the main
purposes of the CPM is to advise all parties in advance of when they need
to provide support for the work next being performed, a great deal of con-
tingency spread throughout the project increases the likelihood that the
anticipated date of performance will be a self-fulfilling prophesy. As pre-
viously noted, contingency for events with dates of occurrence that cannot
be specifically determined in advance but may generally be anticipated,
such as the likelihood of a hurricane in Florida in October, or the known
chance of delay to repair a crane, belong at the end of the project as a proj-
ect contingency. Remember that the CPM calculates the early start, or the
earliest date on which the activity may first be performed, and advises
all parties to prepare to capitalize upon that target if it is achieved.
210 The Tools of CPM Planning and Scheduling

14.2. Deliverable and Responsible Entity:

SHT1, SHT2, RESP, and SUBC
Most construction projects are tied to a set of contract documents includ-
ing plans and specifications. Presumably the activities that will be
included in the CPM logic network will be determined by reviewing these
plans and specifications. Remembering that activity descriptions are
merely abbreviations of these discrete “scopes of work,” it only makes
sense to provide a tie back to the construction plans or specifications
being reviewed as that abbreviation was worded. In most cases in the con-
struction world, a reference to one plan, or perhaps one plan and one sec-
tion, is sufficient to point the users of the CPM to those documents with
the scope that is included in the activity description. In some cases, such
as highway work where there is a specified sequence of construction, a
third set of drawings or traffic control plans (TCPs) may be referenced.
Or perhaps a separate set of environmental drawings need to be refer-
enced to properly perform each activity. In the manufacturing and main-
tenance turnaround worlds, there may also be the need for multiple
instructions for each activity. For example, in one logic network for main-
tenance procedures required in refueling of a nuclear power plant (where
downtime can cost $50,000 per hour), the manuals of seven separate
agencies had to be referenced for each activity. This is not to suggest that
every drawing reviewed needs to be noted. Where a feature is covered
by one drawing spread over several sheets, perhaps only the first in the
series needs to be noted. Moreover, referencing of the engineer’s draw-
ings compliments the engineer and indicates that the contractor intends
to honor the engineering effort included in the contract documents.
Who is responsible for performing the activity is also important. If nei-
ther the owner nor the contractor takes responsibility for seeing that some
activity is performed, the chances that it will be performed in a timely
manner are somewhat reduced. A responsible party code should be lim-
ited to very few possible choices: the owner, the prime contractor, and per-
haps a third-party utility or other agency. In some states, multiple prime
contractors may be at this level. The next level is the subcontractor level
and may include a multitude of entities. By assigning a code for a sub-
contractor to each activity, it will then be possible to print out a schedule
including the activities to be performed by that subcontractor. This, in turn,
may further encourage the subcontractor to use the CPM as a tool.
Similarly, specifying the engineer as a subcontractor to the owner
reinforces for all parties the correct lines of authority and ultimate
responsibility for deficiencies of subordinates. Encouraging the engi-
neer to replace the generic “ENG” code with codes noting specific depart-
ments or even individuals who will perform the stated tasks will also
aid to make the CPM a tool of the entire project team.
 Choosing Codes 211

14.3. Key Resources: CRTY, CRSZ, MHRS, SUPV and EQUIP

The durations estimated by a superintendent of team member also will
be based upon the resources that will be brought to bear upon per-
formance of the work. It is important here to not merely copy labor
hours or equipment hours from an estimate since the scope of work
defined by an activity may differ from that covered by an item on the
estimate. While the estimator may have premised a rate of productiv-
ity upon a certain crew size or certain equipment usage, it is now for the
team to determine if this should be the appropriate crew or equipment
that can be made available and be supported in the stipulated locale.
(It is not yet time to determine if the crew is available in this time
frame. We have not yet calculated the time frame at this stage of prepar-
ing the logic network and the fine tuning of leveling labor and equip-
ment usage is best deferred until we have completed the logic network
and determined the needs of the entire project.) A number of factors go
into estimating the duration of the activity, including the size and com-
position of the crew to perform the work

14.4. O.T., Night Work, and Special Supervision or Inspection

It is the job of the professional Scheduler to record all of the detail upon
which the estimation of duration is based. If a project is anticipated to
have periods of scheduled overtime or special shifts, it may be benefi-
cial to record these in a separate code field for ease of review and sum-
marization. Here again, the fine line between an engineer who exercises
discretion and a technician who merely performs by rote can be noted.
The choice between noting that certain work will entail an extended day
or will be performed at night by use of a dedicated calendar, or by nota-
tion to a dedicated code field or by notation to a log, note, or memo field,
should be based upon whether the special condition will impact the cal-
culation beyond the error inherent in the original estimate of duration
and the potential need for members of the project team to highlight or
summarize such activities.

14.5. Quantities and Rates of Productivity

The durations estimated by a superintendent or team member also will
be based upon the quantity of work to be performed. It is important here
to not merely copy quantities from an estimate since the scope of work
defined by an activity may differ from that covered by an item on the
estimate. To some extent, there may be some overlap and there may be
212 The Tools of CPM Planning and Scheduling

some portions skipped in this process. The purpose of recording this

information is to verify the schedule and not to make the schedule verify
the estimate.

14.6. Location, Location, Location

Since a good part of the rationale for preparing the CPM is to assist the
team members in the field to perform their specified portions of the
works, it may be very helpful to provide additional coding so that you
can filter or select portions of the total works. Thus, in addition to noting
the responsible team member or subcontractor to perform a specified
scope, it may be useful to note the location of the work in a large proj-
ect by area. This may include area, sub-area, alternate breakdown of
area, phase, stage, step, etc. In setting up appropriate codes for location,
different team members need a varying degree of specificity, thus, you
may desire to have codes for both broad areas and more detailed sub-
areas. You should code for both physical (e.g., second floor) and functional
locations, such as “high pressure steam system.”
For certain work, you may even wish to have an alternate designa-
tion for location, especially if the design drawings (and specialty sub-
contractors using them) use an alternate designation for location. (This
often occurs in highway work where the environmental drawings are
prepared by a separate design consultant than those of the traffic
control plans and those of highway design.) Similarly, you may wish
to provide alternate codes for the designated phasing and staging
envisioned by the disparate design teams involved in preparing the
contract documents.
However, it is of utmost importance that your pure logic network not
merely mimic the sequences noted from these various contract docu-
ments, but reflect actual independent review of each step to deter-
mine that each has a predecessor based upon physical reality and all
necessary resources to be performed. In fact, logic loops requiring “a
pipe to be placed before excavating an area before placing the pipe” are
often discovered by the designers’ proposed sequencing during this
process.
It is strongly suggested that you do not rely upon these code fields as
the primary description of the activity. In many instances, a printout or
screen view may not provide the code information and the description
of the activity must be ascertained completely from the 24 / 36 / 48 char-
acters of the given description. Abbreviation may be necessary, but each
activity description must include information on location as well as on
the scope to be performed. A good rule of thumb is that no two activity
 Choosing Codes 213

descriptions should be exactly the same (otherwise known as the unique

description rule of many specifications).
A fair degree of thought is often required prior to assigning the
appropriate coding for location. In some software systems, such as
Primavera P3 Planner, individual codes are independent and may not
be subdivided nor summarized. Thus separate alternate codes are
required to summarize work in the NW quadrant of the 5th floor of one
of several buildings on the site. A separate code will be needed for the
building, for the building’s 5th floor, for the NW quadrant of the build-
ing and for the NW quadrant of the 5th floor of the building. (It is likely
that some trades will want to summarize on NW quadrants for all
floors of the building or only the 5th floor of the building, but not for
the 5th floor of all buildings on the site). Other software systems, such
as P3e/c, allow using some logic within a code (such as location =
A5NW to refer to Building “A,” 5th floor, NW quadrant), to avoid dupli-
cate data entry. Of course the important point is to collect all the infor-
mation any member of the team may possibly want at some future time
since going back to add information to the network is time consuming
at the very least.

14.7. Budget Codes for Cost of Labor, Equipment, and Materials

It cannot be overly stressed that cost accounting, or even cost estimat-
ing, views the detail breakdown of a project from an entirely different
viewpoint from that of planning and scheduling. The primary difference
from both a mathematical and practical viewpoint is that both account-
ing and estimating feature a hierarchical breakdown that may be sum-
marized at various levels leading to one project cost. Accounting codes
may be concerned with who is performing discrete portions of the proj-
ect, for which department is this person working, to which division does
this department belong, etc. Estimating codes may be concerned with
the type of material quantity being installed or erected in place, the sub-
contract or trade that will bid and later perform the work, the CSI divi-
sion of the specification that describes the material or type of work to
be performed, etc.
Planning and Scheduling is not hierarchical. It may be suggested
that the schedule for a 1-year project be divided into months, weeks, and
even days, but a summary for scheduling purposes (rather than cost)
rarely will be by month. And when looking at a pure logic plan, prior to
calculation, the only summary desired is for a rough grouping of the activ-
ities between major interim milestones of the project. Even here, a sig-
nificant portion of the total number of activities will bypass summarization
214 The Tools of CPM Planning and Scheduling

in this fashion. While a schedule view may organize and perhaps sum-
marize activities by subcontract, the CPM is not the sum of these sum-
maries but rather the interplay of the individual activities both within
and between the various subcontracts.
A similar issue of viewpoint arises where other forms of hierarchical
design are imposed upon a CPM plan. Whenever there is a requirement
that each activity must be properly coded to one and only one resource
code, cost code, WBS, OBS, or EPS, there will be activities that simply
do not fit and the “fudge” imposed to meet the specification will mean
that the schedule will not properly model how instructions are provided
and how work is performed in the field. The indivisible base unit of a
CPM logic network is the activity, “a set of instructions given to a com-
petent foreman.” Activities must not be broken apart or combined to
meet the needs of reporting codes.
With all of this in mind, budget codes may be assigned to activities if
it is understood that it must be the reports to the departments of
resource allocation, estimating, and accounting, and those organized by
WBS, OBS, and EPS that are permitted a degree of approximation.
Otherwise, there is the risk that the project will fail, dragging all who
are involved into Chapter 11, but all under perfect control. The key is
to bend the codes and not the activities.
If the software product to be used (or specification) permits only one
budget or cost code per activity, the most significant code must be chosen,
with the recognition that costs that should go to other codes will be mis-
allocated here. Even if multiple cost codes are permitted, there is a limit
to how many codes reasonably can be assigned to an activity. If multi-
ple cost codes are assigned to one activity, cost reports will list the activ-
ity separately under each cost code (Figures 14.7.1 and 14.7.2).

Figure 14.7.1 P3 multiple cost codes.

 Choosing Codes 215

Figure 14.7.2 Multiple cost codes for one activity means activity is listed several times
in cost reports.

14.8. A Word About Codes Tied to the Activity ID

It is important to always remember that the codes that you associate with
an activity are for the primary purpose of allowing the users of the CPM
to best view the plan and calculated schedule information. The codes are
not actually part of the logic and are not used by the software in the cal-
culations performed. The purpose of the schedule is for scheduling and
if the arbitrary coding structure conflicts with a proper depiction of logic,
it is the logic that must prevail. Thus, if a location code indicates 1st, 2nd,
3rd, 4th, 5th floor, and a specific shear wall spans two floors, it must be
entered into the logic network as one activity and not as two activities.
The Scheduler will have to decide which floor code (2nd or 3rd) is to be
used for summarizing work. It is okay that the summaries may be “off.”
It is not okay to compromise the validity of the CPM logic.
This type of issue typically becomes more acute when efforts are made
to place “intelligence” within the Activity ID. Keeping in mind that both
the Scheduler and field personnel will be required to either handwrite
or key in the Activity ID number on numerous occasions, it is suggested
that the Activity ID number be kept simple and as short as practicable.
Since project personnel may desire to locate a particular activity upon
the pure logic diagrams initially prepared, it may be useful to provide
some coding within the Activity ID to locate the activity on the net-
work. However, if the pure logic network diagrams are not regularly ref-
erenced, or if the logic is modified during the project, such efforts at
coding will be wasted. Perhaps the best advice is to code the Activity IDs
to make reference by field personnel easy.
216 The Tools of CPM Planning and Scheduling

Figure 14.9.1 Primavera Alias described.

14.9. A Word About Aliases

Suppose that two location codes are used. One may be by floor, (e.g., 1st,
2nd, 3rd, 4th), while the other may be by quadrant (e.g., NE, NW, SW,
SE). P3 and SureTrak, unlike the more sophisticated P3e/c and
Primavera Contractor, do not support intelligence within a code or sub-
codes (e.g., 1NE, 1NW, . . . , 4SW, 4SE). Reports must either be organ-
ized by floor then quadrant or quadrant then floor. If it is desired that
the reports list all choices at the same level of summary, it is necessary
to use a workaround provided by Primavera called an “alias.”
In Primavera P3 Project Planner, a means exists to get around the
“Building ‘A,’ 5th Floor, NW Quadrant” coding noted in the preceding
section by the use of “Aliases.” (Figure 14.9.1 and 14.9.2)

14.10. Summary
Choosing the codes to be assigned to each activity is the first step in data
acquisition for preparing the CPM. The proper choice of codes is impor-
tant and will greatly impact the usefulness of the CPM. Some codes, such
as activity type, calendar, and responsible entity, must be coordinated
with the activity scope, description, and duration. Others, such as key
resources, overtime, and productivity, may be used to validate the dura-
tions chosen. Still others, such as location and cost, may be used to
enhance the data stream from the CPM effort, but must yield accuracy
to the “set of instructions given to a competent foreman” that may tran-
scend the sharp boundaries of such codes.
 Choosing Codes 217

Figure 14.9.2 Primavera Alias illustrated.

This page intentionally left blank
 Chapter

Acquiring Information
15
to Initial Schedule

The process of acquiring the information for the initial CPM logic net-
work appears relatively simple, but it is actually a fine dance between
the Scheduler, project manager, and other members of the project team.
The Scheduler must alternatively be a supportive assistant, a bit of a
psychologist, a “jack,” if not master, of many fields of construction tech-
nique and a nudge. Once the discussion over the choice of coding struc-
tures has been completed, the process starts with the seemingly
innocuous question, “What is the first activity to be performed after
notice-to-proceed?”
The Scheduler must carefully determine data for all of the code fields
for this activity—engineering drawing sheet number, crew size and com-
position, equipment requirements, quantity of work to be performed,
best estimate of duration, physical and functional location on the proj-
ect, resource codes, resource units and cost. And, then, the second ques-
tion: “What must you or others do before you can start this first activity?”
Invariably, the project manager will have skipped several steps between
NTP and the “first” activity. For example, the project manager may
state the first activity is the building foundation. But what about the
excavation for the foundation? What about clearing and grubbing before
excavation? What about erosion control before clearing and grubbing?
What about the submittal and approval of the erosion and sedimenta-
tion control plan? What about ...?
After what may be an annoying half hour or so, the team will finally
get all the way back to the notice-to-proceed. And when the Scheduler
is assured that each activity has all of its predecessors identified—phys-
ical, crew, equipment, forms, material, access, etc.—the next step can

219

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
220 The Tools of CPM Planning and Scheduling

take place. “Starting from the new first activity after NTP, what is the
NEXT activity that is to be performed other than the one(s) already
noted?” Assuming a positive answer, this must be chased back to NTP
in a similar fashion.
The process must then be repeated until reaching project comple-
tion—a long and tedious endeavor. Next, perhaps after a coffee break,
the process should be repeated starting from the first activity: “What
must others do before this activity may start? What must you do before
this activity may start?” During this step, some logic links may be
deemed to be superfluous and should be removed. The process should
be repeated until all members of the team are satisfied with the logic.
This is the practice of planning and scheduling. The computerization
of this information is merely a technical detail.

15.1. The Activity Description—a Gross Abbreviation

We have defined an activity as “a set of instructions . . . .” The activity
description, or title, used to describe an activity will, by necessity, be a
gross abbreviation. After all, depending upon the software chosen, the
description must fit within 24, 36, 48, or 64 characters. Even where an
extended description is supported by software, invariably only the first
48 to 64 characters can be viewed in a tabular listing of activities. And
yet, a whole paragraph may not be enough to fully inform the foreman,
or person to be responsible for the activity, of the scope to be performed.
Thus another skill required by the Scheduler is to be able to squeeze that
whole paragraph down to 48 characters.
The Scheduler should consider who will read these descriptions. It is of
utmost importance that the foreman, the project superintendent, and the
resident engineer understand the descriptions. Thus, abbreviations of scope
and location should be chosen based upon usage by these individuals.
Often, the abbreviations used in the contract drawings may be incorporated.
However, the choice of terminology (e.g., whether electrical installation is
split into “Conduit,” “Cable,” and “Connect,” or into “Pipe,” “Pull,” and
“Terminate”) is best left to the project manager or superintendent in charge
of that scope of work.
It is useful if the upper management of the contractor, owner, and other
interested parties can understand the description, but the CPM is meant
as a tool to help the contractor construct the project and this primary pur-
pose should not be compromised. It is important that each description be
unique and not appear elsewhere in the schedule. Although one view of
all the activities may organize all of the activities by floor, another view
may organize the activities by function. Having five activities titled
“Install light fixtures” for a 5-story structure is less than fully descriptive.
 Acquiring Information to Initial Schedule 221

15.2. Activity ID, Activity Codes and Logs

In the world of construction, not every craftsperson has immediate
access to a computer interface. Information is provided by the superin-
tendent to the foremen not by email to a desktop, laptop, or PDA, but
rather on a piece of paper. Thus, although the activity number or ID may
include some code or “intelligence,” and although additional codes and
logs or notebooks may be attached to an activity, only the information
that is on the printed page will be useful to the primary end-user. As
noted previously, the activity description must, therefore, be capable of
telling this person the location of the work and enough about the work
to distinguish it from other activities.
However, judicious use of codes, logs, and short notebook entries that
fit on the printed page in a readable manner may add to description. The
best example is tying the activity to the specific sheets of the plans or
specifications via a code. For those individuals working in fields where
team leaders are expected to have immediate access to at least a PDA,
the ability of products such as Primavera P3e/c to create a hyperlink
between the activity and underlying engineering documents is very
useful. Coding by physical location is also helpful, but this information
should also be in the description.

15.3. The Activity Further Defined by Resources Assigned

The assignment of resources to an activity, either on a nominal basis via
an activity code or a more quantitatively correct resource code, further
defines the activity to the scope of work requiring these resources. As
noted previously, it is important to remember that the resource is help-
ing to further define the activity and not expand or limit the activity
scope to the accounting code of the resource code.

15.4. The Activity Further Defined by Predecessors and Successors

We have noted that its predecessors and successors also control the def-
inition of an activity. The scope of the activity is limited to that work
that may be performed after completion of all of its predecessors and
to that work that must be performed prior to the start of any of its suc-
cessors. Although not every minor task on the project need be incorpo-
rated into an activity, some effort should be made to be assured that
any significant scope of work is included in some activity—and that just
because the title sounds right, does not mean that this activity includes
that scope.
222 The Tools of CPM Planning and Scheduling

15.5. The Checklist of Sub-tasks

Taking this principle one step further, if a list of tasks is attached to an
activity, it is important to check that each task of this list requires the
completion of all of the predecessors of the activity and is required for
all of the successors of the activity. The logic of the CPM, which is the
main strength of the methodology, should not be allowed to become
fuzzy just to allow somebody to claim that every task is included in one
of the activities.

15.6. The Checklist of Sub-deliverables (Events)

On the other hand, if it is the purpose of the project to provide a number
of deliverables and not simply one completion, it is suggested that a mile-
stone is provided for each of these sub-deliverables. For example, if con-
struction of segments of a roadway project require each mini water
basin to be secured prior to disturbance, a clear milestone indicating that
all erosion and sedimentation control activities have been completed for
each segment is suggested.

15.7. Summary
Acquiring activity information for a CPM logic network is more than just
compiling a list of activities. The first step is always to determine who
may be using the CPM and to choose appropriate codes to permit easy
dissemination of the activity information collected. The resources
assigned to the activity as well as by the predecessors and successor to
that activity will also define the scope of the activity. If an activity con-
sists of a number of discrete tasks, these may be listed in a log or note
to the activity or as Steps in P3e/c. However, it is important to list as
such tasks or subtasks only those of such scope as is within the origi-
nal set of instructions to be given the foreman or other line-level respon-
sible party.
 Chapter

16
Acquiring the Durations

The time to add the durations is as the scope of the activities are being
determined. Although some older texts suggest first making a list of activ-
ities, placing them in order, and only then determining durations, assign-
ing resources and codes, and finally adding costs, it should be obvious that
much of the work of determining the scope of the activity will be performed
not once but several times by this approach. By the time all of the infor-
mation is collected, the chances of the activity scope initially envisioned
being the same is small.
Thus, this is also the time to record all of the activity codes, resource
codes, cost codes, rough estimates of quantities, and “first rough draft” of
costs. All of this information is in the mind of the project manager as “the
next” activity is contemplated. The skill of the Scheduler is to elicit all of
this information and properly record it at this time.

16.1. Best Estimate with Utilization of Resources Envisioned

So at this point the project manager has stated “the next” activity and the
team has at least quickly glanced at the engineering drawing or specifi-
cation that most clearly describes its scope. The project manager suggests
which crew is to perform this work with what equipment and other
required resources. Although we should already know the physical
restraint leading to this “next” activity, we may also now record where the
crew and equipment and forms and other resources are coming from.
Having noted all of these resources, the project manager can then provide
a best estimate for duration for this one activity.

223

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
224 The Tools of CPM Planning and Scheduling

16.2. Compare with PERT Durations:

Optimistic, Most Likely, Pessimistic
Often the project manager may give a range of durations. This should be
encouraged. After duly recording the high and low estimates (perhaps
even encouraging an increase to the pessimistic duration), the Scheduler
should ask for the “most likely” duration. Compare this with the practice
of PERT, where the scope of the work is much fuzzier and a range of
durations is a necessity. Whether the high and low durations are recorded
to custom code fields or simply kept as part of the Scheduler’s notes, this
exercise helps alleviate the anxieties of the project manager, leading to
the provision of less padded and more accurate information.

16.3. Schedule Durations versus Estimating Durations

Although the Scheduler should strongly dissuade the project manager
from consulting the bid estimate during preparation of the logic network
and recording of durations of the activities therein, once the information
is collected for the CPM, it should be compared to the bid estimate. Either
by use of custom code fields or by exporting to spreadsheet software such
as Excel, the durations of the CPM can be roughly converted to mandays
and the mandays to manhours by the formula:

duration × crew size × 8 = approximate manhours

If the total number of manhours of the CPM is within 5 percent of the total
of the bid estimate, this should be considered a “good fit.” Keeping in
mind that the crew size code field is for a nominal size crew (which may
actually be larger or smaller or vary as the work progresses), ancillary
manpower for support may or may not be included in the nominal crew
size, the rounding to days (1 day is the minimum unit) and other round-
ing errors, an exact correlation is unlikely. It is not unusual for the bid
estimate to have more manhours than the CPM to cover support tasks
that are not otherwise included in the CPM. Subtotals by discipline are
also possible, although the level of acceptable variance should be
increased.
If the variance between the number of manhours of the CPM and bid
estimate appear to be too large, a careful review of each must be per-
formed. It is as likely that a bid error has been made as an error in prepar-
ing the CPM. Although such a situation is never welcome, it is better to
know this information going into the project than at a point 30 percent into
the project.
 Acquiring the Durations 225

16.4. Estimated Durations versus Calculated Durations

The differences in definitions and tolerances between the estimating and
scheduling disciplines should likewise point out the problems that may be
expected by systems that estimate durations based upon available
resources. In many fields, but especially construction, the choice of crew
size is an art and often, a 4-person crew may get no greater production than
a 3-person crew. Even when this is not true and a 4-person crew may be
expected to produce 33 percent more than a 3-person crew, the different
tolerances, noted previously, make it unlikely to have an exact match
between the durations calculated from assigned resources and those deter-
mined during the bid estimate.

16.5. Do We Add Contingency Here?

We encourage high and low estimates of duration, however, we do not
want to add contingency to the activity. If the project manager believes
that there is a significant chance of a major variance to the “most likely”
duration, this should be recorded (perhaps in a custom code field or per-
haps only in a log or note). For example, if the project manager estimates
duration for excavation by stating “most likely 1 week to 10 days, unless
we hit rock, in which case it could be 3 to 4 weeks,” the Scheduler should

Figure 16.5.1 Pertmaster analysis measures likelihood that a change in duration of one
activity may shift the critical path
226 The Tools of CPM Planning and Scheduling

record all of this comment in codes and logs but enter 6 work days or 8
calendar days as the duration. If a SPERT type of software is available
(such as Primavera’s Monte Carlo or Pertmaster), a separate run using
this stored information can indicate the likelihood that this contingency
will impact the project.

16.6. Estimated Durations versus Expected Completion Dates:

“as Good as the Promise”
One special problem in determining duration is when control of the activ-
ity is out of the hands of the project manager. Such a situation may exist
where a vendor promises delivery by a certain date (based upon a stipu-
lated not-to-exceed release date) but can provide no further detail or
means to check delivery progress until this activity is complete. Obviously,
the Scheduler can count the number of days to this date or use a constraint
for the software to calculate this duration. However, it must be remem-
bered that the expected completion date is only as good as the promise.
This may not seem like a significant issue at the time the CPM is pre-
pared, but it can become a more serious problem during the course of the
project since updates may then also be based upon this initial promise
unless a substantial additional effort is made with each update to verify
the promise.

16.7. Productivity
If quantities of work to be performed are recorded as the activity and
its duration are determined, yet another crosscheck may be performed.
Again, either using a custom code field or export to Excel, an attribute
of productivity or quantity per day can be calculated. Then, after group-
ing activities of similar scope, a quick visual comparison can be made
and any activities having a significantly different productivity from
the others can be examined to verify the given duration. Obviously,
this definition of productivity (being divorced from the size of crew or
other resources assigned the activity) will differ from that used by the
estimating department.

16.8. Durations and the Project Calendar or Calendars

In determining durations, it is important to understand the project cal-
endar or calendars. Is the duration for a 1-week activity 5 days or 7 days?
Should we increase durations to account for holidays? Should we increase
durations to account for seasonal weather? The use of multiple calendars
 Acquiring the Durations 227

is complicated as a result of the desire to utilize the calculated value of

float and by the additional level of complexity it causes.
If using only one calendar, the process is somewhat simplified.
Situations such as performing outside work in the “off ” season can be dealt
with by adjustment of durations accordingly; but with log or memo notes
to remind the Scheduler and others that readjustment may be required
if the project is delayed for an extended period. If using two or more
calendars, the process is somewhat more complex. If the architecture of
the software product provides for a global calendar, holidays common to
all (or even most) calendars should be provided here. Depending upon the
sophistication of the software, repeating holidays may or may not require
multiple entries from year to year.
For example, in SureTrak, it is necessary to enter each holiday sepa-
rately, such as 01JAN05, 01JAN06, 01JAN07. In Primavera P3, P3e/c and
Primavera Construction, it is possible to enter a repeating date, such as
01JAN, with a check mark in the adjoining repeat box (Figure 16.8.1).
Users moving from P3 to SureTrak and back often have difficulties, as
SureTrak will interpret the repeating holiday as a one-time occurrence
in the first year only. In the MSCS software of yesteryear, it was possi-
ble to enter as a holiday the fourth Thursday of November (Thanksgiving
in the US), but this feature is now a “lost art” and does not appear to be
available in the popular software products currently available, instead
requiring hand entry of 24NOV05, 23NOV06, and 22NOV07.
Since the global calendar will include all common holidays, the stan-
dard 5-day per week calendar should not. It is important to not duplicate

Figure 16.8.1 Global calendar and global holiday list.

228 The Tools of CPM Planning and Scheduling

Figure 16.8.2 5-Day week calendar and holiday list. Note that January 2 is calculated
as a holiday since January 1 is a Sunday.

holidays in both the global calendar and individual calendars since many
software products, such as those of Primavera, will helpfully move the hol-
iday already on a non-work day to the next non-work day (Figure 16.8.2).
Thus if 01JAN is set as a holiday in both the global and 5-day per week
calendars, and if 01JAN06 occurs on a Sunday, the global calendar will

Figure 16.8.3 7-Day calendar and holiday list exceptions.

 Acquiring the Durations 229

declare 02JAN06 a non-work day, and Calendar #1 will declare 03JAN06

as a non-work day. For the 7-day per week calendar, the Scheduler should
set the exceptions to global holidays as 01JAN to 31DEC (Figure 16.8.3).
If creating a weather-restrained calendar, it is very important to list
as non-work days only the average or even the minimum number of bad
weather days anticipated. Put another way, the weather-restrained
calendar should include as work-days the maximum number of days
that may be reasonably anticipated. The number of non-work days
should be significantly fewer than that set by various public agencies
as the maximum beyond which the contractor is entitled to an exten-
sion of time. (It should be noted that such maximums, deemed to be
beyond what the reasonable contract may expect, should be at least one
standard deviation greater than the average number of weather days
for the time period.) (Figure 16.8.4).
Care should be taken to not list as non-work days those dates imme-
diately before or after a holiday on the global calendar as noted previously.
Care should also be taken to not add contingency to contingency. Since
the calendars are not usually shown in the standard tabular printouts or
standard screen view of the schedule, such things represent another
“hidden” element to the schedule calculation and its interpretation. It is,
therefore, important that care is taken in setting up the calendars and
that the initial narrative describing the logic network and schedule dis-
cusses the calendar(s) used.

Figure 16.8.4 Winter restricted calendar standards and holiday list.

230 The Tools of CPM Planning and Scheduling

16.9. Summary
Durations for individual activities should be estimated as the activity
scope is determined along with the recording of crew size and other
assigned resources. Durations should not be based upon information in
the bid estimate. The project manager should be encouraged to give
ranges of duration concluding with a “most likely” duration. Contingencies
for foreseeable, but not expected, situations should not be factored into
the duration, but should be noted and recorded separately. Once all activ-
ities and durations have been provided and recorded, the durations can
be crosschecked against the bid estimate and against the durations of sim-
ilarly scoped activities for the purpose of validation.
 Chapter

Specifying the Relationships

17
Between Activities

A ToDo list has activities. An ordered ToDo list includes some inkling of
the relationship between items on the list. A bar chart, obviously,
includes some thought behind the ordering of the bars, but such infor-
mation is rarely recorded in a systematic way. The primary benefit of
CPM is the logic network of relationships between activities.

17.1. Mandatory and Discretionary Physical Restraints

It is a basic tenet of CPM that each activity (other than the first in the
network) must follow some other activity. It is a basic tenet of the ADM
variant of CPM that an activity may start only when each of its prede-
cessors is 100 percent complete. The same rule is applicable to the PDM
variant of CPM in a modified format, that is, an activity may start only
when some definable (if unspecified) portion of each of its predecessors
is 100 percent complete.
In the real world this means that an activity may start only when the
physical infrastructure upon which it will be built is in place and nec-
essary resources are made available for its performance. Part of the
day-to-day planning process may include finding and allocating the
resources necessary for a project and making those last minute decisions
of “who goes first” when resources are scarce, however; the requirement
for a physical infrastructure precedent to performing an activity is gen-
erally immutable. It is generally understood that each activity must be
preceded by at least one physical restraint.
In practice, this must be part of the interview process in preparing the
CPM logic. When a project manager suggests that a crew will perform

231

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
232 The Tools of CPM Planning and Scheduling

activity #1, then #2, then #3, and so forth in adjacent locations, the
Scheduler must slow the project manager down and require physical
logic (not scheduling of the crew) to be the primary consideration. Thus,
in location #1 a sequence for a foundation may be “excavate, form, rebar,
pour,” and so forth.
Perhaps as this is being done some of the resource logic may be penciled-
in, but all parties must be aware that even in the planning stage, it may
be discovered that excavation for location #2 will be more extensive and
that it is not the intent of the project manager to lay off the formwork
and subsequent crews while this more extensive excavation is being
accomplished. That is, a more economical plan may be to work #1, #3,
#4, #5, and then #2.
While the rule requiring at least one physical restraint prior to each
activity is often stated as, “You cannot put up the roof until the walls
are up,” this is subject to some discretion. There may be an instance
where a large item of equipment is prefabricated onto a skid at the fac-
tory (to reduce cost of installation) and will be delivered to the project
on a date beyond that when the walls and roof would otherwise be com-
plete. In such a case, the project manager may plan to leave a hole in
the wall and (if necessary) provide falsework to allow timely erection of
the roof prior to this late delivery date. This option requires some degree
of prior planning based upon prior notice of the problem, this being best
provided by the process of preparing the CPM.

17.2. Mandatory and Discretionary Resource Restraints

Only after all of the physical logic of the CPM has been recorded in the
interview process, and either hand calculated or keyed to the software
package of choice and the project been given its first schedule run, is it
possible to check those “penciled-in” crew restraints and add others
based upon the needs of the physical logic. Other resource restraints,
including construction equipment, forms, and materials, may be added
at this time. However, unless the resources specified are expected to be
scarce, it is probably better to leave such determinations until shortly
before the work is actually to be performed.
On the other hand, if there is an economic plan to shift forms from one
structure to the next, or move a crane the fewest number of times, or to
utilize the “A” team crew on certain activities, now is the time to add
these to the plan. The contractor is entitled as part of the contract to
marshal his forces and other resources to their best economic advantage.
This is the time to give notice of the “plan of execution” that all parties
to the contract are expected to support.
 Specifying the Relationships Between Activities 233

17.3. Mandatory and Discretionary Timing Constraints

The earliest implementations of CPM did not have the capability to
“lock-in” specific dates on which an activity may or must start or finish.
Therefore, constraints, with some exceptions, may be provided for by the
standard use of logic or restraints, such as a “timing activity” of some
number of days between NTP and a vendor delivery. This older method
did have the disadvantage of requiring manual computation of remain-
ing duration with each update, the problem being addressed by the use
of a constraint rather than restraint. However, since the hallmark of
CPM is that it uses a logic network to calculate a schedule of dates, any
use of constraints to “lock-in” such dates without the benefit of the logic
must be viewed with some degree of skepticism.
A SNET (start-not-earlier-than) constraint may be mandatory to set
the start of the project NTP (notice-to-proceed). This may be accom-
plished in some software when initiating a new project by setting the
project start date, and assuming that NTP is the first activity. A SNET
constraint may also be appropriate if access to some areas of the proj-
ect is to be delayed but is promised in the contract to be in place by a
date certain.
In the event that certain events are not expected to take place until a
promised date and further details of the activities leading to such events
are outside the control of the project manager, such as vendor deliver-
ies, a SNET constraint may be chosen as the means of entering this logic
to the CPM. However, if the Scheduler intends to use an update strat-
egy incorporating a look-ahead report, it should be understood that the
use of this type of constraint will hide the status of the delivery until
just before it is expected to arrive. The project manager is then entirely
at the mercy of the vendor’s promise. It is suggested that for this appli-
cation, to use an expected finish constraint (assuming the software sup-
ports such) to an activity of fabricate and deliver. This has the benefit
of placing the continued fabrication on each look-ahead report to alert
the project manager (or subordinate) to call the vendor at least once a
month to verify that all is going well.
A FNLT (finish-not-later-than) constraint may be mandatory to set
the deadline for project completion and also for project milestones
required in the contract to be completed by a specified date. A FNLT con-
straint may also be desired by a contractor to set internal deadlines.
Other timing constraints, such as SNLT (start-not-later-than), FNET
(finish-not-earlier-than), and those that require the activity to start or
finish on a specific date without regard to the logic are even more sus-
pect and should be carefully reviewed to determine if they are truly
appropriate.
234 The Tools of CPM Planning and Scheduling

17.4. The Misuse of Restraints and Constraints:

“Nailing the Bar Down Where It Belongs”
It must be emphasized that the purpose of restraints (between activi-
ties) and constraints (to an activity) is not to produce a schedule, but
rather to produce a logic from which the CPM software algorithm can
calculate a schedule. Use of restraints to schedule an activity to occur
on or about a specific date is inappropriate. One example noted some
years ago was a CPM submittal that tied installation of an elevator to
installation of a water cooler at the other end of the building. When the
contractor was asked the logic of this restraint, the reply was “We want
to start the elevator in mid-November; this was the only activity that
finished around then.”
Use of SNET and FNLT constraints simply to “nail down the activity
bar in the correct place on the chart” is similarly flawed. There must be
a coherent reason why the activity constrained cannot start or must
finish by the specified date that is unconnected to the status of other pre-
ceding or succeeding activities. Likewise, the use of expected finish con-
straints clearly states the project manager’s lack of knowledge about the
process leading to the “promised date” and should raise suspicions
among all who use or review the CPM accordingly.

17.5. The Need to Document the Basis of

Each Restraint and Constraint
The Scheduler does not create but rather records the logic provided by
the project manager and project team. It is, therefore, important for the
Scheduler to document not only the restraints and constraints, but the
reasoning behind each. Is a restraint due to a physical requirement or
simply to allow reuse of formwork or to suggest the preferred (but not
certain) sequence of assignment for the crew performing the work? The
software may not have an appropriate slot to record this information
(as descriptive, log and note fields are usually assigned to activities and
not to the relationships between activities) but the data may be noted in
the Scheduler’s notes or in electronic format outside of the CPM software
database. (For example, in an expanded spreadsheet having fields for
activity, successor, relationship, lag, and reason, as noted in Chapter 2.)

17.6. Choosing the Type of Relationship between Activities

Life used to be so simple when there was only one type of relationship. That
was the finish-to-start relationship provided in the original ADM version
of CPM. Theory by Dr. Fondahl and others in the late 1950s and early
 Specifying the Relationships Between Activities 235

1960s, and software by various academicians, private companies, and

computer service bureaus in the late 1960s and early 1970s, addressed the
expansion of the tools to address other types of relationships, specifically
the case where two activities overlap. Software supporting additional
means of showing two activities overlap was developed for the mass market
(on PCs rather than mainframes) in the 1980s. However, the use of non-
traditional relationships is more than being able to show the overlap of two
activities. The theory concerns the “how and why” of the overlap.

17.7. The Case for Restricting Relationships to

Traditional “FS” Without Lag
The sad fact is that the limits of the ADM variant of CPM meant that
users did not have to know or appreciate the theory behind CPM. On
the other hand, the newer non-traditional types of relationship per-
mitted by the PDM variant of CPM require this knowledge and appre-
ciation. The axiom of CPM remains, that is, an activity may start only
when some definable (if unspecified) portion of each of its predecessors
is 100 percent complete and an activity may finish only when some
definable scope of work is 100 percent complete.
If the project manager states that “Activity B may start when Activity
A is 90 percent complete,” the software does not care if the project man-
ager can articulate the tasks comprising the 10 percent of Activity A that
is not necessary for Activity B. The Scheduler, however, must demand
this information. Since the Scheduler works for the project manager, and
the project manager may not desire to be bothered with thinking out
(much less reciting) the detail required, the demand is often not met,
and after all, this detail is not required by the software so why bother.
The abuses caused by this lack of understanding, as well as by those who
deliberately misuse the power of PDM, have led many owners and engi-
neers to place restrictions in their contract documents and specifications.
For example, in the early 2000s, PADOT promulgated a new guide spec-
ification that mandated use of software that runs only PDM, but restricts
the type of relationships to the traditional finish-to-start without lag. This
is not quite an example of punishing all for the misdeeds of a few, but comes
close. Usually, the professional engineers of PADOT will use engineering
discretion and allow limited use of non-traditional relationships if the
Scheduler can demonstrate that there is actually logic behind the overlaps.

17.8. The Need for Non-Traditional Relationships

Use of ADM in the late 1950s, 1960s, 1970s, and 1980s often required
fudges and workarounds to make the mathematical logic network prop-
erly model the real world. The additional power of PDM often was the
236 The Tools of CPM Planning and Scheduling

Excav 1st 50 feet Stone 1st 50 feet Pave 1st 50 feet

Excav 2nd 50 feet Stone 2nd 50 feet Pave 2nd 50 feet

Excav 3rd 50 feet Stone 3rd 50 feet Pave 3rd 50 feet

Figure 17.8.1 ADM.

Excav 1st Stone 1st Pave 1st

50 feet 50 feet 50 feet

Excav 2nd Stone 2nd Pave 2nd

50 feet 50 feet 50 feet

Excav 3rd Stone 3rd Pave 3rd

50 feet 50 feet 50 feet

Figure 17.8.2 PDM.

Excav 1st Stone 1st Pave 1st

50 feet 50 feet 50 feet

Excav Stone Pave

remainder SS 2 remainder SS 2 remainder
FF 2 FF 2
Figure 17.8.3 PDM with some overlap.

Excav Stone Pave

1000 feet SS 2 1000 feet SS 2 1000 feet
FF 2 FF 2
Figure 17.8.4 PDM with more overlap.

Excav 1000 P–1st/last 50 feet Stone 1000 P–1st/last 50 feet Pave 1000 P–Required Open section
feet-10 days SS 2 & FF 2 feet-10 days SS 2 & FF 2 feet-10 days R–M highway
ove
mac
hine
Next
Type of restraint – P = physical, R = resource – and reason for restraint and duration between activities location

Figure 17.8.5 RDM—Rationale of restraints recorded.

 Specifying the Relationships Between Activities 237

best answer to these issues. The Scheduler in preparing the CPM must
not only properly model the real world for the initial schedule submis-
sion, but must do so in a manner that makes using and updating the
CPM simple enough that such will be willingly embraced by the field.
Note Figure 17.8.1, an example of a 1000-foot long highway project
in ADM. Three separate crews (possibly three separate subcontractors)
are involved in constructing the highway. Crew #1 excavates the first
50 feet, then continues while crew #2 places stone. When crew #2 com-
pletes the first 50 feet, crew #3 may begin paving. (Even in this hypo-
thetical example, it is recognized that crew #3 may but need not begin
immediately.) Note that to prevent false logic, a large number of “dummy
activities” or logic restraints are required.
In the 1960s, to prevent the CPM from becoming a huge mass of
meaningless detail, various practitioners would create various fudges,
the most usual to create one activity of “Excavate / Stone / Pave” after
graphically detailing on the side the detail of Figure 17.8.1. However,
this fudge effectively prevented selecting or sorting by crew or subcon-
tractor or properly cost loading the network. Each such situation had
to be solved in its own special way.
The same information can be recorded in PDM as in Figure 17.8.2.
Note that only traditional finish-to-start logic without lags is used, but
the need for the “dummies” is removed. However, the number of activ-
ities is the same and the job of updating the CPM (recording the actual
start and finish for each activity) is not trivial.
Now we can begin to condense the network to show the dependencies
between crews without overstating the obvious. Looking at Figure 17.8.3,
note how a minimum number of tasks of larger activities need be spelled
out to fully describe the relationships between crews.
However, now that overlaps are possible, they can be misused.
Figure 17.8.4 shows the logic fully collapsed by used of non-traditional
logic and lags. Referring back to Chapter 2, how will the software cor-
rect for reduced durations if larger crews are used? How will the soft-
ware treat an update situation where progress on excavation is slowed
by worse than anticipated conditions?
Further extension of PDM theory and software, as suggested by the
RDM approach, as shown in Figure 17.8.5, may allow the network to be
similarly condensed but still retain the information implicit in the orig-
inal ADM model.

17.9. The Desire for Non-Traditional Relationship and

Resulting Misuse
While the use of PDM non-traditional relationships and lags, as shown
in Figures 17.8.3 and 17.8.4, are perhaps better “fudges” than the ADM
238 The Tools of CPM Planning and Scheduling

logic of Figure 17.8.1, the availability of the shortcut, plus placement of

the tool in the hands of “screen jockies” who understand the software
but not the theory of CPM has led to abuses. It is so easy to say, “When
concrete is 30 percent done, we will start mechanical installation and
when mechanical installation is 30 percent done we will begin electri-
cal work,” rather than perform the work necessary to prepare a proper
CPM logic network.

17.10. Non-Traditional Relationships Supported by Popular Software

Non-traditional relationships supported by Primavera include the
following.
■ Start-to-start, with lag counting days from the actual start date
recorded (or data date if PCT > 0 and no date is recorded)
■ Finish-to-finish, with lag counting days from the actual finish date
recorded (or data date if PCT = 100 and no date is recorded)
■ Start-to-finish, with lag counting days from the actual start date
recorded (or data date if PCT > 0 and no date is recorded)
■ Matching start-to-start and finish-to-finish restraints, with matching
or non-matching lags
■ Matching start-to-start and finish-to-start, finish-to-finish and finish-
to-start and other combinations are neither rejected nor flagged for
probable error
■ P3 lags are based on the calendar of the predecessor, P3e/c lags are
based on the calendar of either predecessor or successor as chosen on
a project wide basis in user option settings

Non-traditional relationships supported by Microsoft Project include

the following
■ Start-to-start, with lag counting days from the actual start date
recorded
■ Finish-to-finish, with lag counting days from the actual finish date
recorded
■ Start-to-finish, with lag counting days from the actual start date
recorded
■ A maximum of one relationship may be placed between any two
activities
■ Lag durations are based upon the calendar of the successor activity
 Specifying the Relationships Between Activities 239

17.11. Minimum Restrictions for Proper Usage of PDM

The provision of power to a saw made the work of sawing easier but per-
haps made the work of cutting harder. Not only is it much more impor-
tant to measure and mark before cutting, but the frame of the power saw
makes it much more difficult to see the markings. The power of PDM
as implemented by modern software causes a similar problem in that
the logic ties between activities or even summary activities are much
easier to add, but it is more difficult to see what is to be added and what
has already been added.

17.12. Review the Strengths of ADM: Expand the Definitions

A set of minimum rules for PDM should start with remembering that
although the software allows the user to abbreviate the detail between
two summary activities (or two activities with constituent tasks that are
interrelated), the Scheduler should be cognizant that at some level some-
thing has to be 100 percent complete before something else can start.
Thus, the Scheduler may understand that a SS5 restraint indicates
that some definable portion of Activity A (requiring 5 days of the pre-
sumably larger duration of A) must be complete before starting Activity
B. And the Scheduler should record this knowledge even if only in his/her
notes. With this backup, the Scheduler should be able to persuade a
knowledgeable engineer to allow non-traditional relationships, even if
such are prohibited by the specification.

17.13. Start of Each Activity Must Have Predecessor

A second minimum or immutable rule is that the start of each activity
(other than the first) must have a physical predecessor. In other words,
that something had to be in place before this activity may be performed.
Even if it is obvious to all that this activity should not be performed until
near the end of another activity, there is something that must be in
place before it can start.

17.14. Finish of Each Activity Must have Successor

A third minimum rule is that the finish of each activity (other than the
last) must have a physical successor and that this activity must be 100
percent complete for something else to occur, even if it is turning over the
keys to the owner. The successor may be the start of another activity
240 The Tools of CPM Planning and Scheduling

(finish-to-start) or to the start of some definable portion of another activ-

ity (finish-to-finish.)

17.15. Real World Relationships between Activities

Given that popular software may not support all of the real world rela-
tionships that the project team may convey, it is the job of the Scheduler
to not only provide the necessary “fudge” to input information to the com-
puter, but also to both provide appropriate interpretation to all the var-
ious users of the CPM. The Scheduler must set reminders for other
project control personnel to adjust update information accordingly.
The project manager may say, for a 10-day activity, that a specific suc-
cessor may start when 30 percent has been completed. If using Primavera
products, the Scheduler must choose to enter an SS3 relationship, which
states B may start 3 days after A has started without regard for progress,
or a FS-7 relationship, which states that B may start when the remain-
ing duration of A has been reduced by 3 days from 10 to 7.
Of these two options, it is recommended that the Scheduler not use
the one requiring negative lag as such use may create a “hidden” open
end (as discussed in Chapter 11) and will generally increase the level
of skepticism among reviewers of the CPM. The Scheduler must there-
after be always vigilant to adjust the lag if changing the original dura-
tion and during each update involving this activity, perhaps even to the
point of reserving an activity code to flag all activities having such lag
durations between activities.
Similarly, if the project manager states that the last 5 days of a 20-day
Activity C usually cannot be performed until Activity B is complete, the
“fudge” of an FF5 relationship must be followed by vigilance to make the
necessary adjustments if the remaining duration of C unexpectedly falls
below 5 days. But, if the project manager states that the last 5 days of
C cannot be performed until 30 percent of A is complete, the best a
Scheduler can do with the tools available is utilize a SF8 relationship and
place an appropriate explanation as a log or note to the activity, since
notes cannot be attached to a relationship.

17.16. The Final Forward Pass

We return to the instruction given at the end of the first section of
Chapter 15. After a short coffee break, the interview process should be
repeated starting from the first activity, asking for each activity if all
necessary predecessor logic, physical, crews, forms, equipment, mate-
rials, access, inspectors, and so forth, is either explicitly or implicitly in
place. It may be useful to invite all major members of the project team,
 Specifying the Relationships Between Activities 241

including major subcontractors, vendors, and the owner and engineer,

to this exercise.

17.17. The Final Backward Pass

One final pass should be made, starting from the last activity working
backward, to delete any logic that has accidentally crept in. For exam-
ple, it is a common error that a project manager may assign a crew to
Activity A, then B, then C, then off to Activity X. Later, the project man-
ager may assign another crew to Activity D, then E, then F, then off to
Activity X. Obviously, Activity X requires only one of these two crews and
the restraint from either Activity C or F (or both) should be cut. If the
intent is the first crew available can do this work, neither should be
listed as a hard restraint. Remember that the purpose of the CPM is to
plan the work and not schedule full utilization of the resources. There
are other software tools for that exercise.

17.18. Choosing the Algorithm for the Initial Schedule

The choices of algorithm for the initial schedule are fewer than those
available for updates. However, some care must still be used in making
these decisions. Moreover, this may be the best time to plan what choices
will be made for future updates to the schedule. If it is desired to perform
individual activities subject to a FF relationship without interruption, and
moreover to not tell field forces that they may begin such activities ear-
lier (but then be subject to interruption), the contiguous schedule dura-
tion switch should be chosen. However, if the crew foreman is to be given
early notice that the activity may start, even if it may be interrupted, leav-
ing the operational decision of exactly when to start such work to the fore-
man, choose the interruptible schedule duration switch.
Unless it is desired to have the computer close off open ends to the net-
work that the Scheduler has left in error, the “show open ends as critical”
switch should be used. Finally, if the use of non-traditional relationships
may cause the LF – EF = TF to not equal the LS – ES = TF, it is gener-
ally recommended to print as the total float the more critical of the two.
Planning for the future, retained logic provides the more conservative
(if often too conservative) result, but since the CPM is meant to act as
a guide and not as explicit instructions to the field, the more conserva-
tive warning is usually recommended. Finally, unless coordinating
between several software systems, such as P3 and SureTrak (which
does not support the option), it is best to utilize the calculate start-to-
start lag from actual start. If sharing files between the home office and
field, running on either P3 or SureTrak (or between P3 and Microsoft
242 The Tools of CPM Planning and Scheduling

Figure 17.18.1 Schedule calculation algorithm switches.

Project), it is best to use the early start option. These switches are
shown in Figure 17.18.1.
If using the latest Primavera flagship software, P3e/c or Primavera
Construction, other options must be addressed such as shown in
Figure 17.18.2. Choosing to ignore relationships to and from other proj-
ects begs the question why such were applied. If HQ required such to
be shown, but not to be used in calculating the impact of such on this
project, the option may be checked. A possible use may be for running
various “what if ” scenarios.
Similarly, if expected finish constraints are used properly, there should
be no reason to not include their use in the calculation of the schedule.

Figure 17.18.2 P3e/c advanced schedule options.

 Specifying the Relationships Between Activities 243

If they are now outmoded, it may be best to delete them rather than
show a logic file that says one thing and has results based on another.
Here again, the option may be used to good effect for “what if ” planning.
Finally, the Scheduler must specify what calendar convention was
used when specifying durations between activities or lags, either that
of the predecessor (as in P3 and SureTrak), successor (as in Microsoft
Project and Open Plan), a 24/7/365 calendar (having no holidays or non-
work periods), or the default project calendar (however set by the
Scheduler).
In the choices available for updates, a new option is available for cal-
culating the finish date for work being performed out-of-sequence. The
actual dates option allows the Scheduler to enter, as an actual date, the
date on which it is now anticipated that the activity will be complete.
Thus, rather than requiring the project manager to state that Activity
A is 70 percent complete, or has 3 days remaining duration, the
Scheduler can now accept “I plan to have it complete by next Tuesday.”
Since this is poor scheduling practice, it is not recommended.

17.19. Summary
Setting and recording the relationships between activities is the step
that distinguishes CPM from a ToDo list or a Gantt chart. It is impor-
tant that the start of each activity (other than the first) be preceded by
a relationship from another activity representing a physical dependency.
Similarly, it is important that the finish of each activity (other than the
last) be succeeded by a relationship to another activity representing a
physical dependency. Assuming unlimited resources that would be
enough; to account for less than unlimited resources, additional rela-
tionships may be placed between activities to communicate the pre-
ferred flow of such resources. Constraints, or locked in dates, should be
used sparingly and should be properly documented for need.
It is recommended that the use of non-traditional relationships and
lag durations between activities be kept to the minimum necessary to
make the CPM easier to use in the field and not simply to make the
preparation of the CPM easier. If used, a check must be performed to
assure that the start of each activity has a predecessor and the finish
of each activity has a successor. The Scheduler must work with the tools
of the software being used to “fudge” what is said about relationships
into what the software will accept, but then be vigilant in remember-
ing and explaining the inaccuracies this causes.
A final walk through the project from start to end, and then from end
back to start, is a good way to check that the CPM logic is correct. The
Scheduler must take care in choosing what algorithm will be used in cal-
culating the CPM so as to not have his/her careful work negated.
This page intentionally left blank
 Chapter

Example Project:
18
The John Doe Project

In this chapter, a basic network is planned for the construction of a

combination plant-office-warehouse for a small industrial firm, the John
Doe Company.
A plan of the entire complex is shown in Figure 18.3.1, and a perspec-
tive of the building and exterior elevations are shown in Figure 18.3.2.
Figure 18.3.3 shows a site plan section of the electrical service and
sewer. The floor plan for the plant is shown in Figure 18.3.4; the office
in Figure 18.3.5; and the warehouse in Figure 18.3.6. The list of activ-
ities is broken down by building area where applicable. Exterior eleva-
tion views of the building are shown in Figure 18.3.2, and interior
sections are shown in Figures 18.3.7 and 18.3.8.

18.1. Acquiring Information to Initial Schedule

Creation of the logic network, as discussed in this chapter, may be slightly
atypical in that a transcript of the interview process is not provided;
rather, the results of the interview are presented. Also, since the reader
may not be familiar with the construction techniques discussed, several
lists of activities will be first itemized and only later placed in a logical
order; a step not usually undertaken in the real world by a professional
Scheduler.
The basis for the interview process includes the drawings as depicted
in this chapter, additional drawings and specifications as may be envi-
sioned, and facts and opinions known only to the project manager and
other team members. For example, funding of the project is to be from
current income of the owner and we will see the impact of this fact upon
the logic network.

245

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
246 The Tools of CPM Planning and Scheduling

18.2. Choosing Appropriate Codes

What information may we wish to obtain from the completed CPM?
Provision of the calculated dates (ES, EF, LS and LF) and total float
(TF) are attributes of the algorithm and do not require additional
coding. The project has several buildings and other structures, there-
fore, perhaps we want a code for location. There are several drawings
and we certainly would want to note which drawing most exemplifies
the scope of work being discussed. Several subcontractors will be
employed, and the owner has indicated that he wishes to perform some
portion of the work, possibly by other contractors. Since funding is an
issue, an approximate cost for each activity may be useful; however, it
is not contemplated that payment will be made from the CPM so that
an exact cost is not necessary.
Durations will be, in part, determined by crew size and these assump-
tions should be recorded. Also, if the labor market in the locale of the proj-
ect is tight, it will be important to know if the total manpower required
exceeds the supply. For purposes of illustration in this text, these codes
can be added to the logic network at varying times. However, in the real
world, it may be close to impossible to get the team together again to add
one more code for each activity and, therefore, as far as practicable, all
such information would be collected for each activity at once.

18.3. Activity List

The site is in a low area overgrown with scrub timber and bushes; the
soil is a sand and gravel mixture overlaid by clay. Cast-in-place piles
will be driven to about 30 feet for the plant and warehouse foundations.
The office building will be on spread footings. No water supply is avail-
able, so a well and a 50,000-gallon elevated water tower will be
installed. Sewage and power trunk lines are 2000 feet away. Power con-
nections will be by overhead pole line, up to 200 feet from the build-
ing; from this point in, the power line will run underground. The sewer
will pass under part of the power line. The activities representing
these areas are
■ Survey and layout
■ Drill well
■ Clear site
■ Install well pump
■ Rough grade
■ Install underground water supply
 Example Project: The John Doe Project 247

Figure 18.3.1 Site plan, John Doe project.

■ Drive and pour piles

■ Excavate for sewer
■ Excavate plant and warehouse
■ Install sewer
■ Pour pile caps
■ Set pole line
■ Excavate office building
■ Excavate for electrical manholes
■ Pour spread footings
■ Install electrical manholes
■ Pour grade beams
■ Energize power feeder
■ Install power feeder
Figure 18.3.2 Building, John Doe Co., with elevations.

Figure 18.3.3 Electrical ductbank section XX. (See Figure 18.3.1.)

248
Figure 18.3.4 Plant floor plan.

Figure 18.3.5 Office floor plan.

249
250 The Tools of CPM Planning and Scheduling

Figure 18.3.6 Warehouse floor plan.

The plant and warehouse structures are to be structural steel with

high-tensile bolted connections. The plant will have an overhead
craneway running the length of the building; the warehouse will have
a monorail. The roof system will be bar joists and precast concrete
planks covered with 20-year built-up roofing. The siding of both build-
ings will be insulated metal panels with insulated glass upper panels

Figure 18.3.7 Interior section AA. (See Figure 18.3.4.)

 Example Project: The John Doe Project 251

Figure 18.3.8 Interior sections BB and CC. (See Figures

18.3.5 and 18.3.6.)

to admit light. Both buildings will have concrete floor slabs, which will
be poured on compacted sand. The activities representing this work are
■ Erect structural steel
■ Apply built-up roofing
■ Bolt up steel
■ Compact slab subgrade
■ Erect craneway
■ Install underslab plumbing
■ Erect monorail track
■ Pour floor slabs
■ Install underslab conduit
■ Erect bar joists
■ Erect roof planks
■ Erect siding
252 The Tools of CPM Planning and Scheduling

When the plant and warehouse shells are erected, interior partitions
(offices, bathrooms, etc.) will be made of concrete block. The interior ceil-
ings are hung with integrated HVAC and fluorescent light fixtures;
the loading docks will be reinforced concrete. The railroad siding must
be brought in from a spur line one mile away. This adds the following
activities:
■ Masonry partitions
■ Grade and ballast
■ Office ceilings
■ Railroad siding
■ Piping systems
■ Form and pour truck loading dock
■ Power conduit
■ Form and pour railroad loading dock
■ Branch conduit
■ Install boiler
■ Install electrical load center
■ Install fuel tank
■ Install power panel boxes
■ Install plumbing fixtures
■ Install power panel insides
■ Crane
■ Monorail
■ Heating and ventilating units (roof)
■ Paint interior
■ Ceramic tile (lavatory and lunchroom)
■ Pull wire
■ Exterior doors
■ Electrical fixtures
■ Interior doors
■ Floor tile (offices)
■ Ductwork

The office building is designed as a precast concrete structure with

masonry walls. The roof system is designed as precast planks with
single-ply roofing. The partitions are to be metal studs with drywall.
 Example Project: The John Doe Project 253

The ceiling is to be hung. The building will have a self-contained air-

conditioning unit. The activities include
■ Erect precast structure
■ Roofing
■ Erect roof
■ Exterior masonry (cavity wall)
■ Windows and glaze
■ Interior doors
■ Paint interior
■ Plumbing fixtures
■ Paint exterior
■ Ceramic tile (lavatory)
■ Lighting panel
■ Metal studs
■ Wiring
■ Trim and millwork
■ Flooring
■ Hung ceiling
■ Exterior doors
■ Drywall

The project outside work includes

■ Fine grade
■ Seed; plant shrubs and trees
■ Flagpole
■ Pave parking area
■ Access road
■ Area lighting
■ Perimeter fence

18.4. Could We Prepare a Bar Chart?

At this point, having a detailed list of activities and perhaps (from the
bid estimate) durations, we could easily prepare a bar chart. We can
place each activity as a bar of duration length in a position following
those other activities already placed. Of course, as we move down the
254 The Tools of CPM Planning and Scheduling

list of several hundred activities, the desire to check whether each new
activity may have some impact upon an activity bar previously placed,
and thus the requirement of determining all the other bars that will then
have to be moved, decreases dramatically. If we are told, “good news, our
subcontractor says that the activity can be done in half the time we
expected,” we may simply choose not to go back and realign all the bars
that could benefit from this news. Perhaps there is a better way to do
this. Perhaps we should try using a CPM logic arrow diagram.

18.5. Network Logic in ADM

The first rough arrow diagram usually becomes the activity list. For a
number of reasons, this owner elects to proceed in a definite fashion. To
expedite the project, the site preparation and utilities work are to be put
out as a separate package to be accomplished before the foundation con-
tractor moves onto the site.
The foundation contract is to include pile driving, excavation, and all
concrete for the plant, warehouse, and office.
Since the owner expects to finance the building from current income,
the warehouse and plant areas must be completed before any work on
the office building starts. Steel erection is to start after the slabs are
poured. The office will be temporarily located in the warehouse while
the office building is in construction.
Figure 18.5.1 represents the site preparation and utilities portion of
the project. Note that the events have been numbered according to the
traditional j > i and by the horizontal method.

Figure 18.5.1 CPM network of site preparation and utilities.

 Example Project: The John Doe Project 255

Event 0. The project starts.

0–1 Clear site. Necessary before any survey work can start.
1–2 Survey and layout. Cannot start before the site is cleared; otherwise,
many of the survey stakes would be lost in the clearing operation.
2–3 Rough grade. Cannot start until the area has been laid out. This
activity ties up the whole site with earth-moving equipment.
3–4 Drill well. Cannot start until the rough grading operation is completed.
4–5 Install well pump. Cannot be done until well is completed and cased.
5–8 Underground water piping. Although this might be started earlier, the
site contractor prefers to work from the pump toward the building site.
3–6 Water tank foundation. After the rough grading, these simple foun-
dations can be installed.
6–7 Erect water tank. The water tank cannot be erected until the foun-
dations are poured.
7–8 Tank piping and valves. Cannot be fabricated and erected until the
tank is completed.
8–13 Connect piping. The water piping cannot be linked up until both sec-
tions are completed.
3–9 Excavate for sewer. Can be started after rough grading.
9–11 Install sewer and backfill. Immediately follows the sewer excavation,
working from the low point uphill.
3–10 Excavate for electrical manholes. Can start after rough grading.
10–11 Install electrical manholes. Cannot start until the excavation is
completed.
11–12 Install electrical duct bank. Is started after the electrical manholes are
complete. The start of this also depends on the completion of the sewer
line, because that line is deeper than the duct bank.
3–12 Overhead pole line. Can be started after the site is rough graded.
12–13 Pull in power feeder. Can start after both the duct bank and the over-
head pole line are ready to receive the cable.

Event 13. The site preparation and utilities work are complete.
Figure 18.5.2 represents the foundation and concrete work for the
John Doe project.
13–14 Building layout. Necessary before foundation work can start.
14–15 Drive and pour piles. After layout, this is the first step in the plant
and warehouse foundation work.
15–16 Excavate. Follows piping, including fine grading to finish grading.
16–17 Pour pile caps. Starts after the fine grading.
17–18 Form and pour grade beams. These are poured across the exterior pile
caps in this project.
256 The Tools of CPM Planning and Scheduling

Figure 18.5.2 CPM network of foundation contract.

18–21 Form and pour railroad loading dock. This dock is essentially an
extension of the grade beams.
18–22 Form and pour truck loading dock. This dock, at the opposite end of
the building from the railroad dock, also backs on the grade beams.
18–19 Backfill and compact. Cannot start until the grade beams are ready
to contain the fill.
19–20 Underslab plumbing. Cannot be installed until the backfill is complete.
20–22 Underslab conduit. Is installed after the plumbing because the plumb-
ing lines are deeper.
22–29 Form and pour slabs. The loading dock sides and underslab prepa-
ration must be completed before the slabs are poured.
14–23 Excavate for office building. Can start after the building layout work
is complete.
23–24 Spread footings. Can be placed after the excavation is done.
24–25 Form and pour grade beams. Are poured on top of the spread footings.
25–26 Backfill and compact. Is done after the grade beams are finished.
26–27 Underslab plumbing. Is installed in the backfill.

Figure 18.5.3 CPM network: close-in, plant, and warehouse.

 Example Project: The John Doe Project 257

27–28 Underslab conduit. Is installed on top of the plumbing lines.

28–29 Form and pour slabs. Can be done after the underslab preparations
are complete.

Event 29. The foundations and concrete contract are completed.

Figure 18.5.3 represents the erection of the framework for the plant
and warehouse and also the closing-in of those buildings.
29–30 Erect structural steel. Follows the completion of foundations.
30–31 Plumbing and bolt steel. Cannot be done until the steel has been erected.
31–32 Erect craneway and crane. Can be done after the steel is bolted up.
To make rigging easier, it is planned before the installation of the bar
joists system.
31–33 Erect monorail track. Although this is not as difficult to erect as the
craneway, it is convenient to erect it before the bar joists.
33–34 Erect bar joists. Can start after structural steel and major rigging are
erected.
34–35 Erect roofplanks. Cannot be done until the bar joists system is complete.
35–37 Single ply roofing. Goes on top of the roof planks.
35–36 Erect siding. Follows the roof planking for safety reasons and because
the flashing detail makes it more practical.

Event 37. The building is closed in, and interior work can start.
Figure 18.5.4 represents the interior work for the plant and warehouse.
At this point, the general, mechanical, and electrical contractors can
initiate activities.
37–38 Set electrical load center. Located on the slab in the warehouse. This
is a package unit.
37–43 Power panel backing boxes. Can be mounted on the masonry walls and
structural steel.
38–43 Power conduit. Main runs start after the electrical load center is set
in place.
43–49 Install branch conduit. These runs follow the installation of the main
conduit runs and the backing boxes for the power panels.
49–50 Pull wire. Follows completion of the conduit system.
50–54 Terminate wires. These are terminated after the panel internals are
in place.
55–56 Ringout. After the wiring is connected, the circuits are checked out.
45–51 Room outlets. Start after branch conduit and drywall are complete.

Logical restraints 49–45 and 44–45 operate as spreaders. If 44–45

were not there, “ceramic tile” would depend on “branch conduit.” If
49–45 were not there, “pull wire” would depend on “drywall.”
258 The Tools of CPM Planning and Scheduling

Figure 18.5.4 CPM network: interior work, plant, and warehouse.

51–56 Install electrical fixtures. Follows the completion of the room outlets.
37–39 Masonry partitions. Start as soon as the building is closed in.
39–42 Hung ceiling. Is supported on the masonry partitions.
37–42 Exterior doors. Can be hung after the building is closed in, but must
be installed prior to the drywall.
42–44 Drywall. Cannot start until the building is weather-tight and the
partitions are framed out. (Includes studs and door bucks.)
44–58 Hang interior doors. Can follow drywall installation.
44–48 Ceramic tile. Can follow drywall.
48–53 Paint rooms. Follows the drywall and ceramic tile installation.
53–57 Floor tile. Should be held off until room painting is complete.
 Example Project: The John Doe Project 259

57–58 Furnishings. Are installed last.

53–58 Plumbing fixtures. Are installed after painting.
37–46 Install heating and ventilating units. Can be installed after the built-
up roofing; they are on the roof.
46–52 Ductwork. Can be installed after the heating and ventilating units
and room drywall are complete.
52–58 Insulate heating and ventilating ducts. Cannot be done until the duct-
work is in place.
37–41 Erect boiler and auxiliaries. Equipment is in the warehouse, and
erection is best done after the warehouse is closed in. The unit is
small enough to move through the regular shipping door.
41–47 Preoperational check. A routine check after the boiler is installed.
37–40 Fabricate piping systems. Can be done after the building is closed in.
40–47 Testing piping. Follows completion of the piping systems.
37–47 Install fuel oil tank. Is planned to start after the building siding is on
so that the excavation will not interfere with the siding work.
47–58 Light off the boiler. Cannot be done until the piping systems are
tested, boiler is checked out, and fuel oil tank is ready.
37–58 Install monorail. Can be done any time between the close-in and
completion of the building.

Figure 18.5.5 represents the structure and interior work for the office
building. At the owner’s request, this follows the completion of the plant
and warehouse, which occurs by event 58.
58–59 Erect precast. The first operation in the office building, since the foun-
dations were previously prepared.
59–60 Erect roof. Must follow the erection of the structure. Because it uses
the same crane rigging, it follows closely.
60–61 Exterior masonry. Follows the roof erection.
60–76 Package air-conditioning. Can be set as soon as the roof is completed.
61–77 Ductwork. Can commence when the building is closed in. If started
earlier, this operation would interfere with the masonry scaffolds.
61–63 Built-up roofing. Follows masonry so that the roofers are not mopping
tar on the masons, which might be called preferential logic—the oper-
ation could physically commence at event 60.
61–62 Exterior doors. Installation must wait for the door bucks, which go up
with the masonry.
61–68 Glazing. Is done in the windows, which went up with the exterior
masonry.
61–64 Piping installation. Can start after the exterior masonry is closed in.
61–65 Install backing boxes. Since the boxes mount on the masonry and
structure, the installation can start after the masonry is placed.
63–80 Paint exterior. Starts after the roofing is on and the doors are installed.
260 The Tools of CPM Planning and Scheduling

Figure 18.5.5 CPM network: office building.

Figure 18.5.6 CPM network: site work.

 Example Project: The John Doe Project 261

64–67 Test piping. Follows the piping installation.

65–66 Install conduit. Follows backing boxes, since this is smaller branch
conduit rather than a main feeder.
66–74 Pull wire. Done after the conduit is in place.
67–67 Metal studs. Follow the piping tests and the conduit installation because
portions of these systems are embedded in or behind the drywall.
68–69 Drywall. Cannot start until the building is weather-tight (“glaze,”
“roofing,” and “exterior doors”) and the metal studs are installed.
69–70 Restraint.
69–73 Ceramic tile. Also follows drywall.
70–71 Wood trim. Placed after the drywall.
71–72 Paint interior. Follows the wood trim.
72–80 Floor tile. Follows the painting in order to protect the tile.
73–80 Lavatory fixtures. Installed after the interior painting and ceramic tile
in order to protect the fixtures.
74–75 Install electrical panel internals. Follows the pulling of wires.
75–79 Terminate wires. Follows the installation of panel internals.
76–79 Electrical connections (air-conditioning). Follows the air-conditioning
equipment installation and the electrical panel installation.
77–78 Install ceiling grid. Is preceded by ductwork and the drywall.
78–80 Acoustic tiles. Can be installed after the ceiling grid is installed and
the interiors are painted.
79–80 Ringout. Of electrical systems; comes after systems are complete.

Figure 18.5.6 represents the site work, which starts when the struc-
tural work is completed (event 37). Note that random numbering was
used for this diagram because all digits up to 80 had been used in pre-
ceding sections of the diagram. All of the following can commence when
the structural contractor moves off the site.
37–93 Area lighting.
37–92 Access road.
37–91 Grade and ballast railroad siding.
37–90 Pave parking areas.
37–80 Perimeter fence.
91–58 Railroad siding. Follows grading and ballast of the bed.

The access road, parking, and railroad siding have to be ready by the
completion of the plant and warehouse (event 58). The final activities
for the office building include
58–80 Erect flagpole.
58–94 Fine grade.
94–80 Seed and plant.
262 The Tools of CPM Planning and Scheduling

In preparing the six sections of the CPM description of the John Doe
project, the standard routine of considering the overall project by its sev-
eral physical components was followed. This family of individual net-
works can be effective. If drawing space is a limitation, the drawings
could be sheets one through six of one network.

18.6. Logic Changes Examples

If the initial logic is incorrect or the situation changes, the network
is changed by adding to, deleting from, or revising the logic network.
For instance:

Example 1. What changes to the John Doe network would be required

to run the office building in parallel with the plant and warehouse?

Solution. To run the office building in parallel with the plant and
warehouse, only two activities must be changed:
28–29 Connects directly to the start of the office. To do this, change 28–29
to 28–99.
58–59 Must be unconnected from the warehouse completion. Change 68–59
to 99–59.

Example 2. If the sewer passes under the water tank location, what
work sequence changes are necessary?

Solution. If the sewer passes under the water tank foundations, activity
9–11, install sewer, will have to precede 3–6, tank foundations. Do not
do this with restraint 11–3 or you will have a loop. First, add a spreader
restraint between event 3 and the start of the tank foundations.

Example 3. If the plant building underslab plumbing is deeper than the

office building sewer, how is the restriction shown?

Solution. If the plant plumbing is deeper than the office sewer, a

restraint activity, 20–26, might be in order.

Example 4. If the electrical load center is to be masonry-enclosed, show

the changes required.

Solution. To show the electrical load center enclosed, a restraint from

event 38 to the start of masonry partitions is necessary. Activity 37–39
must be preceded by a restraint to avoid a loop.
 Example Project: The John Doe Project 263

Example 5. If the boiler is too large for the building doors, how are the
necessary logical changes shown?

Solution. If the boiler is too large for the building doors, activity 35–36,
erect siding, must be amended to leave an opening for the boiler in the
warehouse section. Then an activity, 47–42, must be added to close in
the building before drywall is erected.

Example 6. If the primary power feeder is to be pulled in by the build-

ing contractor, what changes are necessary?

Solution. If the power feeder is to be pulled in by the building con-

tractor, activity 12–13 must be replaced by a restraint, 12–13. Also, an
activity 37–66, power feeder, must be added.

Example 7. If “boiler test” depends on regular power, what changes are

required in the diagram?

Solution. If “boiler test” (activity 47–58) depends on power availability,

a restraint from 56–58 completion to event 47 is necessary: Activity
56–58 must be followed by a restraint to avoid a loop.
In these examples, the changed logic is always tested for loops. This
is especially true when the revised logic requires a connection from a
lower j to a higher number i. It is permissible to violate the j > i rule when
necessary, but doing so increases the opportunity for loops.

18.7. Network Logic in PDM

The same network in PDM format is shown in Figures 18.7.1, 18.7.2,
and 18.7.3. Figure 18.7.1 depicts the logic up to the point where the two
buildings could be built one-at-a-time or concurrently. Figure 18.7.2

Figure 18.7.1 John Doe in PDM initial site work.

264 The Tools of CPM Planning and Scheduling

Figure 18.7.2 John Doe in PDM Plant-Warehouse.

Figure 18.7.3 John Doe in PDM Office.

 Example Project: The John Doe Project 265

Contract/Category
1. Site work
2. Foundation
3. Close-in PW
4. Office
5. Procurement

Figure 18.8.1 Contract

Category Codes.

depicts work on the warehouse and Figure 18.7.3 depicts work on the
office. Whether the logic between activities in this straight-forward con-
version (involving only traditional finish-to-start restraints without
lags) is more or less understandable than the ADM diagrams, is left to
the individual student’s opinion.

18.8. Populating the Codes

Assume the project team desires only two codes and that these include
Contract/Category and Trade/Subcontract. A code dictionary could be
created, as suggested in Figures 18.8.1 and 18.8.2. Pure CPM calcula-
tions then can generate a printout, as shown in Figure 18.8.3.

18.9. Checking the Output

The computer output should be checked for errors. This is quite impor-
tant, because CPM data are susceptible to error when transferred from
the network to computer.
Failure to check the computer output has caused embarrassment
more than once. In one instance, the head of a school board received a
telegram stating “Good news!,” which went on to advise him that his
project end date had improved by 3 weeks. This was followed several

Trade/Subcontract
1. Excavate and backfill 10. Water tank 19. Drywall
2. Survey and layout 11. Piles 20. Tile
3. Concrete 12. Siding 21. Doors
4. Electrical 13. Roofing 22. Paint
5. Plumbing 14. Masonry 23. Floor tile
6. Structural/rigging 15. Fencing 24. Furnishings
7. Precast 16. Paving 25. Glaze
8. HVAC 17. RR siding 26. Carpentry
9. Well 18. Hung ceilings 27. Site work

Figure 18.8.2 Trade subcontract codes.

266 The Tools of CPM Planning and Scheduling

Figure 18.8.3 John Doe Output-Ordinal Dates.

 Example Project: The John Doe Project 267

Figure 18.8.3 (Continued)

268 The Tools of CPM Planning and Scheduling

Figure 18.8.3 (Continued)

hours later by another telegram that should have been in red ink (to match
the consultant’s face). It noted that an error in the run had been over-
looked, and the project date had really been delayed by one week.
The computer can be programmed to locate many mechanical errors,
but it will not object to a statement that the moon is made of green
cheese, nor can it pass on the practicality of CPM results.
 Example Project: The John Doe Project 269

The human factor is indispensable, which is one advantage to manual

computation. Although people may make many small errors, they are
not likely to miss a big mistake. For instance, in a hand calculation, a
loop will not just slip by, but it fools the computer every time.
It is also a good idea to trace the critical path on the CPM network.
To assist in checking this, a list of activities in order of total float is
useful. First, the critical activities are listed and then the float is
listed in ascending order. The list is also useful for a fast review of the
project by management. Figure 18.9.1 shows the sort by total float for
the John Doe project.
Another popular list is the early start sort where the activities are
listed in order of early start (ES) times. This list exhibits the activities
in the order in which they could start. The activities for each date are
listed, starting with the critical and low-float activities. Figure 18.9.2
shows the early start sort for the John Doe project. Figure 18.9.3 shows
the John Doe project listed by work category.
Although these and other sorts can be useful, it is important not to
get bogged down generating great amounts of data. Large amounts of
data are more likely to alienate field people than impress them. CPM
is only half as effective if people in the field do not actively participate
in the preparation and use of the information.
To work effectively with field people, find out what information they
want and the form in which they want it in. One field superintendent
asked, “Will CPM shorten my scheduling work?” We gave him a hardy
“yes”; then he noted that it would take him a considerable length of
time just to page through the 2-inch stack of paper that was the early
start sort for his project.
As a result of his constructive criticism, we began furnishing him
only the listing of work for the next 2 months in both early- and late-
start formats. There was no need to supply CPM information for the next
year when we were furnishing a new computer run each month.
For management, the early start sort is usually too detailed, making it
difficult to see the forest for the trees. A sort of critical and near-critical activ-
ities are sufficient to report on the project status in clear and concise terms.
Another caution about computed CPM information: It will be no
better than the input of network information. A soil mechanics pro-
fessor had a similar caution about soil strength formulas. He advised
against formulas integrating, differentiating, and extrapolating field
information to the nth degree. His premise was that there is an inher-
ent danger in cloaking rough field data in polished mathematical
formulas.
In one refinery application, the field was unresponsive even to the abbre-
viated early start sort. One of the plant engineers had an inspiration and,
with scissors, cut out the description list (less all the computed activity
270 The Tools of CPM Planning and Scheduling

Figure 18.9.1 Partial sort by total float for the John Doe project.
 Example Project: The John Doe Project 271

Figure 18.9.1 (Continued)

times and i–j numbers). Once the output was reduced to a plain list, the
field people were willing to work with it.
There is often a psychological barrier to anything associated with
a computer. In some cases, it is justified. Periodically, computer spe-
cialists come up with their own breakthroughs in network analysis.
272 The Tools of CPM Planning and Scheduling

Figure 18.9.2 Partial early start sort for the John Doe project.

For instance, at least three different computer-oriented groups have

advocated methods of generating computer outputs similar to CPM
without drawing a diagram. Such a computed result is naturally sus-
pect. First, if people in the field have strong reservations about the
computed results of an arrow diagram, how would they react to a com-
puted schedule not based on a diagram or their tangible plan? Second,
if the CPM computation must be carefully checked for errors, what can
the diagram-less computer output be checked against?
It is possible to generate an output without a diagram to support it.
As an expedient in high-rise work, we have prepared the basic CPM plan
for one floor and then regenerated it to suit the total number of similar
floors. The same method was effective in a dormitory renovation with
eight similar wings and for the KKMC military complex in Saudi Arabia.
 Example Project: The John Doe Project 273

Figure 18.9.3 John Doe project output by work category (partial).

In both cases, however, we prepared a finished CPM diagram to support

the computations.
Proponents of diagram-less schedules see the arrow diagram prepa-
ration as drudgery. Granted, it can be tedious, but the value of doing it
more than justifies the effort because it offers a graphical representa-
tion of the planners’ thoughts.
Proponents of computerized techniques believe the planner must visu-
alize an arrow diagram without the aid of paper and pencil and with-
out the benefits of the record furnished by the diagram. The planner
is also likely to miss many of the subtle connections that the arrow dia-
gram shows. Although efforts to do without diagrams are sincere and
apparently offer useful results to those who advocate them, they would
appear to have limited application.
274 The Tools of CPM Planning and Scheduling

Figure 18.9.4 John Doe project output by contract.

An extension of the diagram-less computer output is the generation

of a diagram by a computer based on the CPM output, which is discussed
in a later chapter.
Figure 18.9.4 is the John Doe project output by contract.

18.10. Calendar Dates

So far, the lists given in this chapter have been in terms of project days.
Is a project calendar necessary to use outputs? No. The computer, in a
relatively easy step, can calendar-date the output. Figure 18.10.1 is the
 Example Project: The John Doe Project 275

Figure 18.10.1 John Doe project calendar.

project calendar for the John Doe project. It assumes a June 1 start date
and skips weekends and holidays. For activity 4–5, install well pump,
the ES is 22 and the LF is 24. From the project calendar, the ES is July 5
and the LF is July 7, 2000. The activity times list is equivalent to the
list of calendar times shown in Table 18.10.1.
Although the calendar-oriented information is more useful, the addi-
tion of as many as eight more digits per line does make it more difficult
276 The Tools of CPM Planning and Scheduling

TABLE 18.10.1 Calendar Dates Replace Project Days

Duration Float
Activity (days) Description ES* EF* LS* LF* (days)

0–1 3 Clear site 7–1 7–3 7–1 7–3 0

1–2 2 Survey 7–3 7–8 7–3 7–8 0
2–3 2 Rough grade 7–8 7–10 7–8 7–10 0
3–4 15 Drill well 7–10 7–31 7–10 7–31 0
3–6 4 Water tank foundations 7–10 7–16 7–13 7–17 1
3–9 10 Excavate sewer 7–10 7–24 7–16 7–30 4
3–10 1 Excavate electrical 7–10 7–13 7–29 7–30 13
manholes
3–12 6 Pole line 7–10 7–20 8–3 8–11 16
4–5 2 Well pump 7–31 8–4 7–31 8–4 0
*
Numbers in column refer to calendar dates; e.g., 7–1 means July 1.”

to read the activity list. Since early start and late finish are the two dates
usually referred to, the EF and LS columns are often omitted in the
calendar-dated summary of activity times. The float column is the fastest
way to pick out the critical path.
If a project starts on July 29 instead of July 1, will you need to construct
a new calendar? The difference in project days between July 1 and 29 is
20 − 1, or 19. Look up the date for project day 10 under 19 + 10, or 29,
and the date is August 11. Therefore, one project calendar can be used
for a number of projects.
To use the calendar to determine project days between two dates,
enter the table at each date and subtract the reference numbers to get
net project days. Conversely, the table can be entered at any date and
calendar days can be added (or subtracted) to identify a date separated
from another date by a set number of days.
The project calendar can also be generated day for day (i.e., 365 days
per year or 366 in a leap year). The result will schedule work on hol-
idays and weekends. Although seemingly illogical, this calendar is
useful for contracts in which schedules (and extensions) are expressed
in calendar days.

18.11. Summary
In this chapter, a sample light industrial project was planned with CPM.
The activities involved in each section of the project were defined and
the CPM network for each section was drawn. In describing the network
construction, an index or dictionary approach was used. This can be very
useful in CPM, but it is not often employed because of the additional
effort required.
 Part

The Practice of
4
CPM Planning

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
This page intentionally left blank
 Chapter

Equipment and
19
Workforce Planning

Time and cost dimensions have been discussed in connection with plan-
ning and scheduling projects. Workforce and equipment have been
assumed to be available as needed. This, of course, is not usual. Planners,
superintendents, and/or engineers responsible for projects keep their
forces level by juggling float activities. In doing so, they must work crit-
ical and low-float activities first; the activities with more float are worked
as fill-in jobs. As the project progresses, the float values change, which
makes regular updating important in scheduling activities.

19.1. Workforce Leveling

Assume that phase 1 of the John Doe project is to be done overseas by
Seabees; one category (e.g., jack-of-all-trades) of workforce is then
assigned to each activity. Assume also that equipment is available as
needed. See Table 19.1.1.
To determine the workforce requirements for the project, draw the
arrow diagram to scale and plot the workforce against time. The first
step is to draw the critical path, 0–1–2–3–4–5–8–13, and plot the
critical workforce. This must be the initial step, because this portion of
the workforce requirements are fixed. Figure 19.1.1 shows the plot of
critical path and associated workforce.
In the float paths, there is flexibility in plotting the workforce. To get
a planning datum of maximum needs, first plot all of the float paths,
starting at the early start times. The first path plotted is the low-float
path, 3–6–7–8. Since workforce is plotted on early start, the result is
an early peak of workforce requirements. The peak requirement is 31
workers if all activities start early, and it occurs on day 11.

279

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
280 The Practice of CPM Planning

TABLE 19.1.1 Resources Required for John Doe Project

i–j Activity Workers

0–1 Clear 4
1–2 Survey 5
2–3 Grade 4
3–4 Well 3
3–6 Tank foundations 4
3–9 Excavate sewer 6
3–10 Excavate manhole 2
3–12 Pole line installation 6
4–5 Pump 2
5–8 Underground pipe 8
6–7 Tank 10
7–8 Tank pipe 6
8–13 Connect 4
9–11 Install sewer 8
10–11 Electrical manhole 6
11–12 Duct bank 10
12–13 Feeder 5

Figure 19.1.1 Peak workforce requirements based upon early start.

 Equipment and Workforce Planning 281

Figure 19.1.2 shows a similar workforce plot based on the starting

float activities and their late start dates. The peak workforce
requirement in this case is 34 workers, and it does not occur until day
24. Figure 19.1.3 shows both the early start (light line) and the late
start (heavy line) workforce curves. Area A is common to both curves.
Areas B and C are under the early start curve only. Areas D and E are
under the late start curve only. The areas under the curves represent
workforce (workers × project time). Since the workforces under the
curves must be equal, the differences between the late start and early
start curves must be equal. That is:
Since
A+B+C=A+D+E
then
B+C=D+E
In this case, B + C = 108 worker-days = D + E.

Figure 19.1.2 Peak workforce requirements based upon late start.

282 The Practice of CPM Planning

Figure 19.1.3 Combined peak workforce requirements for both early and late start dates.

Having estimated the peak, or worst cases, how can you level the
workforce requirements? In this simplified example, it is relatively
easy. Looking at Figure 19.1.3, the minimum level must be in excess of
20 workers. Since the early start curve is the more level of the two,
work from it. By shifting the 3–12 activity to start on day 13 instead of
day 7, the workforce can be built up slower and held to under 25
workers. See Figure 19.1.4. Because the estimated crew size is fixed,
the job superintendent can level beyond the graph of Figure 19.1.4 only
by further shifting the crew sizes. When shifting activities to the
workforce, keep in mind that the logical sequence must not be violated.
(Note this leveling routine assumes activities may be interrupted and
later resumed without additional time and cost for remobilization.)

Figure 19.1.4 Leveled workforce.

 Equipment and Workforce Planning 283

Having worked out a level workforce plan for the Seabees, assume
that only 20 workers will be assigned to the project. Figure 19.1.5 plots
one solution to this problem (there is no single correct solution). The
particular solution of 40 days is the minimum time in which this project
can be completed with only 20 workers.
In arriving at this solution, a number of factors should be noted. First,
there is no longer a critical path. Every path through the network now
has interruptions during in which the workforce is unavailable. Because
there is no critical path, the critical activities do not have to be carried
out in immediate succession. However, the critical path is a good
starting point for scheduling activities because you cannot complete
the project in less than 34 days. If you do follow the “old” critical path,

Figure 19.1.5 Limited workforce.

284 The Practice of CPM Planning

you cannot complete it in 34 days. Even though there is no critical path,

no activity can be started before this early start because the work must
still be accomplished in the same logical sequence.
In meeting the workforce restrictions, activity splitting is allowed.
That is, you can start an activity, leave it, and come back to complete
it. This occurred in activity 3–12, pole line installation. Note also that
certain impractical scheduling tends to occur. For instance, activity
9–11, install sewer, follows 3–9, excavate sewer, by 2 days. Unless the
climate is quite dry, the field superintendent will not likely hold fast to
this schedule. The superintendent will start installing sewer on day 17
with the 7 workers available rather than the 8-person crew specified. If
this is done, activity 3–12, pole line installation, will probably be
delayed until day 27, which will still allow completion by day 40 with a
slower build-up to the full crew.
Although there is an advantage to having Seabee jack-of-all-trades
as workers, there is one slight complication. Keeping the same total
work crews, you must specify the number of petty officers and con-
struction men for each activity. See Table 19.1.2.
Figure 19.1.6 is similar to Figure 19.1.1 except that the workforce is
broken into the two categories of petty officers and construction workers.
Adding the two curves together results in the same total use
requirements as Figure 19.1.1 (10 petty officers plus 21 construction
workers on day 11 equals 31, etc.).

TABLE 19.1.2 Multiple Resources Required for John Doe Project

Number of
Number of petty construction
i–j Activity officers workers

0–1 Clear site 4 0

1–2 Survey and layout 2 3
2–3 Grade 4 0
3–4 Drill well 1 2
3–6 Water tank foundations 1 3
3–9 Excavate sewer 2 4
3–10 Excavate manhole 1 1
3–12 Pole line installation 2 4
4–5 Well pump 1 1
5–8 Underground piping 1 7
6–7 Erect water tank 3 7
7–8 Tank piping 2 4
8–13 Connect piping 2 2
9–11 Install sewer 1 7
10–11 Electrical manhole 2 4
11–12 Duct bank 2 8
12–13 Power feeder 1 4
 Equipment and Workforce Planning 285

Figure 19.1.6 Peak workforce requirements (two categories).

If the 20 Seabees are made up of 5 petty officers and 15 construction

workers, what is the effect on the schedule? When handling more than
one type of workforce, the graphical plot becomes too unwieldy; another
graphical approach is used to level the resources.
The first step in this method is to list all of the activities in
ascending order of their end event j. The list is shown in the first
column of Figure 19.1.7 (the event numbers must be assigned in the
classical order j > i). First on the list is the first activity 0–1, and last
on the list are the two terminal activities, 8–13 and 12–13. The others,
being in order of end events, are arranged in proper logical order.
If the workforce is scheduled in this order, you will be observing the
network logical order. The second column has the activity durations.
The third and fourth columns list the workforce requirements. With
this information, you can schedule the project without further
recourse to the network.
286

Figure 19.1.7 Workforce calculation: Limits 5 petty officers, 15 construction workers.

 Equipment and Workforce Planning 287

Starting at the top line in Figure 19.1.7, schedule four petty officers
for the first three days for activity 0–1. The next activity, 1–2, cannot
logically start until 0–1 is completed; the heavy line represents the
logical restriction. Two petty officers and three construction workers
are assigned to the activity. Proceed in this manner until activity 3–9;
it logically could commence at day 8, and it does. On day 12, however,
there are not enough petty officers, so the activity is interrupted until
day 22. This interruption is represented with an X on the days that a
workforce is not available.
The procedure, then, is simple: Consider each activity in order;
determine the logical point at which it could start; and then schedule
the activity as soon as the workforce is available. In this example,
dividing the workforce into two categories lengthened the project from
40 to 44 days. Although basic CPM networks of several hundred
activities can easily be manually computed, manual techniques are
slow and complicated. A network of perhaps 50 activities is the
practical limit for manually calculating workforce requirements. It will
also vary considerably with the complexity of the network and the
number of different categories of skill and equipment to be scheduled.

19.2. Computerized Resources Planning

Computer analysis is much more economical of time and money than
manual analysis for most workforce studies, and a number of programs
have been developed for workforce leveling. Two of the earliest programs
were resource and manpower scheduling (RAMPS) by CEIR and resource
planning and scheduling method (RPSM) by Mauchly Associates.
However, the two pioneer programs were designed for computer hard-
ware that is now obsolete.
Other major systems developed to handle resources as well as basic
schedules included PMS by IBM and integrated civil engineering systems
(ICES) by MIT. McDonnell Automation (McAuto) was part of the original
PMS team, which utilized that experience in its development of the
McAuto Management Scheduling and Control System (MSCS) program,
which handled all phases of resource planning and scheduling.
The PROJECT/2 system (Project Software & Development Inc.)
developed by Robert Daniels of the original ICES group had compre-
hensive resource capabilities.
MSCS and PROJECT/2 provided the best resource capabilities during
the 1970s. There were other systems, including resource planning and
control (RPC) by the author and MDC Systems, in development and use
since 1966. RPC gave results similar to those of MSCS and PROJECT/2,
288 The Practice of CPM Planning

but it used resource parameter variation rather than automatic leveling.

This optimized human direction, using the computer to test results.
The systems typically have three phases: The first is a CPM nominal
time run; the second is a resource compilation called unlimited run;
and in the third, the resources available are limited and two outputs
are generated. One of which is a table of resources vs. time. The second
is the schedule needed to achieve that use; it employs the logical
sequence of the network but is no longer time-limited. The resource-
limited project duration will be greater than (or possibly equal to) the
normal time duration.
A computer program’s unlimited phase would generate the usage
shown in Table 19.2.1 for the two categories of peak requirements given
in Figure 19.1.6.
A typical program can generate a schedule for the unlimited resources
phase, but it would contain the same information as the CPM output.
Accordingly, the schedule is not usually printed out for this step.
In the next step, with the petty officer supply limited to 5 and the
construction workers to 15, the use calculated in Figure 19.1.7 is as
shown in Table 19.2.2. The schedule of this manpower is as shown in
Table 19.2.3.
This output is kept on disk so that any desired sort or listing can be
furnished. The usual ones are i-j, start-end. In this case, the sort is j-i,
which is unusual, but it matches the order of activities given in

TABLE 19.2.1 Resource Usage Table

Time P* C* Time P* C*

1 4 0 18 5 16
2 4 0 19 5 16
3 4 0 20 5 16
4 2 3 21 5 16
5 2 3 22 4 13
6 4 0 23 5 13
7 4 0 24 5 13
8 7 14 25 5 19
9 8 17 26 4 15
10 8 17 27 4 15
11 8 17 28 4 15
12 10 21 29 4 15
13 10 21 30 4 15
14 6 13 31 3 11
15 6 13 32 1 7
16 6 13 33 2 2
17 6 13 34 2 2
*
P = petty officer; C = construction worker.
 Equipment and Workforce Planning 289

TABLE 19.2.2 Resource Usage Table

Time P* C* Time P* C*

1 4 0 23 5 9
2 4 0 24 5 9
3 4 0 25 5 15
4 2 3 26 5 15
5 2 3 27 5 15
6 4 0 28 5 15
7 4 0 29 5 15
8 5 10 30 5 15
9 4 9 31 5 15
10 4 9 32 2 14
11 4 9 33 5 15
12 4 9 34 5 13
13 4 9 35 5 13
14 4 9 36 3 11
15 4 9 37 4 12
16 4 9 38 4 12
17 4 9 39 2 8
18 4 9 40 1 4
19 4 9 41 1 4
20 4 9 42 1 4
21 4 9 43 1 4
22 5 10 44 1 4
*
P = petty officer; C = construction worker.

TABLE 19.2.3 Resource-Limited Schedule (Based on Five Petty Officers, Fifteen

Construction Workers)

Duration, Workforce*
i–j days Description P C Start End

0–1 3 Clear site 4 0 3

1–2 2 Survey and layout 2 3 4 6
2–3 2 Grade 4 6 7
3–4 15 Drill well 1 2 8 22
4–5 2 Well pump 1 1 23 24
3–6 4 Water tank foundation 1 3 8 11
6–7 10 Erect water tank 3 7 12 21
5–8 8 Underground piping 1 7 25 32
7–8 10 Tank piping 2 4 22 31
3–9 10 Excavate sewer 2 4 8 11&
3–9 10 Excavate sewer 2 4 22 27&
3–10 1 Excavate manhole 1 1 8 8
9–11 5 Install sewer 1 7 32 36
10–11 5 Electrical manhole 2 4 28 31&
10–11 5 Electrical manhole 2 4 33 33&
3–12 6 Pole line installation 2 4 33 38
11–12 3 Duct bank 2 8 37 39
8–13 2 Connect piping 2 2 34 35
12–13 5 Power feeder 1 4 40 44
*
P = petty officers; C = construction workers; & = split activity.
290 The Practice of CPM Planning

Table 19.2.2. Note that there is no critical path or float. This is the
schedule that must be followed to achieve the level usage. Look at
activities 3–9 and 10–11. The activities are split, scheduled at two
separate times. This is indicated by the ampersand.

Resource applications
Analyzing and planning workforce and equipment by a network should
be preceded with using the basic CPM technique. Often, basic CPM
techniques are sufficient to meet all planning and scheduling needs of
a project. On the other hand, there are some applications in which CPM
alone is inadequate and resources must be analyzed, such as jobs requir-
ing heavy-equipment when constructing earth-fill dams and highways.
Careful scheduling of equipment across one or several projects has
an immediate payoff. Contractors owning equipment are usually in a
constant rental quandary. Should they rent out their idle equipment or
will they have to rent extra equipment themselves in the near future?
In heavy construction work, equipment (not time) is the limiting factor.
In one highway project of 220 working days, the addition of 5 pieces
of equipment shortened the project by 40 days. The time reduction of
almost 20 percent was achieved by means of an equipment increase
of less than 10 percent.
In a water treatment plant project, a series of resource vs. schedule runs
were made to measure the minimum number of tradespeople required per
contractor. In addition to the numbers of tradespeople needed, a second
concern was crowding in work areas with a high density of piping,
equipment, and controls. A maximum number of tradespeople per con-
trolling area was posed as a limit. The runs identified at least two
instances in which the minimum levels of tradespeople required by all the
contractors together reached the cumulative population allowable for
crowded areas.
On that same project, there was a concern that the electrical con-
tractor was understaffed. Figure 19.2.1 is a histogram showing the
projected electrical workforce based on early activity starts in the
project’s finishing stages. The plot demonstrates that a leveled force of
about 25 electrical workers could readily complete the project on time.
However, the late start histogram (Figure 19.2.2) shows that if the
float is used up and the electrical work is not commenced until April
(a 3-month slippage), a peak force of about 40 electricians will be
required. By using the early start approach and a crew of 20 elec-
tricians, the schedule was leveled and the work was completed on time.
Most production processes that stay on-stream for long periods of
time cannot be maintained during the production cycle. When the unit
is shut down, either on schedule or because of a malfunction, the plant
 POTOMAC RIVER WATER SUPPLY PROJECT
Resource plot/01 RESOURCE ALLOCATED CPM AS PREPARED BY OBRIEN-KREITZBERG

Resource code is EL Schedule plotted is (Early-*) Incremental

50.00 50.00

45.00 45.00

40.00 40.00

35.00 35.00
Resource quantity

30.00 30.00

25.00 25.00

20.00 20.00

15.00 15.00

10.00 10.00

5.00 5.00

0.00 0.00
Date 06Jan81 25Feb81 16Apr81 05Jun81 25Jul81 13Sep81 02Nov81 22Dec81 10Feb82 01Apr82 21May82
BCU 371.0 421.0 471.0 521.0 571.0 621.0 671.0 721.0 771.0 821.0 871.0
OKA PROJECTED ELECTRICAL MANPOWER HISTOGRAM - BASED ON ACTIVITY EARLY START DATES

Figure 19.2.1 Projected electrical workforce histogram based upon early activity start
dates.

POTOMAC RIVER WATER SUPPLY PROJECT

Resource plot/01 RESOURCE ALLOCATED CPM AS PREPARED BY OBRIEN-KREITZBERG

Resource code is EL Schedule plotted is (Latf-*) Incremental

50.00 50.00

45.00 45.00

40.00 40.00

35.00 35.00
Resource quantity

30.00 30.00

25.00 25.00

20.00 20.00

15.00 15.00

10.00 10.00

5.00 5.00

0.00 0.00
06Jan81 25Feb81 16Apr81 05Jun81 25Jul81 13Sep81 02Nov81 22Dec81 10Feb82 01Apr82 21May82
Date 371.0 421.0 471.0 521.0 571.0 621.0 671.0 721.0 771.0 821.0 871.0
BCU
OKA PROJECTED ELECTRICAL MANPOWER HISTOGRAM-BASED ON ACTIVITY LATE START DATES

Figure 19.2.2 Projected electrical workforce histogram based upon late activity start
dates.

291
292 The Practice of CPM Planning

maintenance department performs maintenance work on the unit. The

work is usually pushed around the clock because downtime is costly. The
time from off-stream to on-stream is usually referred to as turnaround
and is particularly applicable to chemical and refinery units. However,
maintenance of large power generation stations, boilers, and similar
plants or equipment also is in the turnaround category. Power or pro-
duction plant outages are handled similarly. The payback for time
savings can be tremendous. On one nuclear power plant outage, the
cost/loss was $800,000 per day.
A utility company used resource planning for the scheduled main-
tenance of a special super-pressure turbine. Studies were carried out 2
months prior to the scheduled shutdown. One month prior to that date,
the machine developed bearing noises and had to be shut down early.
The maintenance forces were committed to another turnaround
already in progress and the workforce originally scheduled for the new
turbine could not be assigned. While the unit was down, the company
estimated an out-of-pocket cost of $25,000 per day because of the lower
efficiency of the standby units used to replace it.
While the unit was in its four-day cooldown, new computer analysis
were made reflecting the reduced initial workforce levels. The schedule
that was initially generated retained the original maintenance project
length by shifting work that could wait to a later time in the project
when a larger workforce would be available. The project schedule was
updated regularly.
In the second week, subcontracted work was identified as the critical
path. A workforce analysis indicated that there was no need for the main-
tenance force to work on Easter weekend, which resulted in considerable
money savings in addition to an earlier online time for the unit.
Refineries also have the problem of fixed workforce and limited time
to accomplish substantial turnaround assignments. Resource planning
has been used to reduce downtime, but even the best schedule can
achieve only a limited time reduction.

19.3. Multiproject Scheduling

Figure 19.3.1 shows five concurrent subordinate networks intercon-
nected to produce one major NASA project network. In this case, each
subordinate network is termed a fragnet (fragmentary network). To
compute the major network, it would be necessary only to interconnect
the nine unconnected initial networks by using nine logical restraints
tied back to a starting event of node. Similarly, each of the concluding,
or terminating, events would have to be interconnected to provide a
continuous network from start to finish. Calculation could then be by
basic CPM program.
 Equipment and Workforce Planning 293

Figure 19.3.1 Multiproject networks.

If the calculation is performed on this basis, there will be one critical

path through the longest project, with each of the others showing float.
Also, the calculation on this basis will show equivalent calendar starting
dates for all projects.
In order to bring the projects into line with reality, the starting
restraints are assigned times to reflect the staggering of the actual project
parts, or fragnets. Similarly, lag, or concluding time, durations can be
assigned so that the phasing will apply to the completion event.
To assist in establishing time values for lead and lag arrows
(activities with no work activity by time duration at the start of a project
are termed lead; those at the conclusion are used to establish phasing
or lag), the fragnets can be calculated before they are interconnected.
Often, the interconnection points are dictated to some degree by the
time values. That is, when there is the possibility of a choice, as in pref-
erential logic, the interface between two areas—particularly functional
areas—is established by completion time. Figure 19.3.2 shows more
summary fragnets and more complex interconnections.
Fragnets, or individual networks, do not have to be physically
connected in order to be computed on a common basis. The connections
can be made by merging nodes or adding logical restraints to the input.
To interpret results, however, it is advisable to note them, at least on
the summary network.
294 The Practice of CPM Planning

Figure 19.3.2 Functional multiproject network.

Multiproject scheduling is one of the best bases for project resource

planning and scheduling, because a special skill or resource must often
be mobilized and utilized across many networks simultaneously. Also,
in multiproject network scheduling, there is often substantial con-
currency of activities and flexibility in completing the schedules of some
subprojects. This can provide greater float opportunities and reduce
peak resource projections.
Perini Construction, in an early CPM application, used multiproject
resource planning to schedule special equipment for installing piling
for bridge piers, each of which required 100 piles. Overall project time
available for piling was less than a year between spring flood seasons.
A total of more than 700 pilings was needed, and each had to be drilled
and placed to an average depth of more than 200 feet at an average
drilling rate of 10 feet/h.
To accomplish this, 2 special drilling rigs were designed and manu-
factured in France at a cost of $500,000. Each machine served several
purposes: drilling holes, placing caissons, placing piles, and removing
caissons after placing piles (extraction). CPM was used to evaluate, in
great depth, the placing of a set of piles. Setup and moving times were
included in the calculations, and an average cycle of 36 h was predicted
and subsequently confirmed by field information.
 Equipment and Workforce Planning 295

At this rate, with an allowance for a seven-day week, the piling could
not be completed prior to the flood season. The detailed analysis
pointed out that, not only was additional equipment required, but that
it could be a specialized caisson extractor rather than a full-fledged
combination unit. The new extracting machine cost only 20 percent of
the multipurpose machine and resulted in a reduction of more than
25% of the overall project duration.
Figure 19.3.3 shows three networks that make up a program for three
design projects. Each has three types of design personnel: mechanical
engineer m, a breadboard designer b, and an electronics technician t.
The networks represent concurrent work on three different projects by
one functional design area. Note that it is not physically required for
these networks to be joined by arrows. In this case, the connection at
the conclusion is by two or three lag arrows, and the common zero
starting node establishes the initiation point. The calculation will
determine the minimum reasonable time span for the three projects
with the use of the design workforce available. Note that since these are
sample networks, descriptions are not written on the activities as they
normally would be.
The problem was solved with the RPSM calculation, and the results
are shown in Table 19.3.1. The first stage of calculation indicated that

Figure 19.3.3 Design networks.

296 The Practice of CPM Planning

TABLE 19.3.1 RPSM Usage for Unlimited and Limited Manpower

(a) Unlimited Manpower (b) 10 t; 10 m; 10 b (c) 12 t; 12 m; 12 b

Time b m t Time b m t Time b m t

1 9 10 12 1 3 10 10 1 9 10 12
2 9 10 12 2 3 10 10 2 9 10 12
3 17 12 13 3 10 6 9 3 12 12 11
4 21 15 12 4 10 10 10 4 12 11 9
5 18 15 11 5 10 10 10 5 11 12 11
6 16 18 11 6 5 9 5 6 11 12 7
7 13 13 13 7 5 5 10 7 12 12 10
8 13 16 11 8 8 10 4 8 11 12 8
9 13 12 11 9 9 10 4 9 11 11 11
10 14 8 11 10 9 9 6 10 12 8 11
11 15 9 16 11 8 10 9 11 12 11 11
12 11 9 18 12 9 10 9 12 12 11 11
13 4 11 14 13 10 8 10 13 12 8 12
14 3 11 13 14 10 5 10 14 11 1 12
15 3 12 7 15 10 4 9 15 11 6 10
16 6 14 3 16 9 5 10 16 9 12 12
17 9 7 2 17 8 4 10 17 9 11 12
18 7 7 2 18 10 6 9 18 11 11 6
19 7 4 7 19 10 7 9 19 11 11 11
20 4 3 10 20 10 9 10 20 4 10 10
21 3 5 21 10 7 7 21 3 7 5
22 9 9 7 22 7
23 9 7 7
24 9 7 7
25 8 7 2
26 4 7
27 8 3
28 4 3
29 3 5

the projects could be completed in 21 weeks by using a maximum of 21

designers, 18 mechanical engineers, and 18 technicians. In the next
step, the computer was instructed that the department had only 10
people available in each class, and it, therefore, noted a time extension
from 21 to 29 weeks. At that point, a solution was attempted by
determining an optimum resource use within the basic CPM time. Part
c indicates that, with a 20 percent increase in workforce, a 33 percent
time reduction could be achieved.

19.4. Turnaround Application

Maintenance operation in the petrochemical industry offers one of the
most typical illustrations of multiproject operations. Many individual
mini projects go on concurrently with one or more major projects. The
 Equipment and Workforce Planning 297

use of CPM has been well established in preplanning these highly coor-
dinated operations.
One such turnaround was planned and implemented at a refinery in
Puerto Rico. Key personnel had previously been exposed to CPM
through courses, seminars, and literature. Management decided to use
CPM to plan a major maintenance turnaround of the No. 2 crude dis-
tillation unit, including a catalytic cracker. Two months prior to the
scheduled turnaround, the first CPM networks were prepared by the
conference method. Key process, maintenance, contract, and engi-
neering personnel for the turnaround met to discuss the work items to
be included.
As the scope of each work item was discussed, a network defining
the logical sequence of work was developed on a blackboard, and the
information was subsequently transcribed to a reproducible drawing.
Normal crew sizes were assumed, and time and workforce estimates were
added to the networks to complete the arrow diagram. The individual
subnetworks were linked together to form a multiproject plan.
Figure 19.4.1 shows a summary of CPM for the crude heater
overhaul, which determined the longest major job in the turnaround.
Figure 19.4.2 shows the typical overhaul plan for three similar units,
and there were more than 40 such plans for different pieces of
equipment.
In the planning, there were several major sequential operations,
but the majority could occur concurrently. The establishment of a
reasonable working schedule required either resource allocation or
the introduction of preferential logic. In this case, computerized

Figure 19.4.1 Summary plan for crude heater overhaul.

298 The Practice of CPM Planning

Figure 19.4.2 Summary plan for overhaul of three units similar to

the plan illustrated in Figure 19.4.1.

resource planning was used to establish the role of more than 400
people assigned to the 3-week operation.
During the actual turnaround, the CPM group assigned a repre-
sentative to each shift to work directly with the shift coordinator. The
representatives’ role was to assist the coordinators in using the CPM
information and also to collect status information on completed work
and work in progress. On a daily basis, the completed activities were
noted in the project computer input and a new CPM and resource cal-
culation was made. The resulting resource-usage tables forecasted
workforce trends.
An interesting characteristic of the trend forecast is that, for it to be
effective, the workforce estimating didn’t have to be accurate, just con-
sistent. Accordingly, if the resource computation called for 40 workers
and 50 were assigned, it could be anticipated that the workforce
requirements would show a downward trend as the 50 gain on the work
time estimated for a 40-person crew. If this daily trend remained steady,
it was inferred that the original workforce requirements were too low and
a crew of 50 was the proper size. On the other hand, if the estimated crew
was being used and the workforce requirements trended downward, it
was assumed that the estimates were too conservative.
In the project, the first several daily reports confirmed the
forecasted 18-day project duration. On the sixth daily report, it was
reported that a noncritical area could be completed 2 days early, and
 Equipment and Workforce Planning 299

the 10th report confirmed all earlier trends, which were that all work
would be completed 4 days early. On the 13th day, the unit was
turned over to process and daily reports ceased.
Thus, the trend analysis method was effective. The first four reports
indicated an adequate workforce, which was actually somewhat below
the original projected requirement. On the fifth report, a downward
trend was noticed. Further, it was noted that a shortage of cleaned
bundles for exchangers was causing an excess of available boil-
ermakers. And it was also evident that when bundles became available,
the trend in the craft would reverse and so create a workforce shortage.
The sixth report recommended that a workforce reduction could start.
The next two reports noted that the shift of some of the workforce to
another, unexpected, shutdown would not impede the progress of the
job at hand. With further workforce analysis, it was determined that
on the July 4th holiday, only critical jobs had to be worked, which saved
substantial overtime. Figure 19.4.3 shows an actual report used in the
turnaround.

Figure 19.4.3Turnaround analystical report. (Courtesy: Hydrocarbon Processing and

Petroleum Refiner.)
300 The Practice of CPM Planning

19.5. Examples of Resource Loading on John Doe Project

As another example, the John Doe project, after resource loading, would
generate a graphical resource analysis as shown in Figure 19.5.1.
The number of craftspersons required per day is noted upon the left-
hand scale and also above the bars representing weeks. The cumulative
number of mandays required are noted upon the right-hand scale. What
appears as obvious is that the spikes in late November and early
December are unrealistic, both in terms of season and numbers of craft-
persons. However, if the various crafts are broken down separately, as
shown in Figure 19.5.2, then this initial conclusion is refuted. Although
the total number of craftsmen approaches 80, the maximum for any one
craft is 15.
On the other hand, what would you do if only 10 electricians could be
provided to this site? Figure 19.5.3 shows the use of electricians using
standard CPM scheduling calculations. Figure 19.5.4 shows the use of
electricians if leveling software routines are used limiting the number
to 10. Note that while the number of electricians has been reduced to
the stipulated limit of 10, the project completion date has slipped from
20JUL01 to 03AUG01.

Figure 19.5.1 John Doe project resource histogram and cumulative curve for early dates.
 Equipment and Workforce Planning 301

Figure 19.5.2 Resource histogram with stacked resources.

Figure 19.5.3 John Doe project—unleveled use of electricians.

302 The Practice of CPM Planning

Figure 19.5.4 John Doe project—leveled for a maximum of 10 electricians.

19.6. Resource Leveling Significance

Network planning had its genesis from 1958 to 1964. At that time,
proven computer programs from not only CPM but also PDM and
resource leveling were available. Resource leveling was considered an
option, and it was often not used because of the money, time, and effort
required. However, John W. Fondahl (Civil Engineering Emeritus of
Stanford University) believed that considering resources in scheduling
was so important that he devoted more than half of his Peurfoy
Construction Research Award Lecture* to that topic. Referring to net-
working techniques, he stated:
“They offer examples of our failure to effectively implement techniques
after they have been introduced. Consider the topic of resource leveling.
Almost all construction projects are affected by the availability and eco-
nomics of usage of resources. In most cases, the importance of resource lev-
eling is a matter of reducing costs by avoiding peaks and valleys in daily
requirements. However, in many cases it is essential because there are

*Presented at the November 5, 1990 ASCE National Convention, recorded in the Journal
of Construction Engineering and Management, vol. 117, no. 3, September 1991, pp. 380–392.
 Equipment and Workforce Planning 303

availability limits that must be met. Even where such limits don’t exist ini-
tially, they can be introduced once decisions on plant and equipment capac-
ities have been made or major purchases have occurred.”

While resource-leveling techniques have been available since the

early 1960s, for many years they were largely ignored in CPM appli-
cations. A schedule based on early start dates for each activity was
generally issued. Such a schedule is almost always uneconomical and
is often completely impossible.
Professor Fondahl presented an example problem in which his con-
struction planning and scheduling classes have addressed for more than
25 years. The project involves a small warehouse project with 30
activities. Sequence relationships, durations, and resource requirements
for the resource are given, and the students plot the network diagram
and make the basic scheduling calculations. The resulting project
duration (without resource limitation) is 28 days.
An early-start-based calculation shows a high (20 units) and irregular
requirement for resource C. Sufficient resources are available for
resources A and B. The students manually level the network and
determine that the network can be done with resource C limited to 4
units in 28 days. However, the schedule is now very tight.
In 1977, Professor Fondahl added a contract clause that limited the
extensions of time to changes that exceeded the total float on channels
involved. (Also, time extensions were limited to the amount by which
the total float path was exceeded.) In the problem, a change order that
was issued before the project started affected four activities and
increased their durations. In each case, the duration increase did not
exceed either activity total float or free float. The project was subject to
the same resource limitations, and it was assumed that the contractor
had performed a conventional network analysis without resource
leveling. On the basis of the CPM calculations, with extended
durations, the 28-day duration was not exceeded and no time extension
would be allowed.
By using the same procedures as before, the students determined
that the project duration was extended to 50 days. (Professor Fondahl
notes that resource-leveling answers are heuristic, and that “eyeball”
solutions using “judgment manipulations” can bring the answer to a
rock-bottom solution of 47 days.)
Professor Fondahl then had the students

level both the original problem and the change order problem using one of
the more powerful commercial software packages. The results obtained
are 31 days and 52 days, respectively. An apparent reason for the poorer
results is that, even though this is a sophisticated program, it lacks the abil-
ity to interrupt activities that are interruptible.
304 The Practice of CPM Planning

He described the concept of a “resource critical” activity in this way:

“If critical activities are performed at their estimated duration but start
late, the project duration will also be delayed. One reason for starting late
is that sufficient resources are unavailable in the resource pool. These
resources will only be available when activities using them are completed
and, therefore, those activities are able to release a sufficient number of
the required resource units. Often, those activities that must release
resource units may not be critical in the sense of having zero float. However,
if they fail to release the resource units needed by a critical activity, they
delay that activity and, hence, the project. Therefore, an activity having pos-
itive float can still be “resource critical,” since it will delay completion if it
fails to release resources on time. In more complex networks, there may be
several activities that release resources to the resource pool on a given
date and, therefore, to a critical activity that needs some of these resource
units from the pool. In these cases, there may not be a single specific activ-
ity in the group that must release its units and that, therefore, can be
identified as being “resource critical.”

Some of Professor Fondahl’s conclusions in regard to resource

leveling are:
“The results shown in these simple problems indicate that the conven-
tional concepts of float time break down in a resource-constrained project.
Float times may be much less than computed or may not exist at all, and
project duration may not be determined by the conventionally calculated
critical path. Again, since almost any construction project either must be
resource-leveled to achieve a feasible solution or should be resource-leveled
to achieve an economic solution, we have a problem that is almost always
present but seems to be universally ignored. Some, and often many, of the
conventional CPM data are not valid. The originally calculated network
data, including float data, are useful as a basis for establishing and apply-
ing priority rules in heuristic leveling procedures. However, once leveling
has been performed, float times may have little meaning.
A resource-critical activity is only critical based on the current leveled
schedule. If its duration increases, project duration will increase if the job
program remains the same. However, a new leveling run or a leveling run
with different priority rules may produce a different job program, which
may or may not show a longer project duration. Since it has not yet proved
practical to use mathematically optimum solutions, we must depend on
heuristic solutions whose results, in turn, depend on the particular “fit” that
is achieved on a given leveling run. Thus a 1-day delay of a resource criti-
cal activity might, if releveling were performed, produce a one-day project
delay or no delay, or more than a 1-day delay.
In summary, on this subject of implementation, I am using this example
to say that, after 30 years, very few practitioners, or even those teaching
the subject, seem to be aware of some of the basic shortcomings of widely
used network scheduling techniques.”
 Equipment and Workforce Planning 305

19.7. Summary
Resources such as workforce, equipment and money, can be assigned to
CPM activities. For a simple network, maximum workforce require-
ments can be forecast and leveled by two manual techniques. When
resource limits are set in, the project duration can be lengthened. Manual
techniques are limited and cannot handle large networks. Large net-
works are an excellent area for computer application.
When a schedule is resource-constrained (i.e., certain resources are
not available to support critical activities), the network float concept no
longer controls identification of controlling activities. In multiproject
systems, such as turnarounds, the identification of the critical path is
often less important than the cataloging of all the work to be done.
This page intentionally left blank
 Chapter

20
Procurement

Materials are integral to any construction project. If materials are deliv-

ered early, a particular activity usually cannot be sped up because the
progress of other activities controls the activity’s early start time. Failure
to deliver the materials for an activity can, however, delay the activity
indefinitely. Thus, the project purchasing agent or materials coordina-
tor has a difficult job. If materials are delivered late, the project is
delayed. If they are delivered early, the field group complains about
extra handling and storage. The problem is most acute in urban areas,
where project supervisors generally want to unload materials immedi-
ately from truck or railcar to final location.

20.1. Scheduling Materials Procurement

Just as subcontractors complain that the general contractor neglects
their situations, most purchasing agents complain that their own com-
panies fail to keep them informed about material needs. Obviously, this
problem can be solved with CPM. Because almost every activity requires
materials of some sort, someone would have to review all of the activi-
ties in order to control the delivery of all the materials.
A practical method of reducing the workload is to separate materials
into two classes: commodities and key materials. Materials that can be
ordered out of stock for delivery in a week or less can be classified as
commodities, and the schedule for the first shipment of any type of com-
modity is useful. Key materials are those with long delivery times or
those that involve custom orders.
Reviewing the network computer run can furnish all of the necessary
information about materials, particularly the order in which key mate-
rials should be requisitioned. If an activity is added to the diagram for

307

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
308 The Practice of CPM Planning

Figure 20.1.1 Deliveries for John Doe project site preparation; zero delivery.

each key delivery, necessary information about the delivery is generated

as part of the computer run.
Figure 20.1.1 shows the site preparation network for the John Doe
project with the delivery activity shown in Table 20.1.1.
Note that well pump and water tank would definitely be key deliveries.
The other materials could be commodities or custom items depending on

TABLE 20.1.1 Material Delivery Activities

for John Doe Project

Activity Delivery

0–4 Well pump

0–5 Underground water pump
0–6 Water tank parts
0–7 Tank valves and piping
0–9 Sewer pipe
0–10 Manhole frame and cover
0–11 Conduit
0–12 Power cable
0–150 Poles, crossbars, guys, insulators
 Procurement 309

TABLE 20.1.2 Calculations if All Materials Available

Duration Float
i–j (days) Description ES EF LS LF (days)

0–4 0 Well pump 0 0 22 22 22

0–5 0 Underground pipe 0 0 24 24 24
0–6 0 Water tank 0 0 12 12 12
0–7 0 Tank valves 0 0 22 22 22
0–9 0 Sewer pipe 0 0 21 21 21
0–10 0 Manhole frame and cover 0 0 21 21 21
0–11 0 Conduit 0 0 26 26 26
0–12 0 Power feeder 0 0 29 29 29
0–150 0 Pole line materials 0 0 23 23 23

the specifications to be met. If it is assumed that all materials are on

hand (for instance, if the owner is furnishing them), the time duration
for the activities is zero. The computed information for the deliveries is
shown in Table 20.1.2.
Because materials are not usually available at the start of a project,
a reasonable delivery time estimate is assigned to the delivery activities,
as shown in Table 20.1.3. These durations are added to Figure 20.1.2.
Event times are computed on the diagram. The activity times for
deliveries are shown in Table 20.1.4.
Introducing delivery times increases this portion of the project from
34 to 52 days. The critical path has shifted; it is now through events
0–6–7–8–13. The old and new event times are shown in Table 20.1.5.
Out of a possible 28 event times, 21 have changed. Using the new late-
start information, the purchasing department can deliver the materi-
als in the order shown in Tables 20.1.6 and 20.1.7.

TABLE 20.1.3 Durations for Procurement of Materials

Duration
Activity Assume (days)

0–4 Well pump Stock delivery, 4 weeks 20

0–5 Underground water pipe Mechanical joint, 6 weeks 30
0–6 Water tank parts Standard size, 6 weeks 30
0–7 Tank valves Standard gate valves, 4 weeks 20
0–9 Sewer pipe Terra cotta, 1 week 5
0–10 Manhole cover Stock, 1 week 5
0–11 Conduit Stock, 1 week 5
0–12 Power feeder Special order, 8 weeks 40
0–150 Pole material Stock order, 2 weeks 10
310 The Practice of CPM Planning

Figure 20.1.2 Delivery times for John Doe project site preparation.

Although the list gives the order in which materials should be ordered,
it has two distinct weaknesses. First, although the late start dates for
ordering are important, they are extremes. If the order is placed this late,
all activities following the delivery will be critical. Second, the early
start times have very little value. In this example, the purchasing
department could initiate nine orders the first day of the project. What

TABLE 20.1.4 Calculations Including Procurement Time for Materials

Duration Float
i–j (days) Description ES EF LS LF (days)

0–4 20 Well pump 0 20 20 40 20

0–5 30 Underground water pipe 0 30 12 42 12
0–6 30 Water tank 0 30 0 30 0
0–7 20 Tank valves 0 20 20 40 20
0–9 5 Sewer pipe 0 5 34 39 34
0–10 5 Manhole cover 0 5 34 39 34
0–11 5 Conduit 0 5 39 44 39
0–12 40 Power feeder 0 40 7 47 7
0–150 10 Pole material 0 10 31 41 31
 Procurement 311

TABLE 20.1.5 Impact of Procurement Durations

Early event times Late event times

Old New Event Old New

3 3 1 3 21
5 5 2 5 23
7 7 3 7 25
22 22 4 22 40
24 30 5 24 42
11 30 6 12 30
21 40 7 22 40
32 50 8 32 50
17 17 9 21 39
8 8 10 21 39
22 22 11 26 44
25 40 12 29 47
34 52 13 34 52
— 10 150 — 41
Changes 7 Changes 14

if an enthusiastic buyer orders the sewer pipe and conduit on the first
project day? The conduit will arrive on site about 8 weeks before it is
needed; the sewer pipe will be 7 weeks early. The field group will have
a storage problem and develop a poor opinion of the office group.
These problems have often discouraged the use of CPM for coordi-
nating materials procurement. The real defect in the system is that the
early start time is unrelated to the field work. Leaving the delivery
arrows to represent delivery times, adds another set of arrows to rep-
resent the actual movement of the material from storage to the job site.
The “on-site material” arrows have zero time duration and the same late
finish times as the delivery arrows.
Figure 20.1.3 shows the nine new arrows. Because they have a zero
time duration, early start equals early finish and late finish equals late
start. The ES, LF, and float times are shown in Table 20.1.8.
Note that the late finish times for these activities are the same as the
late finish times for the delivery arrows. However, the early start times
and float times are now related to the field progress. On this basis, the

TABLE 20.1.6 Most Critical Procurement

Activity Description Late start

0–6 Water tank 0

0–12 Power feeder 7
0–5 Underground water pipe 12
312 The Practice of CPM Planning

TABLE 20.1.7 Less Critical Procurement

Activity Description Late start

0–4 Well pump 20

0–7 Tank valves 20
0–150 Pole material 31
0–9 Sewer pipe 34
0–10 Manhole cover 34
0–11 Conduit 39

Figure 20.1.3 On-site delivery times.

TABLE 20.1.8 Calculated On-site Delivery Times

Activity Description ES LF Float (days)

4–104 Well pump at site 22 40 18

5–105 Underground pipe at site 30 42 12
6–106 Water tank at site 30 30 0
7–107 Tank valves at site 40 40 0
9–109 Sewer pipe at site 17 39 22
10–110 Manhole cover at site 8 39 31
11–111 Conduit at site 22 44 22
12–112 Power feeder at site 40 47 7
150–152 Pole material at site 10 41 31
 Procurement 313

TABLE 20.1.9 Priority of Material Procurement

Position
on first Delivery as Delivery no
Priority order list early as later than Float (days)

1. Water tank 1 30 30 0
2. Tank valves 5 40 40 0
3. Power feeder 2 40 47 7
4. Underground pipe 3 30 42 12
5. Well pump 4 22 40 18
6. Sewer pipe 7 17 39 22
7. Conduit 9 22 44 22
8. Manhole cover at site 8 8 39 31
9. Pole material 6 10 41 31

priority of ordering is as shown in Table 20.1.9. Note that all but two of
the items are in a different position of priority on the second list.
In addition to the time required for material delivery and the determi-
nation of the delivery time, which should be specified on the order, a number
of other steps in materials procurement are time consuming and must not
be neglected. These can include approving shop drawings, the architect’s
review of the shop drawings, a resubmittal time for any shop drawing cor-
rections, and review by other agencies. These steps can sometimes be accel-
erated for critical activities (when they are, in fact, identified as critical).
However, there is a tendency to minimize the impact of routine steps, so
take care to properly reflect them on your diagram.
Figure 20.1.4 shows the interrelation between two material orders
(hardware and door bucks) before either material reaches the job site.
Note that in this example the door buck delivery has 5 days float because

Figure 20.1.4 Typical material procurement cycle.

314 The Practice of CPM Planning

of the time required to prepare hardware templates. Larger equipment

might require additional time for the submission of formal bids. In
Figure 20.1.1, the addition of 9 simple delivery arrows almost doubled
the network size. In this network, the number of arrows showing the
total materials procurement could easily be more than double the
number of arrows showing the associated field work.
Because the average project requires several separate sheets to rep-
resent its network, it is recommended that the materials procurement
work be on its own sheet to avoid confusion between the office and field
functions. Of course, the “materials at site” arrows must remain with
the field portion of the network. Figure 20.1.5 shows the materials por-
tion of the John Doe project, and some typical material lead times for a
process plant project are shown in Table 20.1.10.

Figure 20.1.5 John Doe project material procurement.

TABLE 20.1.10 Material Lead Times for a Process Plant

Approval of Anticipated delivery

drawings, (after approval and release),
weeks weeks

Building

Enclosure
Structural steel 4–6 8–13
Steel joists 2–4 8–10
Siding 3–4 13–26
Mechanical
HVAC-fans 2–4 13–18
HVAC-chillers 4–6 18–26
Agitators/mixers 6–8 26–32
Centrifugal blowers 4–6 20–26
Compressors (packaged centrifugal) 8–10 26–39
Compressors (packaged reciprocating) 6–8 26–30
Electrical equipment
Motor control centers 8–10 26–40
Switch gear (low voltage) 8–10 36–40
Switch gear (high voltage) 8–10 40–52
Transformers (low voltage) 6–8 30–39
Transformers (high voltage) 6–8 40–52
Motors (to 150 hp) 6–8 16–26
Motors (over 150 hp) 6–8 26–39
(dependent on horsepower)
Turbines 8–10 40–50
Power cable (600 V) N/R 30–52
(dependent on quantity)
Bus duct 6–8 26–36
Cable tray 6–8 18–26
Conduit (rigid aluminum) N/R Stock–28
Conduit (E.M.T.) N/R Stock–26
Emergency generators 10–12 26–30
Architectural
Hollow metal frames 8–10 12–18
Hardware 10–12 18–26

Process equipment

Pressure vessels (carbon steel)

Small (noncode) 4–6 18–26∗
Small (—under 20,000 lb) 4–6 26–36∗
Large (code—over 20,000 lb) 6–8 36–40
Towers (w/o internals/trays) 6–8 46–50
Towers (with internals/trays) 8–10 52–60
Jacketed vessels/tanks 8–10 52–60
Field-erected tanks 8–10 40–52
(includes erection)
Heat exchangers
Shell and tube (small) 4–6 18–20
Shell and tube (large) 6–8 36–46
Fintube 4–6 18–26
Plate type 4–6 36–40
Air-cooled exchangers 4–6 26–36

(Continued)

315
316 The Practice of CPM Planning

TABLE 20.1.10 Material Lead Times for a Process Plant (Continued)

Approval of Anticipated delivery

drawings, (after approval and release),
weeks weeks

Process equipment

Conveyors
Pneumatic 6–8 26–30
Screw 6–8 24–30
Live roller and drag 6–8 24–28
Vibrating 6–8 26–30
Bucket elevators 6–8 26–30
Belt 6–8 30–34
Pumps
Centrifugal 4–6 20–26
Centrifugal (horizontal) 6–8 26–32
Centrifugal (turbine) 6–8 24–30
Metering 4–6 20–34
Positive displacement 4–6 20–24
Vacuum 6–8 26–30
Reciprocating 6–8 26–30
Dryers, filters, and scrubbers
Instrument air dryers 8–10 24–30
Filters 6–8 20–26
Dust collectors 6–8 30–40
Fume scrubbers 6–8 20–30
Control valves 3–4 20–24
Instrumentation
Displacement-type flowmeters 3–4 18–26
D.P. transmitters 4–5 16–22
Liquid level gauges 3–4 18–20
Transducers 3–4 14–28
Level switches 3–4 12–16
Pressure switches 3–4 16–18
Controllers 4–5 18–20
Recorders 4–5 18–20
Thermometers 3–4 14–16
Pressure gauges 3–4 16–20
Pipe, valves, flanges, and fittings N/A Stock to 52 weeks

Materials handling equipment

Monorail hoists 4–6 18–26

(dependent on capacity)
Traveling/trolley cranes 4–6 30–42
(dependent on capacity)
Forklift trucks 4–6 26–30
∗
Add 4 weeks for stainless.
 Procurement 317

20.2. John Doe Example

As noted previously, it is not usual to incorporate commodity or stock items
in the CPM network. Many projects have a three- to nine-month excava-
tion, foundation (piles), and foundation concrete phase with a short cycle
startup for the design mix and rebar delivery. In this typical situation, the
site and foundation work schedule has float built into it for the procurement
process.
In the John Doe site example, the owner would do well to provide
the water tank and well pump. Another approach would be to evalu-
ate the requirement to provide all site services prior to event 34. If the
site activities could be put in parallel with the foundation work, more
time would be available for site equipment procurement.
The site equipment procurement has been treated previously. To con-
sider procurement for the balance of the John Doe project, the network
should be modified to create more definitive delivery points. For
instance, event 37 is a common starting point for all plant activities.
Accordingly, it is not the best delivery node. Adding logic spreaders
between event 37 and key delivery points will establish more definitive
delivery information. This is shown in Figure 20.2.1.
Table 20.2.1 lists the sequence of procurement activities (i.e., submit
and approve shop drawings, fabricate, and deliver) for 14 items. These 14

TABLE 20.2.1 John Doe Project Procurement

Submit Approve Fabricate

shop shop and
drawings, drawings, deliver,
Starting work work work
Item event days Event days Event days Event

Foundation rebar 0 10 210 10 211 10 16

Structural steel 0 20 212 10 213 40 23
Crane 0 20 214 10 215 50 31
Bar joists 0 20 216 10 217 30 33
Siding 0 20 218 10 219 40 35
Plant electrical load
center 0 20 220 10 221 90 300
Power panels—plant 0 20 222 10 223 75 301
Exterior doors 0 20 224 10 225 80 303
Plant electrical
fixtures 0 30 225 15 226 75 51
Plant heating and
ventilating fans 0 20 227 10 228 75 304
Boiler 0 20 229 10 230 60 306
Oil tank 0 20 231 10 232 50 305
Precast 0 40 223 10 234 30 58
Packaging A/C 0 30 235 10 236 90 60
318 The Practice of CPM Planning

Figure 20.2.1 John Doe plant with added delivery points.

items were added to the computer master file and a new computation was
made. The procurement portion of the John Doe project (after the site
work) is listed by late start (in order of float priority) in Figure 20.2.2.
Note that the example procurement times are in the expeditious
range. If the times, especially for switch gear, were taken from the prior
typical procurement time tables, procurement would control the sched-
ule. Assuming this is unacceptable, the owner has two choices: either
expedite (i.e., shorten) the procurement dates or preorder (i.e., order
before selecting the contractor) key equipment, such as the well pump,
water tank, electrical switch gear, and steel.
Figure 20.2.3 is a partial sort of the John Doe project by specifica-
tion section. It can be used by the purchasing department when
preparing subcontracts or purchase orders to determine the scope of
work under each specification section.
Figure 20.2.2 Procurement activities sorted by late start.

Figure 20.2.3 Partial list of John Doe project, sorted by specification section (Second
code field).

319
320 The Practice of CPM Planning

20.3. Summary
If procurement is ignored in the scheduling process, materials and
equipment deliveries can become the controlling factors by default. In
most major projects, there is enough nonmaterials-oriented front-end
work to allow time to order materials through the contractor. However,
in special situations (renovations, overseas projects, and/or fast-track
projects) it might be necessary for the owner to preorder equipment or
materials.
 Chapter

21
Preconstruction

In practice, and in this book, we have emphasized the use of CPM in

planning and implementing construction. If a project is considered in
terms of its construction phase only, the application of CPM can save
both time and money. When used only in the construction phase, CPM
is used as a control. However, many advantages other than just con-
struction activities can be achieved by the earlier use of CPM in a
project.
Construction is that time when the iceberg emerges and the entire
project can be viewed and understood by many people. Problems are evi-
dent and activity is manifest. In most projects today, however, the time
spent on construction is equaled by the time spent on the precon-
struction design phase. Further, in public projects, the administrative
review cycle often equals in time the design and construction periods.
Thus, in the public or quasipublic sectors, the preconstruction project
time (following identification of the project in a budget) is often twice
the actual construction period.
Obviously, if disciplined project control techniques are applied early,
substantial time can be saved and, therefore, subsequent cost savings
can be achieved at a relatively low unit cost. In fact, the preconstruc-
tion phase of a project is the most probable area for applying cost-
optimization techniques. Inputting additional funding during this phase
can result in tremendous reductions in time.
Figure 21.0.1 illustrates the typical cash flow over 62 months between
the budgeting and the opening of an elementary school. Although the
school authorities on the date of budget approval might have the feel-
ing that they have spent or committed the entire amount of money for
the school, they have actually spent less than 20% of the overall budget
for the next 40 months.

321

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
322 The Practice of CPM Planning

Figure 21.0.1 Cash flow typical elementary school.

In this example, in-house staff costs are not recognized, but they
should be, because they can add another 10% or so to the cost of the proj-
ect. Also, these costs are more heavily drawn on in the project’s first
20 months than later, usually diminishing as others carry the project
to completion. This example is a perfect illustration of why it is not easy
to spend money even when the decision to do so has been reached.
The typical building project has four major phases or categories of
progress:

1. Predesign. Predesign activities is the period between budget prepa-

ration and approval and the initiation of design. Usually the owner
has primary responsibility for progress, including programming, in
this phase.
2. Design. This is the phase in which the architect, engineer, A-E, or in-
house design staff is primarily responsible for the project’s progress.
3. Construction. In this phase it is the contractor or in-house construc-
tion force responsible for project progress.
4. Furnish or move-in, with the owner or contractor having primary
responsibility.
 Preconstruction 323

21.1. Predesign Phase

One of the least-defined, intangible, and time-consuming phases of a
project is the predesign portion. During this period, the owner, with
technical staff and/or consultants, is very busy with a number of impor-
tant roles, often performed by omission or default rather than carried
out in a rigorous, planned manner. The seeds of many project problems
are planted in this field of neglect.
Most all major projects seem to result from an evolutionary type of
aggregate thinking from many sources that gathers pressure, both polit-
ical and personal, until the project has been articulated. Key charac-
teristics are power structure and consensus. Actually, this phase of the
project should go through the following stages: establishment of goals,
means of accomplishing goals, decision to proceed, identification of a
funding source, and budget approval.
The decision to proceed requires identifying specific projects and the
development of preliminary cost estimates, usually on the basis of gross
estimating factors, such as costs per square or cubic foot. After a proj-
ect has been given a budget and funding is available to meet the budget,
the predesign phase moves to other stages.
Site selection, such as for a hospital addition or a school replacement
or other finite location situation, is often part of the basic decision to pro-
ceed with a project. In many cases, however, a new site should or must
be considered. Usually site consideration precedes the selection of a
designer, because the design should be a function of the site. A number
of nontechnical factors may funnel the choice of a site into a specific
direction, such as
1. Encumbrances. Are there tenants who will have to be relocated? Are
there structures to be removed?
2. Land cost. What are the economic values and factors?
3. Transportation. Is the location adequately served, and is it served by
media suitable to the character of the facility’s needs?
4. Utilities. What are the availability? Are there potential problems?
5. Neighborhoods. Is the environment suitable for the facility? Is the
facility suitable for the environment?
6. Zoning. Does local zoning conform to the use intended?
7. Community. How will the community react to the facility?
8. Subsurface conditions. Will the foundations require unusual sup-
port? Are there unusual problems to be overcome?

There are other factors, but it is clear that when choosing a site, many
factors must be carefully evaluated and considered. Unfortunately, many
324 The Practice of CPM Planning

site considerations are often considered in hindsight rather than at the

proper time.
The last predesign activity is developing a specific program that
identifies the owner’s intent regarding the functional use of the proj-
ect. This philosophical statement is important to the designer, but it is
often presented in such a perfunctory, nonspecific fashion that the
designer, through trial-and-error, ends up establishing the philosophy.
It is clearly the owner’s responsibility to establish these requirements
and to interpret them in terms of cost impact prior to selecting a
designer. Programming is a unique talent, requiring a combination of
a consultant’s knowledge and expertise.
Functional planning requires the availability or the assembly of per-
tinent information regarding the project. Demographic sources, such as
the U.S. Census Bureau, city and state planning organizations, and in-
house sources should be reviewed. Information should be arranged and
stored so that it is accessible for reviewing future projects or for recon-
sidering the project underway. If information is stored in a computer
databank, such exercises as modeling, gaming, or simulation of various
alternatives, can be used to test the results of different potential
approaches.
The functional programming effort should be tied back to the budg-
etary estimate, which it should either affirm or revise. Since the func-
tional program incorporates the policy in regard to any project, it should
be approved by the owner or proper authority.
A concomitant to the functional program is the architectural pro-
gram, to which the functional program is necessarily related. The archi-
tectural program can be incorporated in the schematic design phase by
the architect or furnished to the architect.
Typically, projects do not have formal program documents, which
results in uncertainty at the beginning of the design phase. Because
designers are not compensated for uncertainty, their only defense is to
proceed slowly during the early stages of design, developing a program-
type statement that can be confirmed or revised by their clients at a later
time. Unfortunately, clients often change their minds almost constantly.
From the design point of view, this is not only time-consuming but
expensive. Virtually the only defense designers have to carefully control
the progress of the design, holding off every activity until a high level
of definition has been achieved. This is expensive to designers and to
owners, because the true design work is placed in too short a time for
economical implementation.
The predesign phase is a frustrating one for schedulers and schedules.
Many factors influence the viability of a project. In most cases, timing
is not the controlling factor, but in a few situations it is paramount. The
1976 bicentennial celebration was a fixed time frame. World fairs and
 Preconstruction 325

Olympic games have a similar challenge of well-publicized fixed com-

pletion dates.

21.2. Design
Designing a project involves a relatively complex series of activities that
become increasingly detailed as a project moves through the various
design phases of schematic development, preliminary design, and work-
ing drawings.

Schematic development. This is also called the sketch phase, during

which “concept” plans are developed by the architect and the basic engi-
neering system is made. Design criteria are specified, and schematic
drawings are prepared. A set of perspective sketches, or renderings, is
usually prepared. The basic budgetary cost is confirmed, but only in very
broad terms.

Preliminary design. This phase, also called design development, occurs

after the approval of the sketch or schematic phase. The drawings are
refined to a degree sufficient to permit the development of dimensioned
space layouts. Heating and ventilating systems, main feeders or ducts,
electrical main feeders, and dendrite dimensions of the structural frame-
work are identified in this phase. Utility requirements are also defined
and specific requirements are determined. A preliminary cost estimate
is prepared, relatively firm at this stage.

Working drawings. This phase is also termed contract documents or

final design and includes about two-thirds of the design work but fewer
of the decisions, as well as a disproportionate amount of the design
period (usually about half). The design, as defined in the prior stages,
is developed in complete detail, including dimensions, so that it can be
priced by prospective contractors. The contract documents include both
drawings and specifications.
As the project design proceeds, each change becomes more difficult and
expensive to implement. Each revision requires many more reviews
and changes in related items. The range of changes that can be accepted
gracefully narrows and costs much more as the design phase proceeds.
This is illustrated in Figure 21.2.1.
In most cases, the design phase is essentially unscheduled and unco-
ordinated, even by the designer. This is partially understandable
because specific interconnections among phases and disciplines, such as
structural, mechanical, electrical, and plumbing, are difficult to express.
Nevertheless, the design phase should be closely coordinated and inter-
connected at all stages so that field work will run smoothly. Otherwise,
326 The Practice of CPM Planning

Figure 21.2.1 Design change funnel.

Figure 21.2.2 John Doe schematic design.

Figure 21.2.3 John Doe preliminary design.

 Preconstruction 327

Figure 21.2.4 Contract documents: plant & warehouse.

it could result in a fantastically large scheduling network. The usual

compromise is the scheduling of concurrent activities in broader terms
with an implied understanding that continual physical liaison must be
carried out.
Figure 21.2.2 is a CPM network for the schematic phase of design for
the John Doe project, and Figure 21.2.3 is the network for the following
preliminary design phase. Note that the design splits into two pack-
ages—plant-warehouse (P/W) and office—in the preliminary design
phase. Figure 21.2.4 is the CPM for preparing contract documents for
the plant-warehouse, and Figure 21.2.5 is the CPM for preparing
contract documents for the office. Figures 21.2.6 to 21.2.9 are Figures
21.2.2 to 21.2.5 with CPM calculations added. Figure 21.2.10 is a

Figure 21.2.5 Contract documents: office.

328 The Practice of CPM Planning

Figure 21.2.6 John Doe schematic design.

Figure 21.2.7 John Doe preliminary design.

Figure 21.2.8 Contract documents: plant & warehouse.

 Preconstruction 329

Figure 21.2.9 Contract documents: office.

summary level CPM that shows the entire preconstruction plan for
the design of the John Doe project.
Figure 21.2.11 shows a network representing the design stages of a
city school project. In this case, the project is located within city limits,
so that the usual agency reviews are required. Note the “rejection cycle,”
which is a loop and cannot be computerized. It is in shorthand to indi-
cate that the full schematic design cycle sequence is represented (pre-
sumably with shorter durations). Because projects such as John Doe are
placed in industrial parks, fewer reviews are required, and any reviews
are generally required by the state more often than by a township or
county. However, the site development of an industrial park is not inex-
pensive and should itself be planned as illustrated in Figure 21.2.12.
During the design stage, there is a continual interplay between the
designer and the owner. The owner reviews the design at major points
in development and should be available daily for information. Quite
often, the owner is furnishing or specifying special equipment that
requires his or her attention. Both architect and owner are involved in
various agency or company reviews.

Figure 21.2.10 Summary level CPM for John Doe project.

330

Figure 21.2.11 Network for city school design phase.

 Figure 21.2.11 (Continued).
331
332 The Practice of CPM Planning

Figure 21.2.12 CPM plan for site development.

The design phase offers a tremendous potential for time gains or

losses. When an owner is handling many projects concurrently, it is
good practice to use resource allocation for design and management staff
so that project progress of all projects is calculated and structured rather
than left to chance.
Just as the designer typically exercises patience until external pressures
force a decision from the owner, the owner, assuming that the designer
can work around problems without losing substantial time, typically
delays decisions.
With so many people concerned about and responsible for a project,
substantial periods of time are usually spent in review and administra-
tive planning. Often, these activities are overplayed as each individual
tends to see his or her own part in the project as the most important and,
therefore, is willing to take more than their fair share of project time in
arriving at an important decision or confirmation. Also, in the early plan-
ning stages, people generally do not regard the planning time they use
as really affecting the final delivery date because it seems so far away.
In reviewing one Department of Labor administrative project, a net-
work was established that resulted in the review cycles being reduced
and a better chain of responsible personnel established. However, one
startling fact that emerged was the physical handling of the documents
that required review. It was found that the internal mailing system was
so slow that it took up 20% of the preconstruction phase of the project.
Because of the importance of the project, this was changed and all doc-
uments hand-carried to and from project personnel.
 Preconstruction 333

21.3. Summary
To achieve the real benefits of logic and control through network analy-
sis, project management should be instituted as early as practicable,
preferably about the time a project is identified in a budget. Installation
and implementation of CPM in the actual construction phase is of great
importance, but many opportunities to save time and money will be
missed if this control is instituted too late.
This page intentionally left blank
 Chapter

22
Evolution of the Project Schedule

CPM separates planning and scheduling, and once project information

is collected and expressed as a network plan and activity time estimates
assigned, CPM calculations can be made. Planning ceases and schedul-
ing starts when the first computation is performed that shows a project
duration. The project duration is then compared with the desired sched-
ule and scheduling begins. The first comparison is for end date.

22.1. Preliminary Schedule

The owner sets the schedule using the advice from the designer and other
confidants, but generally undertakes establishing the schedule person-
ally. The typical schedule is usually a tight one—intentionally and unin-
tentionally—and reflects the requirements that the project fulfill for the
owner. (Often by the time the design is completed, much of the time orig-
inally available has been used in the preconstruction stages.) The acci-
dentally short construction schedule occurs when the owner is not
knowledgeable of, or realistic about, the time necessary to construct the
project at hand.
Completion of a construction project is not only key to the owner but
to the contractor as well. The contractor must have some definite opin-
ion about the overall length of the project in order to make a meaningful
bid on it. Inevitably there are additional costs, such as increased wages
and basic overhead, that are tied to the length of the job rather than to
the specific level of field activity the job entails. Although some of the con-
tractor’s overhead can be spread throughout the job, some contingent
amount must be included in the bid to cover possible risk or exposure to
an extended contract.

335

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
336 The Practice of CPM Planning

22.2. Preconstruction Analysis

An owner who includes only a completion date in the contract has very
little control of the progress of the job. To establish a more feasible sched-
ule, many owners are using a preconstruction evaluation completed by
their staff, consultants, or the construction manager, if one has already
been assigned, to use as the basis of evaluating their schedule. A pre-
construction study by knowledgeable staff can inform the owner that a
reasonable contractor under normal circumstances cannot meet their
date. The owner will then have a number of alternatives.
One alternative is to identify the contract time as a tight one and pro-
gram overtime into the project on a preset basis, such as 6 or 7 days a
week. Another option is to require that the contractor work double shifts,
although this can result in severe budget ramifications. Also, such an
approach must be evaluated in terms of area work practice. Some labor
unions require full premium for double shifts, whereas others impose
only a nominal increase. Some unions will not work on an accelerated
basis, regardless of salary premium. Another alternative is for the pre-
construction schedule study team to establish a phased, projected series
of dates at which parts of the project can be taken over by the owner. Often,
this option meets the owner’s true requirements: The phasing is made part
of the contract, and no additional cost is programmed into the project.
When a preconstruction evaluation has been made, the owner has two
basic approaches. First, he can state that the study validates the required
completion dates and include a contractual that stipulates damages are
based on reality and will be imposed. Second, the owner can furnish the
preconstruction evaluation to all of the bidders. In the first instance, the
scheduling information provided to the contractor is only a narrative
statement. The owner does not include the results of the study as part of
the contract documents.
The recommended approach is to include a summary network and/or
computer run of the network for the bidding contractors to use. This
section can be marked “Information only,” but it gives the contractors a
rapid method of evaluating one way in which the project can be accom-
plished. When more detailed scheduling information is offered, it should
also be conditioned in this way. Including a network does not mean that
the contractor must perform the project in a specific manner. Instead, it
suggests one way to do it. The owner is, after all, attempting to buy the
contractors’ innovative thinking as one of his basic skills.

22.3. Contractor Preconstruction Analysis

In most cases, bidding contractors do not make a serious evaluation of the
contractual time requirements unless the requirements are unusually
and obviously stringent. Twenty years ago, liquidated damages assigned
by engineers were usually a slap on the wrist of $100 per day. (Compare
 Evolution of the Project Schedule 337

this with the hospital that had a $200,000 per month, or $6,700 per cal-
endar day, time damage.) Even today, liquidated damages are generally
set fairly low.
A contractor who responds to a bid by including a condition is defi-
nitely found to be nonresponsive by public agencies and may be found non-
responsive by private organizations. The contractor whose bid questions
or conditions the time frame of a contract is usually rejected. Therefore,
most contractors will not do so, but they may state their reservations
about the projected dates after the award of the contract.
Experienced contractors know that there will be unforeseen conditions
and unexpected situations for which time extensions will be allowed.
Contractors also expect changes by owners and anticipate that either the
owners will relax end dates or, if need be, they will successfully handle
any delay claims by the owners. Further, liquidated damages have tra-
ditionally been set too low by owners who are unaware that their claims
for damages are usually limited to the liquidated damages specified.

22.4. Milestones
A preconstruction schedule can be used to develop something more than
an end date. A network evaluation can identify key milestones. The analy-
sis tells the owner that if certain things do not occur by certain stages of
the project, there is no way in which the end date can be met. Therefore,
the section in the contract on scheduling can establish the milestones as
specific days following the notice to proceed.
Normally, the only scheduling requirement included in a contract is the
end date by which the contractor agrees to complete the project. Although
general language is usually included that stipulates the contractor must
remain on schedule, when contractors run behind, they can always allege
that they are going to increase the workforce, work overtime when
required, or bring more subcontractors onto the project. There are usu-
ally no definite means of establishing that they have failed to meet their
contractual obligations.
Establishing milestones as a contractual requirement helps the owner
to control the project’s progress, and it provides a definite area to control
the performance of the contractor. The contract language should, however,
be flexible enough to permit the owner to adjust milestone dates if a con-
tractor requests it and can demonstrate a realistic means of readjusting
the schedule. Requests such as this should be in writing and require the
signature of the owner.
Typical milestones include the completion of foundations, structures,
close-ins and watertightness of structures, start of temporary heat,
complete basic air handling system, complete permanent heat, and com-
plete lighting systems. Milestone dates can also be established by area.
Thus, in a hospital, certain areas could be designated for acceptance by
338 The Practice of CPM Planning

the owner in stages. Typical initial areas are the ambulatory care and staff
administrative spaces. If the owner intends to take phased occupancy, the
decision should be made early in the design stage so the layout of the facil-
ity will reflect the incremental occupancy intended. Also, the mechanical
and electrical systems may require controls by local area.

22.5. The John Doe Schedule

Figure 22.5.1 shows, on a summary level, a basic 429 workday schedule
suggested by the John Doe networks. If the end date of the initial plan is
later than the desired date, the first area to examine is the critical path.
There are two distinct methods for shortening the critical path. First,
examine the path for series sequences that could be parallel. For
instance, in the John Doe project, the critical path could be shortened
by 74 days if the company were able to revise its ground rule about
doing the office building after the warehouse. If that is not possible, other
possible areas of overlap should be studied:

1. In the foundation contract, perform the pile caps (16–17) and grade
beams (17–18) in parallel rather than in series. Time savings, 5 days.
2. In the foundation contract, perform the underslab plumbing (19–20) and
conduit (20–22) in parallel rather than in series. Time savings, 5 days.
3. Perform the floor slabs (22–29) in parallel with the structural steel
and craneway erection (29–33). This would be possible by working
from opposite ends of the building. Time savings, 10 days.
4. Start siding erection earlier, at event 33 instead of 35. Time savings,
5 days.

If all of these changes were implemented, it would result in a total

time savings of 25 days. The plant-warehouse area does not offer much
opportunity for time savings because several paths would have to be
shortened. As noted in Chapter 20, Procurement, 18 days can be saved
by prepurchasing the well pump and water tank.
If the time reduction of 43 days is not sufficient, re-examine the crit-
ical activities with longer durations. Perhaps by adding equipment and

Figure 22.5.1 John Doe schedule without expediting summary level.

 Evolution of the Project Schedule 339

increasing the workforce, the time for some of the critical activities
could be shortened. For instance:
15–16 Shorten “excavation” from 5 to 3 days.
16–17 Shorten “pour pile caps” from 5 to 3 days.
17–18 Shorten “grade beams” from 10 to 5 days.

And so on. Take care not to arbitrarily shorten durations. An unfor-

tunate tendency is to be optimistic when estimating the project time
required for an activity. Some people fall into the trap of using the best
time they have ever experienced. Further, it is easy to overlook the time
inevitably lost in coordinating many activities. Experienced estimators
include this factor in their estimates.

22.6. Resources
The CPM calculation assumes unlimited resources, that is, enough
people and equipment available to do each activity. This is fairly rea-
sonable and can usually be maintained for critical activities. However,
the superintendent must use float time as a guide in spreading out crew
assignments. Although it is theoretically possible to call workers out one
day and lay them off the next, no sensible contractor wants this repu-
tation with craftsmen or small subcontractors.
To set up the CPM schedule to take this situation into account, crew
scheduling arrows can be added. In Figure 22.6.1, which shows the John
Doe project site work, note the access road and parking lot. If they are to

Figure 22.6.1 Lead and sequence arrows; site work.

340 The Practice of CPM Planning

be done by the same contractor, it would be reasonable to schedule them in

series rather than in parallel. This can be done by changing activities 92–58
to 92–115 and then adding one sequencing arrow 115–90. That schedules
the access road before the parking lot and allows the same paving equip-
ment to be used on both activities. It also reduces total float by five days.
Other examples can be seen in the John Doe project foundation work
(Fig. 22.6.2). For instance, the general contractor has to call in the
plumber and the electrician for underslab work. One reasonable method
is to schedule the critical work first. To do this, a CPM computation must
be made before the schedule sequence arrows are added. In the network,
the plant-warehouse work is critical, so the addition of sequence arrow
20–26 will schedule the office building underslab plumbing after the crit-
ical plant-warehouse underslab plumbing. This will reduce the float in
path 26–27–28–29 from 30 to 6 days.
To sequence the conduit work, do not add arrow 22–27, because this
will make an illogical sequence. It will make the concrete work in the
two loading docks precedent to the office underslab conduit, 27–28. Add
a logic spreader to separate event 22 into two events. For instance,
22–116, where 116 is the completion of the loading docks and precedes
the slab pour and 22–27 will then provide a proper crew sequence.
A useful technique in sequencing crews is to bar graph the CPM
output by trade or specialty, such as plastering, painting, or concrete.
This could be quite a task if you were to attempt to bar graph the entire
CPM output, but it is not unreasonable to do it by hand selecting only
key categories. The bar graphs can also be generated by computer. Using
the bar graph “family” for a category, the best sequencing can be deter-
mined; then the schedule arrows necessary to set the sequence added
to the network.
Schedule arrows can be very effective in changing CPM from a pure
plan into a workable schedule, but a strong note of caution. Schedule
arrows are pseudo rather than true logic and are more likely to go awry

Figure 22.6.2 Work sequence arrows; foundation contract.

 Evolution of the Project Schedule 341

as the project progresses. This can produce some very illogical results,
which the field group is usually the first to note. The bad impact on field
workers is difficult to erase. The use of schedule arrows is recommended,
but with discretion.

22.7. Fast Track

From the standard John Doe networks, the minimum time from start
of design to close-in is:
Design 105 workdays
Bid/award 20
Site and procurement 52
Foundations 54
Close-in 36
267

Using a compressed fast track (Figure 22.7.1), close-in milestone can

be reached in:
Start design 35 workdays
Structural design 20
Bid/award steel 20
Fab/deliver steel 90
Close-in stage 36
201

This is a time savings of 66 days, or 15% of total. If the owner can

reconsider the delay in starting the office building, the time savings with
fast track will be:
Total time 429 workdays
Less fast track (66)
Less parallel office building (78)
Net time 285, or 33 percent reduction

Figure 22.7.1 Fast-track network.

342 The Practice of CPM Planning

22.8. Responsibility
Construction scheduling is usually the responsibility of the contractor and
the concern of the owner. When there is a single construction contract, the
contractor is the key to all scheduling problems and solutions. In certain
cases, the owner is either required to or chooses to undertake contracts with
several prime contractors. In this case, the owner becomes the coordinat-
ing contractor.
Although it would appear obvious that the owner must take positive
management control steps, that usually just does not happen. In most sit-
uations, the owner hopes for the best and, except in the very worst cases,
the individual contractors usually accept the poor level of coordination even
though they might have legal grounds for action because of delays caused
by other prime contractors. Owners who recognize their responsibilities
often retain either project managers and/or construction managers to
carry those responsibilities out.
Usually, the contractor does not preplan or schedule a project when bid-
ding on it. The reason is economy, since contractors can expect to win only
between 10 and 20 percent of the jobs they bid on and the money spent on
planning jobs not acquired is wasted. This reality points to a very signifi-
cant advantage in using the construction manager (CM) approach. The CM
can apply preplanning using a preconstruction working plan to identify
problems and set an environment of thoughtfulness in regard to the con-
struction schedule. The owner, or CM, can also use the same preconstruc-
tion study to establish a reasonable schedule or to develop special
construction phasing or work-arounds. Each costs more, and the owner
should expect to pay more for the service. Of course, the final working plan
and schedule are developed after the successful contractor has been deter-
mined.
Often, completion of the contract provisions is assumed to be com-
pletion of the project, but the turnover of the facility from the contrac-
tor to the owner often includes a punch list of items remaining to be
finished. The items may be trivial or they may involve substantial
additional labor. The relationship between the owner and the contrac-
tor at the conclusion of a project can be directly influenced by the
number of and difficulty of completing the items on the punch list.

22.9. Schedule versus Calendar

After a suitable end date has been realized by adjusting the network, look
at the practicality of the dates computed. For instance, in the John Doe
networks, all foundation work is to be finished in late fall. This is rea-
sonable, and it allows a little room for the unexpected. In an actual high-
rise project, however, the plumbing tests were predicted for November.
They were delayed for a few weeks, and the hard-freeze period set in.
What should have been a one-week test took six weeks to accomplish.
 Evolution of the Project Schedule 343

So look at the dates computed and compare the activity with the weather
you might expect at that time of year. This is an area in which you are
much better equipped than the computer.
What if you find that the concrete, earthwork, and so on are going to occur
at an unfavorable time of year? First, face up to the fact that a winter job
will cost more or try to delay the project until spring. If your end date is
acceptable, set up your schedule on this basis. If you cannot afford the
delay, consider applying overtime, extra crews, and so on at the start of the
project to complete as much as possible before the onset of bad weather.
What do you do about work that must necessarily be done during a
period of bad weather? The question was perhaps best answered by a
Pennsylvania Dutch concrete superintendent when asked what he would
do if it rained during a big slab pour: “I’d just let it rain.” If you must work
through seasonal bad weather, add project time to account for lower work-
ing efficiency. The factor will vary from Canada southward, of course.
You do not have to alter each time estimate to account for the weather
factor. This would obscure the facts. A practical method is to use weather
arrows. Assume, for instance, that the portion of the John Doe project net-
work between events 29 and 37 is to be accomplished in January and
February. The total durations for the sequence of work is 36 days. In the
middle Atlantic states, we could assume an efficiency of 60 percent; 3 days
work accomplished in each 5 project days. To introduce this factor into the
network, add an activity “weather factor” (29–37) with a duration of 60
days. In Montana, the efficiency factor might drop to 40 percent; in Alaska,
it might be even lower; in Texas, the schedule could be almost normal.
The Texas Department of Highways and Pennsylvania Department
of Transportation (PennDOT) have published their own schedules of
productive days anticipated per month for highway construction subject
to weather influences. The PennDOT schedule is shown in Table 22.9.1.

TABLE 22.9.1 PennDOT Anticipated Workdays Per Month

Conversion Factor,
Cumulative Workdays to Calendar Cumulative
Month Workdays Workdays Days Calendar Days

Jan. 2 2 15.50 31
Feb. 2 4 14.00 59
Mar. 7 11 4.429 90
Apr. 12 23 2.500 120
May 18 41 1.722 151
June 18 59 1.667 181
July 18 77 1.722 212
Aug. 18 95 1.722 243
Sept. 18 113 1.667 273
Oct. 15 128 2.067 304
Nov. 5 133 6.00 334
Dec. 2 135 15.50 365
344 The Practice of CPM Planning

22.10. Contingency
Achievement of the end date desired is not necessarily an acceptable sched-
ule. CPM is not a crystal ball. Even though the activity and time estimates
used in the network are based on experience, a project rarely finishes
ahead of its computed end date. Poor weather, difficult site conditions, labor
disputes, change orders, and so on, are unavoidable and unpredictable.
There is a definite tendency for the actual completion date to exceed the
first CPM end date. It is, then, reasonable to allow for some contingency
between the CPM end and the actual desired completion dates.
There is no definite answer on how much contingency to allow for,
because it will vary with the specific circumstances of the project.
However, if you need a 12-month period for completion of the project,
set your CPM goal at about 11 months, and so forth. Some people have
been reluctant to set a flat contingency at the end of the schedule.
Contingency can be buried in the activity estimates, but if it is, you will
not be able to separate true estimates from contingency.
Another approach is setting contingency based on anticipated site con-
ditions or any predictable problems that can be projected with some rea-
sonableness. Then, in a fashion similar to the weather arrow, a specific
contingency can be identified and assigned only to that area it impacts.
For instance, the availability of space to shake out structural steel will
impact the time frame in which the structural steel is erected; difficult
site access is solved after construction roads are in place; and storage of
equipment and materials becomes less of a problem when foundations
are ready and the equipment and materials can be set in place.
A more mathematical approach to this issue is to look at the conse-
quences of merge bias upon the amount of contingency required. Assume
that each “Most Likely” duration given is an estimate somewhere between
a “best case” or Optimistic duration and “worst case” or Pessimistic dura-
tion. We would expect that the “worst case” estimates are further from the
“Most Likely” than the “best case” estimates. Let us assume only a slight
skewing of this nature, such as the “best case” is 15 percent better than
the “Most Likely,” but the “worst case” duration is 20 percent greater than
the “Most Likely.”
Now consider the John Doe project. If a subset of activities are extracted
such that there is only one linear chain of activities forming the network,
and the durations are randomized to between 15 percent less than esti-
mated to 20 percent more than estimated, it would be expected that this
skewing to the high side would tend to make the project take longer than
the simple CPM calculation (Figure 22.10.1).
And in fact, due simply to this skewing between “best case” and
“worst case,” there is only a 14 percent chance of completion by the
21DEC04 date calculated by the CPM algorithm. Similarly, there is only
 Evolution of the Project Schedule 345

Figure 22.10.1 Critical Path for a one path linear project (based on John Doe project).

a 50 percent chance of completing by 28DEC04 (1 week late), a 95 per-

cent chance of completing by 07JAN05, and a 99 percent chance of
completing by 12JAN05 (3 weeks late). Thus, even for the simplest
logic network of a single linear path, there is a 5 percent likelihood that
this 10-month project will run over by more than 12 work or 17 calen-
dar days (Figure 22.10.2).

100
9

90
8

80
7

70
6

60

% 5

50

40

3
30

2
20

1
10

0 0
6 13 20 27 3 10 17
DEC DEC DEC DEC JAN JAN JAN
04 04 04 04 05 05 05

Latest date 10 Jan 5 Plot Date 27Oct04 6001MPRJ.MC Sheet 1 o 1

Probability Analysis Date Revision Checked Approved
Expected date 27 Dec 4 (c) Primavera systems, Inc. John Doe Critical Path Activities
Earliest date 0 Dec 4 Standard −15%/+20% Distribution
Target date 21 Dec 04 Finish date of Project

Figure 22.10.2 Probability for a one path linear project.

346 The Practice of CPM Planning

Now consider only a slightly more complex logic network. When there
are multiple paths to a network, there is the statistical chance that the
critical path durations may all be lower than anticipated, but that the dura-
tions of a near-critical path may be higher than anticipated. In such a case,
what was the near-critical path may become critical and extend the total
project duration, notwithstanding the reductions of the original critical
path. The project just cannot win—if the durations of the critical path are
higher than anticipated, the project takes longer—and if they are shorter
than anticipated, another path may yet make the project take longer. This
bias toward the project as always taking longer whenever two or more logic
paths merge is known as the merge bias. The greater the number of near-
critical paths that merge, the greater the bias.
For the John Doe Project, there is a 66 percent chance that the criti-
cal path will be 0–100–1–2–3–8–13–14 . . . , a 33 percent chance that
the critical path will be 0–100–1–2–3–401–6–7–8–13–14 . . . , and a 1
percent chance that the critical path will be 0–100–1–2–3–402–
9–11–12–13–14 . . . . (Figure 22.10.3). In this instance, the chances of
completion on time due to the merge bias are reduced from 14 percent
to 11 percent, the 50 percent chance of completion is extended to
23DEC04 and the 95 percent chance of completion is extended to
08JAN05. The 99 percent chance of completion remains at 12JAN05,
although not all logic networks are as forgiving (Figure 22.10.4). In gen-
eral, the more complex the logic network and the greater the number

Figure 22.10.3 Alternate Critical Paths for a project without long duration procurement
activities.
 Evolution of the Project Schedule 347

100 10

90 9

80 8

70 7

60 6

50 5

40 4

30 3

20 2

10 1

0 0
6 13 20 27 3 10 17
DEC DEC DEC DEC JAN JAN JAN
04 04 04 04 05 05 05

Latest date 17 Jan 5 Plot Date 27Oct04 6002MPRJ.MC Sheet 1 o 1

Probability Analysis Date Revision Checked Approved
Expected date 27 Dec 4 (c) Primavera systems, Inc. John Doe without Procurement
Earliest date 7 Dec 4 Standard −15%/+20% Distribution
Target date 21 Dec 04 Finish date of Project

Figure 22.10.4 Probability for a project without long duration procurement activities.

of near-critical paths, the greater the chance that a merge bias will
cause an overrun.
If long lead procurement items are also included in the base network,
the chance of late completion becomes even greater. The long lead item
“fabricate and deliver packaged air conditioners,” of 90 days’ duration,
should be a source of real concern in a real world situation. Here too,
when the default bracket of durations of −15 percent to +20 percent is
applied, the calculated total float of only 8 days is quickly consumed, cre-
ating a potentially overriding new critical path. The likelihood of this
occurring in our model (17 percent) matches that of the real world,
reducing the chance of timely completion to only 2 percent, with only a
50 percent chance of completion by 07JAN05 and only a 95 percent
chance of completion by 25JAN05, more than a month after the calcu-
lated completion date for this nominally 10-month project (Figure 22.10.5
through Figure 22.10.7).

Figure 22.10.5 Alternate Critical Path for a project with long duration procurement
activities.
348 The Practice of CPM Planning

100
7

90

6
80

70 5

60

% 4

50

3
40

30
2

20

1
10

0 0
13 20 27 3 10 17 24 31 7
DEC 04 DEC 04 DEC 04 JAN 05 JAN 05 JAN 05 JAN 05 JAN 05 FEB 05

5OEOMPRJ.MC Sheet 1 o 1 CPM in construction management - 5th editi

Latest date 7 Feb 5 Plot Date 27Oct04 O'brien Kreitzberg & Assoc., INC.
Expected date 8 Jan 4 (c) Primavera systems, Inc. Date Revision Checked Approved
JOHN DOE PROJECT PDM VERSION
Earliest date 14 Dec 4 Standard −15%/+20% Distribution
Target date 22 Dec 04 Finish Date of Project

Figure 22.10.6 Probability for a project with long duration procurement activities.

The general rule here is that activities having longer (and thus less
detailed) durations are more likely to have serious overruns (as 20 percent
of 50 days is larger than 20 percent of 10 days) and are more likely to have
their float consumed and become an alternate critical path and extend the
completion date of the project. If it is important to complete the project by
a specific date, some level of contingency must be allowed at the end. The
amount of contingency required will vary from network to network and can

O'brien Kreitzberg & Assoc., Inc. Monte Carlo Page 4

John Doe project PDM version CPM in construction management - 5th edition Report date 27Oct 4
Criticality Path Report : Second critical path Run No. 5

Activity status: * Completed + Underway

Relative
PCT Predecessor REL PCT Free
Activity Description CRIT Activites Description Type CRIT Float

0 Start milestone 100 None

135 Submit packaged A/C 17 0 Start milestone FS 0.0 100 0.0

235 Approve packaged A/C 17 135 Submit packaged A/C FS 0.0 17 0.0

236 Fab/Del Packaged A/C 17 235 Approve packaged A/C FS 0.0 17 0.0

60 Exterior masonry office 17 59 Erect Precast Roof Office FS 0.0 0 30.3

236 Fab/Del packaged A/C FS 0.0 17 0.0

418 Install piping Office 9 60 Exterior masonry office FS 0.0 17 0.0

64 Test piping office 9 418 Install piping office FS 0.0 9 0.0

67 Metal studs office 17 64 Test piping office FS 0.0 9 0.6

65 Install conduit office FS 0.0 9 0.6

68 Drywall 17 61 Exterior doors office FS 0.0 0 14.8

67 Metal studs office FS 0.0 17 0.0
420 Glaze office FS 0.0 0 14.8

69 Drywall 17 68 Drywall FS 0.0 17 0.0

70 Wood trim office 17 69 Drywall FS 0.0 17 0.0

71 Paint interior office 17 70 Wood trim office FS 0.0 17 0.0

72 Floor tile office 10 71 Paint interior office FS 0.0 17 0.0

80 End of network milestone 100 37 Perimeter fence FS 0.0 0 84.8

58 Erect flagpole FS 0.0 0 6.0
63 Paint exterior office FS 0.0 0 68.8
72 floor title office FS 0.0 10 13.5
73 Toilet fixures office FS 0.0 0 18.6
78 Acoustic tiles office FS 0.0 8 13.5
79 Ring out FS 0.0 0 34.5

Figure 22.10.7 Probability of this Alternate Critical Path.

 Evolution of the Project Schedule 349

be calculated by SPERT style software. However, for construction projects

a general rule of thumb still calls for 1 month for a 12-month project, “and
so forth.”

22.11. Schedule Manipulation

Contractors can use the conversion of the basic plan into a schedule as
an opportunity to manipulate the schedule in their favor. In one major
hospital project, the contractor submitted a network plan for a 4-year
project that showed a very easy and extended schedule for foundations
and structure spanning more than 50 percent of the 4-year time frame
available. All the mechanical, electrical, and finish work was crowded
into less than half the time allotted for the project. To the scheduling
reviewer, it was clear that the contractor intended to set up a schedule
that would be easy to meet during the front end, thus keeping the proj-
ect management team off his back, while claiming that the final portions
of the very complex project could be achieved in record time. The con-
struction manager had imposed an extensive CPM specification, which
unfortunately failed to establish interim milestones. The lack of mile-
stones allowed this contractor’s hybrid approach (i.e., slow start, fast
finish) to meet the letter, even when clearly not meeting the spirit of the
specification.
Because it is important to have a baseline as-planned schedule, the
CPM consultant recommended accepting the schedule while pointing out
its obvious weaknesses to the contractor. Further, it was clearly noted
that if the contractor did not perform more quickly than his schedule
called for during the first 2 years, there was no practical probability that
he would complete the project on time.
Manipulating a schedule can be a two-edged sword, as it was in this
case. The contractor did have delays during the early portion of the
project, but his schedule did not support any delay due to changes caused
by unforeseen conditions. Accordingly, he was properly denied time
extensions that he might otherwise have been allowed.
In the same project, the contractor also attempted to have the network
defined as 100 percent critical. Given enough time and effort, he doubt-
less could have succeeded, but what he ended up with was a network
showing approximately 80 percent of the activities as critical. This is
clearly incorrect from a logical viewpoint. In accepting the network as an
as-planned baseline, the construction manager pointed out that it
appeared to be resource-balanced and that it violated the industry defi-
nition of “critical.” Therefore, the network would not be an appropriate
basis for determining activities in which delays could readily be overcome
by doubling relatively small work crews in special craft areas.
In effect, the contractor had submitted a plan in which he hoped would
identify every activity as critical, and so managed to avoid defining the
350 The Practice of CPM Planning

truly critical activities of the project. Thus, he lost a valuable tool for eval-
uating delays and assigning responsibility.
Another type of manipulation that is becoming more frequent is the
short schedule, where the contractor submits a project plan that involves
substantially less time than the scheduled time required by the owner.
The shortfall is usually substantial, often as much as a year in a three-
to four-year project. The contractor asserts that the bid for the job was
on the basis of the short schedule and that any failure on the part of the
owner to completely support the short schedule will itself be a proper
basis for delay claims.
In reviewing a short schedule, the owner should be certain that suffi-
cient time is allowed for shop drawing approval and other managerial
reviews required by the specifications. Also, it is appropriate to question
the considerations for weather and any unusual conditions included in the
plan. In a multiple prime project, the schedule should be reviewed to be
certain that primes other than the general contractor have sufficient
working time and that, in a general construction contract, that the major
subcontractors have sufficient time to complete their work. The contrac-
tor submitting the short schedule should be required to certify that the
other major primes or major subcontractors have reviewed and agreed to
the plan.
If the contractor submitting the short schedule persists in claiming
it is long enough, one suggested approach has been to issue a change
order at no cost and, thereby, change the end date for completion to
agree with the short schedule. However, if the owner believes that is an
unrealistically short date, calling the contractor’s bluff may have built-
in legal problems.
Short schedules can also be addressed directly in the scheduling spec-
ification. The specification can state that any schedule that is substan-
tially shorter than that required (i.e., 10% or more) will be considered
unrealistic. It could also state that a foreshortened period will be con-
sidered a scheduling contingency and that the owner will make his best
effort to support the short schedule without foregoing any prerogatives,
such as the mandated time for review of shop drawings or the right to
review only priority shop drawings.

22.12. Working Schedule

After the adjustments discussed previously, the CPM schedule is estab-
lished. But is it really a schedule? The critical activities have definite
start and completion dates, but what about the activities with float? For
activities having a float of 10 days or less, the CPM dates are fairly def-
inite, but it is not reasonable to consider an activity with 100 days float
as scheduled.
 Evolution of the Project Schedule 351

A number of attempts to make the CPM schedule more definitive

have been instituted. One method is a computer routine that allocates
the total float to each activity, which can be done by either a flat allo-
cation per activity or an allocation proportioned to the activities’ dura-
tions. Although there is nothing particularly wrong with this routine,
there is also nothing particularly useful either. Because the float allo-
cation is arbitrary, it only clouds the network information. Also, there
is no judgment factor in machine-handled allocation of float. Obviously,
some activities should receive a larger proportion of float than others
(if the method of float allocation is to be used). Contractors usually
prefer to retain all the actual float unallocated and try to work as close
to early start dates as practical.
The character of the network affects its tightness. The John Doe net-
work would be described as very tight. Tightness, or lack of large float
values, is the result of the network tying back into strong events or
nodes, which results in a definitive schedule and is desirable. But do not
force it. Don’t introduce fake logic to achieve a definitive schedule,
because the results will be a fake schedule.
It is entirely reasonable to introduce lead or lag arrows into float
paths to schedule certain items. For instance, three activities in the
John Doe site work that could start in early January are 37–91, grade
and ballast railroad siding; 37–90, pave parking area; and 37–92, access
road. Although those three items could logically commence at event 37,
none should be scheduled in January in the northern part of the coun-
try. Because each activity has more than 12 weeks float (37–90 has 73
days; 37–91 has 63 days; and 37–92 has 68 days), the 3 could obviously
be scheduled for a more suitable time of year by introducing a lead
arrow in front of each activity, as shown in Figure 22.6.1.
The lead arrow is essentially a restraint with time added. If 12 weeks
duration is assigned to each lead arrow, the 3 activities will have early
start times in late March. (In many programs, an activity can be
assigned a “not earlier than” date to restrain the start. Similarly, activ-
ities can be assigned “not later than” dates or an activity start can be
locked in place.)
Fine grading and seeding could start as early as mid-April or as late
as mid-August. A choice should be made whether to use a spring or fall
seeding. Because a fall planting would follow the completion date,
assume a spring planting. To do that, use a lag arrow after the seeding.
This does not affect the early start date for seeding, but it does bring
the late start date to an earlier time.
If all of the available float time is assigned to either a lead or a lag
arrow, the activities in that chain become critical. On some occasions,
it is useful to force certain paths to become critical. In a high-rise build-
ing project, for instance, concrete work was no longer critical because it
352 The Practice of CPM Planning

Figure 22.12.1 Schedule slippage over 11 updates. Latest date to stay on schedule:
project date.

had been expedited and, thus, another path had become critical.
Although concrete work was no longer critical for the overall project com-
pletion, it was still critical in regard to the schedule. The roof pour was
scheduled for late November, and the completion was a race against tem-
perature. The concrete crew won. By completing the last slab before
winter protection was needed, many thousands of dollars were saved.
This was more than enough to pay for the overtime required to main-
tain a fast schedule.
Under pressure by the owner, a scheduling consultant manipu-
lated the schedule of the contractor working on a large water pollu-
tion control plant by applying the various methods necessary to
shorten the schedule projections so that the end date projections would
meet (or appear to meet) the needs of the project. Figure 22.12.1 shows
a series of 11 updates for the project, over which time a key milestone
date actually slipped 6 months but appeared to slip only 2.5 months
because of schedule manipulation. At the time, the “schedule embez-
zlement” was performed with the best of intentions. It was also recorded
in the narrative reports provided with the monthly schedule updates.
However, the shortening and paralleling of activities were accomplished
independently by the scheduling consultant and, therefore, were not
truly part of the contractor’s plan.
Experience has taught a clear lesson that the approved schedule plan
should be changed only with the permission of the project manager and
should be documented in the narrative update at the time the change
 Evolution of the Project Schedule 353

is made. Further, the basis for the change or changes should be clearly
and rationally explained (e.g., additional equipment brought in, addi-
tional crews added, or other logical reasons).

22.13. Summary
The first CPM computation is a plan, not a schedule. After adjusting the
project completion date by changing sequences and time estimates, an
end date is determined. The date should precede the desired completion
with a suitable contingency. The intermediate dates should be reviewed
with the realities of seasonal weather. Seasonal factors can be accounted
for if necessary. The schedule at this point can still be rather loose.
Lead and lag arrows can adjust float to position activities within the
range of their CPM dates. Schedule sequence arrows can be used to
provide a schedule using fewer crews. Scheduling arrows can add to the
effectiveness of the CPM results, but one error in their use can far out-
weigh their benefits.
The owner sets the overall schedule dates, often on the basis of unin-
formed intuition. Preconstruction schedule analysis by a construction
manager can bear important results. Milestones help in controlling the
schedule. The basic schedule can be expedited, but a major time reduc-
tion requires either changes in basic policy or major efforts, such as
fast tracking.
This page intentionally left blank
 Chapter

23
CPM and Cost Control

So far, project planning has been discussed in terms of the time dimension
only. Although the original Remington Rand-DuPont team tied money to
the network in a very sophisticated fashion, the construction industry
was not ready to assimilate two new concepts at the same time. Just as
you cannot learn to run until you have learned to walk, a cost system
based on CPM could not be useful until CPM was accepted.

23.1. CPM Cost Estimate

The first, and perhaps most difficult, step in using CPM for cost control is
estimating costs by activity. The traditional method of estimating begins,
of course, with the takeoff of material quantities from the drawings and
specifications. Then, based on cost records, material unit costs are assigned.
Finally, there are overhead costs, including estimates of anticipated super-
vision and equipment costs as well as a factored portion of the home office
costs. Adding all of these costs together results in an accurate bid price.
CPM does not offer a replacement for this type of cost estimate.
To use CPM in project cost control, a cost must be assigned to each
CPM activity. Contractors can expect to be the low bidder on an aver-
age of perhaps 10 percent of the work they bid on, and since CPM should
only be done on low bids, they can expect to have to do a CPM cost
breakdown on only 10 percent of their project estimates.
One way to prepare the CPM cost estimates would be to undertake
a second quantity takeoff by activity. This way, the same unit costs
and overhead factors would be assigned to those quantities as before.
The total resulting activity costs should equal the contract price. The
re-estimate would cost about 50 percent more than the original esti-
mate, though, and the adjustments required to equate it to the contract
price would be an accountant’s nightmare. Moreover, this method is a

355

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
356 The Practice of CPM Planning

compromise between traditional estimating procedures and breaking

down projects into CPM activities.
Preparing a cost estimate by CPM activity can be inexpensive, fast, and
sufficiently accurate if done properly. Keep in mind that the contractor
knows the answer that his CPM estimate must achieve. Why not start
with that answer, the contract price, and work backward? Actually that
is even easier than estimating by quantity takeoff. Almost every contract
requires that the bid include a cost breakdown which, when approved by
the owner, becomes the basis for progress payments. The specified cost
breakdown should be in categories compatible with CPM analysis:
Clearing Roofing
Rough grading Hung ceilings
Excavation—general Structural steel
Excavation—utilities Bar joists
Footings Insulated metal panels
Foundation work Masonry—exterior
Grade beams Masonry—interior
Floor slabs Windows
Underfloor plumbing Glazing
Underfloor conduit Doors—exterior
Major equipment (by item) Doors—interior
Ductwork Heating plant
Power conduit Water piping
Branch conduit Insulation
Switchgear Air conditioning
Wiring Plumbing fixtures
Drywall Hardware
Paint

The Construction Specifications Institute’s (CSI) 16 specification stan-

dard divisions, which have become construction industry standards, a def-
inite, industrywide shift to a common terminology has been established.
The sub-breakdown of the 16 CSI divisions MASTERFORMAT into
about 250 BROADSCOPE categories and an unlimited number of
NARROWSCOPE categories provides the means to identify all estimat-
ing factors in common terms. With the increasing use of computers to
write specifications, the use of MASTERFORMAT is increasing. Further,
that increased use of MASTERFORMAT is resulting in an increasing
number of estimates structured on the same numbering system. As of
2004, CSI is promulgating a new breakdown of 50 divisions to be refer-
enced by a standard six digit code that replaces the current five digit code.
 CPM and Cost Control 357

Especially with computerized estimates, categories can be summa-

rized from the standard estimating sheets without re-estimating quan-
tities or recalculating costs. Also, the architect or construction manager
can review the cost breakdown by using quantities from the control
estimate. Thus, the CPM cost estimate has stayed within the boundaries
of usual estimating practices.
Once the architect, construction manager, and/or owner have approved
the broad category cost breakdown, the next phase is to further break
down costs into activity costs. This can be done informally, for example,
by assigning project time to the activities. The cost assignment will be
realistic and accurate because of the detailed breakdown of the dia-
gram. For instance, in the John Doe project, if the cost for the category
“foundation concrete” is $144,300, we can list all the activities involv-
ing foundation concrete by just sorting and listing under that code, as
shown in Table 23.1.1.
The yardage breakdown by activity can be approximate, but the total
should equal the exact figure taken from the original detailed estimate.
If the actual yardage for the office grade beams (24–25) was 57 yd3 and
the plant slab (22–29) was 1490 yd3, the effects of such differences on
the total cost would be insignificant.
The breakdown of costs by activity will take additional time; and
since time is money, the contractor will incur a cost. With practice, how-
ever, this cost should become nominal.
Another approach to keeping the effort and cost of assigning costs low is
to utilize the cost code feature of many popular software products. As the
logic network is initially being developed, each new activity added should
be assigned a cost code derived either from the categories listed previously,
or from the contractor’s own estimating system. Many of the popular soft-
ware products support multiple cost codes to be assigned to one activity,
but keep in mind that the cost coding is for general categorization only,

TABLE 23.1.1 John Doe Project Concrete Costs

Approximate
i–j Description Cubic Yards Cost ($)

3–6 Water tank foundation 20 5,000

16–17 Pour pile caps 200 50,000
17–18 Grade beams, plant-warehouse 200 70,000
18–21 RR loading dock 50 15,000
18–22 Truck loading dock 50 15,000
23–24 Spread footings, office 100 17,500
24–25 Grade beams, office 60 21,000
22–29 Slab, plant-warehouse (12-in reinforced) 1500 225,000
28–29 Slab, office (6-in mesh) 120 18,000
Total 2300 436,000
358 The Practice of CPM Planning

rather than for accounting purposes, and should be kept as high level and
simple as possible. The goal should be to have fewer than 100 distinct cost
codes for this exercise.
At this point, and again as each activity is being added to the network,
the project manager should be encouraged to throw a “wild guess” for cost
at each activity. If multiple cost codes have been assigned, then a “wild
guess” cost estimate must be entered for each cost code. The total of all of
the “wild guess” costs for each cost code may be higher or lower than that
for each category of the cost breakdown or bid estimate, but it can be
assumed that the error for each activity should be consistent.
It is then a simple matter for the Scheduler to factor costs by cost code, up
or down, to match the desired total. This can be done by the use of global
changes to the costs for each cost code, or by first exporting the costs and cost
codes to an Excel or other spreadsheet file, factoring, and then importing the
adjusted costs. The Scheduler can also take this process one step further by
adjusting the cost of the first of a string of similar activities up (for example
by an industry standard of 7 percent) to account for the learning curve.
Obviously, the Engineer can challenge any gross adjustments that cannot be
supported as attempts at front loading. This process can be repeated to
adjust all costs to include a proportionate share of overheads and profit.
The suggestion is often made to defer cost loading until the activities and
logic of the network have been approved since the addition or deletion of
activities would require repeating this balancing process. Either way, the
goal is to reduce the effort required by the project manager by shifting the
burden that the “the numbers must add up correctly” to the Scheduler.

23.2. Progress Payments

Figure 23.2.1 shows the first sheet of the John Doe project printout
with costs added for all activities. (Even when costs are in the master
file, they can be excluded from the printout.) A primary use for CPM cost
data is as the basis for progress payments. Figure 23.2.2 shows the cost
summary for one trade (electrical). CPM places progress evaluation on
a well-defined basis-activity completion rather than the traditional per-
centage estimates. Because agreement on project status can be imme-
diate with CPM, progress payment invoices based on that status should
be approved for payment with no delay. Figure 23.2.3 shows a sample
CPM-based invoice for update 2.
To the contractor, faster payment of invoices represents definite cash
savings. If the approval time for invoices is shortened by 2 weeks, the
savings in interest on a $300,000 invoice would be approximately 2/52
× 10% × $300,000, or $1154. On a $9 million project, the savings would
be $34,620. It would be reasonable to expect these savings to average
0.5% of the project cost. Additional, intangible, savings would be the
lower cost of preparing and justifying invoices.
 CPM and Cost Control 359

Figure 23.2.1 John Doe project printout, Lotus format: first sheet of i–j.
360

Figure 23.2.2 Cost summary for electrical trade.

 CPM and Cost Control 361

Figure 23.2.3 CPM-based invoice for update 2 (electrical work).

362 The Practice of CPM Planning

The owner’s (as well as the construction manager’s and architect’s)

tangible savings stem from the shorter time required to approve invoices.
This frees staff for other work. More important to the owner, however,
is the assurance that the invoices paid represent a correct and equitable
portion of the contract.

23.3. Cost Forecasting

The costs of the activities for the first portion of the John Doe network
are shown in Table 23.3.1. Time and cost dimensions can be combined
to forecast the rate of spending on a project. If the project is on sched-
ule, the contractor will earn the cost of an activity somewhere between
the early finish and the late finish dates. Plot the cumulative cost of
activities completed against project time, and cost against early com-
pletions will give the maximum amount of money required on any proj-
ect day. On any project day x, the plot determines a maximum-minimum
range of funds required. The actual amount will be somewhere between
the two. For the contractor, this is the forecast of his earning rate on the
project. Working backward, the contractor can borrow just the amount
of money needed to finance the project until sufficient cash is derived
from invoices to make the project financially independent.
The contractor’s savings will depend on his mode of financing. If the
sum is being borrowed outright, a specific savings in interest is achieved
by borrowing less. If the contractor is working against a credit com-
mitment, this approach will define the number of projects that can be
handled within that amount.

TABLE 23.3.1 John Doe Project Activity Costs

Activity Cost, $

0–1 Clear 28,000

1–2 Survey 2,500
2–3 Rough grade 10,000
3–4 Drill well 10,000
3–6 Water tank foundations 5,000
3–9 Sewer excavation 21,500
3–10 Excavate for manhole 1,000
3–12 Overhead pole line 30,000
4–5 Well pump 11,250
5–8 Underground pipe 18,000
6–7 Water tank 225,000
7–8 Tank piping 40,000
8–13 Connect piping 2,000
9–11 Sewer 65,000
10–11 Electrical manholes 6,500
11–12 Electrical duct bank 7,500
12–13 Power feeder 6,000
Total 489,250
 CPM and Cost Control 363

The owner’s savings from the cash forecast are even more definite. If
financing the project from securities, the owner can liquidate at the latest
time practical and thus earn interest for the maximum length of time and
maintain the principal at its largest practical value. If the owner receives
the total construction fund in one lump sum, as in a bond issue, the
greater portion can be scheduled for higher-interest, long-term invest-
ments and only the part that must be held for near-term use has to be
placed in lower-interest, short-term investments.
Figure 23.3.1 shows a plot of the John Doe project site preparation
costs based on early finish times. Figure 23.3.2 gives a similar plot of

Figure 23.3.1 Project days. Cost vs. time: early-finish basis.

364 The Practice of CPM Planning

Figure 23.3.2 Project days. Cost vs. time: late-finish basis.

money vs. time but is based on late finish times. Figure 23.3.3 shows both
curves on the same plot. In larger network samples, the early and late
finish cost curves tend to parallel to each other. Also, the curves are usu-
ally smooth and have very few inflection points. The time scale is usually
in weeks or months, which is of more concern for the broad financial con-
trol of a project.
The cost forecast is meaningful because it is plotted to a true time
scale, whereas the example plot was done manually, the computer can
 CPM and Cost Control 365

Figure 23.3.3 Project days. Comparison of early- and late-finish costs vs. time curves.

provide computer-generated curves when the network is cost-loaded.

Figures 23.3.4 and 23.3.5 are the computer-generated cost forecasts,
and Figure 23.3.6 is the cost basis for the John Doe i-j sort. As the proj-
ect moves slightly ahead of or behind schedule, curves to reflect those
conditions can easily be generated. The recommended updating fre-
quency for the cost curves is quarterly.
366 The Practice of CPM Planning

Figure 23.3.4 Computer generated cost forecast for the John Doe project—Early & Late—
Cumulative.

Figure 23.3.5 Computer generated cost forecast for the John Doe project—Early dates—
Cumulative and Histogram.
 CPM and Cost Control 367

Figure 23.3.6 Cost basis for the John Doe project (i–j).

When an owner is forecasting the finances of a fixed-price contract,

unless a change order is added to the contract, the time may change but
the costs will not. What if a major change is made or the contract is cost
plus? In this case, the cost changes can be introduced by changing the
cost values for the affected activities.
368 The Practice of CPM Planning

The cost savings possible using CPM cost forecasting are difficult to
assess. However, the uncommitted construction funds, which are 100%
of the project cost at project day 0 and 0% at the end of construction,
roughly average 50% over the life of the project. Figure 23.3.7 represents
a project cost vs. time curve.
The uncommitted area (cost × time) approximately equals the
money payable to the contractor. If an owner has the total construc-
tion fund at the start of the project, part of it can be invested in long-
term bonds yielding about 6% interest. Another portion can be held
in short-term notes yielding about 4%. Over the life of the project, an
average interest of about 5% can be realized on the uncommitted
funds (or about 50% of the project cost). For a 1-year project, this
could amount to 2% of the cost. A 2-year project would be more nearly
average, and the total earnings on the uncommitted funds for that
period would be about 5%.
The use of accurate CPM cost forecast curves to predict how much
money will be needed each month to pay for a project can guide the
owner in investing any as yet uncommitted funds. However, the con-
tingency cash required at any given time will be 2 to 4 percent lower
because of accurate CPM cash forecasting. The additional peace of mind
such accuracy will provide for the owner’s investment counselor can be
counted only for its intangible value.

Figure 23.3.7 Project cost vs. time curve.

 CPM and Cost Control 369

23.4. Network Time Expediting

The cost assigned to each activity is the normal cost, that is, the cost of
doing the activity with a normal crew under normal conditions. But
there are cases in which owners want projects completed on an expedited
basis or contractors must expedite their efforts to complete projects on
schedule. The traditional approach is to put the entire project on a crash
overtime basis. This is quite expensive for two reasons. First, it usually
occurs late in the project when the workforce is at a peak. Second, most
of the work activities done on a crash basis are float jobs, the comple-
tion of which does not shorten the project by even a day.
Crash is defined as the shortest time within which an activity can
be accomplished by using a larger crew, overtime, extra shifts, or any
combination of the three. Some activities which it might appear that
expediting cannot affect are crashed by using such special techniques
as high early-strength cement. By definition, normal time must be
longer than crash time (Figure 23.3.7).
To shorten an activity duration from normal to crash, the activity costs
are inevitably increased. The increase results from premium time costs,
inefficiency of larger crews, and increased material costs (such as extra
forms and high early-strength cement). The cost that is associated with
the crash time is known as the crash cost. For the John Doe site prepara-
tion network, the crash times might be like the ones shown in Table 23.4.1.

TABLE 23.4.1 John Doe Project Crash Times

Normal Crash
time, time,
i–j Description days Method of expediting days

0–1 Clear 3 Overtime 2

1–2 Survey 2 Extra crew, overtime 1
2–3 Rough grade 2 Extra crew, overtime 1
3–4 Drill well 15 Double shifts 8
3–6 Water tank foundations 4 Extra crew, overtime 3
3–9 Excavate sewer 10 Extra equipment, overtime 6
3–10 Excavate manhole 1 1
3–12 Pole line 6 Extra equipment, overtime 4
4–5 Well pump 2 Extra crew, overtime 1
5–8 Underground pipe 8 Extra crew, overtime 6
6–7 Erect tank 10 Overtime 8
7–8 Tank piping 10 Extra crew, overtime 8
8–13 Connect piping 2 Overtime 1
9–11 Sewer 5 Extra crew, overtime 3
10–11 Electrical manhole 5 Extra crew, overtime 4
11–12 Duct bank 3 Extra crew, overtime 2
12–13 Power feed 5 Third shift 2
370 The Practice of CPM Planning

Figure 23.4.1 Full-crash plan for the John Doe project.

Figure 23.4.1 shows the calculation of crash event times. Note that it
is just a basic CPM calculation. The normal and crash results are com-
pared in Table 23.4.2.
Note that the total crash duration of the project is 10 days shorter than
its normal duration. Note also that the critical path has shifted.
Estimated crash costs are shown in Table 23.4.3.
The crash cost to pick up 10 days appears to be $160,750, or $16,075
per day. However, what if you do not need to expedite the completion by
the full 10 days? Figure 23.4.2 shows a plot of normal versus. crash
times and costs for the activity drill well. For a cost difference of $7,000,
the operation can be expedited in 7 days, an average extra cost of $1,000
per day. A linear connection between normal and crash points is gener-
ally a reasonable assumption; minor variations tend to cancel out. How
much would drilling the well cost if it were to be done in 11 days? The
answer, from Figure 23.4.1, is $14,000. The cost of expediting a partic-
ular activity is a linear plot of the crash and normal costs, but that
assumption does not apply to the costs of expediting an overall project.
To cut 1 day off the 34-day John Doe site preparation project, cut 1
day off the critical path from any one of the following activities for the
costs listed in Table 23.4.4.
The best choice for 1 day would clearly be activity 3-4, drill well, at
$1,000. At that point, both the well and tank paths are critical.
Expediting along the tank path would cost as shown in Table 23.4.5.
 CPM and Cost Control 371

TABLE 23.4.2 Normal vs. Crash Times

Normal Crash
Normal Normal float, float,
Activity ES Crash ES Crash LF LF days days

0–1 0 0 2 3 0 0
1–2 3 2 3 5 0 0
2–3 5 3 4 7 0 0
3–4 7 4 16 22 0 4
3–6 7 4 7 12 1 0
3–9 7 4 17 21 4 7
3–10 7 4 16 21 13 11
3–12 7 4 22 29 16 14
4–5 22 12 17 24 0 4
5–8 24 13 23 32 0 4
6–7 11 7 15 22 1 0
7–8 21 15 23 32 1 0
8–13 32 23 24 34 0 0
9–11 17 10 20 26 4 7
10–11 8 5 20 26 13 11
11–12 22 13 22 29 4 7
12–13 25 15 24 34 4 7

If the path is to be shortened by another day between events 3 and

8, the well driller continues to be a bargain. Activity 3-6, water tank
foundations, is the next best buy, since activity 6-7, erect tank, and
7-8, tank piping, are very expensive at both normal and crash costs.

TABLE 23.4.3 Normal vs. Crash Costs

Normal Crash
i–j costs, $ Description Source of extra costs costs, $

0–1 28,000 Clear Overtime 34,000

1–2 2,500 Survey Second crew, overtime 5,000
2–3 10,000 Rough grade Second crew, overtime 15,000
3–4 10,000 Drill well Double shifts 27,000
3–6 5,000 Water tank foundations Crew, overtime 8,000
3–9 21,500 Excavate sewer Equipment, overtime 32,000
3–10 1,000 Excavate manhole 1,500
3–12 30,000 Pole line Equipment, overtime 40,000
4–5 11,250 Well pump Crew, overtime 14,000
5–8 18,000 Underground pipe Crew, overtime 27,000
6–7 225,000 Erect tank Overtime 275,000
7–8 40,000 Tank piping Crew, overtime 60,000
8–13 2,000 Connect piping Overtime 4,000
9–11 65,000 Sewer Crew, overtime 90,000
10–11 6,500 Electrical manhole Crew, overtime 7,500
11–12 7,500 Duct bank Overtime 10,000
12–13 6,000 Power feeder Third shift 10,000
Total 489,250 650,000
372 The Practice of CPM Planning

Figure 23.4.2 Cost-time relationship for activity 3–4, drill well.

TABLE 23.4.4 Choice of Activities to Crash by 1 Day

Difference Difference
Critical between crash Normal Crash between crash
path and normal duration, duration, and normal Extra costs
activity costs, $ days days duration, days per day, $

0–1 6,000 3 2 1 6,000

1–2 2,500 2 1 1 2,500
3–4 7,000 15 8 7 1,000
4–5 2,750 2 1 1 2,750
5–8 9,000 8 6 2 4,500
8–13 2,000 2 1 1 2,000

TABLE 23.4.5 Expediting Along the Tank Paths

Difference Difference
between between
crash and crash and
normal Normal Crash normal Extra costs
Activity costs, $ duration, days duration, days duration, days per day, $

3–6 3,000 4 3 1 3,000

6–7 50,000 10 8 2 25,000
7–8 20,000 10 8 2 10,000
 CPM and Cost Control 373

Figure 23.4.3 Cost-time cash planning factors for John Doe project shown as costs per day
per activity.

The minimum cost to cut 2 days off between events 3 and 8 involves
activity 3-6, water tank foundations ($3,000), plus 3-4, drill well
($1,000), for a total of $4,000.
Figure 23.4.3 shows the John Doe site network with the potential accel-
eration per activity and the costs per day to accelerate those activities.
Using that information, the optimum expediting for the initial 3 days
would be as shown in Table 23.4.6.
Thus, 3 days of a possible 10 can be expedited by using the logic and
information at hand. The results are impressive: a 30% gain in time at
an average cost of $1,833 per day versus maximum projected crash costs
of $16,076, for a 9:1 advantage.
Candidates for expediting the 4th through the 7th days would be
those shown in Table 23.4.7. Taking activity 3–4, drill well, and 3–6,
water tank foundations, together, day 4 can be expedited for $4,000.
Days 5 and 6, taking activity 0–1, clear, and 2–3, rough grade, will cost
an average of $5,500 each to expedite, or more than three times the aver-
age cost of expediting the first 3 days.

TABLE 23.4.6 Optimum Activities to Crash by 3 Days

Day 1 Drill well (3–4) $1,000

Day 2 Connect piping (8–13) $2,000
Day 3 Survey (1–2) $2,500
374 The Practice of CPM Planning

TABLE 23.4.7 Candidates for Crashing Day 4 Through 7

Clear (0–1) $6,000

Rough grade (2–3) $5,000
Drill well (3–4) $1,000
$4,000
Water tank foundations (3–6) $3,000
Erect tank (6–7) $25,000

TABLE 23.4.8 Crashing by 7 Days

Normal Crash Costs/

time, time, ∆T, day to
i–j Description days days days ∆$ expedite, $

0–1 Clear 3 2 1 6,000 6,000

1–2 Survey 2 1 1 2,500 2,500
2–3 Rough grade 2 1 1 5,000 5,000
3–4 Drill well 15 8 7 7,000 1,000
3–6 Water tank foundations 4 3 1 3,000 3,000
3–9 Excavate sewer 10 6 4 10,500 2,625
3–10 Excavate manhole 1 1 — — —
3–12 Pole line 6 4 2 10,000 5,000
4–5 Well pump 2 1 1 2,750 2,750
5–8 Underground pipe 8 6 2 9,000 4,500
6–7 Erect tank 10 8 2 50,000 25,000
7–8 Tank piping 10 8 2 20,000 10,000
8–13 Connect piping 2 1 1 2,000 2,000
9–11 Sewer 5 3 2 25,000 12,500
10–11 Electrical manhole 5 4 1 1,000 1,000
11–12 Duct bank 3 2 1 2,500 2,500
12–13 Power feeder 5 2 3 4,000 1,333

Float
Expedited Original used path
day Activities expedited Costs, $ path 3-4-5-8 3-9-11-12

1. Drill well 1,000 1 1

2. Connect piping 1,000 0 1
3. Survey 2,500 0 0
4. Drill well/water tank 4,000 — 1
foundations
5. Rough grade 5,000 0 0
6. Clear 6,000 0 0
7. Tank piping/drill well 11,000 1
8. Same activities as for day 7 13,625
plus 3-6, excavate sewer
9. Erect tank/drill well/excavate 28,625
sewer
10. Same activities as for day 9 28,625
Total 101,375
 CPM and Cost Control 375

And day 7, taking activity 7–8, tank piping, and again 3–4, can be
expedited for $11,000. Expediting beyond this requires consideration of
the paths through the sewer, duct bank, and pole line, because the normal
float of 4 days following event 3 will have been used up. See Table 23.4.8
for a summary.
Note that this selective approach to expediting the project costs
$45,000, or only 54 percent of the costs resulting from the total-crash
approach ($83,350).
To get maximum 10-day acceleration, if planned by CPM, the cost is
63 percent ($101,375) of maximum 100 percent crash ($160,750).
Further, if 80 percent of the maximum acceleration (i.e., 8 days) is
acceptable, the acceleration is only 27 percent of full crash. That is, the
first 8 days of acceleration averages $5,516 versus $10,938 averages for
a 10-day maximum acceleration.

23.5. Minimum Cost Expediting

Why is the owner building? Obviously, to use the facility. In the case of
a hotel, a hospital, a manufacturing facility, or a restaurant, the owner
can realize a definite cash payoff for every day gained in the comple-
tion of the project, or basically a linear payoff. Considerable losses can
result in the same fashion. For instance, a school gains nothing (except
considerable peace of mind) by opening early. However, if the school
opens late, the cash costs can be calculated for extra buses, rented
quarters, etc.
Combining the direct cost curve with a straight-line indirect cost curve
creates a third curve, the total cost curve. This combination is shown in
Figure 23.5.1. Note that, at some time between crash and normal, the
total cost curve dips to a minimum point. In any project, it is worth some
cash outlay ∆$ to expedite the project. That will save time (∆T) and
money (∆$). In the John Doe site network, assigning values to the indi-
rect costs:

Contractor’s supervision: Two persons, total $4,000 per week

Equipment: Shacks, power, telephone, and so on, $2,500 per week
Owner: Project engineer, $2,500 per week
Production advantage: $2,000 per day

The combined indirect costs and savings per day are

$2,500 +$2,500 +$4,000)

+ $2,000 = $3800
5
376 The Practice of CPM Planning

Figure 23.5.1 Combination of direct and indirect cost curves.

In Figure 23.5.2, the direct and indirect costs are shown combined.
Table 23.5.1 is a summary of the costs for expediting, combined with
indirect cost savings. This approach is realistic, but it has not been used
widely for four reasons: First, since it is CPM-based, only a company
already using CPM can consider it. Second, it requires the assignment
of two costs to activities and there is a psychological barrier to the assign-
ment of even one cost. Third, in most construction projects, it is not prac-
tical to put just certain crafts on overtime. If you do, the other trades will
usually make their objections felt in a number of ways. Fourth, only one
computer program has been available for the calculation. That one, by
James E. Kelley, Jr. for the GE 225 computers, is now obsolete.
The first barrier (CPM usage) is rapidly falling away. The second
(cost assignment) will crumble as other CPM cost systems are adopted.
It is easy to assign crash costs and times at the same time as normal
costs and times (adding perhaps 10 to 20 percent to the normal effort
 CPM and Cost Control 377

Figure 23.5.2 Combined direct and indirect cost curves for John Doe project.

TABLE 23.5.1 Summary of Crashing up to 10 Days

Expedited day Expedited activities Construction costs, $

Normal time 489,250

1. Drill well 490,250
2. Connect piping 491,250
3. Survey 493,750
4. Drill well/water tank foundations 497,750
5. Rough grade 502,750
6. Clear 508,750
7. Erect tank/drill well 519,750
8. Same activities as for day 7 plus 533,375
excavate sewer
9. Tank piping/drill well/excavate sewer 562,000
10. Same activities as for day 9 590,625
378 The Practice of CPM Planning

required to make the assignments). The third problem (not being able
to put a project on partial overtime) cannot be completely overcome.
However, when there is a choice, expedite in early activities, such as sur-
veying and clearing, when the number of people involved in the project
is lower. Usually, the lower costs of expediting those areas will direct the
computer solution to the same areas anyway. However, the fourth bar-
rier (the lack of a program for a currently viable computer) is significant,
and it will remain so until solved. The calculation is not suitable for the
manual mode.
Inventory planning, as calculated by industrial engineers, offers an
inspiration in regard to an expedited approach. For a given category of
material, equipment, or spares there is a minimum amount of each
item which must be kept in stock. There is also some larger amount
which can be purchased at a lower price. When the costs for the items
in a category (such as pump impellers) are summarized, a curve simi-
lar to the direct cost curve is achieved, which indicates that it costs
more per unit to purchase fewer units at a time. Now the indirect costs
of handling one item can be added in. It will be a linear relation of cap-
ital costs, including the costs for the warehouse staff, utilities, account-
ing, inventories, and so on. When the indirect costs are combined with
the direct cost curve, a minimum cost point can be estimated for the cat-
egory and eventually, by extension, for the entire inventory. This is
demonstrated in Figure 23.5.3.

Figure 23.5.3 Example of inventory

optimization.
 CPM and Cost Control 379

23.6. Summary
A cost breakdown of the CPM network is best done by activity and best
carried out immediately after the award of the contract. The cost
breakdown should be within the framework of the bid, and it must be
realistic. An important use of the activity cost breakdown is making
progress payments.
Cash requirements of the project can be forecast on a time basis by com-
puter with the use of the CPM cost estimates. The forecasts can guide
owners in investing the construction funds to realize the highest yield and
contractors in determining their financial needs and methods.
The cost of expediting a project can be accurately estimated by using
a CPM-based cost system. There are even cases in which a project can
be completed early at a lower cost through carefully directed expediting.
The promise of cost expediting has not been fully realized, principally
because existing cost collection and accounting systems do not relate
directly to construction activities.
This page intentionally left blank
 Part

The Practice of CPM

5
Scheduling

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
This page intentionally left blank
 Chapter

24
Let’s Look Under the
Hood at the Engine

Primavera Systems former flagship Primavera Project Planner or P3,

until the time it was supplanted by Primavera’s P3e/c product, was retail-
ing at $4000 while both Primavera’s SureTrak and Microsoft’s Project 98
were retailing at $400. Despite the variance in cost, the additional features
of P3 are mostly related to multiproject applications and large company
or remote site interconnectability. The screen configurations, formats, and
even click or keystrokes, are much the same for SureTrak as for the more
expensive P3 software, and most if not all of the screen-based tabular and
graphic reporting features are offered in the lower-priced package.
Therefore, comparisons between Primavera Project Planner P3 and
Microsoft Project are not “apples to oranges” or “high cost to low cost”
alternatives, but could be made equally between the similar cost
SureTrak and Microsoft Project. Similarly, since the primary focus of this
text is on the “engine” that performs calculations with only a secondary
consideration of graphic capabilities and tertiary concern for database
compatibility with “the accounting department,” all of the products
reviewed may be considered on an equal basis. Thus, this “tutorial” in
navigating through screens for the P3 and P3e/c products should be
viewed as the general manner in which such programs are organized,
with the understanding that competing products or even next year’s
release of this software may have small differences.

24.1. Primavera Project Planner P3

Upon initiating Primavera, the user is given the choice to start a New
project or Open an existing project. Following Windows’ standards, users
may also choose from a list of most recently accessed projects. The default

383

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
384 The Practice of CPM Scheduling

number of recently opened projects is four, but this may be changed by

modifying the P3.INI file in the main \WINDOWS subdirectory.
However, this modification process is not automated and erroneous mod-
ification may cause serious problems (Figures 24.1.1 through 24.1.3).
Opening an existing project by the Open command presents a list of
all projects on the default PROJECTS subdirectory for point and click
choosing. Most users do not go further than the default subdirectory. For
those who may wish to store their projects elsewhere, perhaps in a sep-
arate subdirectory for each project (along with other project specific
files), the user may choose to change subdirectories or even drives.
However, summarization across projects is limited to projects stored in
a common subdirectory.
A limitation is that each project (or saved variant or update thereof)
must be represented by a 4-character abbreviation or code. A usage tip
is to use only three of these characters to identify the project, reserving
the last for variants and updates. Thus for the John Doe Project, the
baseline or initial schedule is DOE0, with updates being saved as DOE1,
DOE2, . . . , DOE9, DOEA, DOEB, . . . , DOEZ. If more than 36 updates
are required, it is probable that some of the earlier updates may be
deleted from disk or saved (in compressed format) to another area of the
disk. The “retired” suffix numbers may then be reused.

Figure 24.1.1 Primavera opening screen.

 Let’s Look Under the Hood at the Engine 385

Figure 24.1.2 Primavera add a new project screen.

Figure 24.1.3 Primavera quick open an existing project screen.

386 The Practice of CPM Scheduling

As noted in Chapter 9 Primavera Project Planner software allows

two means to designate a mandated completion date (FNLT) for a proj-
ect. Entering such a deadline in this opening screen, or in the Overview
screen, will impose such date for float calculations and, if the date is
greater than that calculated for the early completion of the project, the
software may calculate a positive total float for activities on the critical
path. If the FNLT field in the OVERVIEW screen is left blank, but the
same FNLT constraint is entered for the ending activity in the network,
the software will calculate a total float of zero for activities on the crit-
ical path (Figures 24.1.4 and 24.1.5).
After highlighting a specific project, the user may choose to view an
Overview of the project, which details more complete data on the proj-
ect name, mandated project start and finish dates, number of activities,
and other project information. This screen is similar to that used to ini-
tiate a new project. The Open dialog box also allows the user to set
restrictions on Access so that other project members may be permitted
to view but are restricted from changing the schedule. Although multi-
ple users may view the network simultaneously and even make changes
to their respective sections simultaneously, the limitations of the data-
base utilized requires that only one user have access if the unique

Figure 24.1.4 Setting a project finish date will cause a project completing earlier to show
a positive, non-zero total float for the critical path.
 Let’s Look Under the Hood at the Engine 387

Figure 24.1.5 Setting a FNLT constraint upon the last activ-

ity will cause a project completing earlier to show a zero total
float critical path.

activity number for a specified activity is to be changed. If the user

desires this capability, he or she must click the Exclusive box in the
dialog box. This Exclusive checkoff box also may be chosen within the
Project Overview screen (Figure 24.1.6).
The default opening screen for a new project is a combination tabu-
lar entry sheet (or spreadsheet) and bar chart template meant to dis-
play activities as entered. This is somewhat unfortunate insofar as this
is not the easiest format for entering new data. However, the user may
then choose the screen format (or View Layout) of his or her choice,
which may be customized to the user’s desires.
When opening an existing project, the opening screen is in the format
of the layout in effect when the project was last accessed. (Primavera
provides the option of not saving the layout when exiting. In this case,
the last layout saved will be the new opening screen.)
The default opening screen is shown in Figure 24.1.7. Following
Microsoft standards, the top line, or TITLE BAR, reflects the software in
use (Primavera Project Planner) and the 4-character project code. The
second line, or MENU BAR, includes the main menu headings, which may
be accessed either by point and click or by <ALT> plus the underlined letter
388 The Practice of CPM Scheduling

Figure 24.1.6 Checking the exclusive box permits changing Activity IDs.

Figure 24.1.7 Opening screen for a new project.

 Let’s Look Under the Hood at the Engine 389

of the menu. In addition, for many of those menu selections, which may
only be located by moving down multiple menus, Primavera provides a
third means of access, a <CONTROL> plus designated hot key. Several
of the hot keys are also assigned to the FUNCTION keys. Unfortunately,
the assignment of hot keys does not appear to be intuitive and may
require frequent use to be remembered (Figure 24.1.8).
The third line, again following Microsoft standards, is the toolbar. This
toolbar may be customized by the user. As with other Microsoft and
Microsoft compliant software, the icons of the tools are often difficult to
recognize, much less understand. However, resting the mouse pointer
over any icon for a moment will generate a small description below that
icon. One failing of Primavera is that the “?” icon, usually reserved for con-
text sensitive help, instead brings up Primavera’s standard help screen.
Below this third line is then presented a choice of the combination tab-
ular and bar chart layout noted above or a pure logic format (improperly
using the acronym PERT) for review of activity data and relationships.
Finally, at the bottom is a status line, split into three segments, which
notes what the computer is doing, the name of the layout being used,
and the name of the filter (or selection criteria) being used.
Primavera programmers spent a great deal of effort adding features
to the bar chart view. Additional columns of data may be added to the

Figure 24.1.8 Point and Click versus <ALT> keys versus <CTRL> hot keys.
390 The Practice of CPM Scheduling

spreadsheet half, adjusting the portion reserved for the bar chart accord-
ingly. Individual columns may be customized for type of data, width of
column and size, and font of the text. The scale of the bar chart may be
modified and two separate periods may be shown at once by splitting
the bar chart portion in two. Similarly, if a large number of activities
are in the project, the layout may be split vertically in two to permit
showing two portions of the list of activities simultaneously. The size and
font of the text may be modified and the color, size, and shape of bars
and end points may be modified by the user (Figures 24.1.9 and 24.1.10).
The user controls these several options by use of the Format menu
selections, Columns, Bars, Summary Bars, Timescale, Sight Lines, Row
Height, and Screen Colors.
The pure logic (or PERT) view also has several customizable options.
These include the size and data to be included in the Activity Boxes, and
color and type of line for showing relationships. A special feature is the
Trace Logic view. This feature, accessible through the View menu, allows
showing any activity and its immediate predecessors and successors in
a family tree format through one or more “generations.” Unfortunately,
this feature is not available in the bar chart view in Version 2.0 but is
available in the 3.0 and P3e versions released in 1999 (Figure 24.1.11).

Figure 24.1.9 Bar chart portion of screen split vertically.

Figure 24.1.10 Bar chart portion of screen split horizontally.

Figure 24.1.11 Pure logic or “PERT” view of network with Primavera’s trace logic high-
lighted, with two generations shown.

391
392 The Practice of CPM Scheduling

24.2. Project and Activity Codes for Organizing Project Data

One of the strongest features of Primavera Project Planner is its capa-
bility of organizing, summarizing, and depicting project data in a myriad
of ways. Version 2.0 permits 64 characters for unrestricted coding, plus
it permits the use of the activity identifier (an additional 10 characters)
for additional coding. Common codes fields assigned are for responsi-
bility, area of the project, and type of work to be performed. Other codes
used by this author include subcontractor, nominal crew size and com-
position or type, and the drawing number of the engineer’s plans that
most clearly depicts the work included in this activity.
These code fields are extremely useful for sorting, filtering, and fur-
ther explaining the scope of the activity. A common error by many new
users of this software is to reduce the number of usable codes by failure
to properly abbreviate the coded information. Thus, for Responsibility,
one user may use “Owner,” “Civil,” and “Electrical.” Another may prop-
erly use “O”, “C,” and “E,” while defining such codes in the activity code
dictionary. In the first case, 10 characters of the 64 available are used
to define responsibility, in the second, only 1.
As noted in Chapter 9, choosing the coding structure is extremely
important. In the event that the users’ organization desires to summa-
rize details over several projects (for example, concrete pours) a common
company-wide coding structure must be in place. A shortcoming of
Primavera is that the program looks to the position and not the title of
the code field when summarizing across projects. Thus the user should
place common codes, such as subcontractor, near the top of the activity
code list, while codes customized by the individual project manager,
such as area of the site, should be near the end of the list. If a company
wishes a report on all work by a specific subcontractor, the software
selects on, for example, positions 2 and 3 in the 64 character code field
rather than codes fields titled “SUBC.”
Although both Primavera personnel and practitioners extol the impor-
tance of determining the code fields prior to entering data, Primavera
Project Planner defaults to entry of activity data upon specifying a new
project. After escaping from the activity data entry field, access to cre-
ating the new project activity code structure and dictionary may be had
through the Data menu, choosing Activity Codes.
Other important data to be entered at this time includes Calendars,
WBS coding, Resources, and Custom Data Items. Custom Data Items
allows up to eight additional coding fields (either in character, numeric,
or date formats) for display, but limited sorting and no summarization
capability.
A common error of new or infrequent practitioners is to over rely upon
coding placed into the activity identifier (ID). The primary purpose of
 Let’s Look Under the Hood at the Engine 393

the activity ID, it should always be remembered, is to designate a unique

activity. Subsequent referral of the activity in the field will often be by
this activity ID, and data entry for updates will require keying such ID.
If the ID is too long or complicated (mixing letters and numbers, for
example), errors in transcription or a disinclination for use by field
forces may result. This author typically uses numeric IDs only, reserv-
ing the first one or two digits for the sheet number of multiple-sheet,
hand-drafted pure logic network diagrams. The “0” or “00” or blank
sheet number is reserved for submittals and procurement activities.
A further practical restriction on the use of coding within the activ-
ity ID, is Primavera’s reservation of two of the ten character spaces for
designating subprojects. These two characters are automatically placed
at the front of the activity ID and a code field “SUBP” is created in the
activity ID code dictionary. (As “SUBP” is thus a reserved code name by
Primavera, the user should not use the same code name elsewhere even
if not using subprojects.)
While most practitioners do not have the need to break their projects
into subprojects, the use of more than eight of the ten characters available
for the activity identifier should be discouraged. A second use for these two
reserved spaces may be for merging of multiple projects. Accessible through
the Tools, Project Utilities, Merge menu selection, this option creates a new
“superproject,” which may include as many of the user’s projects as desired.
This is accomplished without affecting the original individual projects and
is useful for “rolling-up” resource usage across projects.
To assure uniqueness of activity identifiers during the merging
process, Primavera must add a “project identifier” to the existing activ-
ity ID. These additional characters may be placed at the front or end of
the remaining characters available for unique activity designation.
Reserving one character for this purpose effectively limits the organi-
zation to 26 projects, while reserving two should suffice for all but the
largest organizations.

24.3. Entering Data

Primavera has three internal methods of entering activity data and
also accepts data imported from a separate spreadsheet or database
program. The first method, called by Primavera the Barchart method,
uses the combination tabular entry sheet (or spreadsheet) and bar chart
template. After entering the first activity identifier and establishing a
desired coding structure, the user should format a desired entry layout
by clicking or keying Format, Columns or keying the function F11 key
and creating a tabular layout desired. Data columns that may be desir-
able include Activity ID, Activity Description, Original Duration, user
394 The Practice of CPM Scheduling

defined codes, user defined custom data items, and (up to one) resource
and cost field. A useful feature is that the user-assigned column width
need not be sufficient to display the entire data item, so that multiple
log entries (each formatted to display only 10 characters) may fit on one
line or row. Unfortunately, inclusion of even one successor activity ID
is not an option using this method.
After determining the data columns desired, move the split line
between the tabular and bar chart portions of the layout to display as
many columns as desired. Data may then be entered directly in a spread-
sheet-style method with some minor caveats. The first is that, by default,
in moving the cursor to a new row, Primavera will automatically create
a new activity ID, sequenced a user defined increment, and move the
cursor to the second column of the layout. This “assistance” can be shut
off by clicking or keying Tools, Options, Activity Inserting and then
clicking off the Automatically Number Activities box. The second is that
once an activity ID has been entered, it cannot be changed unless the
user has checked the Exclusive box in the Open New Project or Overview
dialog box as described earlier in this chapter.
As noted previously, this method does not include assigning rela-
tionship links between the various activities. To accomplish this, the user
must either click or key View, Activity Detail, Successor (and
Predecessor) or key <Control> plus J (and <Control> plus E) to bring
up the Successor and Predecessor detail entry boxes. These can be
moved on the screen to appropriate locations. Data entry will now
require flipping between the main screen and the Successor detail entry
box. Another means of relationship link entry is via the bar chart half
of the layout screen, where the user may click and link the activities
shown on the screen. However, other than for the smallest of networks,
this can become tedious.
This problem can be partially alleviated by use of Primavera’s
Autolink function. By choosing this function, via clicking of keying
Insert, Autolink, or keying <Control> plus L (toggling on/off), each new
activity is linked with a finish to start relationship from the last entered
activity. Where this creates an improper relationship, or where multi-
ple links are required, such may be addressed using the Predecessor and
Successor data entry boxes.
The second method, called by Primavera the PERT method, uses the
pure logic network diagram format of Primavera. Here the user clicks
a location for a new activity after clicking or keying Insert, Activity or
keying <INS> and move the new activity box to a location desired, or
merely double click at the desired location for the new activity. This then
brings up Primavera’s Activity Form, which via a combination of point
and click and typing, may be filled in for the activity ID (remembering
to deactivate automatic numbering), activity description, and codes.
 Let’s Look Under the Hood at the Engine 395

Additional data entry boxes, for successors, resources, costs, custom

data items, constraints, logs, and so forth, may be called from this
Activity Form and be moved around the screen to accommodate data
entry. Relationship links may be made by using the Autolink function,
Predecessor and Successor data entry boxes, or by point and click drag-
ging logic connection lines between boxes.
The third method, called by Primavera the Activity Form method,
similarly enters data using the Activity Form from the combination tab-
ular/bar chart layout. Here the user must first click or key Tools,
Options, Activity Inserting, and “Use activity form when inserting an
activity.” Then for each new activity, key the {down arrow} key, add the
new activity in the Activity Form box and click OK, or key <ENTER>.

24.4. Debugging and Diagnostic Tools: Improving the Logic Network

To err is human, and errors in data entry are a fact of life in any data-
intensive system. Primavera, as well as its competitors, is fairly good
at providing context sensitive error trapping. If a user attempts to enter
an activity ID as its own successor or predecessor, Primavera will refuse
to accept such input. If a user attempts to enter an activity ID that
does not currently exist as a successor or predecessor, Primavera will
notify the user and prompt the user for a title or activity definition
for such new activity. However, if a logic link is incorrectly made to a
legitimate activity ID, or if a desired link is missing, such an error will
not be caught by the software. To address these types of errors,
Primavera has various diagnostic reports, which are some of its most
sterling features.
One such diagnostic report is generated each time the user schedules
the project. A word must be noted here related to the process of sched-
uling a project. Primavera does not schedule on the fly. In other words,
if a user changes a duration or constraint, the individual bar repre-
senting the activity may be elongated, shortened, or moved, but other
activities impacted by that activity will remain unaffected until the
user specifies a desire to reschedule the project. Many other software
products do reschedule on the fly, including Primavera’s Suretrak soft-
ware. However, as Primavera Project Planner or P3 projects are typi-
cally larger, the delays for recalculation required for each change may
be significant, even with the most powerful hardware. Thus the user
may make several changes (for example during an update process) and
then reschedule once.
A downside of this discretionary rescheduling process is that parties
receiving printed Primavera reports may find it necessary to rerun such
reports as a check. If a project is scheduled, showing an acceptable end
date, then the user changes the durations of activities, but does not
396 The Practice of CPM Scheduling

reschedule, printed reports will report the new durations alongside the
previously calculated dates. On the other hand, this feature has an
upside when merging multiple projects which may have been updated
on differing dates. Although Primavera will prompt the user to resched-
ule before printing “roll-up” reports, such need not be done, giving
greater flexibility to the individual project managers as to update fre-
quency while permitting company-wide “roll-ups.”
Returning to the Primavera diagnostic report, such a report is gener-
ated each time the project is scheduled. A project may be scheduled by
one of three actions, clicking the “clock” icon, clicking or keying Tools,
Schedule, or by keying the function F9 key. In the first two cases, the
report will be either printed or sent to the screen, in the third case, it is
still generated but is by default saved to file c:\p3win\p3out\p3.out. The
diagnostic report starts by noting the name of the registered user of the
software currently being used, its serial number (required to obtain tech-
nical help from Primavera), and the 4-character project designation.
The second section lists artificial constraints. As these may override
the calculated logic, they should be carefully reviewed by all users and
recipients of the CPM. Also listed here are any activities designated as
milestones, flags, or hammock activities.
The third section lists open ends. As noted previously, a project usu-
ally should have but one starting activity and one ending activity. If more
than one of each is listed, these should be checked to determine if a data
entry error has occurred.
The fourth section becomes useful after the project is underway and
being updated. It lists activities that have been started or completed
“out-of-sequence” or before their predecessors have been completed.
When an activity is reported being performed “out-of-sequence,” the
reason for such should be investigated. Either the original logic was
wrong, or field conditions permitted an informal change to the logic, or
more seriously, the work reported not complete in the predecessor or
reported started in the noted activity is not part of the intended defini-
tion of the predecessor or activity. Finally, the predecessor activity may
be complete but such was missed during the update process.
The fifth section lists the various options used for calculation, such
as how floats are calculated, whether “out-of-sequence” work is handled
by “retained logic” or is totally cut from the logic with “progress over-
ride,” whether activities must be continuous or may be interrupted by
finish-to-finish restraints, and other calculation options.
The sixth and final section lists statistics on the network including
the number of activities, number of critical activities, the number of
activities started and/or completed, percent complete, the number of
logic restraints in the network, the start date, data date for the most
recent update, and the latest calculated early finish.
 Let’s Look Under the Hood at the Engine 397

In the event a loop is detected (“activity A follows activity B which fol-

lows activity A”), the second section is a loop report, which is sent to the
screen. In such a case other sections are not generated until the loop is
corrected and the project rescheduled.
Other useful diagnostic tools include the standard Schedule Report
including Predecessors and Successors. (By default, such is generated
by keying Tools, Tabular Reports, Schedule, choosing SR-06, and keying
Run.) This report lists each activity as well as its predecessors and suc-
cessors and can be modified by the user to include constraints and
resource information.

24.5. Viewing Output

This author has always advocated that the most important aspect of
project controls is the preparation of the pure logic network of the CPM.
If at that point no further calculations were performed and the entire
CPM and all notes for its creation were destroyed, the project person-
nel who participated in its development would still have acquired 90 per-
cent of the value that the CPM can provide. But now that the effort has
been expended, how can that additional 10 percent of value be achieved?
The extra value of preparing a CPM, that portion which is best
assisted by today’s modern personal computers, is conveying the infor-
mation within the CPM to various members of the project team in a vari-
ety of views tailored to highlight all aspects and inferences of the
information collected. The ability to display the information in various
levels of detail with summarization in an almost unlimited number of
viewpoints is a hallmark of Primavera’s software.

24.6. On-screen Formats

The combination tabular and bar chart layout, which is the default
opening screen of Primavera, is highly useful for various levels of the
management team who wish to acquire information about the project.
By choosing appropriate data fields suited to the needs of the viewer for
the tabular half of the screen, and careful tailoring of information to be
conveyed by the bar chart half, the user can convey a great deal of
knowledge without having such lost in irrelevant (to the particular end
user) data. For example, at the beginning of the project, remaining
durations (RD) will equal original durations (OD) and percent com-
pletes (PCT) will be uniformly zero. If some degree of crew leveling has
been included in the initial schedule, late dates (LS and LF) will be irrel-
evant, although the reviewer may wish to know the amount of total
float (TF) to afford the degree of importance of meeting individual activ-
ity target dates. Once the grand logic has been reviewed and approved,
398 The Practice of CPM Scheduling

users of the CPM will care to focus upon activities for which they are
responsible, or which may have an immediate impact upon them.
Finally, the end user may wish to focus upon those activities that are
coming up in the near future as opposed to those several months away.
A sample report prepared for such a user may be produced on screen
(and then may be printed directly from screen for further distribution),
as shown in Figures 24.6.1 and 24.6.2.
To obtain such a report, the user need only key or click a few instruc-
tions. First, click or key Format, Columns or key the F11 key. Using the
John Doe project located on the enclosed CD (file “JDOE”) or a project
of the user’s choosing, click or key to Column information. Click
“Remaining duration” and then the “-” box or key <DEL>. The high-
lighter will automatically move to the next field, “Percent complete”
and the user may again click the “-” box or key <DEL>. Click the field
“Resource” and the “+” box, then the “{down arrow}” box and scroll to and
click “Total Float” then repeat for “Free Float.” Click OK.
Click on the bar between the tabular and bar chart halves, and move
the bar to the right to display all relevant columns. Next, right click
(using the right button on the mouse) on the calendar bar on the bar

Figure 24.6.1 Sample mixed summary and detail screen from point and click exercise,
setup.
 Let’s Look Under the Hood at the Engine 399

Figure 24.6.2 Sample mixed summary and detail screen from point and click exercise,
results.

chart half. Left click on the Density button and slide to the left until
September of 2001 is visible. Click OK. In this case, the entire John Doe
project is now viewable by scrolling down the screen. Finally, to reduce
screen clutter, key Format, Bars or key <Control> F11, and click off the
box calling for the “Float Bar” to be “Visible.” Click “Close.”
Next, assuming we are interested in only contractor number 5. Click
or key Format, Filter, Add, “OK,” then the “+” button, the {down key
underlined} box, scroll to “CONTRACTOR” and click, click the “Is”
column and key “EQ,” click the “Low Value” column and key “ 5” (space
then “5”) and again click “OK,” then click “OK” yet again, and then click
“Yes.” The user now has a detail schedule for the paving and landscape
subcontractor.
Next, to show the general relation of all subcontractors, key Format,
Filter, “All,” “OK,” “Yes,” then key Format, Organize, the “+” button
in the “Organize by” section, the {down key underlined} box, scroll to
“CONTRACTOR” and click, and click “OK.” Each contractor may now
view his individual tasks. Click or key View, Relationships, or the
function F3 key or click the pitchfork icon to show the relationships
400 The Practice of CPM Scheduling

between activities. Note that critical relationships are highlighted with

red lines, non-critical but driving relationships with solid black lines,
and non-driving relationships with dashed black lines.
Next click or key Format, Summarize All, Summarize to, the {down
key underlined} box, scroll to “CONTRACTOR” and click, and “OK.”
Double clicking on any of the summary bars will now toggle showing
detail for that one contractor. The power of CPM is in the preparation
of the logic. The power of software is in the variety of presentation
options it may provide.

24.7. Viewing Output: Tabular Formats

Primavera provides a number of standard reports for review of the ini-
tial schedule and for subsequent project monitoring. These are accessed
through Tools, Tabular Reports, and then from a list including Schedule,
Resource, Cost, and various custom report styles. A standard report
that should be reviewed at the start of each project is Primavera’s default
report SR-06 listing each activity, its predecessors and its successors
(Figures 24.7.1 and 24.7.2).

Figure 24.7.1 Standard report SR-06 setup screens.

 Let’s Look Under the Hood at the Engine 401

Figure 24.7.2 Sample page from standard report SR-06.

24.8. Viewing Output: Graphical Formats

Primavera also provides a number of standard graphical formats, cus-
tomizable by the user. A series of bar chart formats largely duplicates
the WYSIWYG tabular and bar chart layout, which is the default open-
ing screen. However, as the computer code for these bar chart views were
written prior to Primavera’s migration to a Windows platform, there are
differences. One report worth reviewing is the summary bar chart with
detail provided (Figures 24.8.1 through 24.8.3).
Various resource and cost reports in graphical format are also largely
duplicative of the newer Windows screen reports. The main set of graph-
ics not otherwise available is of the Timescale logic (Figures 24.8.4 and
24.8.5).
One important caveat in preparing any of the graphic reports in
Primavera is to set the printing or plotting device prior to preparing the
graphic. This is accomplished by File, Print Setup, Specific Printer.
Failure to do this before preparing the graphic will limit the smallest
text font on the graphic to that which is readable on the screen or on a
81/2 × 11 sheet of paper.
402 The Practice of CPM Scheduling

Figure 24.8.1 Summary bar chart with detail provided, setup and dialog boxes.

Figure 24.8.2 Summary bar chart with detail provided.

 Let’s Look Under the Hood at the Engine 403

Figure 24.8.3 Blowup of section of summary bar chart with detail provided.

Figure 24.8.4 Timescale logic graphic report, setup.

404 The Practice of CPM Scheduling

Figure 24.8.5 Timescale logic graphic report, results.

24.9. Entering Update Data

Primavera is forgiving in most of its demands upon users to follow
proper procedure in preparing or updating a schedule. Failure to prop-
erly tie all activities to designated ending activities will not preclude cal-
culation. (Such a condition will be reported in the diagnostics, however.)
But Primavera insists that activities that started are given an actual
start date and activities that are complete must be given an actual
finish date. Failure to provide an actual finish date will result in activ-
ities that are reported 100 percent complete and have zero remaining
duration to be displayed as having an incomplete portion thereof remain-
ing to be performed on the data date.
In most cases, this creates merely a problem of display. However, if
two activities are connected by a relationship with a lag, a more seri-
ous problem can occur. In the case of a start-to-start relationship with
a lag, the default algorithm is to decrementing the lag from the reported
actual start date. If no date is given, the entire lag will be included in
schedule calculations even if 90 percent of the activity is reported com-
plete. If neither an actual start nor actual finish date are given for an
activity reported 100 percent complete with zero remaining duration, the
software will still require the entire lag duration from the data date of
 Let’s Look Under the Hood at the Engine 405

the update until the activity restrained by that relationship may begin.
Therefore, the project manager who desires a quick update, with actual
dates to be provided later, is required to manually reduce each lag when
attempting such a “quick” update.
This problem is exacerbated in SureTrak, which does not make
allowances for activities started but not completed, even if an actual
start date is given. If an activity is reported as complete (with actual
start and actual finish dates), the lag will zero out. Otherwise, activi-
ties in progress require manual updating of the remaining lag of each
relationship as well as the remaining duration of the activity. If the
activity has been started out-of-sequence, the problem is made even
worse, with the start of the entire lag deferred until the calculated early
start of the remaining portion of the work. To provide compatibility
between systems, Primavera permits use of the SureTrak algorithm, by
keying Tools, Schedule, Options, and Early Start to the dialog question
“Calculate Start-to-Start lags from.”

24.10. Primavera Project Manager: P3e/c

This tour through the P3/SureTrak class of software has been meant to
be illustrative rather than a tutorial. In 2004, Primavera Software
Systems retired these two products that are still supported (for the
350,000 licensed users worldwide) but are no longer sold. However,
the general class of instructions, screens, menus and architecture of
these products are indicative of how other software products by
Primavera and its competitors are designed. The release of P3e/c 5.0,
coincident with the publication of this Sixth Edition, includes features
not previously available and modifications to the screens and menus of
prior releases. Indubitably, next year there will be a release 5.1 that has
further changes and within the next few years there will be another
major change to a release 6.0 or even to another brand new flagship
product. Similar to the experience of the past, it may be expected to go
through several new generations before there is the need to again update
this text to a Seventh Edition. Therefore, having provided a tour of gen-
eral steps for one class of software (as well as a working copy on the
enclosed CD for purposes of familiarization and practice exercises,) we
leave to the practitioner to read the user manual and utilize the tuto-
rial of the software product de jour.
However, before leaving this chapter, the administrative steps
required of the Primavera’s new flagship P3e/c (and thus of other soft-
ware vendors that do have a tendency to copy each other) should be
addressed. The old model was based upon the primacy of a project man-
ager (using Primavera Project Manager P3) who was to work inde-
pendently upon a personal computer. The new model is based upon the
406 The Practice of CPM Scheduling

hierarchy of projects within an organization, sharing organizational

resources amongst many project managers who report to a program
manager. The allocation of these scarce resources, as well as real time
tracking of production, are under the oversight of this program manager
as all team members (using Primavera Program Manager P3e/c) work
together on laptops and desktop computers tied via the internet to the
main organization database in the home office.
We have learned that before launching a new project logic network,
it is important to determine all the players involved with a project and
to set up the code fields necessary for the systematic collection of infor-
mation on each activity so that we can provide the necessary filters,
grouping and sorting of output information for these many audiences.
As we go to organizational reporting – considering the needs of the man-
agers of the asphalt batch plant and heavy equipment yard as well as
the manager of each individual project, there is a little more setting-up
exercises required before we start drafting out the pure logic network.
These are best illustrated by the new opening screens for the starting
of a new project under P3e/c.
The first step in designing the codes data that will be collected while
planning a project mimics the process of opening a new project in the
software. The first step is to choose an organization, the program man-
ager of which will be overlooking the project manager for this project
(Figure 24.10.1). In the real world, this means it is a good idea to deter-
mine to whom the project manager will report before choosing a coding
structure to meet this person’s preferred style of report. (Of course, in
the construction world, there are usually two EPS’s to consider – that
of the owner and that of the contractor.)
If the new project will be supervised by other than one of the exist-
ing programs, it will be necessary to add a new program to the
Enterprise Project Structure before adding the new project. This will
be added initially below the last project already in the system, but may

Figure 24.10.1 Opening P3e/c

screen to add a new project.
 Let’s Look Under the Hood at the Engine 407

Adding and placing

Figure 24.10.2
a new Enterprise.

then be moved to a location of choice upon the menu of programs and

projects as well as elevating or decrementing the level in the hierarchy
as shown in Figure 24.10.2.
At this point the user can finally name the project and assign a proj-
ect ID – fortunately with more than four characters since, with one
common database rather than separate subdirectories for each proj-
ect, each project manager cannot name his baseline schedule “BASE
(Figure 24.10.3).” Unfortunately, this may require some level of organ-
ization-wide coordination, if for example, PennDOT has two highway
projects on I-80 at different ends of the state under separate district engi-
neers but all under one Secretary of Transportation. The two projects
cannot both be named “I80 Baseline,” nor can that name be reused in a
few years for the next section unless the first project has been totally
purged from the database.
Six screens later, after assigning end dates, the responsible manager,
resource rate types and choosing whether to use a Wizard to plan the
whole project “on the fly,” the user now has added a new project to the
project directory and may then add activities to the new project. At this
point, after opening the program manager and choosing “Projects,”
the user can scroll down to open the desired project. (See Figure
24.10.4)
Having expended the effort just to add a new project does have a pay-
back. While the older P3 software had a limit of 20 activity codes shar-
ing a maximum of 64 characters, the newer P3e/c allows up to 500 codes
(of an unlimited number of characters) for each project and an unlimited
number of codes shared between projects at the global level. In addition,
it is now possible to sort or select within a code, for example for a code
running from 0 to 99 choosing all the activities coded in the 20s.
408 The Practice of CPM Scheduling

Figure 24.10.3 Process of adding a new project to the database.

Figure 24.10.4 Opening the Program Manager and choosing a project.

 Let’s Look Under the Hood at the Engine 409

The difference between global and project level codes is important to

recognize. Many projects may share resources such as heavy equipment
and other support facilities. Many projects may share a stable of sub-
contractors. Many projects may use a common specification coding struc-
ture. These are therefore candidates for global codes.
However, locations are usually specific to one project. In preparing the
code dictionary for a highway project, station numbers may be used for
locations, for a building the coding may be by floor. The drop-down menu
for choice of location on a highway project should not normally include
“5th floor.” With all of this additional power does come the requirement
to properly set up the codes and code dictionaries. (See Figures 24.10.5
and 24.10.6.)
The concept of global codes for standardization amongst projects can
be very useful so long as all projects within the group have some com-
monality. P3e/c version 5.0 recognizes this issue and has therefore added
an intermediate “global” coding capability for projects sharing the same
EPS code. (See Figure 24.10.7) This intermediate level of “global” coding
is essential for any consultant or other user who may have multiple
clients, each of whom may have their own “global” codes.

Figure 24.10.5 Selecting global or project level codes.

410 The Practice of CPM Scheduling

Figure 24.10.6 Including codes in the tabular layout.

Figure 24.10.7 Three levels of codes.

 Let’s Look Under the Hood at the Engine 411

24.11. Summary
As can be seen from this cursory review of only one software product,
the implementation of various options with regard to each of the exten-
sions of the traditional, simple ADM model are fraught with the danger
of accidental or intentional misuse. The marketplace has demanded of
high-end systems a level of power that requires study, care, and integrity
in its use. Conversely, the marketplace requires simplicity in applica-
tion of powerful and often only partly understood tools. The software
products available today, struggle to provide that which is desired by the
marketplace, often of a higher complexity than understood by the user.
This page intentionally left blank
 Chapter

25
Converting the Team Plan
to the Calculated Schedule

The project team has met and a pure logic network model of the con-
tractor’s “plan of execution” has been prepared. This process may have
been spread over several (preferably contiguous) days and may have
involved anywhere from 3 days (24 contact hours) for a complex $10 mil-
lion wastewater treatment plant to 5 days (40 contact hours) for a rel-
atively repetitive $100 million highway project. Now the Scheduler is
left with getting the plan information into the computer, getting the first
rough draft of the schedule out of the computer, meeting with the proj-
ect team for one more day to resolve any issues, and then preparing
reports that will help the project manager and the entire team get the
project out on time and under budget.

25.1. Data Entry Made Easy

The first step is to open a new project and define the activity codes that
have been recorded during the interview process. As previously dis-
cussed, it is generally recommended to not enter a “project must finish by”
date in the opening screen, but rather to assign a FNLT constraint to
the activity or milestone representing substantial completion of the
work. Figures 25.1.1 and 25.1.2 depict the screen shots for adding a
new project and establishing activity codes in P3.
Next, set the columns for easy entry of the activity and code infor-
mation collected. If the little extra time is taken to populate acceptable
values for each code, entry can possibly be made easier. This spreadsheet
style of entry may be far easier than entering only the activity ID,
description, and duration, and then having to go back to enter code
information on a detail screen for each activity.

413

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
414 The Practice of CPM Scheduling

Figure 25.1.1 Opening a new project in P3.

Entry of relationships is a little more difficult and may require use of

the mouse as well as the keyboard to navigate between the main screen
and predecessor or successor dialog boxes. In P3, there is a handy fea-
ture to create a finish-to-start restraint (without lag) for each newly
entered activity from the last activity entered. Keying the Autolink
option of the Insert menu, as shown in Figure 25.1.3, sets this feature.
In the event that a different relationship is desired, such may be accom-
plished by editing within the appropriate dialog box. A similar layout
may be created in P3e/c, as shown in Figure 25.1.4.

25.2. Check and Set Schedule Algorithm Options

At this point, the plan has been recorded and transposed to the software
for calculation. One last step is required, setting the proper algorithm
for the software. In SureTrak, the number of options are few, that is, the
choice of retained logic or progress override and whether open ends are
critical or not. Since the initial schedule will not include progress, the
choice of retained logic or progress override is irrelevant. However, it is
a good idea to make this decision now, prior to submitting the CPM and
running updates. The choice regarding open ends is more important
since making such not critical is in effect revising the logic that was
 Converting the Team Plan to the Calculated Schedule 415

Figure 25.1.2 Adding activity codes in P3.

Figure 25.1.3 Autolink.

416 The Practice of CPM Scheduling

Figure 25.1.4 P3e/c Predecessor and Successor detail.

Figure 25.2.1 Setting scheduling options in Suretrak.

Figure 25.2.2 Setting scheduling options in P3.

 Converting the Team Plan to the Calculated Schedule 417

Figure 25.2.3 Setting schedule options in P3e/c.

entered. Finally, setting the float to be declared critical only impacts the
graphic output, that is, whether an activity bar on the Gantt chart is
red or some other color. (See Figure 25.2.1.)
In P3, the choices are increased to also include the algorithm for cal-
culating lag when an activity has been started, determining if schedul-
ing durations are contiguous or interruptible and determining what will
be reported as the total float for each activity. As noted previously, the
initial schedule does not have progress, thus the choice of actual or early
start for calculating lag is irrelevant, but it is still a good idea to make
this decision before submitting the CPM to the owner. The choice of con-
tiguous or interruptible will impact the early start date of an activity
whenever it is driven by a finish-to-finish restraint. The choice to report
total float as either LF − EF or LS − ES or the more critical of the two
calculations will also have to be made at this time. (See Figure 25.2.2.)
In P3e/c, the options are different. Added options include the choice
whether to include or not include relationships to other projects in the
calculation, whether to use or ignore expected finish constraints in
determining remaining durations of affected activities, and choosing
which calendar will be used for lag durations between activities. The
choice of interruptible duration is not supported, probably as this sched-
ule concern conflicts with P3e/c’s strong linkage between schedules,
resources, and costs. (see Figure 25.2.3.)

25.3. First Run and De-bugging the Logic

Regardless of the care taken in data acquisition and entry, the first
attempt to run the schedule may result in the uncovering of various
errors. Perhaps these will be of a technical nature, such as a logical loop.
Or they may be of a situational nature, such as performance of sea-
sonal work out-of-season.
A logical loop may be due to an entry error, and solvable by the
Scheduler. Or there may truly be a misunderstanding or dispute
amongst the project team members. It is even possible that the loop orig-
inates with the contract documents, as in the example in Chapter 4.
418 The Practice of CPM Scheduling

Primavera Scheduling and Leveling Calculations -- Scheduling Report Page: 1

This Primavera software is registered to EnProMaC.
Start of schedule for project TEST.
Serial number...16660017
User name FRED .

Loop report -- Scheduling Report Page: 3

Loop # Activity Description
-------- ---------- ------------------------------------------------
1 A1000 first activity
A1010 second activity
A1020 third activity
A1000

Figure 25.4.1 P3 diagnostic report of a loop.

25.4. Loop Detection and Correction

Most software systems have some form of loop detection and reporting.
In P3, an attempt to schedule a project containing a loop will result in
an error message with details of the loop included in the diagnostic
report, as shown in Figure 25.4.1. This is viewed using the view program
included with Primavera P3.
P3e/c includes its loop report with the error message, saving the user
the trouble of switching to a separate program to read the diagnostic.
Often there may be several loops contained within each other and it may
require some time to unravel the error. Having the hand-drafted pure
logic drawing can be of great assistance in determining how and where
to cut the loop and retie the logic. It is important to do this first on the
hand-drafted diagram so as to understand the full consequences of such
manipulations and not to end up with a network that runs, but provides
incorrect information (Figure 25.4.2).

Erroneous loop detection and work-around. It was noted in Chapter 11

that many, if not most, software products will report an erroneous
report of a loop where a short activity is straddled by a larger activity,

Figure 25.4.2 P3e/c loop error report.

 Converting the Team Plan to the Calculated Schedule 419

Drywall
both sides
A+C7
7 days

R/I
electric
SS 2 B5 FF 2
5 days

Figure 25.4.3 Erroneous loop report.

such as in Figure 25.4.3. As previously noted, this type of situation

should properly be addressed by combining the short and longer activ-
ity into one activity, under the supervision of the drywall foreman, who
then directs the electrical foreman where and when to work. The one
activity can have both resources assigned to it, and possibly a special
code field may be required to alert the electrical superintendent or sub-
contractor that it is to perform ancillary work around the drywall
work. In P3e/c, the use of a “Step,” with its own associated codes, may
be of particular use in this situation. Depicting the two tasks as two
concurrent activities is not technically correct, but is sometimes the
“fudge” used, especially if in-artfully crafted specifications require sep-
arate coding or “complete” summary rollups for each craft.

25.5. Technical Review: The Primavera Diagnostic Report

The diagnostic provided by Primavera’s P3 software is unsurpassed in
the industry. Although the newer flagship P3e/c diagnostic is not quite
as good, the newer software has other features that make up for any dif-
ference. However, it is hoped that this “oldie but goodie” may be resur-
rected in a future release of P3e/c.
The P3 diagnostic is made up of seven sections. Each addresses a different
concern for the contractor reviewing its first run of the schedule and even-
tually for the engineer reviewing the final submission of the initial CPM.

Section 1: honesty. It is not the intent of the authors of this text to push
for the sales of software, nor is it the intention to act as an auxiliary police
force for the benefit of the creators and vendors of such software.
However, if the software printout notes the serial number and owner of
the software license assigned to the copy of the software being used, it
420 The Practice of CPM Scheduling

is strongly suggested that any practitioner of project controls verify that

a legitimate and properly licensed copy of the software is being used. As
has been noted, the Scheduler is in the unique position to truly ascer-
tain the truth of how a project is progressing with respect to time, much
like a CPA is in a position to determine the financial status of a project.
And with the power that has been provided in modern software pro-
grams to override the traditional algorithms and calculations of CPM for
special or odd situations, comes the responsibility to use such features
truthfully. Most casual users of CPM software are aware that it is pos-
sible for a misguided user to “cook the books.” Thus, to a large extent,
the weight given to the reports generated by the software is going to be
based upon the reputation of the Scheduler who prepared such reports.
Primavera Scheduling and Leveling Calculations—Scheduling Report Page: 1
This Primavera software is registered to EnProMaC.
Start of schedule for project 1234.
Serial number...16660017

User name FRED.

Section 2: constraints. The use of constraints is shorthand for the proper

use of restraints. Thus, for most valid uses of a constraint, it is possible
to say the same thing by means of one or more restraints. For example,
rather than using a SNET constraint, one may use a restraint with lag
from the start of the logic network to the activity to be constrained.
Obviously, the chore of counting the number of days for the original dura-
tion of this restraint for the initial network and for the remaining dura-
tion for each update is something that all parties would want to avoid.
Other types of constraint override the mathematics of CPM to some lesser
or greater extent and may have a use in special situations where the
mathematics of CPM are incapable of modeling the real world. Therefore,
the use of constraints should be viewed with some level of skepticism and
the user should be prepared to show how a restraint could be used instead
or why there is a need to override the mathematics of CPM.
Constraint listing -- Scheduling Report Page: 2
Activity Date Constraint
---------- ------- ---------------------------
99999 30AUG05 Late Finish Constraint

Section 3: open ends. Classically, each logic network had to start from
one node or event and end at one node or event. Many of the earlier com-
puter programs also required that the first node feed only one START
activity and the last node be fed from one END activity. However, the
real world does call for situations where multiple starting and ending
activities are desirable. On the other hand, unless a specific activity is
desired to be a starting or ending activity, the inclusion of such open ends
announces an error. As activities are added to a network and logic ties
 Converting the Team Plan to the Calculated Schedule 421

are added and deleted, it is all too possible for such an error to occur.
This section of the diagnostic notes all of the open ends for an ADM net-
work or a PDM network restricted to finish-to-start relationships.
The diagnostic does not note open ends of activities succeeded only by
start-to-start or preceded only by finish-to-finish activities as noted in
Chapter 11. Perhaps this section of the software was written before
PDM became popular. Thus there is a need to check for these additional
open ends using other tools.
Finally, the equivalent diagnostic from P3e/c (Click Tools Check
Project Integrity item numbers 36 and 37) lists activities that do not
have a predecessor or successor, but does not include in these lists those
activities that have connections to other projects but do not connect to
the end of the project in which they reside. Thus, the diagnostic for the
advanced software does not catch the type of error that the diagnostic
for the older P3 software was designed to report.
Open end listing -- Scheduling Report Page: 3
----------------
Activity 10 has no predecessors
Activity 76550 has no predecessors
Activity 76570 has no successors
Activity 76590 has no predecessors
Activity 99999 has no successors

Section 4: progress and actual dates. Section 4 of the diagnostic should

be blank for the initial schedule. This section normally notes work that
has been completed out-of-sequence and other irregularities relating to
reporting data to an update. Since the initial schedule should be based
on matters prior to beginning work, no progress should have been
reported. Occasionally, a Scheduler may improperly enter the actual
date for the start of the project or other actual progress that has occurred
prior to completing the initial CPM. Since the initial CPM data date will
be start of the project, this type of error is often reported in this section.
Remember also when updating that the data date should be at least one
day after the latest actual date reported.

Section 5: choice of algorithm. Since there may be more than one possi-
ble algorithm to calculate the schedule, and each algorithm may calcu-
late a different answer, it is important to document what options have
been chosen. It is also useful to note when the calculation was performed.
Scheduling Statistics for Project 1234:
Schedule calculation mode - Retained logic
Schedule calculation mode - Interruptible activities
Schedule calculation mode - Make open ends critical
Float calculation mode - Use more critical float from start
or finish dates
SS relationships - Use early start of predecessor
Schedule run on Mon Aug 01 12:47:26 2005
Run Number 232.
422 The Practice of CPM Scheduling

Section 6: statistics. Section 6 provides a number of statistics on the net-

work and calculated schedule. These include total number of activities and
number of activities on the critical path. Obviously, if the percent of crit-
ical activities begins to push or exceed the limit set by the Corps of
Engineers guideline, there should be some level of concern and perhaps
this should be addressed in the narrative. Also, as a general guide to the
health of a schedule is the comparison of number of relationships to
number of activities. This is generally around 1.6 in a healthy network.
Number of activities.................. 873
Number of activities in longest path.. 12
Started activities.................... 843
Completed activities.................. 822
Number of relationships............... 1236
Percent complete...................... 97.6
Number of expected finish activities... 2
Number of late constraints............ 1

Obviously, for an initial schedule, there should not be any activities

started nor finished. Although constraints have already been addressed
in Section 2, the use of such should raise such concern that the diagnostic
again notes their inclusion here.

Section 7: dates. Finally, in Section 7, we get the “beef,” that is, when
is the project expected to be finished based upon a specific start, and how
does this compare to the date required by the contract.
Data date............................. 01AUG05
Start date............................ 01NOV04
Imposed finish date.................... 31AUG05
Latest calculated early finish......... 30AUG05

25.6. Beyond the Primavera Diagnostic

This text started by noting that planning and scheduling is a field of
engineering. The problem of orphaned open ends to the logic network,
undetected by even the top diagnostics of P3, does not mean the prob-
lem should be ignored. Either the professional must review each activ-
ity’s relationships to be assured that each is preceded by a
something-to-start restraint and is succeeded by a finish-to-something
restraint, or the professional must obtain or write the necessary soft-
ware to have a computer perform this rote task.
Figure 25.6.1 illustrates the program used for many years by
EnProMaC to supplement the P3 diagnostic report. A copy of this pro-
gram is included also on the diskette accompanying this text. The pro-
gram is written in dBase programming language, leaving to the student
the option to rewrite for use by Microsoft Access or other relational
database software.
 Converting the Team Plan to the Calculated Schedule 423

∗P3orphan - to locate PDM orphaned activity logic

∗prior to running - export activities to actlist.dbf, relationships to pslist.dbf
∗file pslist0 is modified with fields atitle and stitle, maybe more
∗file actlist0 is extended format “ACT2, pcount..., scount...”
use pslist
copy to temp1
use pslist0
copy stru to pslist
use pslist
appe from temp1
repl all act2 with act
index to pslist on act2
sele 2
use actlist
copy to temp2
copy to temp1 structure extended
use temp1
appe from actlist0
∗goto 1
∗n=field_type
∗m=field_len
∗loca for field_name =’ACT2’
∗repl field_type with n
∗repl field_len with m
use
create actlist from temp1
appe from temp2
repl all act2 with act
index to actlist on act2
sele 1
set rela to act into b
repl all atitle with b->title, aod with b->od && more
set rela to
set index to
repl all act2 with suc
index to pslist on act2
∗sele 2
∗actlist already indexed on act2
∗sele 1
set rela to act2 into b
repl all stitle with b->title, sod with b->od && more
close all
∗file pslist is now has all activity data added
∗
∗
set talk off
∗count number of succs for each activity
clear all
use actlist

Figure 25.6.1 P3orphan extended diagnostic software code. This code is avail-
able in computer readable form on the CD provided with this book.
424 The Practice of CPM Scheduling

sele 2
use pslist
sele 1
r1=1
do while .not. eof()
r1=recno()
nact=act2
sele 2
count to ncountSS for act=nact .and. rel=’SS’
count to ncountFS for act=nact .and. rel=’FS’
count to ncountFF for act=nact .and. rel=’FF’
count to ncountSF for act=nact .and. rel=’SF’
count to ncount for act=nact
sele 1
goto r1
repl scountss with ncountss,scountfs with ncountfs,;
scountff with ncountff,scountsf with ncountsf,scount with ncount
skip
if substr(str(recno(),5),5,1)=’0’
disp recno()
endif
enddo
∗
∗count number of preds for each activity
1
r1=1
do while .not. eof()
r1=recno()
nact=act2
sele 2
count to ncountSS for suc=nact .and. rel=’SS’
count to ncountFS for suc=nact .and. rel=’FS’
count to ncountFF for suc=nact .and. rel=’FF’
count to ncountSF for suc=nact .and. rel=’SF’
count to ncount for suc=nact
sele 1
goto r1
repl pcountss with ncountss,pcountfs with ncountfs,;
pcountff with ncountff,pcountsf with ncountsf,pcount with ncount
skip
if substr(str(recno(),5),5,1)=’0’
disp recno()
endif
enddo
∗
set talk on
return

Figure 25.6.1 (Continued)

 Converting the Team Plan to the Calculated Schedule 425

The program combines the information from the activity file and rela-
tionship file that may be exported by P3 (or, with modifications, other
CPM software) to create an expanded activity file and relationship file.
The expanded activity file includes the total number of predecessors
(pcount) as well as number of each type of predecessor (SS, FS, FF, and
SF) and the total number of successors (scount) as well as number of each
type of successor. The expanded relationship file includes the title of
each predecessor/successor pair and the duration of each. (Figure 25.6.2)
The two files are then linked to prepare an expanded diagnostic report
via a second program (diaglist.prg) provided in Figure 25.6.3 The result-
ing report provides a listing by exception of suspect activity durations and
relationships for further review by the Scheduler, or later by the engineer.
This starts with an expanded “open end” report highlighting “open
ends” orphaned by misuse of PDM, continues by listing activities with
duration greater than 22 days (long durations) and equal to zero days
(often signifying a typographical error or missed entry), update reporting
errors of showing progress without providing an actual start date or show-
ing completion without providing an actual finish date, and concludes with
a listing of relationships where the lag duration between activities exceeds
the duration of the predecessor or successor. (Figure 25.6.4)
The diligent student may note that the programs discussed in this sub-
chapter would not catch the open end created by the use of a finish-to-start

Structure for database: actlist.dbf Structure for database: actlist0.dbf

Field Field Name Type Width
1 ACT Character 10 Field Field Name Type Width
2 TITLE Character 48 1 FIELD_NAME Character 10
3 RESP Character 1 2 FIELD_TYPE Character 1
4 SUBC Character 3 3 FIELD_LEN Numeric 3
5 CRSZ Character 2 4 FIELD_DEC Numeric 3
6 CRTY Character 2 ** Total ** 18
7 other codes
20 OD Numeric 5 Display content of database: actlist0.dbf
21 RD Numeric 5 Record# FIELD_NAME FIELD_TYPE FIELD_LEN FIELD_DEC
22 CAL Character 3 1 ACT2 C 10 0
23 ES Date 8 2 PCOUNT N 2 0
24 ESA Character 1 3 PCOUNTSS N 2 0
25 EF Date 8 4 PCOUNTFS N 2 0
26 EFA Character 1 5 PCOUNTFF N 2 0
27 TF Numeric 5 6 PCOUNTSF N 2 0
28 ECON Character 2 7 SCOUNT N 2 0
29 ECOND Date 8 8 SCOUNTSS N 2 0
30 LCON Character 2 9 SCOUNTFS N 2 0
31 LCOND Date 8 10 SCOUNTFF N 2 0
32 CON Character 2 11 SCOUNTSF N 2 0
33 COND Date 8
34 LOG1 Character 48
35 LOG2 Character 48
other logs Structure for database: pslist.dbf
39 ACT2 Character 10 Structure for database: pslist0.dbf
40 PCOUNT Numeric 2 Field Field Name Type Width
41 PCOUNTSS Numeric 2 1 ACT Character 10
42 PCOUNTFS Numeric 2 2 SUC Character 10
43 PCOUNTFF Numeric 2 3 LAG Numeric 4
44 PCOUNTSF Numeric 2 4 REL Character 2
45 SCOUNT Numeric 2 5 ATITLE Character 48
46 SCOUNTSS Numeric 2 6 AOD Numeric 5
47 SCOUNTFS Numeric 2 7 STITLE Character 48
48 SCOUNTFF Numeric 2 8 SOD Numeric 5
49 SCOUNTSF Numeric 2 9 ACT2 Character 10
** Total ** 472 ** Total ** 143

Figure 25.6.2 File structure for relational databases for extended diagnostic software.
426 The Practice of CPM Scheduling

*diaglist.prg
*program to list all orphanned activities and other suspect activities and relationships
set alte to diaglist
set alte on
use actlist
list off act,title,od,cal,'has no predecessor' for pcount=0
count for pcount=0
list off act,title,od,cal,'has no START predecessor' for pcountss+pcountfs=0 .and. pcount#0
count for pcountss+pcountfs=0 .and. pcount#0
list off act,title,od,cal,'has no successor' for scount=0
count for scount=0
list off act,title,od,cal,'has no FINISH successor' for scountff+scountfs=0 .and. scount#0
count for scountff+scountfs=0 .and. scount#0
list off act,title,od,cal,'has duration > 22 days' for OD>22
count for OD>22
list off act,title,od,cal,'has duration = 0 days' for OD=0
count for OD=0
list off act,title,od,cal,'has RD > OD' for RD>OD
count for RD>OD
list off act,title,od,cal,'has RD=0 but no AF' for OD#0 .and. RD=0 .and. EFA=' '
count for OD#0 .and. RD=0 .and. EFA=' '
use pslist
list off act,atitle,aod,suc,stitle,sod,rel,lag,'uses SF relationship' for rel='SF'
count for rel='SF'
list off act,atitle,aod,suc,stitle,sod,rel,lag,'LAG > activity duration' for lag>aod
count for lag>aod
list off act,atitle,aod,suc,stitle,sod,rel,lag,'LAG > successor duration' for lag>sod
count for lag>sod
set alte to
return

Figure 25.6.3 Diaglist extended diagnostic software code.

act title od cal 'has no predecessor'

100 Notice to Proceed 0 1 D has no predecessor
13061 Fab & Deliver Str Stl 110 1 D has no predecessor
2 records

act title od cal 'has no START predecessor'

53353 Construct Retaining Walls 45 1 D has no START predecessor
28925 Term Signal Wiring&Cable 45 1 D has no START predecessor
2 records
act title od cal 'has no successor'
99999 Project Completion Date 0 1 D has no successor
1100 Submit Community Relations Plan 60 2 D has no successor
1200 Submit Utility,Municipal,Other Notif. 30 2 D has no successor
1520 Submit Final Test Plan 45 2 D has no successor
4 records

act title od cal 'has no FINISH successor'

75661 Cut&Fill Sta1363+19-1368+18 1 1 D has no FINISH successor
86310 Build Retn/Grav/Crash Walls 50 1 D has no FINISH successor
28575 Instl Stormwater Systems 30 1 D has no FINISH successor
3 records

act title od cal 'has duration > 22 days'

11001 Grade Sta 252-142 86 1 D has duration > 22 days
11005 Grade/Fill Sta 58-142 35 1 D has duration > 22 days
13061 Fab & Deliver Str Stl 110 1 D has duration > 22 days
53353 Construct Retaining Walls 45 1 D has duration > 22 days
28925 Term Signal Wiring&Cable 45 1 D has duration > 22 days
86310 Build Retn/Grav/Crash Walls 50 1 D has duration > 22 days
28575 Instl Stormwater Systems 30 1 D has duration > 22 days
7 records

act title od cal 'has duration = 0 days'

100 Notice to Proceed 0 1 D has duration = 0 days
99999 Project Completion Date 0 1 D has duration = 0 days
2 records

act title od cal 'has RD > OD'

No records

act title od cal 'has RD=0 but no AF'

No records
…Etcetera… Suspect Lag Report Not Shown Due to Width Limitations…

Figure 25.6.4 Extended diagnostic report.

 Converting the Team Plan to the Calculated Schedule 427

with negative lag, as pointed out in Chapter 11. It is left as an exercise

for that student to modify the programs accordingly.

25.7. First Review of Calculated Output: Reality Check #1

Often a contractor or engineer will foolishly skip the technical review
and go directly to the graphical bar chart to determine if the activity
placement on the schedule appears to be correct. It is important to per-
form the technical review first and correct any technical errors discov-
ered as such may have a significant impact upon the schedule. But
finally, we must look at the activities and their scheduled dates in the
context of actually utilizing the CPM.
Up to this point the data acquisition, entry, and technical review may
be performed by a young engineer, fresh out of school. It is at this point
that a proper review requires some element of experience in the field of
work for which the CPM is being produced. Even here, however, the
Scheduler can assist project personnel by providing appropriate views
of the multitude of activities and other data and reducing the task of
review from that of all activities to significant samplings.

25.8. Detail Views of Output of Schedule Calculations

Having walked through the entire project not once but twice during the
data acquisition process, it may not be realistic to expect the project man-
ager or superintendent to review 100 percent of the scheduled dates of
activities to determine reasonableness. After all, the time of these indi-
viduals is also limited and it is far better utilized in preparing the plan
than reviewing the schedule. Thus, the job of the Scheduler is to high-
light those areas in the schedule that most need this review.
Obviously, the critical path or paths is the first area that should be
scrutinized. Our previous discussions on what constitutes “critical” now
come into play. The Scheduler must determine whether to review only
zero float activities or those having 5, 10, or 15 days of float. A serious
issue here is to unravel multiple paths of float rather than merely pres-
ent a list of activities that may be on one of several “critical” and “near-
critical” paths. As noted, the advent of multiple calendars further
exacerbates this issue as a sort by total float may not include all activ-
ities of a near-critical path.
For purposes of demonstration, assume Activity 236 of the John Doe
project will require 140 calendar days rather than 90 work days for fab-
rication, thus creating concurrent critical paths through the plant and
office structures. A standard screen shot, such as Figure 25.8.1, will
428 The Practice of CPM Scheduling

Figure 25.8.1 Two concurrent critical paths cannot be distinguished in a simple early start
printout.

show all activities having zero float but no means to easily trace the two
concurrent paths.
To provide usable information to the project manager, superintend-
ent, and other construction specialists, the Scheduler must prepare a
proper report. First, it is necessary to add a new code value for segre-
gating the various paths. In Figure 25.8.2, this code is named PLOT
as it may be reused for other graphics in the future. After filtering for
“No Activities,” Find (by keying Control F) Activity #80, the last activ-
ity in the John Doe network. Add to the screen the Predecessor dialog
box (by keying Control E) and notice those activity relationships that
“drive” this last activity, Activities 72, 78, and 94, highlighted by an
asterisk in the column between the Activity ID and Relationship.

Figure 25.8.2 Use of PLOT code.

 Converting the Team Plan to the Calculated Schedule 429

Also note that two activities, 58 and 73, are within the project man-
ager’s “envelope of criticality.” Click activity 72 and then click the Jump
button. Continue to click the Jump button until reaching the start of the
network, making a note whenever there is more than one driving rela-
tionship or other activities having float less than the threshold set for
being near-critical. (Figure 25.8.3.)
Assign a code of 1 (or first critical path to be plotted) to these activi-
ties. Now go back to activity 80, click the next driving relationship, and
click Jump until jumping to an activity already previously chosen, coding
the second path to 2. Repeat until all such paths are traced and coded.
Your notes taken during this exercise will include that activity 67 had
two driving relationships and thus starting from 67 and clicking first
65 then Jump will trace a fourth path.
This process may be repeated for near-critical paths by going to activ-
ity 80 and clicking 58 then Jump, and then, importantly, clicking the
driving relationship to 58 then Jump. Repeating this process will track
additional near-critical paths, including the importance of fabrication
of the plant electrical package. (Figure 25.8.4.)
The last step to preparing this graphic for the project team is to reor-
ganize the bar chart in ascending rather than descending order. Clicking
Format Organize brings up the organize dialog box. Add Group by Plot
Code and Sort by Early Start, as shown in Figure 25.8.5, to complete the
graphic, Figure 25.8.6.
Another area of concern to a project manager would be work per-
formed out-of-season, such as pouring concrete during the winter in the
northern US or during the summer in Panama. One of the codes for
which information was carefully collected and keyed to the system for
the John Doe project was the type of work. Thus, it is easy to create a
filter requesting all concrete work between December 1 and April 1,
such as in Figure 25.8.7. In the John Doe project, this resulted in no
“hits” so the filter was expanded to include all concrete work, as shown
in Figure 25.8.8. Note that the late finish of activity 28 extends past

Figure 25.8.3 Backward pass of one critical path.

430 The Practice of CPM Scheduling

Figure 25.8.4 Backward pass of alternate critical paths.

December 1 and therefore there is actually less than the 38 days of

total float for the string of 23–24–28.
A third area of concern to a project manager coordinating multiple sub-
contractors may be to be assured that drywall activities are deferred
until after “Test Piping” activities. The work may preferably be properly
coded to support such a filter or one may be created based solely upon
 Converting the Team Plan to the Calculated Schedule 431

Figure 25.8.5 Grouping by PLOT code.

Figure 25.8.6 Intelligent report of multiple critical paths.

432 The Practice of CPM Scheduling

Figure 25.8.7 Filter for Winter Concrete.

the words of an activity description. Obviously, if this second option is

to be used, exact and proper spelling is mandated. A filter, such as in
Figure 25.8.9, may be used along with organization (as in Figure 25.8.10)
by location, to provide the desired graphic review for the project team.
A fourth area of concern is to assess whether the impact of seasonal
weather has been properly addressed in the CPM plan. By setting a filter
to include only those activities that may be subject to seasonal weather,
and a few key milestones such as that point at which a building is
weather-tight, a quick review of the calculated early start dates may be
useful to note contractibility. But also it is important to note the late
finish dates, if these are shifted into “bad” weather periods, the calcu-
lated float will be unrealistic and may give the contractor a false sense
of security.

Figure 25.8.8 Report of winter concrete.

 Converting the Team Plan to the Calculated Schedule 433

Figure 25.8.9 Filter.

Figure 25.8.10 Organization.

434 The Practice of CPM Scheduling

Another test for seasonal weather is to bump the start date of the CPM
by 3 months and recalculate. The results may highlight those activities
that could be adversely affected and for which the use of a weather cal-
endar may be useful.

25.9. Time Scaled Logic Diagram

For a good general overview, a properly prepared time scaled logic dia-
gram (as provided in Figure 24.8.5) may be useful. Since this diagram
can show so much, it is important to moderate the amount of informa-
tion on the plot to that which can be easily viewed and understood.
Even for an “E” size drawing, trying to cram in more than 200 activi-
ties may be self-defeating.
Care should be taken once again to set the appropriate envelope of crit-
icality so as to highlight those activities having less than a set limit of
total float. Assuming that proper activity codes have been utilized, dif-
ferent color bars may represent differing crafts, subcontractors, or
responsible parties. For the initial plot, keep in mind that the remain-
ing duration is the same as the original duration and there is no need
to show both, but that the bars are plotted in early date format so it is
desired to display the total float. (Figures 25.9.1 and 25.9.2.)

Figure 25.9.1 Timescale overview content.

 Converting the Team Plan to the Calculated Schedule 435

Figure 25.9.2 Timescale overview pen.

If detail about a specific craft or interaction between two crafts in

a specific location is desired, this is best displayed on a separate plot.
A time scaled logic diagram is a graphic written paragraph. Choose a
paragraph topic and keep the “writing” simple and readable.

25.10. Tailoring Initial Output to the Chosen Audiences

The reviews are complete. The plan and schedule are done. It is time to
prepare the formal reports to all of the audiences that the CPM should
serve. The Scheduler has put on yet another hat—that of the Great
Communicator.

25.11. Whatever Owner Wants, Owner Gets

The provision of funding by the owner, like the charms of Lola in Damn
Yankees, assures the maxim that the owner, whose specification must
be met by the contractor, will get the reports that he or she wants. What
the specification may require may not always be useful to the owner. A
professional Scheduler may point this out to the owner’s representative
or engineer, but the advice is not always accepted. In such an instance,
436 The Practice of CPM Scheduling

document the exchange and provide exactly what is called for if such is
possible. When the owner or engineer declares such unreadable or unus-
able, the contractor should charge an appropriate fee for providing a
more useful submittal.
In one case involving the authors, the specification called for a com-
puter drafted time scaled logic diagram including all activities, that is,
construction, submittals, approvals, fabrication, and deliveries. When
it was suggested that such a diagram might be unreadable, the Engineer
insisted that if such a plot were not impossible that the contractor must
follow the specification without deviation. The final plot, stringently
following the specification and with all sheets pasted together, was over
8-feet tall and 6-feet wide. Following the logic lines from one activity to
another was almost impossible. The ensuing fracas involved extensive
litigation over a period of years and even involvement of a state engi-
neering disciplinary board.
Where a specification calls for an improper product, such as dictat-
ing the contractor “prosecute the work diligently,” complete by a stip-
ulated date, and yet submit a CPM showing that it intends to use all
contract time without a proper contingency at the end, the contractor
may point out the absurdity of such a combination of demands, but if
all else fails, he or she will be required to provide such a submittal. An
attempt may be made to include a “contingency” activity near the end
of the network, or a long duration activity for punchlist, or some other
“fudge” to comply with all three of the demands noted. But, the engi-
neer may reject these efforts and require the contractor to overstate
durations along the critical path to force the answer desired. In such
case, the submittal or accompanying narrative should be appended
with the note:

“This CPM submittal does not represent the Contractor’s intent and deliberately
misstates the Contractor’s intent per direction of the Engineer.”

The contractor should be pleased to remove such a note if requested

to do so in writing.

25.12. “You Can’t Always Get What You Want,

But . . . You Get What You Need”
A specification may require testing of safety equipment on the first day
of each month. The contractor’s insurance policy may require it test its
safety equipment every Monday morning. A good contractor may not like
the extra expense but will comply with both requirements, even if the
first of a specific month starts on a Tuesday.
 Converting the Team Plan to the Calculated Schedule 437

A specification for one project may require use of a special timesheet

or cost recording system for the benefit of the owner and quite different
than that typically used by the contractor. Except for the smallest “Mom
and Pop” contractor, and often even there, the contractor will comply by
having his personnel fill out two timesheets, one for the owner and one
for his own accounting department. Similarly, a good contractor will
implement a proper CPM for each of its projects whether such is required
by the specification or not.
The contractor’s CPM, prepared for the purpose of assisting his per-
formance on the project, may include codes not required by the specifi-
cation. The contractor may choose to update weekly (this being the
recommendation of the authors) to best mimic how the superintendent
and foremen would schedule a project without a CPM. And the CPM
specified by the owner may include requirements that reduce or elimi-
nate the ability of that CPM to be a tool to assist the contractor’s per-
formance. In which case, the contractor should properly submit the
“CPM” reports required in full compliance with the specification while
concurrently using his own CPM to run the project, similar to the usage
of duplicate timesheets noted previously.
This is not a radical idea. The reality is that many projects with a CPM
schedule unsuitable for “running the job” will have a duplicate sched-
uling system based upon 3-week look-ahead bar charts prepared by the
foremen. In some instances the engineer will request such a duplicate
system suggesting “the CPM is for recording purposes only.” In some
instances the specification will even require the duplicate systems. As
discussed in the next section, the update process should include a com-
puter generated 3-week look-ahead. Maintaining a second system for
such a purpose may even be less costly, not to mention the superior
product and other benefits of a proper CPM.

25.13. Reports and Views for the Foreman Performing the Work
The foremen on the project may desire to have a quick overview of the
forest, but for day-to-day operations must focus upon individual branches
of specific trees. In general, a foreman will want a view of activities
within the next 3 to 6 weeks at a maximum, and only those within the
location on the project where the foreman will be working. In addition,
the foreman already has too much paper and paperwork. Anything that
reduces or at least does not add to this burden will be welcome.
Without a CPM, foremen will typically prepare short-term bar charts
of upcoming work. The purpose of the CPM is to assist the natural abil-
ities of the foremen and not to replace them with an unfamiliar system.
Thus a combination tabular and bar chart report, similar to that which
438 The Practice of CPM Scheduling

is the default of every CPM software vendor, is a good starting point.

However, since in construction the foremen rarely have immediate
access to a PC screen, the Scheduler must choose the minimum infor-
mation that will convey the maximum information useful to the foreman
on a one or two (or as few as possible number of) pages.
A suggested template for a turn-around document is that provided in
Figure 25.13.1. Having the information organized (with page breaks) by
a responsible party, then subcontractor, then either location or foreman
and then the other will allow easy dissemination of the proper infor-
mation to the proper foreman. Limiting the activities shown to those that
either have already started or may start within the next 3 weeks sets
a realistic limit to the forward planning required of the foreman.
Providing a bar chart view of the next 4 to 6 weeks also allows the fore-
man to see the near-term substance of this limited activity set. An RDM
system might be able to include on this report information regarding not
yet started or incomplete activities by others that immediately precede
the foreman’s activities.
Tabular information that may be of use to the foreman and should be
displayed includes:
■ the activity number or identifier,
■ the activity title or description,
■ the original duration,
■ the remaining duration at the start of each periodic update period,
■ the calendar upon which the durations are based,

Figure 25.13.1 Foreman’s turn-around document.

 Converting the Team Plan to the Calculated Schedule 439

■ the referenced engineering drawings for this activity,

■ the crew size and composition,
■ other code data as appropriate and as may fit,
■ the early start and finish for the activity,
■ the total float for the activity, and
■ the junior float for the activity (if calculation of such attribute is
available.)

In addition, this same form may be used to collect information for the
next update and therefore tabular columns should be reserved for the
following items.
■ Actual start date if started,
■ Remaining duration at the end of the periodic update period if started
but not finished,
■ optionally the percent complete for progress reporting of longer dura-
tion activities,
■ an alternate cost percent complete if the CPM is cost loaded or the
inspector wants to “hold money” for an activity otherwise complete,
and
■ Actual finish date if finished.

For purposes of simplicity and providing to the foreman only the infor-
mation necessary to assist the crew’s efforts, it is suggested that only
one bar be displayed, indicating the current early dates of the activity.
Alternately, a dotted line or other symbol may be displayed to graphi-
cally display the available float for an activity.
The printout shown in Figure 25.13.1 provides the paving foreman
with most of the scheduling information needed for the next 6 weeks.
The activity 6144 may begin this week, but has over 200 days of float.
Work on Ramps A, B, and AB is near-critical, having only 9 days of
float. The foreman knows to check on drawing “C127” for details on this
work. The foreman may choose to check on activities 3304 and 3305
(whatever that might be), which is noted to the top left of the bar for 3306
as being the activities that must be completed during the coming week
if this subbase activity is to be performed next week. The asterisk on
3305* indicates this is the driving activity.
The key to transforming this computer printout to the scheduling tool
of choice of the foreman is to emphasize that the bars only represent those
activities that are anticipated to be available for performance and do not
represent the short-term schedule of the superintendent or foreman.
The early dates represent when it is anticipated the activity may first
440 The Practice of CPM Scheduling

be performed. The tabular or graphic float entry represents when the

CPM calculates the activity must be finished if the project is to complete
on time without expensive overtime or other fixes. The actual placement
on the short-term schedule is still a decision to be made by the project
team using all data known when the decision is made.
For this, the superintendent and foremen should still meet each
Monday morning or Friday afternoon or time of choice, and mark the
turn-around document with a Sharpie pen or other highlighter. The
superintendent now knows what activities are scheduled for the next
weeks and can schedule resources accordingly. The foremen should now
again review the contract drawings for work scheduled in the coming
weeks to ascertain any last minute RFIs or determine any changes in
condition.
At the end of the week or other reporting period, the foreman can fill
out the actual date and remaining duration information for computer
calculation and printing of the next week’s turn-around document. Any
special problems may be noted next to the bar representing the affected
activity, and the Scheduler can have such entered as a log entry or
“Note” or memo and printed on subsequent turn-around documents
until the activity is completed. (Figure 27.15.1)

25.14. Reports and Views for the Contractor’s Superintendent

The job of the project manager or superintendent, or for that matter any
other line manager, is to see that those who will be producing have all
the tools, materials, and instructions necessary to produce. This key
person must have at all times both a view of the forest and the trees. To
oversee the total project, a good graphic report may be the time scaled
logic diagram used for the final review, limited (per 3-foot width) to the
200 or so most critical activities plus significant milestones whether or
not near-critical. For a project of significant size or duration, an addi-
tional 3-foot panel may be required, but the goal is still to produce a chart
that can be easily used by the superintendent rather than one encom-
passing all “important” information. The skill of preparing graphics
that meet these all-too-often exclusive requirements is part of the art
of being a Scheduler.
The turn-around document prepared for the superintendent should
be similar to that prepared for the foremen with a few significant dif-
ferences. Unless the superintendent will be personally doing a walk-
through of the project for each update to collect information, there is no
need for the actual information columns. (On the other hand, if the
super will collect this information, there is no need to clutter the forms
used by the foremen with these columns.)
 Converting the Team Plan to the Calculated Schedule 441

Additional tabular information that should be on the superinten-

dent’s printout includes the variance between the early finish of the
last update and the current schedule. Additional graphical information
that should be on the superintendent’s printout includes an additional
bar to reference the following:
■ baseline or initial schedule,
■ most recent revision to the baseline (if applicable) and
■ last most recent update
The turn-around document for the superintendent has more informa-
tion on it than the one for the foreman (Figure 25.14.1). Instead of only
one, there are now three bars: the baseline schedule, the last update, and
the current schedule. Legend or memo notes remind the superintendent
that there is a stop work order on activity 3302 while investigating a
Native American artifact, but that this work is now near-critical. As a
result of this stoppage, three subcontractors, the bulk excavator, under-
ground pipe installer, and electrician, have had their work disrupted. The
superintendent should be placing the owner on notice of a claim for an extra
while reminding the subcontractors to record their costs of demobilization
and remobilization in this area. Another note indicates the inspector insists
on holding 10 percent for a damaged inlet, but that this can be repaired
or replaced at a later time and will not delay work in this area.
Although not available in popular commercial software, if in-house or
third-party software provides forecasting capabilities to adjust future
original durations based upon past performance for similar work, the
superintendent’s report should also include the current forecast of the

Figure 25.14.1 Superintendent’s turn-around document.

442 The Practice of CPM Scheduling

adjusted original duration and an additional bar to reference the most

recent forecast.

25.15. Reports and Views for the Contractor’s Upper Management

The job of upper management is to see that the superintendents, or
other line managers, have the management tools necessary to best per-
form their function and are otherwise “in good running order.” A part of
this is to measure the health or performance of the project and to pro-
vide additional assistance even if not requested, (including if required,
changes in line management). This does not suggest knowing every
detail of the workings of the superintendents nor that of the foremen.
Modern physics tells us that when we measure something, we impact
that something and tend to slow it down. A project where every action
is carefully measured may be expected to slow to a standstill.
The Scheduler must strike a balance between the stated desires of
upper management to be informed of “everything” and to not be both-
ered with “inconsequential details.” The careful choice of coding struc-
ture at the beginning of the preparation of the CPM interview process
is the key to providing this balance. For the highway project example,
choosing to summarize by subcontractor or crew can provide one view
of the health of the project, while summarizing by area can provide
another. In Figure 25.15.1, the one week lapse between the second and
third bar for each subcontract indicates that something has stalled the
entire project. The note appended to the printout in Figure 25.15.2 gives
one possible reason for the problem.
How did that note get there? It was added by being “pasted” in place
by the Scheduler using the various graphic tools available to one extent
or another in all of the commercial software on the market. In many
cases this is the best that can be done. While P3 has its log fields for an
individual activity and P3e/c has notes that can attach objects (such as

Figure 25.15.1 Summarized by subcontract.

 Converting the Team Plan to the Calculated Schedule 443

Figure 25.15.2 Summarized by area.

a photograph) to an individual activity, neither has the ability to attach

information to the summary of selected groups of activities. But if this
is what is needed to provide the proper information to the selected audi-
ence, this must be part of the work of the Scheduler in preparing such
presentations.

25.16. The Narrative Report for Each Audience

After providing informative tabular reports and after creating killer
graphics to illustrate exceptions and areas requiring further attention,
the Scheduler typically must write a narrative to describe, in terms
understandable to a nonscheduler, the contractor’s plan of execution
and how the CPM has been implemented to assist the project team
achieve it.
“How to write” could be the topic for an entirely separate textbook; and
there are many excellent texts that do discuss the writing of a techni-
cal report. For purposes of this subsection, there is no universal format
and each Scheduler should work out his or her own style. The Scheduler
should keep in mind for whom the narrative is being written, and pro-
vide separate narratives if necessary for multiple audiences. If the CPM
submittal is to be made to another Scheduler, the use of scheduling
jargon and the absence of long explanations may be appropriate. In
general, the engineer reviewing the CPM submittal is not a scheduler
and must have even simple concepts repeated, such as that the early
dates represent only when the activity may possibly be performed and
do not guarantee, promise, or even suggest that the contractor has any
intent of performing on such dates unless the activities are critical or
near-critical or the schedule has been resource leveled.
A narrative provided for the engineer should highlight all items in the
submittal that should raise issues or call for higher scrutiny. Thus, a
separate listing of those activities utilizing non-traditional relation-
ships, noting the reason why such will make administration of the CPM
easier during the course of the project and with at least a sample of how
one might be traditionally portrayed, even if such is not required by the
specification, can go a long way towards building a rapport with the engi-
neer. If an internal effort was made to verify durations, for example via
a calculated productivity rate by work type, this should be mentioned.
444 The Practice of CPM Scheduling

Any special abbreviations or idioms used by the staff of the contractor,

such as “conduit, cable and connections” or “pipe, pull and terminate,”
should be explained.
Next, the narrative should talk out the critical path and various near-
critical paths. If a critical or near-critical path begins at other than the
start of the project, such at a constraint, highlight this and explain why.
Note the anticipated number of concurrent operations and resources to
be fielded as the project builds momentum through to punchlist. Make
a special effort to highlight any actions by the owner (or the engineer)
or third parties that are reasonably anticipated and the importance to
the schedule of such performance.
A separate narrative to upper management of the contractor may
highlight the usage of scarce resources of the company (such as cranes)
and any special support needs that the schedule assumes. A narrative
to a department of the owner not involved in construction, be it opera-
tions, public relations, and so forth, should highlight what impact the
construction will have on such “third parties.”

25.17. Summary
Going from the step of data acquisition for a logic network to the pub-
lication of the contractor’s schedule involves a number of steps, from data
entry to validation of the plan to validation of the schedule to prepara-
tion of reports tailored to the intended audience. An important point to
keep in mind is that the quality of the product must be kept in the fore-
front and not be subject to dilution by the desires of the project man-
ager, the owner’s engineer, or third parties.
 Chapter

26
Engineer’s Review of the
Submitted Initial CPM

The law states that having power and authority imply an equal obliga-
tion of proper use. The specification gives the engineer the right and duty
to review and approve or reject the CPM submittal. An improper accept-
ance or rejection of a proper submittal for an initial CPM or update or
revision, may place the owner at risk of extras and may even negate the
obligation of the contractor to complete the project on time.
The foregoing statement is not meant to be limited solely to submit-
tals of a CPM, but merely reiterates the general rule of liability when-
ever one party to a contract is required to provide a shop drawing or
other submittal to the other. If an engineer were to reject a shop draw-
ing for a beam because “it looks too thin” but be unable to fault the cal-
culations to show that it meets both the specification and industry code,
there is little doubt that the loss of time and cost caused may be recov-
ered by the contractor from the owner. This is not to fault the engi-
neer’s skepticism; the “feeling” that it is “too thin” should trigger an
additional and perhaps intensive investigation of the matter by the
engineer. However, acceptance or rejection must be based upon the
results of such investigation and not mere “gut” feelings.
In the case of a CPM, the submittal is merely another shop drawing;
nothing more nor less. The submission of shop drawings are often detailed
in a contract specification in section 01300 organized in the CSI format.
Section 01305 details the provision of a log of shop drawings. Section
01310 details the CPM, the contractor’s proposed sequence of construction.
The purpose of any required submittal should be to provide additional
assurance that the contractor can comply with the substantive (or phys-
ical work) obligations of the plans and specifications, to provide early

445

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
446 The Practice of CPM Scheduling

notice of the contractor’s interpretation of the plans and specifications, and

to provide the engineer with additional means of monitoring compliance
with the plans and specifications. Implicit in the planning stage of prepar-
ing the CPM is the need for the contractor to take a second look at the plans
and specifications, this time, by the team that will be performing the work
and separate from that previously performed to bid the job, to ascertain
“exactly how are we going to build this thing.” Requiring the contractor to
take such a “second look” has several advantages for the engineer.
The first two of these advantages may be of more assistance to the con-
tractor than to the engineer or the client, but the contractor’s success is of
benefit to the owner also. The detailed “second look” will often cause the
discovery of errors in the contractor’s bid. Since the method of “take off ” is
by activity (discrete instruction sets to foremen) rather than by quantity,
the effect is much like the practice of checking a land survey by transiting
back on a different path from that initially taken. If errors are discovered
(and every bid has some errors), the contractor has the maximum amount
of time to mitigate the impact thereof. The “take off ” by activity also pro-
vides a constructability review that will hopefully detect any issues of
“it don’t fit” that may be determined from such a detailed review of the
drawings. This determination of errors of the plans and specifications (and
every set of bid documents have some errors) provides the engineer the max-
imum amount of time to mitigate the impact for the client.
The submitted CPM will provide the “plan of execution” established
by the contractor at the time of the bid (updated perhaps by correction
of bid errors noted) and nails down this roadmap to its intended means
and methods to be used on the project. This should include the con-
tractor’s proposed allocation of resources. If the engineer later requires
changes, the contractor cannot say that they prevented performance as
planned, unless the initial CPM submittal shows such a plan.
Perhaps the most important of the three reasons is eliciting the con-
tractor’s interpretation of the contract plans and specifications. This
requires the contractor to “tell me what you think you heard me say”
before going off to perform such instructions. While the contract docu-
ments are not usually overridden by the accidental acceptance of a non-
conforming submittal (although many engineers try to protect their
clients and themselves from this spurious threat by inserting generally
ineffective exculpatory clauses in the contract, specification, and upon
their “accepted but not approved” stamp), should a court be convinced
that the contract plans and specifications contain an ambiguity that may
be reasonably interpreted by the contractor differently than what was
intended by the engineer, the contractor’s interpretation may prevail.
The submitted CPM will also provide to the engineer what the con-
tractor believes are the obligations of other parties outside of his control.
These include those obligations alleged to be that of the engineer, the
 Engineer’s Review of the Submitted Initial CPM 447

client, third parties that are within the control of that party (or for whom
such party accepts responsibility), and third parties outside the control
of the contractor, the engineer or the owner. If the contract documents
clearly state otherwise, such characterization should not modify the con-
tract. If the contract is vague or ambiguous, the failure to reply to such
notice of interpretation may be taken for acquiescence. Assuming that all
parties agree on the scope of and responsibility for such obligations, the
CPM allows the engineer to see how the contractor proposes to coordinate
the efforts of these multiple parties for the mutual benefit of the project.

26.1. Legal Aspects of a Review

The flip side of having the authority to review a submittal is the obliga-
tion to do so properly. Obviously, the owner has a contractual right to a
proper review and in certain circumstances, if the review falls below an
applicable standard of care, may have rights against the engineer notwith-
standing contractual limits of liability. Similarly, although the contractor
would not normally have a contract with the engineer, he/she may have
pass-through rights (via the owner) to a proper review. Depending upon
the state or other jurisdiction, this right may supersede contract language
excusing or indemnifying the engineer. It is but a small step to suggest-
ing that subcontractors may have pass-through contractual rights to a
proper review and from there, to noting that third parties (utilities and
other locally impacted individuals) may have tort rights to a proper review
(including provision for consequential and punitive damages).
Thus, the authority to review the CPM submittal does not empower
the owner or engineer but rather creates an additional obligation to
assist the contractor to perform the scope of work of the contract. It is
important to remember that the acceptance or rejection of the CPM by
the engineer does not change the viability of a schedule. If the submis-
sion is accepted in error, this does not create an automatic “gotcha” for
the contractor, but will merely acquiesce to the contractor’s reasonable
interpretation of claimed ambiguities. However, if a submission is
rejected in error, the contractor may later “submit” to a real judge, with
real consequences! These may include a change order to the contractor
for additional CPM preparation, a change order for changed “means
and methods,” and a loss of credibility of engineer as objective arbiter.
What are the limits for which “acceptance” should be made? An accept-
ance should cover the following.

1. The format, as per specification, e.g., maximum durations, limits to

logic types
2. Owner-controlled activities, including between multiple prime con-
tractors
448 The Practice of CPM Scheduling

The engineer’s acceptance should reserve rights relating to the following.

1. The contractor’s “means and methods”

2. Requiring reasonable additional resources for any change order extra
work
3. Requiring reasonable additional resources for recovery from any delay
4. Inclusion of constraints that may unduly reduce the owner’s or engi-
neer’s float in responding to submittals, RFIs, or changed conditions
(such as to field measurement of renovation work)

When may rejection come back to “bite” the engineer? If the engineer
rejects a format that technically meets the specification, such as
demanding code fields or graphics not in the specification, the contrac-
tor may claim a change order as well as loss of efficiency due to lack of
an approved CPM. If the engineer rejects a duration, specified to be con-
tractor’s best guess, the contractor may claim a change order. However,
the engineer may require supporting information for the duration
and demand a general consistency between durations, resources, and
costs.
If the engineer rejects inclusion of or demands addition of a logic
restraint, there may again be a question of dictating the contractor’s
“means and methods.” However, the engineer may demand a written
explanation of questionable physical logic and may reserve rights relat-
ing to a contractor’s attempt to level or smooth resources. On the other
hand, the engineer may and should demand exclusion of “place marker”
logic that cannot be adequately explained.
The important issue is that the engineer may not reject calculated
output while neither finding fault with the input nor the calculation algo-
rithm. The sole questions for the engineer when reviewing the CPM sub-
mittal are whether it meets the requirements of the contract specification,
as construed against the party that wrote the specification and as may
be modified by the exercise of reasonable engineering discretion.

26.2. Reviewing the Plan

The seminal question in reviewing the CPM plan is whether what is
submitted is a CPM or merely a bar chart by CPM software. To review
the plan, first we must look at the activity descriptions. Activity descrip-
tions must be for a proper level of detail, not too broad (e.g., ALL
EQUIPMENT or ALL ELECTRICAL) and not too narrow (e.g., FLIP
SWITCH). Each description should be unique and used only once. And,
each description should be understood without additional references
(Figure 26.2.1).
 Engineer’s Review of the Submitted Initial CPM 449

Figure 26.2.1 Activity descriptions must not be too broad or too narrow.

26.3. Technical Review

In addition to the educated review of the initial CPM submittal by a
seasoned engineer, a technical review may be performed by a perhaps
less seasoned engineer trained in the theory of CPM and familiar with
one or more of the software packages used for turning the plan into a
schedule. It is suggested that since the “testing” of the CPM may be
“destructive,” a copy of the electronic file or files submitted be saved and
renamed prior to the review.

26.4. Reschedule and Review the Diagnostic Report

Some software programs recalculate the schedule each time a change
is made to the input. The down side of this practice is that calculation
takes a perceptible time and can slow data entry significantly. Other pro-
grams will allow the user to enter data and recalculate the impact of
such changes only upon command (Figure 26.4.1). P3 allows the user to
change both input data (such as durations) and output data (such as
overriding a calculated early start date). Of course, upon clicking the
Schedule button, the output fields will be corrected back to the calcu-
lated date. However, if the submitted CPM report is printed after adjust-
ments and before rescheduling, the output can say anything desired by
the contractor and have no correlation with the logic network.
As a result, it is incumbent upon the engineer to have a copy of the
software used by the contractor and to not rely solely upon the printed
reports submitted by the contractor. After assuring that the file name
will not conflict with other files on its computer, the engineer should copy
the files provided by the contractor and reschedule the project from the
data date. (Initially, this is the notice to proceed.)
450 The Practice of CPM Scheduling

Figure 26.4.1 Choice of algorithm.

Section 1: honesty. The myriad of special features in modern software

programs for overriding the traditional calculations of CPM, combined
with the ability with many systems to directly access and manipulate
the database where the output from these programs is stored, and the
ability to “cut and paste” to official-looking reports, means that no matter
how carefully the submittal is reviewed, there is the possibility of skull-
duggery. Thus, the first line of defense may be to know the reputation
of whoever has prepared the reports that you will be reviewing. If the
specification requires the submission to include a copy of Primavera
P3’s diagnostic report, a quick glance at the top notifies you of the serial
number and licensed owner of the software being used. If this informa-
tion is missing, and is not supplied if subsequently requested, what
other corners is the contractor cutting?
Similarly, as the engineer reschedules the CPM to be assured that the
output is related to the input, this report should be printed, and may
be used to illustrate corrections requested of the contractor to its sub-
mittal. What type of statement will the engineer make if it does not dis-
play a valid license in the analysis?

Section 2: constraints. The use of a constraint in most cases is a short-

cut for the depiction of the proper logic restraints required to plan or
schedule an activity and should be viewed with skepticism. However, in
many cases, the use of a constraint is beneficial in that it will make the
use of the schedule easier, especially during updating. In some special
instances, a constraint may even be used that could not be duplicated
by the use of restraints.
 Engineer’s Review of the Submitted Initial CPM 451

Section 3: open ends. In most cases, a proper logic network should have
only one start and one end. In the event that the contract contains inter-
mediate milestones involving tangible and discrete scopes of work, there
may be more than one ending activity. Similarly, if the contract provides
that access to a portion of the site will be delayed, there may be more than
one NTP or other starting activity. Other open ends not required by the
engineer’s specification should be viewed with suspicion and should be
a reason for rejection if not properly and adequately explained.
However, it is important that the engineer go beyond the printouts, pure
logic diagram, or diagnostic in checking for open ends. The Scheduler, for
the engineer must either perform a manual review of each activity con-
nected to others by non-traditional relationships or perform some auto-
mated form of review using a software tools as discussed in Chapter 25
to check for these types of errors.

Section 4: progress and actual dates. The initial schedule should not
indicate progress and no activity should be assigned an actual date.
Either transgression is a cause for rejection of the submitted CPM.

Section 5: choice of algorithm. The choices of algorithm affecting the

initial schedule are limited and, in any case, the choice of the contrac-
tor, unless otherwise provided in the specification. Although the engi-
neer may desire the choice of interruptible schedule duration, show
open ends as critical, and calculate total float as most critical, the choice
of contiguous schedule durations and printing of a start float or finish
float do convey information that may be useful to the contractor. There
should not be extra open ends in any case (Figure 26.4.1).
Similarly, the choice of retained logic or progress override should be
addressed in the specification, but if not, it is important to note that each
is an exaggeration of the situation when work is performed out-of-
sequence. The choice of whether to calculate start-to-start lag from the
actual start or early start may have more to do with the contractor’s use
of SureTrak in the field trailer than a conscious choice of which method
makes the task of updating the schedule easier.
If the contractor and engineer are using P3e/c or Primavera
Construction, it is important that the specification require careful coor-
dination before including relationships to other projects. Similarly, if the
specification does not mandate the calendar convention to be used for
lag durations between activities, the engineer will need to use care
when reviewing contractor submissions.

Section 6: statistics. The number of activities in the CPM should bear

some relationship to the size and complexity of the project. One “rule of
thumb” is a minimum of one activity for each $10,000 of contract value.
452 The Practice of CPM Scheduling

Another more complex rule might factor out the value of purchased
plant equipment and other large capital expenditures, determine the
average hourly cost of labor, average crew size, and average activity
duration somewhere between the minimum and maximum durations to
be used, and calculate a suggested minimum number of activities.
Number of activities.................. 873
Number of activities in longest path.. 12
Started activities.................... 843
Completed activities.................. 822
Number of relationships............... 1236
Percent complete...................... 97.6
Number of expected finish activities... 2
Number of late constraints............ 1

A healthy network, on the average, should have approximately 1.6

relationships per activity. If the number of critical activities (on the
longest path) is near or exceeds 25 percent, additional scrutiny may be
called for, but the number of near-critical activities (having perhaps 5
to 10 days of float) should also be reviewed. Obviously, these are all
guidelines and require the use of appropriate engineering discretion
rather than be cited as hard and fast rules.
For the initial submittal, there should be no progress: nothing started,
nothing finished, and zero percent complete. The concern over the
misuse of constraints cannot be over emphasized; each one must be
checked as to why it is there and if it is truly necessary.

Section 7: dates. Obviously, the calculated completion date must be

reviewed. Keep in mind that due to the limitations of the mathematics,
exacerbated in complex networks due to the influence of merge bias, the
probability that the project will actually finish by the calculated com-
pletion date is well under 100 percent. The larger the durations of the
activities in the network, and the more complex the network, the lower
the probability that the project will complete by this date. Thus, based
upon the level of detail of the network, the engineer must “feel” whether
an appropriate contingency has been included, or subject the CPM to fur-
ther analysis by a Monte Carlo or PertMaster type of software.

26.5. But Is the Logic Realistic? The Smell Test

But is the logic realistic? The engineer should walk through the critical
and near-critical paths of the logic network without looking at the calcu-
lated dates to see if the logic makes sense. Does each activity have a
predecessor reflecting a physical dependency? Do resource-based rela-
tionships tend to assist or confuse the logic? Do logic connections exist that
appear to have no rationale? Do the durations used appear reasonable,
 Engineer’s Review of the Submitted Initial CPM 453

and if not, do they at least appear to be internally consistent? If ques-

tions arise at this level of the review, it may be best to discuss them with
the contractor rather than blindly issuing a rejection. Once the logic and
durations appear reasonable, then, and only then, the engineer should
walk through the bar chart printout of these same activities to see if the
timing makes sense. If something seems out of order, a more intensive
review may be called for, possibly including an inquiry to the contrac-
tor. Of course, this mere intuition of the engineer cannot be a valid basis
for rejection of a submittal.

26.6. Project Calendar or Calendars

Because the project calendar or calendars are not easily viewed (none of
the popular software products providing a proper screen view of the
entire project calendar), it is important for the engineer to review these
carefully. If multiple calendars are used, the need is much greater since
few casual users of the software truly understand how these operate.
Unfortunately, the only means of review is to open the software files and
drill down through several menus to the non-work days for the global cal-
endar and then to the basis, non-work days, and exceptions to the global
non-work days of each calendar.
Understanding the interplay of global and individual calendars is
important, as are the rules of the software product for observing holi-
days that fall on weekends and non-work days that fall on holidays. Also
important is to guard against adding contingency to contingency, either
through the calendars or activity durations already padded for antici-
pated weather conditions. Finally, if multiple calendars are used, it is
important to educate all who may read the calculated float attribute of
its modified meaning under these circumstances.

26.7. Summary
The engineer cannot and should not attempt to verify that the contrac-
tor can perform the contract work according to the “plan of execution”
provided by the initial CPM submittal. However, the engineer can and
should verify that the submittal is technically correct and that the logic
and durations of the submittal appear “reasonable.” Finally, the engi-
neer should “walk through the CPM” and determine if everything
“smells right,” possibly leading to additional scrutiny. But the review of
the initial submittal of the CPM must be handled in as professional a
manner as any other submittal to the engineer.
This page intentionally left blank
 Chapter

27
Updating the Schedule

How do you stay on a CPM schedule? In early CPM applications, the

CPM network was left to its own devices once the project was in progress.
The planners, confident that they had planned the project more care-
fully than ever before, did not follow up on their careful efforts. The
result was similar to buying an automobile and then letting it break
down because the oil was not changed.
In the sections following, updating the schedule by the contactor team
is described. Traditionally, the project manager or superintendent would
meet with all foremen late each Friday afternoon or early each Monday
morning or at some other convenient time to tweak the short term
schedule based upon performance of the past week and observations of
other conditions that may affect the plan for the coming week. Since the
use of the CPM is meant to support the traditional methods, rather
than merely adding another chore for the project team, a weekly review
or update of the CPM is recommended. Unfortunately, this is not typi-
cally done as many specifications are geared to monthly reporting to
meet the needs of upper management rather than the project, and many
contractors provide only what is required rather than what is in the
interests of their own project teams.
Another reason why specification often requires only monthly
updates is that despite the mantra that “On a project, time is king,”
the reality is that issues of schedule are often an afterthought after
issues of money. (Rarely will a Proposed Change Order from the owner
start with “How might this impact the schedule – and by the way
how much will this cost?”) Thus, usually on a monthly basis, a mutu-
ally agreed upon (between CM/Engineer Team and the Contractor
Team) progress update is performed and is the basis for the monthly
progress payment.

455

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
456 The Practice of CPM Scheduling

27.1. Why Update the Schedule?

How do you determine the status of the schedule? The CPM network can
be used as the basis for monitoring project progress. On the job, the net-
work can be posted on an office wall and progress marked right on it.
The ever-popular colored marker pens are the best way. The field usu-
ally cooperates because plotting progress seems to strike a responsive
chord in most of us. Not only does it result in a current status report for
the project, but the process of keeping the network up-to-date famil-
iarizes the field office with the logic diagram.

Figure 27.1.1 Plot of progress: project day 150.

 Updating the Schedule 457

TABLE 27.1.1 Activities in Progress Having Float

Time
remaining New float Original
i–j Description (days) (days) float, (days)

42–44 Drywall 5 7 13
37–46 Heating and ventilation units 2 10 23
37–47 Fuel tank 2 45 70
41–47 Boiler check 4 43 43
40–47 Test piping 5 42 33

An electrical contractor raised a very practical problem in regard to

posting the diagram on the job site. If craftspeople review the networks,
however casually, some will notice the time estimates for activities on
which they are to work. If the estimates are too long, there will be the
tendency for the crew to take too long on the activities. If the times are
too short, the CPM schedule will be seen as unrealistic. This problem,
if anything, bears out the need to use a frank and realistic approach to
CPM planning. The diagrams will have a big impact on field people, and
some excellent comments about scheduling have been known to origi-
nate from this level.
Figure 27.1.1 is the John Doe network with progress shown by dark
lines. The event times for the activities in progress are shown. The first
path to check is, of course, the critical path. This status is as of project
day 150; a check along the critical path shows that activity 43–49, install
branch conduit, has 10 days to go and is 1 day ahead of schedule. The
activities in progress with float have the status shown in Table 27.1.1
All the float items are within the allowable CPM range. The drywall
and heating and ventilating units should be pushed so that ductwork
installation can start. Figure 27.1.2 is another representation of the
same project status. This format could be used to submit a quick weekly
status report.

27.2. Acquiring the Data for an Update

In this section, updating is described from the Contractor’s point of
view, where the Contractor has the responsibility to perform the update.
However, especially when the update is part of the progress payment
process, the Owner’s Representative (CM, Engineer or both) will either
be part of the update process or will have approval of the process.
In either case, a representative of the contractor must physically
walk the job and eyeball the work in progress. Even in the case of non-
construction projects, this process of personal inspection, rather than
458 The Practice of CPM Scheduling

Figure 27.1.2 Quick status report: sample for day 150.

merely accepting the reporting of progress by line personnel, should be

attempted to the extent possible. Stopping by each foreman to elicit
what has been performed in the past update period (and verified by
eyeball) also has the advantage of bringing any schedule (or support)
concerns of the foremen to the attention of the Scheduler and thus to
the project manager.
As previously noted, this representative is usually not the project
manager or superintendent who was involved in preparing the initial
CPM since the update should ideally be prepared on a regular periodic
 Updating the Schedule 459

basis and the project manager may have other pressing business on the
specific day that this walkthrough should occur. Thus the extra care
taken in preparing the initial CPM with the level of detail, descriptions
and durations geared for easy updating will be rewarded at this time.
Note again that the foremen and others will be much more able to pro-
vide the information needed if collected weekly rather than monthly and
that the time wasted in looking up dates of work performed weeks ago
may exceed the extra effort of weekly walkthroughs.
For those activities that are not amenable to visual review, such as
the progress of procurement, an equivalent process of placing a phone
call or email to each vendor asking about progress or problems of the
past week is in order. Even if the procurement item is in a queue and
awaiting its turn (such as often the case with special fabrication of
steel,) the phone call can confirm that a special order (perhaps with mil-
itary priority) has not bumped this procurement item within the last
update period.

27.3. Distinguishing Updates from Revisions

One major purpose of an update is to gauge the progress made during
the past time period against that anticipated or estimated in the initial
schedule. Based upon the information gained by this exercise, the proj-
ect team may choose to make mid-course corrections to the plan for the
work remaining in the future. But it is important that all parties see this
as a process and not to cover up past poor performance with unrealis-
tic changes to future performance. For this reason, it is important to dis-
tinguish between an update, measuring past performance, from a
revision, a refining of the plan for future action.
An update should measure and record only those facts that have
occurred. These will include whether a specified activity has started
during the past reporting period, and if started whether it has been
completed or is still in-progress as of the date that the update informa-
tion is collected.

27.4. Purpose of an Update

The purpose of an update is to record past performance and to determine
the impact of past performance upon expectations for the future. If we
could do five activities last week, and we did do three of them, how will
this impact our projections for timely completion of the project and for
what activities may be performed in the coming weeks? An important
aspect of the update is to gauge whether progress on the project is meet-
ing expectations, exceeds expectations, or is slipping behind.
460 The Practice of CPM Scheduling

27.5. The Purpose of a Revision

The purpose of a revision is to insert deliberate changes to the existing
“plan of execution,” either to account for changes to the project or to
improve upon the previously “means and methods” chosen. Thus, a revi-
sion should only consider that which may be performed in the future.
Changes in the project may include additional or modified scope, changes
in anticipated conditions, or correction of errors previously made in
interpreting the project plans and specifications. A revision to the logic
is neither required nor desired for a casual deviation from the previously
stated “plan of execution” so long as the contractor’s general intentions
are to continue with that plan. In a majority of cases where a contrac-
tor deviates from a previously published plan, the decision to do so is
involuntary and represents a course of conduct that is more costly than
the original plan.

27.6. Who Should Collect Data for an Update

If the initial CPM is properly prepared, utilizing information from the
people most knowledgeable about the project but organized by a
Scheduler, personnel of lesser experience may collect the data necessary
to update the schedule. As has been noted previously in this text, it is
this skill that distinguishes a Scheduler from a “screen jockey” who
merely runs the software. If the definition of tasks is clearly under-
standable and the junior person collecting data is given some general
guidelines, any errors in data collection will be less than the window of
tolerance surrounding the original duration provided.

27.7. Who Should Prepare Data for a Revision

On the other hand, the level of attention given to a revision requires the
original team (or current personnel fulfilling the functions of the origi-
nal team) to adequately plan and then schedule the remainder of work
to be performed. Even as a small change requires that all subcontrac-
tors sign off on the cost impact thereof, the time impact must also sim-
ilarly be addressed. Let us assume a small change involving minor
relocating and resizing of an HVAC duct. If the prime contractor receives
the cost estimate from the HVAC subcontractor and simply adds its
OH&P (overhead and profit) markups before forwarding to the owner,
the prime contractor may be in for swallowing some serious collateral
costs. It is important that other trades that will have to relocate their
systems around the new configuration also provide cost information. It
is also important that the coordination of these multiple subcontractors
 Updating the Schedule 461

be revisited in the CPM plan, a revision plan be prepared, a schedule

calculated, and the impact in terms of time and incidental or disruption
costs (if any) be noted.

27.8. Information Required for Schedule Control: AS, RD, AF

The information required for an update of the schedule includes only:

■ a determination if the activity has started, and date started if appli-

cable,
■ a remaining duration in work days (comparable to that of the origi-
nal duration) if started but not yet finished, and
■ a determination if the activity has finished, and date finished if appli-
cable.

27.9. Determination of Actual Start and Actual Finish Dates

In determining whether the activity has started, it is well to look beyond
the 24–36–48–64 characters of the description and establish whether the
predecessors are finished and whether the resources assigned to this
activity are in play. Similarly, in determining whether the activity is fin-
ished, it is important to note whether the activities successors have
started and whether the primary resources assigned to this activity
have moved on to other work.

27.10. Determination of Remaining Duration of Activities:

Repeat the Steps of the Master
If the activity has started but has not yet finished, it is necessary to inde-
pendently determine the remaining duration. The determination will
obviously be with knowledge of the original estimate of duration, but to
the extent possible should “repeat the steps of the master” rather than
blindly accepting the original duration. Thus, if visual inspection of a
10-day activity appears “50 percent complete,” a starting point for
remaining duration may be 5 days, but also may be more or less depend-
ing upon how actual productivity compares to that estimated.
A quick review of the assumptions surrounding the original estimate
of duration will include confirming the physical scope of the activity from
the 48-character description and referenced contract drawings, review-
ing the various other codes assigned to the activity, reviewing the
resources assigned to the activity, and reviewing the estimated pro-
ductivity compared to that experienced to date.
462 The Practice of CPM Scheduling

If all the assumptions appear unchanged, the visual percent com-

plete may be a good approximation of remaining duration. If, however,
the reality requires reassessment of the assumptions behind the origi-
nal duration, a new estimate of duration may be required for the remain-
ing scope. Keeping in mind that this new estimate is likely to be
performed by an individual with less experience than the one who pre-
pared for the initial schedule, the error of the estimate of remaining
duration may be larger than that of the original duration. However, the
larger percent of error will be for a smaller number.
It was for this reason that a guideline was set for maximum activity
durations. Notwithstanding the safe harbor thus created, it is still good
practice for the junior scheduler to estimate a remaining duration rather
than a percent complete for most activities. The opposite is true for
those (hopefully) few activities of a bulk or repetitive nature involving
a known large quantity of work, a known productivity, and a resultant
large duration, such as a large bulk excavation. Here, the junior sched-
uler should enter the percent complete calculated from the measured
quantity complete. The junior scheduler, however, should be aware if the
productivity realized differs significantly from that assumed in setting
the original large duration, and call for assistance in reevaluating the
remaining duration by the project manager or appropriate area super-
intendent.

27.11. Expected Completion and Renewing Promises

Similarly, for those activities that are subject to an expected finish con-
straint and where remaining duration is automatically calculated, a
healthy level of skepticism is called for. It is not uncommon that ven-
dors providing products with long lead durations include a “no damage
for delay” clause in their contracts along with their salesperson’s per-
sonal promise of timely delivery. As part of the process of an update, the
Scheduler should call that salesperson, or other responsible party, and
inquire if everything is still “on schedule” or if any “glitches” have
occurred or are anticipated. In the event that there are any problems,
the project manager then has the opportunity to get involved in rectifi-
cation of the issue, or to explore other avenues of mitigating the impact
of a late delivery.
This continual reminder to the vendor of the contractor’s reliance upon
the promise can, depending upon State law, supersede the “no damage
for delay” language if it is found that vendor deliberately, or even neg-
ligently, misrepresented the current status of an open order. For mis-
sion critical deliveries, the contractor may choose to avail itself of the
Uniform Commercial Code UCC 2-609 Right to Adequate Assurance of
 Updating the Schedule 463

Performance by demanding such status updates. This will not protect

the contractor from “last minute glitches” at the vendor, but will cer-
tainly reinforce the importance to the vendor of this particular delivery.

27.12. Automatic Updates

Several of the CPM software vendors tout a feature that will automat-
ically update the schedule. Phrased simply, the user enters the new
data date and the software converts prior early start dates to actual
starts, early finish dates to actual finishes, and calculates remaining
durations for those activities with performances that straddle the data
date. The help screen for this feature then suggests that the user go back
and manually correct the “few” instances where reality did not perfectly
mimic the early start schedule.
In the years that the authors have been involved in scheduling, we
have rarely, if ever, seen projects where even one quarter of the activi-
ties actually start and finish exactly on their baseline early dates. More
importantly, this process negates the main purpose of an update, that
is, to carefully and methodically check the status of each activity that
may or has started. It is the rigor of this process that makes the update
a tool for the contractor’s ability to best perform rather than merely a
tool for printing a report read by none.

27.13. The Forgotten Step:

Determination of Remaining Duration Between Activities
Many contractors believe that the advent of PDM made preparing the ini-
tial CPM logic network and resultant schedule easier. However, the added
but sometimes hidden power of PDM can make updating the schedule
much more difficult. The reason for this is that while the statusing of
the remaining duration of activities is visible and understood by the con-
tractor, the statusing of remaining durations between activities is usu-
ally left to a default, which is both hidden and not clearly understood.
The first area of misunderstanding is to determine what the lag meant
in the initial network. Although this has been addressed earlier in this
text, the difference between various interpretations was largely aca-
demic for purposes of the initial schedule. The bars land in the right spot
regardless of meaning. Now that the contractor is using the CPM to
manage the project, the definitions used are more important.
In again reviewing the start-to-start relationship in Figure 27.13.1,
it is important to note whether the 7-day lag should measure number
of days from the reported actual start or measure when the original
duration is reduced to a remaining duration of 3 days. In a perfect
464 The Practice of CPM Scheduling

Figure 27.13.1 Problems in Updating PDM durations between activities.

world these two measurements may be the same, but such is rarely the
case in the real world. Compounding this issue is that most popular
software offerings do not support the second choice, even though it is
the one more commonly meant by the project manager providing the
original information. Therefore, it is now required that the Scheduler
tailor the input for each update to work around the inaccuracies caused
by this problem.
The general consensus on reporting an actual start date may be that
this is merely historical datum, and can be left blank or “fudged” and
corrected later. However, as noted previously, the actual date is now part
of the calculation for reducing the lag. If an activity is reported started
by entry of a percent complete or a remaining duration but an actual
start date is not entered, the lag duration will not be reduced. If the
activity is reported as 99 percent complete but an actual start date is
not entered, the lag duration will not be reduced. In some software
systems, if the activity is reported as 100 percent complete with zero
remaining duration but an actual start date is not entered, the lag will
not be reduced. Other software systems are written to catch this obvi-
ous “error” but then treat 99 percent and 100 percent complete activi-
ties in entirely different manners.
P3e/c addresses this problem by locking the entry of a remaining
duration or percent complete until after the user has entered a start
date. In the real world, this is more likely to elicit entry of a bogus date
 Updating the Schedule 465

Figure 27.13.2 P3e/c does not allow user to status a reduced remaining duration until an
actual start date is entered.

(which is then not highlighted as “to be correctly entered later”) than

stopping the update process to research a correct date (Figure 27.13.2).
Thus, to truly update a PDM logic network, it is necessary to print
out or view each relationship where a lag is present and to manually
review and update such lag duration. Here again, the orphaned status
of “durations between activities” is shown as popular software sys-
tems do record and distinguish between original and remaining dura-
tion of activities, but can record only one duration between activities
(Figure 27.13.3).

27.14. Save and Rename: Naming Strategies

Update information should never be applied directly to the file for the
initial project or prior update. In each case, it is important to copy and
rename the file, updating only the new file. And it may be important, if
litigation results from a project, to save each and every one of these files.
Every software product has its problems with this process. P3 is limited
to only four characters for the name of the file, effectively limiting the
number of updates based upon the how many of these four characters
are to be reserved for noting the project, and how many to distinguish
466 The Practice of CPM Scheduling

Activity A is followed by Activity B via a Start-to-Start relationship with 15 day lag (all days are calendar days)

Activity A is statused to RD=10 on 06MAY06 (ahead of schedule) but actual start date is not entered

Activity B is still scheduled for 15 days after DataDate 06MAY06

But if actual start of 01MAY06 is entered for Activity A, early start of 16MAY06 is calculated for Activity B

If Activity A is statused at 95% complete without a reported actual start,

early start of Activity B is calculated as DataDate plus 15 days

If Activity A is statused at 95% complete with a reported actual start,

early start of Activity B is calculated as actual start plus 15 days

Figure 27.13.3 Problems with reporting status for durations of activities and durations
between activities.

which update or revision to that project. For example, JDOE may rep-
resent the John Doe project. If we know we are going to have less than
ten updates, perhaps we can name updates JDO1, JDO2, etc. If we expect
more than ten but less than 100 updates (not uncommon for multiyear
 Updating the Schedule 467

If Activity A is statused 100% complete without an actual start of finish date, early start of 21MAY06 is calculated for Activity B

If Activity A is statused 100% complete with an actual start but not an actual finish, P3 illustrates a bar finishing the original
estimated finish date but does not record that date on tabular format, and continues to calculate Activity B lag from 01MAY06

Even if Activity A is completed three weeks early, and all actual dates reported, the early start of Activity B is still calculated as
01MAY06 plus 15 days. Is this what the superintendent meant when saying “Activity A is expected to take 20 days, Activity B
can start when Activity A is done”?

Updating without an actual start date does not reduce the remaining lag

Now that an actual start, but not finish, has been reported, updating reduces the remaining lag

Continued updating with an actual start but without actual finish eventually allows next activity to start

Figure 27.13.3 (Continued).

projects with weekly or biweekly updates), we can name update files

JD01, JD02, etc. If the John Doe project experiences a major change
order or other mid-course correction, we can remember that JD34 is a
revision and not an update, or name the revision JE34, effectively remov-
ing the “JE” prefix for other projects in the same subdirectory.
However, in P3, individual projects may be stored in individual sub-
directories. Thus naming conventions such as JDOE (for the initial
schedule,), U721 (for the update of 2007, February, first update of
month,), R752 (for revision of 2007, May, after the second update of the
month,), U7d2 (for update of 2007, December, second of month), etc.
may be used. Care must be exercised to not mix up the several files
named U721 for different projects, but these will be stored in separate
subdirectories.
468 The Practice of CPM Scheduling

P3e/c and OpenPlan now require all projects to be placed in one master
database file selected by the software. The limitation by Primavera of
only four characters to name a project has been lifted; however, the
maximum number of schedules (including updates) assigned to one
project is set at 10 by default and 50 by user preferences. Thus, although
“older” updates can be downloaded and archived before deletion to make
room for “newer” updates, possibly later to be restored after the “newer”
updates are similarly archived, comparison of update to update for a
complex and lengthy project becomes much more difficult.

27.15. Reports and Views to Assist Acquiring Data for an Update

The foreman’s turnaround document, suggested in Chapter 25, can be
used not only to plan work for the coming weeks but to record past per-
formance for the next update, as shown in Figure 27.15.1. First, the
foreman used a sharpie pen to highlight that work would be performed
on 3306 and 3308 in the upcoming 2 week period. By contrast, 6144
was slated to be deferred until at least mid September. Then, as the
end of the update period approached, the foreman noted that 3306
was started and finished on 8/10, while 3308 took from 8/11 to 8/13.
There is no need to enter duplicate information – if finished, the
remaining duration will be zero and percent complete will be 100%. But
the inspector did not like the chipped inlet to the U-drain and with-
held 20% of the cost associated with this item. Notwithstanding, paving

8/10 8/10
8/11 80% 8/13
chipped inlet – hold 20%

Figure 27.15.1 Input for update of 15AUG05.

 Updating the Schedule 469

can continue so it is 100% complete for schedule purposes. This and

other turnaround forms from other foremen may be collected, faxed to
the home office of the contractor and utilized to prepare the next turn-
around document.

27.16. Electronic Tools to Assist Acquiring Data for an Update

Third-party hardware and software is also available to allow updat-
ing by project control walking the project, including the ability to
download the current schedule to a Palm Pilot or other handheld device
for updating on the go. The updated information can then be directly
uploaded to the scheduling software without the need for rekeying of
information.

27.17. Choosing the Correct Algorithm for Updates

At this point, the choices in algorithm made back in Chapter 17 between
retained logic and progress override and between calculate start-to-
start lag from actual start or early start will become important. The first
choice relates to the type of logic used to schedule activities that are in
progress. If choosing retained logic, the remaining duration of a pro-
gressed activity is not scheduled until all predecessors are complete. If
choosing progress override, network logic is ignored and the activity can
progress without delay. If choosing the new option provided in P3e/c and
Primavera Construction, actual dates, backward and forward passes are
scheduled using the users guesses as to future “actual” dates. Note that
the actual dates option can cause negative total float if the schedule has
work being performed out-of-sequence and “actual” dates before the
data date.
Progress override may show better progress, but may also provide a
false sense of security. Retained logic may cause an uninformed resident
engineer to become nervous, but the authors suggest honesty and edu-
cation. The use of actual start may make data entry easier for the
Scheduler since there will be less adjustments required to the lags of
activities started but not yet complete with this option.

27.18. Scheduling the Update: Interpreting the Results

At this point, it is time to run the update, always remembering to reset
the data date to the end of the past period and start of the new period.
A careful review of the calculated results can assist the entire project
team focus upon the real problems every project has, rather than the
ones that have generated the most yelling and screaming.
470 The Practice of CPM Scheduling

27.19. Technical Review: the Primavera Diagnostic Report

Once again, the first step of review is to go to the diagnostics. Third-party
diagnostics, such as Claims Digger, may be used to compare the new file
to the old file (before or after calculation) to list differences. P3e/c and
Primavera Construction have incorporated this third-party program.
Any changes between the files, down to correction of spelling errors in
an activity description, will be listed. Hopefully, the Scheduler for the
contractor will not run into any surprises and will merely verify that sub-
stantive changes include only the provision of actual start and finish
dates for activities completed, and actual start and remaining dura-
tions for activities started and still in progress.

Sections 1, 2, 3, and 5. The first three sections of the diagnostic report

should remain the same. If a consultant prepared the initial CPM but
updates are processed from the field, a different user name and license
number should be noted. There should be no additional constraints as
part of an update and there should be no additional open ends. During
updating of P3, it is possible to accidentally add a blank new activity.
If this has occurred, the open end report will note this issue. There
should be no changes to the algorithm chosen from the initial schedule
submission.

Section 4. Section 4 of the diagnostic is the most important section for

an update. This section will list some types of data entry errors, such
as the accidental entry of a future date as an actual, but most impor-
tantly it provides a look at whether the contractor is having difficulty
performing to its “plan of execution.” The work performed out-of-
sequence report is an indication of either reporting progress to the
wrong activities or a warning of disruption.
Chapter 36 describes the issue of disruption and how this report may
be augmented to record such issues and quantify the time and cost
impact of them. However, from the viewpoint of proactively managing
the project, the Scheduler should discuss with the project manager each
instance of such work performed out-of-sequence to determine if it
merely represents a temporary assignment of resources to work that is
available, or a more serious reassignment of resources because planned
work is being blocked by some factor.

Section 6. The statistics of number of activities started and completed

and the rough assessment of percent complete are useful for reports to
upper management, but provide little of the detail required by the proj-
ect manager. However, if the number of started activities begins to
 Updating the Schedule 471

markedly increase above those complete, this may be a sign of either

reporting or real trouble.

Section 7. Obviously the key date to watch is the calculated completion

date. If this improves or remains the same, all is well. If it begins to slip,
additional review is necessary to determine why and to correct such
before a small problem becomes a large problem.

27.20. What to Look for when Reviewing the Update

In construction heaven, the update will match the initially calculated
schedule. In such a venue, the use of the automatic update features of var-
ious software vendors may be useful. However, here on Earth, very little
goes according to plan. It is the job of the Scheduler to note where vari-
ances occur and to direct attention to these areas by the project manager.
The key focus must always be upon the critical and near-critical paths.
If there is any delay or disruption of work on the current or near-term
activities on these paths, all audiences should be notified of the need to
address such issues. If new critical or near-critical paths are being iden-
tified as float is expended on previously non-critical activities, this may
be a sign of minor problems now becoming major problems. One impor-
tant metric that should be measured is the number of critical and near-
critical activities compared to the number of activities not yet complete.
If this number is rising over several update periods, it may be a sign that
the available resources are being spread too thin.
As noted from the previous subsection, a rising number of activities
started but not finished may also indicate trouble. The listing of work
being performed out-of-sequence should also be closely monitored. If
items on the list continue to remain incomplete for more than one
update period, there may be a question of inaccurate reporting or of a
more serious issue. Remember that the definition of an activity in CPM
presumes that if a successor can start, that activity must be finished.
Ancillary work associated with the activity but not required for its suc-
cessors is included elsewhere in the logic network in the activity called
punchlist.
On the other hand, if the activity truly is not complete, but the suc-
cessor has been able to start through the use of falsework or other
means of mitigation, the disruption necessitating such should be docu-
mented for either a change order to the owner, a back charge to a sub-
contractor or vendor, or for review by the upper management of the
contractor. The bottom line is that if the plan was carefully prepared,
any work performed out-of-sequence is going to be at a higher cost than
anticipated.
472 The Practice of CPM Scheduling

27.21. Tailoring Update Output to the Chosen Audiences

As noted in Chapter 25, output should be tailored for each audience. The
field personnel will probably best be served by quickly being provided
with an updated turnaround document. The project manager will also
require a copy of the turnaround document, perhaps grouped by area
or location rather than by crew or subcontractor. The project manager
also should be provided with a report detailing issues of work being
performed out-of-sequence with a view toward both corrective action and
documentation for purposes of cost recovery.
One project manager client prefers a copy of the pure logic diagram
pasted across the walls of the field office, with work completed marked
over the logic with a sharpie pen. Primavera’s pure logic view (mis-
named its PERT view) does this by default with a single diagonal line
across the activity box for work started and in progress and an “X”
across the box when complete. Another project manager prefers a time
scaled logic diagram (limited to near-critical and other “high profile”
coded activities) to hang on his wall. Across the top, he places a thumb-
tack holding a string weighted by a small plumb-bob, the thumbtack
moving 1 week rightward each week. Activities are marked with a

Figure 27.21.1 End date projections by update.

 Updating the Schedule 473

sharpie pen as they are completed as a visual of those not complete to

the left of the line indicates falling behind schedule.
Another useful graphic report is the use of a rate chart, such as is
shown in Figure 27.21.1. In the illustration, the 45-degree line repre-
sents the track along which the monthly update falls. Relative to that
track, the end date projections of the updates are plotted. If the plot of
the end date is vertical or leans to the left, the project is on or ahead
of schedule; if it leans to the right, then little or no progress has been
made.
Occasionally a contractor may be faced with an engineer who demands
a submission of an update or revision differing from what the contrac-
tor believes is factual. The owner and owner’s representative, the engi-
neer, must be provided with what is required in the specification.
However, providing additional information (clearly marked FYI only)
relating to problems encountered may help to bring the engineer into
the process of mitigation and correction. If the owner, rather, chooses to
consider this adversarial, then the additional reports provide abundant
notice of the problem.
The maintenance of the schedule in two configurations (e.g., the
approved schedule and the contractor’s version of the schedule) at the
same time is recommended in such situations and is supported by sev-
eral court decisions. (See Appeal of Blackhawk Heating & Plumbing
76-1 BCA (CCH), E.C. Ernst Inc. vs. Koppers 530F.Supp 830 (1981) and
Titian Pacific Construction Corp. ASBCA 87-1.)

27.22. The Narrative Report for Each Audience

The narrative reports that should accompany tabular and graphical output
should also follow this pattern. The report to the owner or engineer should
note the current critical and near-critical paths and any changes or addi-
tions (Figure 27.22.1). The report of a stable percent of near-critical to total
remaining activities should suffice for discussing non-critical activities; but
if the percent is rising, a discussion of the reasons why and corrective
measures planned is in order. The work performed out-of-sequence diag-
nostic should be discussed specifically noting any disruptions caused by
factors outside of the contractor’s control and related associated costs.
The narrative for the contractor’s upper management should also
focus on other issues of disruption, such as problems caused by sub-
contractors or vendors or by downtime caused by equipment or person-
nel issues. The narrative for the owner’s clients (either the upper
management of the owner or third-party “clients” to the owner) should
focus upon the general trends and general types of problems that may
cause or have caused schedule slippage.
474 The Practice of CPM Scheduling

Figure 27.22.1 Narrative report for update.

27.23. Summary
The creation of a process of updating the schedule from the initial plan
was the initial impetus of the Kelley-Walker group at DuPont in devel-
oping CPM. It is still the major difference from static processes, such
as a Gantt chart. The effort and care by the top members of the project
team in preparing the initial CPM plan will be rewarded by allowing
lesser experienced personnel collect update data.
It is important to distinguish between an update and a revision. An
update will only add the date an activity actually started and finished,
or if not finished, will add a new assessment of the remaining duration
of an activity. An update will never modify logic nor allow changes to the
original duration of activities not yet started.
In updating activities connected to others by non-traditional rela-
tionships, it is important to manually review the lag durations between
such activities. Update information should never be applied to an exist-
ing file; the file should be saved and copied to another file to which the
update data may be applied.
Review of the update should focus upon the critical and near-critical
paths and upon variance from the initial “plan of execution,” as may be
determined by review of the work performed out-of-sequence diagnostic.
Tabular, graphic and narrative reports should be prepared with a con-
sideration of their intended audiences.
 Chapter

28
Engineer’s Review of the
Submitted Update

The engineer’s review of the submitted CPM update should mimic,

to a large extent, the review by the contractor’s Scheduler. The pur-
pose of the review is twofold. First, the engineer must be assured that
the submittal is accurate and technically correct and provides a true
assessment of the status of the project. Second, the engineer should
be reviewing the update to determine how the owner can assist the
contractor in getting the project completed on time and what the
owner must do to minimize the extra costs that are always associated
with a project.

28.1. Is This an Update or a Revision?

The first step of review is to determine whether the purported update
improperly includes changes to the activities, durations or logic that
more properly belong as part of a revision. Unfortunately, popular micro-
computer based software accepts keyed data on an item by item basis
and does not mimic the old mainframe concept of batch processing.
Thus, it is difficult to determine that only the information as agreed
between the contractor and engineer has been entered. If a document,
such as the turnaround document noted in previous chapters, is used,
a copy can be consulted. If updates are performed on a PDA during a
walkthrough of the project, even that separate recording of the intended
input is lost.
Third party programs, such as Claims Digger, can list all differ-
ences between two files. It is, then, left to the engineer to perform the
tedious task of checking that each change matches that on the turnaround

475

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
476 The Practice of CPM Scheduling

document or contemporaneous notes (or memory) of the project walk-

through. P3e/c and Primavera Construction include Claims Digger and
seamlessly provide its comparison reports.
Although technically any changes other than AS, AF, and RD (or per-
cent complete) represent a revision, minor corrections of spelling and
other “housekeeping” changes may be permitted as part of an update,
after review by the engineer that they are not creating substantive
changes to the network. Similarly, if a minor change to a duration or
logic of a non-critical activity will not substantively change the plan or
schedule, acceptance of such is within the engineer’s discretion.

28.2. The Technical Review

Since Primavera permits the user to modify the output fields of early
and late dates after calculation and without notice that the now recorded
dates are not the same as those calculated, the engineer should require
an electronic copy of the file and run it on his/her computer. The engi-
neer should then spot check the calculated results against the tabular
printouts provided with the submission. Keep in mind, however, that
slight variances are to be expected if, for example, the contractor is
using SureTrak and the engineer is using P3 or P3e/c.
The engineer should next review the diagnostics, similar to the review
performed by the contractor’s Scheduler. There should not be additional
or modified constraints. There should not be additional open ends. The
options and algorithms chosen should not be different from that in the
submission of the initial CPM or prior updates.
Section 4 of the diagnostic should not list activities given actual dates
beyond the DataDate. The number of activities being performed out-of-
sequence and details thereof should be reviewed very carefully because
each one represents some additional and unplanned cost to the con-
tractor for which the contractor is going to be looking for additional
compensation. Or this could simply be a sign of bad reporting, for exam-
ple, failing to report an activity finished because one shovel full of back-
fill is still to be placed, or reporting a false start because one shovel full
of excavation has been taken from an area scheduled to be worked at a
later time.
The engineer should similarly be concerned if the number of activi-
ties started but not finished is rising. This again points to either bad
reporting or a pattern of interference by some party that is impacting
the contractor’s ability to complete its work. Keep in mind that for a
proper network, where no activity, has a duration greater than twice the
update period, the theoretic limit of activities-in-progress cannot exceed
the number of new activities started this past update.
 Engineer’s Review of the Submitted Update 477

28.3. The Critical Path

The engineer should focus review upon the work on the critical and
near-critical path activities that are expected to be worked upon during
the upcoming update period. This should be a relatively small number
of activities. The purpose of the review of this subset of activities is to
double check that the contractor has provided all necessary submittals
to perform this work, the design department has reviewed them, and
that any other preparations that involve the engineer or the owner have
been made.
The engineer should also compute the number of critical and near-
critical activities and divide by the number of activities not yet com-
plete. Here, the engineer may use a personal definition or comfort zone
of criticality, be it 2 days, 5 days, 10 days, or more of float. If this per-
cent of near-critical activities is rising, then the engineer should inves-
tigate whether the contractor is properly manning the project or has
other resource-based bottlenecks.
The engineer should also be calculating the actual durations of activ-
ities as a comparison to the original durations and tracking actual
resource usage versus planned. If it appears that the contractor is losing
money, there may be strong odds that a claim is brewing, whether jus-
tified or not. And if the engineer can assist the contractor to mitigate
losses without spending the owner’s money or increasing his liabilities,
this will be to the owner’s benefit.

28.4. What to Accept and What to Reject

Acceptance or rejection of the submittal of a CPM update is really not
going to change the validity of the submission. The engineer should be
accepting or rejecting the input data (the actual start and finish dates)
based upon the CPM definitions of the activities. The engineer should
accept a remaining duration as being the opinion of the contractor,
requiring some form of backup verification only if it appears to be seri-
ously in error. When a large duration for quantifiable activity has been
accepted, such as for bulk excavation, the engineer should verify that
the reported percent complete is accurate.
Improper mixing of update and revision information should be sum-
marily rejected. The update is not the place where the contractor should
be adding change orders or making other logic changes. A strict adher-
ence to restricting actual dates to those prior to the DataDate should
be enforced. In reality, the acceptance of the update should be a relatively
simple matter. The information was properly collected from the field,
entered to the computer, and the software properly calculated the impact
of this data.
478 The Practice of CPM Scheduling

So far as comments in the narrative that may be considered editorial,

the engineer should simply note that such are accepted as the opinion
of the contractor and that the review of the submittal is not the appro-
priate forum for agreement, disagreement or rebuttal. Furthermore,
since the CPM update submittal is merely another shop drawing of the
current conditions and calculated projections based thereon, the review
is also not the proper forum to discuss concerns over under-manning or
other perceived ills of the contractor. These deserve a separate letter by
the chief resident engineer.

28.5. Summary
The submitted update of the CPM allows the engineer to determine if
the project is continuing on course or is falling behind. It also may be
used to determine if the contractor is having difficulties that may result
in a proper or improper claim against the owner.
The first check by the engineer should be to be assured that the sub-
mittal is strictly for an update and does not improperly mix an update
with a revision. Next, the engineer should verify that the input to the
update (the actual start and finish dates and remaining durations for
work-in-progress) is correct and matches what exists in the field. The
engineer should run the update file on his/her own computer to remove
the temptation of cheating from the contractor.
The engineer should carefully review the diagnostics to determine if
and where the contractor is deviating from its “plan of execution” and
determine the reasons why. The engineer should review the near-term
for the critical and near-critical paths to be assured that the engineer
will be ready to assist so that this work is performed in a timely manner.
The engineer should also check for signs of under-manning or lack of
progress on non-critical but soon-to-be-critical work.
The engineer’s acceptance of the update should be limited to agree-
ment to the input and to the form of submittal. Opinions of the contractor
as to the cause for delay or disruption should be noted as heard, reserv-
ing the right to respond to a more proper forum. Nonetheless, the engi-
neer should be watching for contractor problems before they are brought
as claims.
 Chapter

29
Revising the Logic Network

Murphy’s law states that if anything can go wrong, it will; and even “the
best laid plans” often go awry. One of the most salient features of the pro-
motion of CPM at DuPont was that design changes requiring a 40 percent
change in the plan recorded in CPM format required only 10 percent of the
effort required by a traditional scheduling team to implement. Thus, if
major changes are required in a project under way, the project team should
be able to prepare a revised logic in a reasonable amount of time.
But for the more common small changes, the revision process is even
more powerful. The contractor and owner may explore the impact of one
or multiple proposed changes. The impact of a CIC (change in condition)
or unresolved RFI (request for information) can quickly be ascertained.
And the impact of a catastrophe or mere misfortune by the contractor
(such as equipment failure) similarly can be calculated.

29.1. “What If” versus Committed Changes

The contractor is often asked to price an estimate or prepare a quote on
various proposed change order work. The difference between the esti-
mate and the quote is that the first is merely an intelligent opinion. The
second case constitutes a legal offer to perform at the stated price. If the
scope of the change is complex and possibly unclear, the contractor may
add a disproportionate contingency premium to the price, or may refuse
to provide a quote at all, demanding that such work be done at its actual
time and material cost.
Similarly, the contractor is often asked to provide a time impact analy-
sis for a proposed change. In the same manner as for the cost proposal,
this may either be advisory and thus the risk of error falls upon the
owner, or guaranteed, in which case, the contractor should certainly

479

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
480 The Practice of CPM Scheduling

provide a contingency commensurate with the risk involved. In either

case, the cost of preparing the estimate of cost or time impact usually
cannot be directly recovered unless the owner goes forward with the pro-
posed change, and thus the contractor has an incentive to keep the costs
of the estimating effort low and the contingencies high. If the rare owner
agrees to pay for this estimating work on his/her behalf, more accurate
estimates and lower contingencies may be received.
The first step in these “what if ” analyses, is to copy the electronic files
for the most recent update and rename the new file (Figure 29.1.1). The
original file should be set as a target of the “what if ” file (Figure 29.1.2).
Placing two bars, one for the last update and the second for the revision,
side by side allows a graphic means to assess the impact of the revision.
Additional activities representing the new scope of the proposed
change should be inserted into the logic network, starting at a point
where the physical infrastructure required on which to place the new
work is to be performed, and ending at an activity that cannot start until
the new work is complete. The durations of the new activities (or mod-
ified durations of existing activities) should be supported by the same

Figure 29.1.1 Copy files command.

 Revising the Logic Network 481

Figure 29.1.2 Project overview—rename version.

backup data provided in the initial CPM, that is, crew size and type,
equipment to be used, and so forth.
Next, the contractor must determine from where the resources to per-
form this additional scope will come. There is a non-trivial cost to bring-
ing in an additional crew or an additional crane to the site. If the
contractor chooses, or is directed to divert resources from non-critical
work, there are still the costs of disruption. If the contractor is directed
to divert resources from critical or near-critical path activities, there will
be both a cost and time impact, even if the proposed extra work would
not be on a “critical path.” If multiple proposed changes are being con-
sidered to be performed during the same time frame, the contractor can
only assume the worst case scenario that extra work on non-critical
paths will divert resources from critical path work.

29.2. Changes: Approved, Constructive and at Contractor’s Cost

Before a change order should be added as a revision to the CPM, it
should be added to the contract. Until the change order is “signed,” or
may be deemed by law to be constructively “signed,” or the contractor
receives a written proceed order, the change should not be incorporated
482 The Practice of CPM Scheduling

into a submitted revision. A constructive change order is one where the

“signature” will be provided by a Court, even if the parties have not for-
mally agreed. For example, if you walk up to a hotdog vendor in New York
City and request one “with sauerkraut,” and it is supplied, there is no
question that you have a contract to pay for that hotdog even if the
words on compensation were never spoken.
There is also the change to the “plan of execution” that is due solely
to the contractor’s failure to include such base bid scope in the initial
CPM. Unless there was some ambiguity that such work was to be per-
formed by the contractor, the acceptance of the initial CPM by the owner
or its engineer will not relieve the contractor from performance.
However, once the bid estimate or CPM preparation error is discovered,
it will modify the contractor’s “plan of execution” and must be incorpo-
rated into the logic network the same as any other change. The only dif-
ference is that this change will be chargeable to the contractor and not
to the owner. A crane failure or other catastrophe should similarly be
added to a revision to the last update prior to the incident.

29.3. Revised Baseline

In some instances, the quality or quantity of changes, from the side of
either the owner or contractor, require a complete re-think of the “plan
of execution” for the remainder of the project. A similar situation might
be caused by the default or termination of a contractor. In either case,
the project team should be planning an entirely new project based upon
the scope remaining. In such an instance, it is appropriate to first deter-
mine the delay (if any) to the project as of the date of the re-think and
then plan a new project from that point in time.
The completion date for this new project may be set at the original
contract completion if the engineer believes that a Court will determine
the causes of delay to be solely that of the contractor (and for which the
owner will pay acceleration costs if the engineer’s belief is unfounded)
or at the delayed date, to which the engineer believes a Court would
find that the contractor is entitled to an extension of time, or to any date
in between. The mechanics of preparing the revised baseline are to
bring all work performed to date to one new milestone activity, and from
this new starting point, begin the planning of all remaining work. In
many cases the project leader if not the entire project team may have
changed, and although it may be easier to reshuffle the existing activ-
ities with new logic, this should not be a requirement. The bottom line
is the desire for a new network from this point forward that demon-
strates the ability to complete the project by the newly agreed or dic-
tated completion date.
 Revising the Logic Network 483

29.4. Update then Revise

If various factors call for a revision during the course of an update
period, the proper order of analyses is to first prepare the update and
determine how the project performed during the past period, and then
to apply the revisions to the new update. Maintaining this order is good
practice, even if it calls for preparing an extra update between the
updates required by a specification. To recap, updates measure the
impact of the reality of the past, revisions measure the impact of the fore-
cast of the future.

29.5. Summary
Revisions to the network are required whenever the assumptions of the
original “plan of execution” are no longer accurate. The first step of a
revision is to copy the last update to a new file. Additional activities must
be placed based upon “physical” logic restraints—leading from work
that must be complete before the new activity may start, and leading
to work that cannot start until the new activity is complete. Serious con-
sideration must be given to the provision of resources to perform the
extra work and from where such resources may be diverted. It is impor-
tant to always update before revising a network.
This page intentionally left blank
 Chapter

30
Engineer’s Review
of the Submitted Revision

Review of a submitted revision of the CPM logic network should be han-

dled similarly to that of the initial CPM submittal with two important
caveats. The first is that the engineer must be particularly on guard for
the risk that the contractor may use this opportunity to “cook the books”
in preparation of a claim for additional time and perhaps money. The
second is that management of the owner may desire the engineer to
improperly use his/her authority to pressure the contractor to “cook the
books” on the owner’s behalf.
One of the first questions should be to determine the type of revision sub-
mitted. The revision may be simply the adding of various unanticipated
events and other causative factors with the purpose of illustrating the
impact of them upon the existing “plan of execution.” This, in turn, may
be the basis for the granting of additional compensation to pay for the costs
of demonstrated disruption or an extension of time for a verified delay.
Alternatively this may be the basis for a request (or demand) for a recov-
ery schedule involving either a minor revision (such as selective overtime
on the current critical path) or a major revision (involving a total rethink
of the “plan of execution” for the remaining scope of the project).

30.1. Minor Revision

The first step in reviewing a revision to the CPM logic network is to
verify that the narrative of the submitted revisions matches the changes
to the computer file. This can be accomplished by use of Claims Digger
software product (included in P3e/c and Primavera Construction) or by
other means.

485

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
486 The Practice of CPM Scheduling

The second step is to review the content of each change to the current
approved baseline CPM schedule. If the change is merely to the scope
of an existing activity, increasing or decreasing the duration of it, the
parties may agree on the change in duration or it may be treated as a
proceed order or directive, leaving the determination of the impact until
after the work is performed.
If the change involves new work requiring a new activity, the activ-
ity should be placed into the logic after the completion of the physi-
cal infrastructure required to perform the new work, but also subject
to a constraint based upon the date when the contractor was author-
ized to perform the additional work. (Authorization in this context
usually requires a signed change order for an agreed price or a pro-
ceed order to perform such work on a time and material basis.) Based
upon whether or not the parties have agreed from where the
resources to perform the additional work will be provided, changes
to logic relating to such resources (craftspeople, equipment, or forms)
may be appropriate. A change order not addressing such concerns or
reserving rights related thereto implies the contractor expects to bring
in fresh resources or utilize some portion of the float of non-critical
activities.
If the change involves a stoppage of work due to an unanticipated con-
dition, the impacted activity may be broken into two portions. Although
theoretically, it may be possible to record a suspend date when the work
was stopped and utilize a start-to-start relationship to the causative
event followed by a finish-to-finish relationship to the remaining por-
tion of work (for which a resume date will be recorded at the appropri-
ate time), the erroneous loop detection issue previously discussed
prevents this more elegant solution.
In many cases the liability for a change to the logic network is dis-
puted. For this reason, the activity descriptions should reflect what is
the issue rather than who is responsible. For example, the contractor
may be required to redo work that he/she claims is correct but the engi-
neer claims is defective. A responsibility neutral description of this new
scope may read, “Demolish and Re-Pour Column 27N claimed defective.”
The engineer’s review and acceptance of the revision to the CPM may
certainly reserve rights relating to characterizations of responsibility for
claimed changes.
But what if the engineer believes that the changes are part of the orig-
inal scope of the project? The same logic applies. Certainly the work dis-
cussed is not already in the schedule and may also not be in the contractor’s
bid. The question of whether the work is in the contract plans is irrele-
vant to the scheduling question. In either case, the work still must be
performed.
 Engineer’s Review of the Submitted Revision 487

30.2. Major Revision

In some instances, the contractor will desire to make a major revision
to the plan. This may be due to the loss or acquisition of major equip-
ment, such as replacing two small cranes with one large crane. The
entire sequence of construction, or “plan of execution,” will be impacted
by this change. The contractor will also expect other parties, including
both his own subcontractors and agents of the owner, to modify their
plans to mesh with the new master plan.
A major revision may also be required to a major change order that
significantly disrupts the original “plan of execution.” In either case, the
most important step to be taken by the engineer in review of such a revi-
sion is to properly update the current CPM and otherwise document the
status of the project and its schedule prior to the abandonment of the
current plan. The new plan should then be reviewed on its own merits
with a view that revisions are for future while updates record the past.

30.3. Recovery Schedule

A special type of major revision is a recovery schedule. The purpose of
the recovery schedule is to modify the current plan because the sched-
ule slipped for whatever reason, with the goal of regaining whatever time
was lost. This may merely involve reduction of selected durations by
means of overtime, larger or multiple crews, or larger equipment. A
slightly more serious revision will require deletion of at least some
resource logic with the intent of working two areas concurrently (with
two crews and two sets of forms), which was planned to be performed
sequentially. The original notation of physical versus resource restraints
previously suggested becomes very useful for this type of exercise.
If the project encountered serious delays, the current plan may need
to be scrapped and the remainder of the work replanned, almost as a new
project. (And if the contractor defaults or is terminated, this is exactly
what will be done.) However, as noted previously, the reason for the
recovery schedule should be responsibility neutral. A great deal of dis-
cord is often raised on projects when the engineer demands a recovery
schedule and the contractor provides an acceleration schedule. As long
as each side makes its position known, the submitted and approved plan
is not going to be judged by a Court based solely upon what it is called.

30.4. Summary
The engineer’s review of a submitted revision to the CPM requires care
in preventing the contractor from “cooking the books” while profes-
sionally avoiding pressure by the owner to do the same for it. The first
488 The Practice of CPM Scheduling

step of any revision is to have a full and complete update of the status
of the project prior to modifying the approved baseline plan. Minor revi-
sions involving adding various changes and unanticipated events to the
last update to determine their impact should be done on a contempo-
raneous basis. Major revisions to the contractor’s “plan of execution”
should be preceded by a full audit and documentation of the current
status of the project. A recovery schedule or acceleration schedule should
be reviewed with responsibility neutral bias.
 Chapter

31
Case Histories

After 40 years of experience with CPM, thousands of case histories can

be recounted. Many are unavailable for publication because of their
proprietary information. The increase in construction litigation in the
past two decades has dampened any enthusiasm for releasing schedul-
ing data, further limiting the availability of case histories.
Some case histories memorialize early network applications and were
used like foundation stones in building the credibility of the approach. In
one application, the Bureau of Labor used networks to plan the publica-
tion of annual statistical results. The first computation of the plan showed
a critical path of 420 days, which was 50 percent longer than the target
schedule. A study of the critical path indicated that more than 25 percent
of the time was absorbed by the interdepartmental mail system. Through
revisions in planning, and by upgrading the internal mail delivery system,
a 269-day work schedule was achieved.
What follows are case histories that reveal the importance of CPM in
construction projects when properly planned and applied.

31.1. Chicago Courthouse

Paschen Contractors, Inc. used network planning to build a $32 million
courthouse and federal office building in Chicago. The company worked
with 17 major subcontractors to develop the basic network for the
courthouse. The immediate benefit of which was a better understand-
ing on the part of the subcontractors about their responsibilities.
An example of the effectiveness of the planning and subsequent analy-
sis of the network involved the elevators and the power transformers
they required. The electrical company had planned to use the elevators
to hoist the power transformers to the roof, however, the elevator

489

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
490 The Practice of CPM Scheduling

company was depending on the same transformers being already

installed to operate the elevators. It also became apparent that, although
steel erection was on the critical path, pouring the concrete floors was
only 2 weeks behind that activity on a near-critical path. Accordingly,
any time gained in steel erection had to be carefully compared with the
progress in floor pouring. The planning group made up a detailed net-
work for carrying out structural work on the basement and 9 of the 30
floors in the building, which was sufficient to provide detailed plans for
the remaining 22 floors.

31.2. Times Tower

The total renovation of the historic New York Times building for use by
the Allied Chemical Corporation was performed with a detailed CPM
plan consisting of 1200 separate activities. Allied decided to remodel the
60-year-old building rather than build a new one because, under the New
York City building and zoning regulations, a new building on the site
would have been limited to 12 stories rather than the existing 23.
The total renovation consisted of stripping down the basic steel frame-
work and then rebuilding, retaining many of the features of the famous
landmark. The contractor, Crow Construction Company, made direct
use of CPM planning. Construction equipment, workers, materials,
and all work vehicle flow had to operate without interrupting the steady
flow of pedestrians and city traffic at one of the busiest intersections
in the world. CPM was a crucial factor in the timely completion of the
project.

31.3. Airport Construction

Highway and pipeline projects depend on the effective use of resources
for timely completion. Building projects, on the other hand, rely more
on the sequence of work activities within the confined areas of the build-
ing(s) being constructed. Airports offer problems in both areas, and
CPM has been used very successfully to solve these problems. During
the early 1970s, many major airport construction programs used net-
work analysis and control in one form or another, including both the
Philadelphia and Pittsburgh International Airports.
Another successful application of CPM was the expansion of a runway
for the Allegheny County Airport in Pennsylvania. The 2-year project
was completed in 18 months, not only because CPM identified oppor-
tunities for saving time but also because of the contributions made by
the balanced team of owner, contractor, and project manager.
The project consisted of a 1000-foot extension to the main runway,
which was complicated by a four-lane highway, three railroad tracks, and
 Case Histories 491

a ravine more than 100-feet-deep in the path of the extension. A rigid-

frame, precast concrete underpass, 144 feet wide × 828 feet long, was con-
structed to house the highway and the railroad tracks. Approximately 1
million yd3 of embankment was used to fill in the ravine. The general con-
tractor, W. P. Dickerson, actively cooperated in the CPM planning and
the implementation of the CPM plan-cooperation that was most impor-
tant to the success of the operation.
CPM established a target completion date of December 9 for the
underpass, and actual completion occurred on December 23. That com-
pletion was a key activity, ensuring that work on the project could con-
tinue through the winter.
The second most critical activity was the relocation of four phases of
the Union Railroad track, which was completed five days ahead of the
CPM target date. With the aid of CPM, the CM and the contractor were
able to coordinate the manufacture and delivery of precast, prestressed
concrete beams after an initial analysis indicated that construction
would be delayed without the beams. The contractor studied his precast
plant facilities and offered to expand them if the county would pay for
the inventory of the beams. The county agreed that having the beams
available to meet the schedule was important and, therefore, underwrote
the inventory costs so the supplier was able to meet the construction
schedule.
CPM was also used to evaluate the effects on meeting the schedule of
a number of other considerations, including relocating a graveyard,
securing additional foundation material, substituting foundation mate-
rials, and a strike. In some cases, the CPM analysis indicated how meas-
ures to expedite could be undertaken; in others, it pointed out that a
longer timing than originally hoped for would have to be accepted.
The major vertical structural members of the tunnel had to be braced
during curing and prior to pouring the tunnel’s roof. CPM was used to
plan the entire forming, pouring, and stripping sequence of activities and
to evaluate special bracing equipment to expedite the entire operation
and ensure its timely completion. The information on bracing that CPM
provided required the contractor to order additional braces, which per-
mitted pouring the tunnel’s roof to proceed as originally scheduled. In
the end, the contract was completed six months early.

31.4. High-Rise Construction

High-rise contractors have used CPM both to study the activities required
on a single typical floor and to correlate the activities in summary fash-
ion for work on all of the floors. In constructing the Chicago Marina City
Towers, the James McHugh Construction Company used CPM to plan
two 60-story towers, including 20 commercial and parking floors.
492 The Practice of CPM Scheduling

Initially, a detailed critical path diagram was developed for a typical floor,
which was then used to study what the rest of the project would require.
It was decided to complete the first few floors at the rate of one per week
and later accelerate to two and a half floors per week in the east tower and
two per week in the west. The complete activity range was regenerated in
detail for all of the project, resulting in a network of 9600 activities that
was used successfully.
In the construction of a high-rise building in the Bronx, the super-
structure was on the critical path, which is a usual situation. Through
intensive coordination, the contractor was able to achieve a 3-day
pouring cycle per floor. As the project progressed, the cycle was cut to
an almost unbelievable 2 days. However, in concentrating the super-
vision of the reduction of time spent on the superstructure work, the
plumbing riser work that followed became critical. CPM highlighted
the need to accelerate the riser work, and if that had not been noted,
the 2-day cycle achieved for superstructure work would have had
little effect on the completion date because the riser cycle was still at
3 days.
In Phoenix, Arizona, the Mardian Construction Company used CPM
scheduling for all its projects and for apartment buildings in particular.
One of the projects was a 22-story Executive Towers apartment build-
ing in which concrete framing was completed in less than 88 days. The
superintendent attributed much of the success to CPM planning.
The prime factor was the development of a feasible forming system
and the choice of a tower crane as a result of early CPM planning. One
floor was scheduled in great detail, and then the information was recy-
cled for the rest of the high-rise. Close monitoring of the project resulted
in a reduction of the basic floor cycle from 4 to 3 days, but the CPM plan
demonstrated that the shoring required to continue the phase was
uneconomical, so the 4-day schedule was reinstated.
In Philadelphia, the Arthur A. Kober Company, a developer-builder,
collaborated with OKA to develop a CPM schedule for its $35 million
Academy House Condominium. This 37-story high-rise, including three
subsurface levels for parking, was built on a congested urban site. The
structure was of reinforced concrete with a brick exterior. The founda-
tion work was complicated by the need to underpin and brace adjoining
structures, including the historic Academy of Music.
The upper 30 floors were residences, and apartment color and
material selections were coordinated with the construction schedule.
The public areas, except for the condominium service portion, were
shelled, and the work leapfrogged up the structure into the living
units. Every tenth floor housed temporary shops, and two cranes were
used.
 Case Histories 493

31.5. NASA
Network analysis was used in all of the major contracts awarded for work
on and at the Apollo launch complex at Cape Canaveral and for similar
space project contracts awarded earlier and later. The level of detail used
in the network systems varied, as did the forms of the networks. One of
the major applications was under the direction of the Corps of Engineers,
Canaveral District, and it included the review of independent contractors’
networks and the correlation of this information into a master analysis net-
work for the vertical assembly building (VAB) and related facilities used
for the Saturn program.
The approach was to require both systems and construction contrac-
tors to provide network schedules. In turn, both NASA and the
Canaveral District Corps of Engineers (under Major General W. L.
Starnes) used network-based PMIS systems to monitor and evaluate the
network input from the contractors.
In the Saturn program, an unused launch complex control room was
turned into a war room (later dubbed the “moon room”) displaying the
many contractors’ networks at various levels of detail. Today, the use of
networks to plan and control space programs has become routine pro-
cedure for most projects, including the space shuttle.

31.6. Housing
The Rouse Company used CPM to plan the engineering and site devel-
opment phases in constructing the new town of Columbia, Maryland.
Activities related to grading, sewers, water, electrical service lines, and
paving were coordinated so that entirely developed areas were ready for
housing construction. CPM was also used to plan the building of the
town center, an engineered lake, and the sewer and water utility con-
nections to service the first completed part of the town.
CPM was credited with the on-time delivery of 300 duplex housing
units for a Navy housing project at the naval station in Rota, Spain. The
CPM program analyzed more than 3100 required operations, including
not only the prefabrication of the housing units, but the distribution of
available workforce and equipment resources, as well as activities relat-
ing to site preparation, utilities, roads, and foundations.
The plan included a sewage station and distribution system, and it was
used to determine the basic field crew size needed to erect prefab units
most efficiently. The crew size decided on was 12 workers, including a
superintendent, a crane operator, a rigger, an electrician, and a plumber.
The study also helped in the selection of such equipment as air-
powered hammers (the need for which was determined after it was
494 The Practice of CPM Scheduling

pointed out that 5000 nails per duplex would be used). Stateside fab-
rication speeded up operations by premanufacturing 80 percent of the
buildings.

31.7. Manufacturing Facilities

Butler Manufacturing used network techniques for planning the con-
struction of a 95,000-ft2 plant in Knoxville, Tennessee, which had to be
completed within 20 weeks. Preplanning by the owner indicated that
deliveries from a sole source would materially shorten the implemen-
tation period. This information was used as the basis for justifying the
sole source purchase of the Butler building. Complete sections of the
building were prefabricated and organized as units and zone-delivered
to the site. The design development identified the need for 26 cranes,
and the cranes were added without a delay of even a day.
Another major systems facility delivered for partial occupancy within
6 months was a 300,000-ft2 building in Georgia developed for use by
Lockheed in constructing its C5A transport plane. The design was car-
ried out by the Atlanta-based firm of Heery and Heery, which had pre-
viously used CPM to complete the Atlanta Braves stadium on time.
In this case, a different form of preplanning was used. The architect-
engineers drew on the preplanned Inland Steel modular systems’
design, which was developed as part of the School Construction System
Development (SCSD) in California. The system was based on a pre-
designed 4 ft2 horizontal module, including structural and ceiling light-
ing systems.
CBS Records was proceeding with design of a plant at a new site to
manufacture and distribute records and tapes. Because of the close inte-
gration needed between the manufacturing/processing and storage
equipment to be used at the facility with the plant’s construction require-
ments, CBS Facilities Engineering decided to develop a CPM network
for the various equipment development, design, and procurement lead
times and decision points. A detailed CPM network illustrated the var-
ious actions required and their interfaces with the design process for con-
structing the facility by the outside designer.
General Electric’s Aerospace Division in Valley Forge, Pennsylvania,
used CPM plans to monitor the performance of developers and con-
tractors in the delivery of more than 20 facilities at the height of its aero-
space programs. In addition to requiring contractor networks, facility
manager John D. Orr had in-house training seminars for his facilities
engineering staff. (Organizations such as Corning Glass and Celanese
have used the in-house seminar approach to either introduce or revi-
talize CPM planning for facilities.)
 Case Histories 495

31.8. SEPTA RailWorks

When Southeastern Pennsylvania Transportation Authority (SEPTA)
assumed operation of the former Pennsylvania and Reading rail lines
in 1983, it inherited a network of bridges, track, and overhead power
lines (catenary equipment) that had already been in service for many
years. Decades of deferred maintenance and virtually no dedicated cap-
ital funding had resulted in a useable but deteriorating rail system.
The commuter tunnel, completed in October 1984, connected the once-
separate rail lines; it allowed all regional rail lines to access the three
center-city rail stations: Market East, Suburban, and 30th Street.
Several months after the tunnel’s completion, an engineering inspection
study found that many of the system’s bridges, some of which have
stood for nearly 100 years, were in critical condition.
The four-mile stretch of track between Wayne Junction rail station
and 9th and Brown Streets in North Philadelphia was listed as a ren-
ovation property. The stretch consists of track and catenary system and
25 rail bridges—a total of 16 track miles—forming part of the main
line, or throat, of the old Reading line. Six SEPTA regional rail lines feed
into this central corridor.
The completed project cost approximately $300 million, secured mostly
from UMTA (now FTA) grants. The project, named SEPTA RailWorks,
entailed major infrastructure rehabilitation of this regional rail corri-
dor. The major components of the work included renovating five bridges,
replacing 20 bridges, replacing all of the track, adding new power lines
(the catenary system), and replacing related equipment, including
switches and signals.
All of the bridges, except one that crosses a stretch of Conrail track at
Wayne Junction, span active highway crossings in a congested urban area.
RailWorks also resulted in two brand new rail stations: the Temple Station
and the Fern Rock Transportation Center. Fern Rock Transportation
Center provided a new, permanent connection to the Broad Street subway
line. During the two summer shutdown periods, R2, R3, and R5 riders used
this line as a temporary transfer point to the subway, allowing continued
access to center city.
RailWorks entailed a 3.5-year construction period, which began in
August 1990. The track shutdown periods were scheduled from April
1992 to October 1992 and May 1993 to September 1993.

Scheduling
Primavera was the base scheduling system used both by the contractors
for project scheduling and at a higher level for program integration by
the CM (OKA).
496 The Practice of CPM Scheduling

A preliminary and final schedule was required to be submitted by each

contractor, and after approval by the CM and SEPTA, cost was loaded
and the resulting document (known as the “value line”) was used for pay-
ment purposes. Monthly schedule updates were used as the basis for con-
tractor payment requests in a conventional manner.
In addition to SEPTA’s normal scheduling requirement, the RailWorks
program also required the contractor to prepare and use a detailed
“window schedule” for managing the work during the construction win-
dows. This also was a Primavera schedule, but it was prepared in much
more detail, was resource-loaded, and was updated weekly. It was not,
however, used for payment purposes.
A program master schedule was prepared and maintained by OKA,
and the contractors provided 2-week look-ahead schedules. The con-
tractors also provided a schedule of any track outages they required and
a detailed schedule of operations during outages.
The emphasis on preplanning paid off on the first shutdown window.
Everyone, particularly the contractors, were focused on the liquidated
damages. These were set at $70,000 per calendar day. Not surprising,
the first shutdown completed 3 days early. The second shutdown was a
shorter window, but the contractors had gained confidence in the first
shutdown. The second shutdown completed 1 to 2 weeks ahead of the
system restart date.

31.9. New Jersey Turnpike Authority 1990–1995 Widening Program

The New Jersey Turnpike Authority (NJTA) had a program to widen the
New Jersey Turnpike between exits 11, New Brunswick, and 14, the
Newark Airport. This stretch of the turnpike is made up of four roadways,
each three-lanes wide. The two inner roadways are used for northbound
and southbound passenger vehicle traffic. The outer roadways are used
as truck and bus lanes. Passenger vehicles are allowed in the outer road-
ways. The project widened the outer roadways to four-lanes wide from
exit 11 through exit 14. The construction value of this project was approx-
imately $250 million.
The NJTA developed a team approach for the management of the
project, with three levels of management between the NJTA and the 27
contractors that actually did the construction. The top level was the pro-
gram manager (Hill International), whose most important role was
reporting the program status to the NJTA. The program manager got
most of its information from another part of the team, the construction
manager (Howard, Needles, Tammin & Bergendorff). The construction
manager had the task of providing continuity among the various design
engineers and oversight engineers.
 Case Histories 497

The program was divided into five sections from south to north. Each
section was assigned to one of five section engineers (one of which was
OKA). The section engineer was responsible for the actual construction
oversight. Responsibilities included monitoring the progress schedules
and progress payments and the inspection of the work itself. Twenty-
seven separate construction contracts were spread across the five sections.
In order to standardize the numerous project schedules, all contrac-
tors were required to prepare project schedules using Primavera Project
Planner. Using a standard scheduling specification ensured that each
of the 27 project schedules would be compatible with all of the others.
Each project schedule was the responsibility of the respective con-
tractor. The first review of baseline submissions, as well as monthly
updates, were provided by the section engineer. In addition, the con-
struction manager reviewed each project schedule for the interrelations
between adjacent contracts. To improve its programwide perspective of
each individual project schedule, the construction manager, through
the efforts of a scheduling consultant, prepared a composite schedule by
merging all of the individual schedules.

31.10. JFK Redevelopment

The JFK 2000 program was started in 1987 for 2000 completion. It was
the first major upgrade to the airport since the 1960s. The upgrade was
planned from 1987 to the groundbreaking in 1989. The Port Authority
of New York and New Jersey (PA) selected OKA as the program man-
ager (PM) and Bechtel/Tishman as the construction manager (CM). The
program, funded at $1.66 billion, included:
■ Airport traffic control tower (tallest in North America)
■ Roadways, including high-occupancy vehicle system (HOV)
■ Airport utilities
■ Early action/preliminary phasing: building 14 (former PanAm hangar)
and preliminary construction phasing
■ Terminals: East Garage and expansion of IAB federal inspection serv-
ices (FIS)
■ Passenger distribution system

Private projects included hotel development and co-generation. Work

plans and schedules were revised in 1992, especially for projects planned
for the later years of the capital plan, such as the passenger distribu-
tion system and International Arrival Building and Federal Inspection
Service expansion.
498 The Practice of CPM Scheduling

The design and construction of the automated, on-airport passenger

distribution system (people mover) was delayed until the conceptual
design and financial issues for the system were coordinated with the
design of an off-airport, transit connection with city and state trans-
portation agencies. During the first quarter of 1991:
■ Contract negotiations and architectural/engineering design develop-
ment for a new $250 million co-generation plant made significant
progress.
■ Construction under roadways contract package 1 continued ahead of
schedule, and roadways contract package 2 went out for bid. Planning
and programming of terminal frontage roadways continued.
■ Concrete for the airport traffic control tower shaft was completed at
a height of 291 feet.

A new program budget was established at $1.66 billion. It included

the $275 million spent as of January 1, 1991; $985 million spent between
1991 and 1995, and $400 million to be spent in 1996 and beyond. The
second phase of the program will incorporate an automated, on-airport
people-mover, or passenger distribution system (PDS).
The PM used ARTEMIS to produce schedule network drawings and
network reports validating logic computer reports submitted by the
CM. The CM used Primavera software to produce schedule network
drawings and network reports validating the logic computer reports
submitted by the contractors. To ensure that schedule data were
transferred from the CM’s Primavera scheduling software to the PM’s
ARTEMIS software, the CM prepared a computer disk file for each
activity within the network.
The CM reviewed, approved, and monitored the contractor’s detailed
construction schedules. The CM also developed and maintained pre-
liminary, baseline, preconstruction, construction, and as-built sched-
ules at the subproject summary level by using Primavera scheduling
software. The CM developed work-around scenarios with the contrac-
tors to ensure final on-time completion.
The PM developed and maintained summary schedules at the pro-
gram, subprogram, and project levels through all program phases, based
on designer and construction manager input, by using ARTEMIS. The
PM also audited the construction manager’s and designer’s project sum-
mary schedules and subsequently provided the management reports
evaluating the subprojects, projects, and overall program.
As the JFK Redevelopment plan continued, the need for an entirely
new International Arrival Building (IAB) was identified. Phase II focus
shifted to this project, valued at more than $800 million.
The ARTEMIS system required full-time programming support. OKA
shifted to its own PM/CS (Project Management/Cost System).
 Case Histories 499

OK-PM/CS consists of off-the-shelf computer programs customized to

respond to the special requirements of large programs. The system is
capable of operating on stand-alone or networked personal computers.
All computer program components of OK-PM/CS are based on a com-
mercially available relational database management system. The nucleus
is a database cost program integrated with Primavera Project Planner
for schedule management and our proprietary software, OK-TRACK,
for contract administration. Many other specialized utility computer
programs are available, if required, to further enhance the system.
The use of a relational database gives the program management
team the capability to create a global information system. Data within the
OK-PM/CS are stored and manipulated within a standard environment,
providing an efficient solution to the problem of integrating data from
different applications. Data is readily available to applications within
the system as well as to outside applications. The use of a relational
database within a microcomputer has the following benefits:
■ The user can substitute and integrate new software products as they
become available.
■ The cost of implementation is low. The system has the ability to grow
incrementally, as program or agency needs increase.
■ The user can build on proven software and hardware.
■ The need for specialized programmers and management information
system (MIS) staff is minimal.

PCs provide the flexibility required to integrate a large and diverse

number of users and resources of data. The configuration can also
include bridges to other computers as well as gateways to remote users.
Key features of the OK-PM/CS system architecture include data gath-
ering, system installation/expansion, system integration, external inter-
faces, and relationship to the owner’s financial system.

Data gathering. At the data level, OK-PM/CS uses coding structures

such as a Work Breakdown Structure (WBS) to allow precise definition
of program data elements. Using common codes for data elements pro-
vides the ability to capture, link, and report data in unison. Each user
provides and maintains a portion of the data but benefits from the data
and information provided by others. The user can generate specialized
reports by using industry-standard inexpensive report-writer programs,
such as R&R Report Writer.

System installation/expansion. At the computer program level, OK-PM/CS

uses an industry standard relational database system that allow data to
be exchanged, combined, and modified by any of a multitude of products.
Integration is easily accomplished through a PC’s local area network
500 The Practice of CPM Scheduling

(PC-LAN). The network will also serve as the bridge to existing owner
mainframe or mini-systems and a gateway to remote locations.
OK-PMICS can be installed and expanded in an incremental, cost-
effective manner. Initial investments in a PC-LAN network for the
basic server and software are small compared to minicomputer hard-
ware and software investments. For small projects or programs in the
initial stages of development, OK-PM/CS can be installed and operated
on stand-alone PCs with little or no computer integration. These pro-
grams can be transferred easily to the LAN when the number of users
expands. The programs can also be installed on preexisting PC-LANs
in an organization. Workstations can be added for a small marginal cost
as the project proceeds throughout the engineering and construction
phases. Connections to field offices and other computer networks in the
owner’s organization can also be made.

System integration. OK-PM/CS provides integration of various compo-

nent applications to achieve a unified, comprehensive reporting system.
This ensures the consistency of information reported and minimizes
the repetitive input of control data. Benefits of integration include:
■ Schedule-Cost Control: Merging of CPM progress updates with cost
control module information can provide up-to-date cash flow forecasts
and performance monitoring reports at various levels of detail.
■ Contract Administration-Cost Control: Current information on pending
changes or potential claim items can be incorporated from the forecast
costs to completion in the cost control database.
■ Estimating-Cost Control: Detailed quantity takeoff and pricing infor-
mation for project elements or specific change orders from the esti-
mating subsystem can be summarized and transferred to the
budgeting and forecast fields of the cost control database.
■ Engineering Management Cost Control: Actual costs incurred from work
packages can be gathered from the owner’s financial systems or engi-
neering management subsystems to update the cost control database.
■ Document Control-Cost and Schedule Modules: All correspondence
concerning a proposed scope change can be related to the specific
CPM activities or contract pay items affected.

31.11. Toronto Transit’s “Let’s Move” Program

OKA developed its Executive Information System (EIS) for the Toronto
Transit Commission’s $7 billion “Let’s Move Program.” The following
sections highlight the key features and benefits of the EIS. Screen
samples (Figures 31.11.1 through 31.11.4) are from an application
 Case Histories 501

Let’s Move Executive Information System

System Projects References Help

Younge Spadine Loop Sheppard Scarborough

Eglinton West
New Subway Yera Sheppard North York

Scarborough RT Ezt

Program Management

Bloor Dantarin West Ezt

Let’s Move Program

Waterfront LRT West

Toronto transit commission’s let’s move program

executive information system
Quit Print Mail Home Back Help

Figure 31.11.1 Opening screen.

Eglinton West WBS Sub-Item Data View

System Help

Project Element Total AS of

Eglinton West Engineering $239,897,000 August 31, 1992

TTC Overhead

Budget Project Management

Environmental Assessment
27% 20%
Preliminary Design
1%

Detailed Design

Construction Management
25%
26%
Forecast
Bar graph
Committed
Pie graph Incurred

Quit Print Mail Home Back Help

Figure 31.11.2 Cost graphic.

502 The Practice of CPM Scheduling

Eglinton West Overview Graph

System Help

Eglinton west overview As of August 31, 1992

5.500
Limit
5.000
Budget
4.500
Incurred
4.000
Forecast
3.500

3.000
× 1.000

2.500
Total
2.000
TTC
$

1.500
Consultants
1.000
Prelim Eng
.500
Other
.0
Apri 91

Apri 92
Dec 90

Mar 91

Sep 91
Oct 91

Dec 91

Mar 92

Sep 92
Oct 92

Dec 92
Nov 91

Nov 92
Jan 91

Jun 91
Jul 91

Jan 92

Jun 92
Jul 92
Feb 91

May 91

Aug 91

Feb 92

May 92

Aug 92

Quit Print Mail Home Back Help

Figure 31.11.3 Time-scaled cost graphic.

Eglinton West Major Activities Chart

System Help

Project Eglinton West Target

Current Last Revised August 31, 1992
1 Project star t-up
2 Document PRJ need alter to & PRJ
3 Demand forecast technology asses
4 Public consolidation no. 1
Public meeting no. 1
5 Undertake initial screening of AL
7 Detailed evaluation
8 Public consultation no. 2
Public meeting no. 2
9 Detailed assessment of preferreda
10 Preliminary design
11 Public consultation no. 3
Public meeting no. 3
12 Prepare EA report
13 Progress meetings
1 Oct 90
16 Oct 90

1 Dec 90
16 Dec 90

1 Mar 91
1 6Mar 91
1 Apr 91
16 Apr 91

1 Sep 91
16 Sep 91
1 Nov 90
16 Nov 90

1 Jan 91
16 Jan 91

1 Jun 91
16 Jun 91
1 Jul 91
16 Jul 91
1 Feb 91
16 Feb 91

1 May 91
16 May 91

1 Aug 91
16 Aug 91

Select to scroll

Through/Items

Quit Print Mail Home Back Help

Figure 31.11.4 Time-scaled schedule graphic.

 Case Histories 503

developed and installed by an O’Brien Kreitzberg project controls

team.

Ease of use. The EIS was designed to be used directly by management

without the assistance of intermediaries. The system is Microsoft
Windows-based and most functions can be accessed by pressing large,
easy-to-use buttons. From the opening screen shown in Figure 31.11.1,
users can select one of the many projects that make up the Toronto
Transit Commission’s “Let’s Move Program.” Additional programwide
screens can be accessed by selecting the “Let’s Move Program” button.
Common functions such as accessing help, quitting the application, going
back to the previous screen, and printing are all accessed by clicking but-
tons that appear at the bottom of every screen.

Graphical presentation. EIS’s extensive use of graphics to summarize and

highlight information often brings out information that might remain
buried in a standard report. Each screen provides the user with graphi-
cal options suitable for the data being displayed. The Cost Graphic screen
shown in Figure 31.11.2 contains options to change the graph type from
a pie chart to a bar chart and shows the subitems of the engineering
budget as a percentage of the total. Figures 33.11.3 and 31.11.4 show time-
scaled Cost and Schedule Data screens. The powerful graphics allow the
manager to adjust the time scale on these screens in a matter of seconds.

31.12. Phoenixville-Mont Claire Bridge

The Phoenixville-Mont Claire Bridge project in Montgomery County,
Pennsylvania is an example of a planned acceleration and use of CPM to
complete a project ahead of schedule. The Pennsylvania Department of
Transportation (PennDOT) had specified a $30,000 per day incentive/dis-
incentive for meeting the specified completion date of September 17,
1987. (This date included a 16-day acceleration from PennDOT’s man-
dated completion date, agreed by the contractor as submitted with its bid.)
The project had two concurrent operations. First, to demolish an exist-
ing bridge between the two communities and rebuild it. Second, to ren-
ovate and replace the utilities under the main state highway, which
ran through the communities over the bridge.
Meeting the schedule, much less beating the schedule, was deemed a
challenge for both portions of the job. Improving the schedule for the high-
way portion was done in traditional schedulers’ fashion, splitting the
work so that two crews could work concurrently instead of sequentially.
Improving on the bridge was a bit more difficult. The demolition
sequence, involving blasting, did not offer much room for improvement. The
sequence of steel erection was mandated to run from one end of the bridge
504 The Practice of CPM Scheduling

TRADITIONAL SEQUENCE
MANDATED CONCRETE POUR SEQUENCE
1 4 2 5 3 5 2 4 1 FORM & POUR DECK
1 2 3 4 5 REBAR
1 2 3 4 5 SHEAR CONNECTOR
1 2 3 4 5 OVERHANG FORMWORK
1 2 3 4 5 STAY IN PLACE METAL FORMS
1 2 3 4 5 R/I UTIL & DRAINAGE
1 2 3 4 5 ERECT STRUC STEEL
MANDATED STEEL ERECTION SEQUENCE
Figure 31.12.1 The traditional sequence.

to the other. On the other hand, once the steel was set, followed by the util-
ities, stay-in-place metal forms, overhang form work, shear connectors, and
rebar, then concrete had to be placed in a pattern of nine, noncontiguous
sections to equalize the weight on the structure. See Figure 31.12.1.
The problem was that, although the first concrete pour could be per-
formed immediately after the completion of the rebar at the end of the
bridge where the steel was started, the second pour required 100 percent
of all preparatory work to be completed for the entire span of the bridge.
Improvement was accomplished by precessing the work between steel erec-
tion and pouring of the deck so that the concrete could be poured as soon
as the rebar was complete for each of the nine segments. Figure 31.12.2.
A graphic of this improvement won the Best Time-scaled Diagram
Award at Primavera’s 1997 Annual Convention. [See Figure 31.12.3].

MODIFIED SEQUENCE
MANDATED CONCRETE POUR SEQUENCE
1a 4a 2a 5a 3 5b 2b 4b 1b FORM & POUR DECK
1 3 5 4 2 REBAR
1 2 5 4 3 SHEAR CONNECTOR
1 2 4 5 3 OVERHANG FORMWORK
1 2 3 5 4 STAY IN PLACE METAL FORMS
1 2 3 4 5 R/I UTIL & DRAINAGE
1 2 3 4 5 ERECT STRUC STEEL
MANDATED STEEL ERECTION SEQUENCE
Figure 31.12.2 The modified sequence.
 Case Histories 505

1997
APR MAY JUN JUL
21 28 5 12 19 26 2 9 16 23 30

THE VALUE OF A GOOD SCHEDULE

SPAN #1 24540 3,7 24740 5,8 24790 5,20 TRADITIONAL SEQUENCE
RB SPAN#1 R/I RB SPAN#1 RB POUR#1A MANDATED CONCRETE POUR SEQUENCE
1 4 2 5 3 5 2 4 1 FORM & POUR DECK
UTIL & DRAINAGE REBAR FORM & POUR DECK
1 2 3 4 5 REBAR

24390 3,7 24640 6,7 1 2 3 4 5 SHEAR CONNECTOR

1 2 3 4 5 OVERHANG FORMWORK
RB SPAN#1 ERECT RB SPAN#1
STRUC STEEL OVERHANG FORMWORK 1 2 3 4 5 STAY IN PLACE METAL FORMS
1 2 3 4 5 R/I UTIL & DRAINAGE
24590 3,7 24690 1,8 1 2 3 4 5 ERECT STRUC STEEL
MANDATED STEEL ERECTION SEQUENCE
RB SPAN#1 STAY IN RB SPAN#1
PLACE METAL FORMS SHEAR CONNECTOR

SPAN #2 24500 2,11 24600 3,10 24700 1,7 24830 3,8 24870 3,7
RB SPAN#2 ERECT RB SPAN#2 STAY IN RB SPAN#2 RB POUR#2A RB POUR#4
STRUC STEEL PLACE METAL FORMS SHEAR CONNECTOR FORM & POUR DECK FORM & POUR DECK

24550 2,10 24650 6,7 24760 4,7

RB SPAN#2 R/I RB SPAN#2 RB SPAN#2
UTIL & DRAINAGE OVERHANG FORMWORK REBAR

SPAN #3 24510 2,11 24610 3,13 24670 4,7 24860 3,7

RB SPAN#3 ERECT RB SPAN#3 STAY IN RB SPAN#3 RB POUR#3
STRUC STEEL PLACE METAL FORMS OVERHANG FORMWORK FORM & POUR DECK

24560 2,10 24730 1,7 24900 3,7

RB SPAN#3 R/I RB SPAN#3 RB POUR#5
UTIL & DRAINAGE SHEAR CONNECTOR FORM & POUR DECK

24780 4,7
RB SPAN#3
REBAR

SPAN #4 24520 2,11 24630 3,10 24720 1,7 24840 3,8

RB SPAN#4 ERECT RB SPAN#4 STAY IN RB SPAN#4 RB POUR#2B
STRUC STEEL PLACE METAL FORMS SHEAR CONNECTOR FORM & POUR DECK

24570 2,10 24680 4,7 24850 2,10

RB SPAN#4 R/I RB SPAN#4 RB POUR#2B FORM & POUR
UTIL & DRAINAGE OVERHANG FORMWORK DECK CURE 02

24770 4,7
RB SPAN#4
REBAR

SPAN #5 24530 2,11 24620 3,11 24710 1,7 24820 3,21 24880 1,9
RB SPAN#5 ERECT RB SPAN#5 STAY IN RB SPAN#5 RB SET STRIP RB POUR STRIP
STRUC STEEL PLACE METAL FORMS SHEAR CONNECTOR SEAL DAMS SEAL BLOCKOUTS

24580 2,10 24660 4,7 24810 2,10

RB SPAN#5 R/I RB SPAN#5 RB POUR#1B FORM & POUR
UTIL & DRAINAGE OVERHANG FORMWORK DECK CURE 02

24750 5,7 24800 3,7

RB SPAN#5 RB POUR#1B
REBAR FORM & POUR DECK

SPAN #1 24540 3,15 24740 5,20 24790 5,20 MODIFIED SEQUENCE

MANDATED CONCRETE POUR SEQUENCE
RB SPAN#1 R/I RB SPAN#1 RB POUR#1A 1 4 2 5 3 5 2 4 1 FORM & POUR DECK
UTIL & DRAINAGE REBAR FORM & POUR DECK
1 3 5 4 2 REBAR

24390 3,15 24640 6,15 1 2 5 4 3 SHEAR CONNECTOR

1 2 4 5 3 OVERHANG FORMWORK
RB SPAN#1 ERECT RB SPAN#1
STRUC STEEL OVERHANG FORMWORK 1 2 3 5 4 STAY IN PLACE METAL FORMS
1 2 3 4 5 R/I UTIL & DRAINAGE
24590 3,15 24690 1,20 1 2 3 4 5 ERECT STRUC STEEL
MANDATED STEEL ERECTION SEQUENCE
RB SPAN#1 STAY IN RB SPAN#1
PLACE METAL FORMS SHEAR CONNECTOR

SPAN #2 24500 2,20 24600 3,18 24700 1,18 24760 4,15 24830 3,15 24870 3,15
RB SPAN#2 ERECT RB SPAN#2 STAY IN RB SPAN#2 RB SPAN#2 RB POUR#2A RB POUR#4
STRUC STEEL PLACE METAL FORMS SHEAR CONNECTOR REBAR FORM & POUR DECK FORM & POUR DECK

24550 2,19 24650 6,15

RB SPAN#2 R/I RB SPAN#2
UTIL & DRAINAGE OVERHANG FORMWORK

SPAN #3 24510 2,21 24610 3,18 24670 4,16 24730 1,19 24900 3,15
RB SPAN#3 ERECT RB SPAN#3 STAY IN RB SPAN#3 RB SPAN#3 RB POUR#5
STRUC STEEL PLACE METAL FORMS OVERHANG FORMWORK SHEAR CONNECTOR FORM & POUR DECK

24560 2,20 24780 15 24860 3,15

RB SPAN#3 R/I RB SPAN#3 RB POUR#3
UTIL & DRAINAGE REBAR FORM & POUR DECK

SPAN #4 24520 2,22 24630 3,23 24680 4,16 24770 15 24840 3,15
RB SPAN#4 ERECT RB SPAN#4 STAY IN RB SPAN#4 RB SPAN#4 RB POUR#2B
STRUC STEEL PLACE METAL FORMS OVERHANG FORMWORK REBAR FORM & POUR DECK

24570 2,21 24720 1,16 24850 2,20

RB SPAN#4 R/I RB SPAN#4 RB POUR#2B FORM & POUR
UTIL & DRAINAGE SHEAR CONNECTOR DECK CURE 02

SPAN #5 24530 2,24 24660 4,15 24820 3,29 24880 1,17

RB SPAN#5 ERECT RB SPAN#5 RB SET STRIP RB POUR STRIP
STRUC STEEL OVERHANG FORMWORK SEAL DAMS SEAL BLOCKOUTS

24580 2,23 24710 1,15 24810 2,19

RB SPAN#5 R/I RB SPAN#5 RB POUR#1B FORM & POUR
UTIL & DRAINAGE SHEAR CONNECTOR DECK CURE 02

24620 3,18 24750 5,15 24800 3,15

RB SPAN#5 STAY IN RB SPAN#5 RB POUR#1B
PLACE METAL FORMS REBAR FORM & POUR DECK

Activity Classification: TYPE OF WORK The Phoenixville - Mont Clare Bridge Replacement over the Schuykill River is part of a PennDOT project

ERECT STRUCTURAL STEEL R/I UTILITIES & DRAINAGE STAY IN PLACE METAL FORMS OVERHANG FORMWORK with a $12,000 per day incentive/disincentive contract clause.

SHEAR CONNECTOR REBAR FORM & POUR DECK

The sequence of steel erection was mandated from one end of the bridge to the other. The sequence of deck
act id rd tf Sheet 1 of 1
Plot Date 14AUG97 Activity Bar/Early Dates concrete pours was mandated to equally load the steel. Work was confined to six days per week.

Data Date 27FEB97 Critical Activity ALLAN A. MYERS

Project Start 20AUG96 Allan A. Myers, Inc. was the contractor. Engineering & Property Management Consultants, Inc. ("EnProMaC",)

Project Finish 25NOV97 * PENNDOT RT 29 SEC. 6B & 6M headed by Fredric L. Plotnick, Esq., P.E., was the scheduling consultant.

ngineering & Property Management Consultants, Inc.

BRIDGEWORK SCHEDULING ART v SCIENCE The project was completed 50 days ahead of schedule. Details of how 12 of those 50 days were earned are

depicted in this graphic.

(c) Primavera Systems, Inc. Jenkintown, PA 215-885-3733

Figure 31.12.3 Impact of the modified sequence upon the traditional approach.
506 The Practice of CPM Scheduling

This example project reminds us that proper use of CPM is not just
mastering a specific software package and that planning remains as
much an art as it is a science.

31.13. Graduate Hospital Diagnostic Building

The phone call was received on Friday, December 22, 1989. “We have a
contract to complete [this project] by December 31st.” After a few laughs
it was determined that the contractor had now estimated a new com-
pletion target of March 1, 1990, but the owner wanted an independent
confirmation of the achievability of that date.
The structure was already in place and most of the work remaining
related to mechanical systems and interior finishes. And yet with 20-some
subcontractors, some in revolt and walking off the project due to a
claimed “lack of coordination,” this was no small scheduling task. Even
the simple task of installing a door required coordination of several sub-
contractors – carpenter, glazer, painter, mechanical opener and electri-
cal connections for the opener.
After three days of effort, and a new CPM network relating only to
interior work of some 3503 activities and 4510 relationships, the ver-
dict was returned – there was no chance of completion by March 1st
without significant overtime, but April 9th was an achievable goal. The
owner set a deadline of April 1, 1990.
Then the real work began – getting the 22 subcontractors into the
same room at the same time and getting them all to agree to follow the
schedule. The two keys to this step were to impress upon all the need
to work as a team and a careful tailoring of reports. Only the scheduler
had the full report (other than file copies for the contractor and owner)
as shown for room 551 in Figure 31.13.1.
Each subcontractor was given a detail report of only their activities
including the calculated free float indicating the slippage that could
occur before the next subcontractor would be impacted. The schedule

Figure 31.13.1 Sequence of work in Room 551.

 Case Histories 507

included both the physical logic of studs, drywall, tape and so forth,
and tracking each crew of each subcontractor through the 5 floors of
the building. The goal was to have at least one day’s work for each
crew available at all times, so that if a crew finished work in one
room early, the crew knew what room they could next move to with-
out waiting for the next day. (See Figure 31.13.2.) Since several of the
more important subcontractors would be visiting each room more
than once, each knew “what goes ‘round, comes ‘round” and that the
impact to another subcontractor would soon leave them with a crew
requiring reassignment.
All of the subcontractors bought into the plan. Each Friday, the sched-
uler would walk the entire project with a clipboard, poke his nose into
each room, and checkoff work performed in that room. On one occasion,
the scheduler asked the prime contractor if the electrical subcontractor
had experienced a problem. Amazed, the superintendent noted that the
subcontractor had pulled all crews off the project for two days for some
“emergency” at another job, and asked how the scheduler knew. “It’s not
rocket science – he’s only done three-fifths of the work this week com-
pared to each week the past month.” After minor adjustment of the
logic, everyone was back on schedule.
The owner was given a E-size summary barchart of 58 “activities” as
depicted in Figure 31.13.3 that was updated each Monday. Although
minor punchlist work continued through June, the first procedure was
performed in the new diagnostic building on April 9, 1990.

Figure 31.13.2 Summary of work on 2nd floor with partial detail shown.
508 The Practice of CPM Scheduling

1990
CODE TOTL JAN FEB MAR APR
VALUE TASK FLR /TASK FLT 2 3 4 5 6 7 8 9 10 11 12 13 14 15
020 20 LOWER LEVEL INSTALL STUDS 23
022 22 LOWER LEVEL INSTALL DRYWALL 43
024 24 LOWER LEVEL TAPE DRYWALL 43
026 26 LOWER LEVEL APPLY FLOORING 18
030 30 LOWER LEVEL INSTALL CEILING GRID 16
034 34 LOWER LEVEL INSTALL CEILING LIGHTING 17
060 60 LOWER LEVEL INSTALL CASEWORK 44
062 62 LOWER LEVEL INSTALL CASEWORK ELECTRIC 46
070 70 LOWER LEVEL INSTALL PLUMBING FIXTURES 48
0EQ EQ LOWER LEVEL INSTALL EQUIPMENT 16
120 20 FIRST FLOOR INSTALL STUDS 25
122 22 FIRST FLOOR INSTALL DRYWALL 28
124 24 FIRST FLOOR TAPE DRYWALL 33
126 26 FIRST FLOOR APPLY FLOORING 35
130 30 FIRST FLOOR INSTALL CEILING GRID 34
134 34 FIRST FLOOR INSTALL CEILING LIGHTING 35
160 60 FIRST FLOOR INSTALL CASEWORK 34
162 62 FIRST FLOOR INSTALL CASEWORK ELECTRIC 36
170 70 FIRST FLOOR INSTALL PLUMBING FIXTURES 49
1EQ EQ FIRST FLOOR INSTALL EQUIPMENT 40
220 20 SECOND FLOOR INSTALL STUDS 26
222 22 SECOND FLOOR INSTALL DRYWALL 27
224 24 SECOND FLOOR TAPE DRYWALL 18
226 26 SECOND FLOOR APPLY FLOORING 43
230 30 SECOND FLOOR INSTALL CEILING GRID 40
234 34 SECOND FLOOR INSTALL CEILING LIGHTING 42
260 60 SECOND FLOOR INSTALL CASEWORK 40
262 62 SECOND FLOOR INSTALL CASEWORK ELECTRIC 43
270 70 SECOND FLOOR INSTALL PLUMBING FIXTURES 47
320 20 THIRD FLOOR INSTALL STUDS 29
322 22 THIRD FLOOR INSTALL DRYWALL 18
324 24 THIRD FLOOR TAPE DRYWALL 20
326 26 THIRD FLOOR APPLY FLOORING 26
330 30 THIRD FLOOR INSTALL CEILING GRID 24
334 34 THIRD FLOOR INSTALL CEILING LIGHTING 26
360 60 THIRD FLOOR INSTALL CASEWORK 42
362 62 THIRD FLOOR INSTALL CASEWORK ELECTRIC 45
370 70 THIRD FLOOR INSTALL PLUMBING FIXTURES 48
3EQ EQ THIRD FLOOR INSTALL EQUIPMENT -5
420 20 FOURTH FLOOR INSTALL STUDS 29
422 22 FOURTH FLOOR INSTALL DRYWALL 26
424 24 FOURTH FLOOR TAPE DRYWALL 30
426 26 FOURTH FLOOR APPLY FLOORING 40
430 30 FOURTH FLOOR INSTALL CEILING GRID 39
434 34 FOURTH FLOOR INSTALL CEILING LIGHTING 40
460 60 FOURTH FLOOR INSTALL CASEWORK 39
462 62 FOURTH FLOOR INSTALL CASEWORK ELECTRIC 44
470 70 FOURTH FLOOR INSTALL PLUMBING FIXTURES 48
520 20 FIFTH FLOOR INSTALL STUDS 25
522 22 FIFTH FLOOR INSTALL DRYWALL 7
524 24 FIFTH FLOOR TAPE DRYWALL 35
526 26 FIFTH FLOOR APPLY FLOORING 21
530 30 FIFTH FLOOR INSTALL CEILING GRID 41
534 34 FIFTH FLOOR INSTALL CEILING LIGHTING 21
560 60 FIFTH FLOOR INSTALL CASEWORK 42
562 62 FIFTH FLOOR INSTALL CASEWORK ELECTRIC 44
570 70 FIFTH FLOOR INSTALL PLUMBING FIXTURES 48
5EQ EQ FIFTH FLOOR INSTALL EQUIPMENT 21

inactive Sheet 1 of 1 TYPE OF WORK

Plot Date 1JUN05
APPLY FLOORING
Data Date 8JAN90 Summary Bar/Early Dates
Critical Designator
LOTZ CONSTRUCTION INSTALL CEILING GRID Date Revision Checked Approved
Project Start 8JAN90 Milestone/Flag Activity
Project Finish 9APR90 GRADUATE HOSPITAL DIAGNOSTIC BLDG INSTALL CEILING LIGHTING
INSTALL CASEWORK ELECTRICAL
9E07 GRADUATE HOSPITAL DIAGNOSTIC INSTALL PLUMBING FIXTURES
(c) Primavera Systems, Inc. INSTALL EQUIPMENT

CODE TOTL
VALUE TASK FLR /TASK FLT
020 20 LOWER LEVEL INSTALL STUDS 23
022 22 LOWER LEVEL INSTALL DRYWALL 43
024 24 LOWER LEVEL TAPE DRYWALL 43
026 26 LOWER LEVEL APPLY FLOORING 18
030 30 LOWER LEVEL INSTALL CEILING GRID 16
034 34 LOWER LEVEL INSTALL CEILING LIGHTING 17
060 60 LOWER LEVEL INSTALL CASEWORK 44
062 62 LOWER LEVEL INSTALL CASEWORK ELECTRIC 46
070 70 LOWER LEVEL INSTALL PLUMBING FIXTURES 48
0EQ EQ LOWER LEVEL INSTALL EQUIPMENT 16
120 20 FIRST FLOOR INSTALL STUDS 25
122 22 FIRST FLOOR INSTALL DRYWALL 28
124 24 FIRST FLOOR TAPE DRYWALL 33
126 26 FIRST FLOOR APPLY FLOORING 35

Figure 31.13.3 The network has 3503 detail activities, but the owner sees only 58 sum-
mary activities.

31.14. Interstate 76, 202 & 422 Interchange

American Infrastructure (“AI”) is an ENR top-400 heavy/highway and
water/wastewater construction company with offices in Pennsylvania,
Maryland and Virginia. Back in the 1980s, AI’s Pennsylvania arm and
founding entity which continues to be known as Allan A. Myers Inc.
(“AAM,”) began using CPM for large public projects being awarded,
quickly determined that CPM was an investment rather than a cost and
made a corporate decision to require a CPM on all major projects (and
eventually on all of its work down to subdivision development.)
To prepare the CPM schedules for its works, AAM hired several con-
sultants providing a majority of work to one, but maintaining a work-
ing relationship with backups. AAM also hired its prime consultant to
provide classes in CPM and P3 software, eventually working out a deal
with Drexel University to provide college credits for such instruction.
 Case Histories 509

Very quickly, AAM began to handle the updates for its projects, utiliz-
ing the consultant only for the initial setup and presentation of the
schedule to the owner and its engineer at pre-construction meetings.
As part of AI’s standard hiring practice, new engineers move through
a training program in several departments including estimating and
projects. As part of one six month module of the program, new engineers
were assigned to work on CPM schedules and updates in collaboration
with the consultant. In addition, AAM requested the consultant to design
a master project resource roll-up scheduling system that anticipated the
functionality of P3e/c (and AAM later became one of the first companies
to migrate from P3 to P3e/c.) Implementation of the system required
matching P3 with the relational database capability of Microsoft Access
but was standardized to be maintained entirely by the existing employ-
ees and new engineers going through the training program.
The decision to move from outside consultant to an in-house sched-
uling team was made and AAM chose to attack the problem not by
starting small but on a “give it all you have” basis. The project, even-
tually growing to four separate phases of a Pennsylvania Department
of Transportation project totaling more than $200 million between
2001 through 2003, was the reconstruction of Routes 76, 202 and 422
located in suburban Philadelphia. The project also faced several major
scheduling constraints including the relocation of an Amtrak bridge
over the Schuykill Expressway (I76) and the maintenance of traffic in
one of the most heavily congested areas in the Philadelphia region. The
newly formed scheduling team was given the challenge and began the
preparation of the detail CPM required for this endeavor.
Each Friday, the consultant was asked to come in and look over the
work prepared during the past week. Constructive criticism was accepted
and ideas exchanged on how to proceed. And then, after several weeks
the CPM was done, presented to and accepted by PennDOT and the
AAM scheduling team was off and running. The project encountered
challenges including persistent problems with sinkholes, but with the
partnership of the owner, engineer, AAM and its subcontractors even-
tually went on to complete the project successfully and profitably and
to win several awards, including a Pennsylvania Governor’s Award of
Excellence.
And the scheduling team continued to plan and schedule other high-
way work as it was awarded. Over the next couple of years, the con-
sultant would be called in to review major changes to the 76/202/422
project, other highway projects as they were won by AAM and for other
questions (including over the conversion to P3e/c) while the consultant
continued to assist project managers on water and wastewater facilities.
Finally, the AAM scheduling team felt comfortable with going beyond
highway work and took over the more complex scheduling effort for
510 The Practice of CPM Scheduling

these more complex projects. The consultant continues to provide train-

ing (through Drexel) every few years as another AI “class” takes the
scheduling module but otherwise no longer constitutes an expense. And
American Infrastructure continues to grow and move up the ranks of
the ENR 400.

31.15. CPM Preparation Time

Experience in time versus size of networks cannot be considered a
definitive guide to how long it might take to prepare a network. Nor
does quantity ensure quality. However, the following case histories are
useful as a reference:

Case A: NASA missile launch site utility system

Cost: $20 million
Construction time: 6 months
Client: Contractor (pre-bid) who was concerned about the short con-
struction period and the high liquidated damages ($5,000 per day)
Planning approach: Executive (contractors, estimator, project engineer,
and CPM consultant)
Results: A network of 900 arrows. The preparation for the compu-
tation phase took about 70 hours of team time. However, this partic-
ular network was deliberately condensed in portions so that workforce
studies could be applied. In perhaps 80 hours, a 1600-activity net-
work could have been developed.

Case B: construction of a new hospital and demolition of the old hospital

Cost: $15 million
Construction time: 24 months
Client: General contractor who wanted a good construction schedule
to ensure on-time completion.
Planning approach: Executive (contractors, superintendent, project
engineer, and CPM consultant)
Results: About 104 team hours were used in preparing a 1200-arrow
network.

Case C: high-rise apartment (40 stories)

Cost: $30 million
Construction time: 14 months
Client: Owner-builder
 Case Histories 511

Planning approach: Executive (owner’s assistant, superintendent,

and CPM consultant)
Results:1500-arrow diagram was completed in 5 weeks. The dia-
gram for the 40 similar floors was based on a detailed study of a typ-
ical floor. This typical floor had 150 activities and connections.
Multiplied by 40 floors, this could easily have been converted into a
6000-arrow network.

Case D: high school

Cost: $15 million
Construction time: 16 months
Client: School board
Planning approach: Executive (CPM consultant worked directly with
the major contractors’ superintendents to prepare networks)
Results: The 1200-arrow diagram was prepared in 4 weeks.

Case E: oil refinery turnaround

Cost: $2 million
Downtime: 7 weeks
Client: Refinery
Planning approach: Conference (group of about 10 people when dis-
cussing boiler overhaul; the group was reduced to 6 when discussing
more routine overhaul work)
Results: The 1200-arrow diagram took about 3 team-weeks to prepare.
The careful planning resulted in less equipment downtime.

Case F: unique process facility

Cost: $2 Billion plus
Owner: Department of Energy
Contractor: Design-Construct-Operate
Results: Project is approaching the end of design. A state-of-the art
Primavera based schedule including all planning, engineering, archi-
tectural, governmental, and construction activities is at 22,000 activi-
ties. Activities presently assigned to construction on a summary basis
are 4000 activities. As between 25 and 45 contracts will bid on this pro-
gram, the anticipated number of activities added to this network will
approach 30,000 activities.
Even though P3e/c can handle this scope, at some point after the bid-
ding, it will be appropriate to record, and split off the preceding historical
block of about 15,000 activities, and establish a new, principally con-
struction and startup window.
512 The Practice of CPM Scheduling

Case G: Port Authority of NY and NJ

The Port Authority has encouraged the use of network based schedules
since the 1960s. (See underground JFK fuel system and also JFK rede-
velopment earlier in this chapter.) In 2004, it was announced that the Port
Authority of NY and NJ won the Primavera Excellence Award (2003) for
its implementation of Primavera’s P3 System to support the Port’s 5-year
$9 Billion Capital Program. Pradip M Mehta, Project Controls Manager
for the PA Engineering Department headed by Frank Lombardi (who
had been part of the JFK Redevelopment Project Management System),
lead the implementation of the new Primavera based management system
for the Engineering Department. Two recent projects completed by the PA
after the 9/11 cleanup: opening the PATH Station at the WTC Site, and
the JFK Air Train connecting to JFK Airport.

Case H: Maricopa County Prison Project

Cost: $500 million
Owner: Maricopa County, AZ
Program Manager: Hunt-Jacobs JV
Using Primavera and demanding rigid adherence to the scheduling
specifications developed with ENPROMAC provided Maricopa County
the tools needed to design, contract for, and build over 500 million dol-
lars in their Jail Expansion Program. The specifications required a
detailed, cost-loaded CPM within 30 days of notice to proceed. The
schedule was the tool used for both the pay application process, sched-
ule progressing process and delay claims analysis.
The result of adherence to the specification was virtually no “end of
project” delay claims. Contractors were able to obtain schedule extensions
when they were due. The owner wrote change orders, including sched-
ule extensions, at a minimum of once a month. The county was able to
accurately forecast the overall program schedule and program cost and
completed the Jail Expansion Program on time and under budget.
Key elements to schedule success and claims avoidance were:
Contractors following the delay notice provisions, the owner requiring
recovery and mitigation plans, and the project teams quantifying and
analyzing potential delays at the time of their occurrence. By early iden-
tification of potential delays, the team was able to mitigate delay and
make decisions based upon the current schedule.
By both the owner and the contractors adhering to the contract
schedule specifications, although the County attorneys were looking
for claims, there were very few to be had on this Jail Expansion
Program.
 Case Histories 513

31.16. Summary
The size of a useful network is almost unlimited. Network analysis is
usually a must in projects valued over $5 million, but it can also be
used in less expensive projects. CPM often exposes undefined planning
factors. The trend is to apply CPM after the award of contract, but this
is not a hard and fast rule. Network analysis done prior to award of con-
tract can provide better construction schedule requirements, and CPM
can also be applied profitably after construction work has started.
Phase 1 of the network preparation is collecting information and the
concurrent preparation of a rough diagram. The information collection can
be made by any of four approaches: conference, executive, consultant, or
staff planning. The second phase of the network preparation is the
rearrangement and redrawing of the rough version into a smooth form.
In any approach, it is vital that the plan reflect the real plans of the
contractor. Subcontractors perform many critical work functions. Their
information must also be incorporated in the network.
It is difficult to set definite time requirements for the preparation of
a network. Familiar projects can be diagrammed more quickly than
unfamiliar ones and noncomplex projects more quickly than complex
ones.
CPM seems to require more time than traditional planning but only
because, with CPM techniques, planning is done in more depth.
This page intentionally left blank
 Chapter

32
Additional Exercises for
Students of Project Controls

32.1. Sample Problems

1 Draft the following network in ADM format:

1. B & C follow A
2. D& E follow C
3. F follows B & D
4. G follows E & F

2 Draft the following network in ADM format:

1. B & C follow A
2. D follows B
3. E & F follow C
4. G follows D & E & F

3 Draft the following network in ADM format:

1. B & C follow A
2. D & E follow B
3. E follows C
4. F follows D & E

4 Draft the following network in ADM format:

1. M & R follow T
2. N & B & F follow R
3. A follows N
4. E follows B
5. F precedes G
6. C follows A & M & E
7. D follows A & G
8. E precedes L

515

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
516 The Practice of CPM Scheduling

9. H & J follow C & D

10. K follows H & J & L

5 Solve by the Matrix Method

A B D F
1 2 4 5 6
5 6 4 5

C E
3
7 3

i/j 1 2 3 4 5 6 TEi max(TEi + Dij) i−j dur ES EF LS LF TF

1
2
3
4
5
6
TLj
min(TLj − Dij)

TEi

TLj

6 Solve by the Intuitive Method

A B C D
1 2 3 5 6
10 EF ES 10 EF ES 10 EF ES 10 EF
EF

E ES F
4
ES 11 EF ES 9 EF

7 Solve by the Computer Simulated Method

act i j dur ES EF LS LF TF
A 1 2 5
B 2 3 6
C 2 4 5
D 3 5 4
E 4 5 3
F 5 6 5
 Additional Exercises for Students of Project Controls 517

32.2. Navigating the Enclosed CD-ROM

Enclosed at the back of this text, you will find a CD-ROM disk which has
been specially prepared by Primavera Software Systems at the request
of the authors. This CD includes demonstration versions of three soft-
ware products, files for the John Doe project example used throughout
the text, a PDF file of the various screen shot figures from this text in
full color, a PDF file of the source code for expanded P3 diagnostics dis-
cussed in Chapter 25, and the two sample CPM specifications discussed
in Chapter 33. The three software products are:

■ the classic P3 Primavera Project Planner,

■ the latest flagship for small contractors, Primavera Construction and
■ MCA Monte Carlo Analyzer—the just-being-released small version of
Pertmaster risk software designed for use with P3e/c and Primavera
Construction.

The three programs are for demonstration purposes only and each
has some limits. The P3 software is limited to 60 activities. Thus, if you
desire to read the John Doe project files, you will be limited to the
JDOE file which has been reduced to only 49 activities (leaving room
for the student to add activities while learning the software.) Should
you desire to review the full 181 activities of the DP00 As-Plan file, or
DPAB As-Built file, both (along with the JDOE file) in “mock-ADM”
format, you will need to use a fully licensed copy of P3 or SureTrak, or
you will need to load and use the Primavera Construction software on
the CD. You may practice with the P3 demonstration software as long
as you desire, and even use it for very small projects, up to a practical
limit of $500,000.
The Primavera Construction software is the same that can be pur-
chased from Primavera and can handle up to 750 activities. Once you
are comfortable using the classic P3 demonstration software, you
should consider moving up to this product. You can use this to run real
projects of significant size; the authors suggest a practical project
limit of $7 million due to the 750 activity limit. However, this demon-
stration software will only work for 90 days. At that time you will have
to decide if you wish to purchase the software. Details on purchase of
a license to convert the 90-day demo to the full product are supplied
on the CD.
The MCA or Monte Carlo Analyzer by Pertmaster demonstration soft-
ware is meant as a learning tool rather than for practical use, and is lim-
ited to approximately 25 activities. It is not limited by time. The MCA
software is specifically designed to work with Primavera Construction
518 The Practice of CPM Scheduling

and may be used to further explore the concepts of risk discussed in

Chapter 38.
To load onto your computer any one of the three software products,
insert the CD to your computer, Browse or Explore to your choice of
P3, Primavera Construction or MCA and click the icon for setup.exe.
This will lead you through copying the appropriate program to your
computer. (Figure 32.2.1.)
To view the John Doe project files, after loading the software, use the
P3 Tools Restore command to explode the condensed file to their 23
component files for each project, as shown in Figure 32.2.2. Remember,
you can open only the JDOE file with the demonstration copy of P3. Once
the condensed John Doe files have been exploded, they can be imported
to the Primavera Construction software.
The four John Doe project files are:
■ JDOE – 49 activity subset of the project, in “mock-ADM” for teaching
purposes
■ DP00 – 181 activity network, in “mock-ADM,” mimicking the CPM as
created in chapter 18
■ DPAB – 181 activity network, in “mock-ADM,” showing progress to
project completion
■ DOE0 – 155 activity network, in true PDM format, to illustrate the
conversion of ADM to PDM and subsequent reduction in activities as
“dummy activities” used as logic restraints are removed.

Figure 32.2.1 Browsing the CD.

 Additional Exercises for Students of Project Controls 519

Figure 32.2.2 Restoring the John Doe project files from the CD.

A common complaint of the previous edition was that computer screen

shots that are reduced to fit the printed page are difficult to read and
that the loss of color to the black, white and grey of the printed page fur-
ther reduced the usefulness of such figures. Therefore, all figures involv-
ing computer screen shots have been saved to a file, Edition6Figures.pdf
and saved to the CD. To view, Browse to “Full Color Screen Shots” and
click the PDF.
Finally, to save readers the trouble of rekeying material meant to be
copied, three additional files are stored as PDF files. These include the
sample specifications of Appendix A and Appendix B, and the extended
diagnostic software source code.
520 The Practice of CPM Scheduling

32.3. Answers to Sample Problems

1 Draft the following network in ADM format:

1. B & C follow A
2. D& E follow C
3. F follows B & D
4. G follows E & F

B A is 1–2
B is 2–4
C is 2–3
A D F G D is 3–4
1 2 4 5 6
E is 3–5
F is 4–5
C G is 5–6
3

2 Draft the following network in ADM format:

1. B & C follow A
2. D follows B
3. E & F follow C
4. G follows D & E & F
B D
3

A E G
1 2 6 7

C
4

F
5

3 Draft the following network in ADM format:

1. B & C follow A
2. D & E follow B
3. E follows C
4. F follows D & E
 Additional Exercises for Students of Project Controls 521

B F
3 5 6

A E
1 2 4

4 Draft the following network in ADM format:

1. M & R follow T
2. N & B & F follow R
3. A follows N
4. E follows B
5. F precedes G
6. C follows A & M & E
7. D follows A & G
8. E precedes L
9. H & J follow C & D
10. K follows H & J & L

M C

T N A L

R B E H K

F G D J

5 Solve by the Matrix Method

A B D F
1 2 4 5 6
5 6 4 5

C E
3
7 3
522 The Practice of CPM Scheduling

i/j 1 2 3 4 5 6 TEi max (TEi + Dij) i−j dur ES EF LS LF TF

1 — 5 0 defined as 0 1–2 5 0 5 0 5 0
2 — — 7 6 5 0+5 2–4 6 5 11 6 12 1
3 — — — 0 3 12 5+7 2–3 7 5 12 5 12 0
4 — — — — 4 12 5 + 6 or 12 + 0 3–4 0 12 12 12 12 0
5 — — — — — 5 16 12 + 3 or 12 + 4 3–5 3 12 15 15 16 1
6 —— —— —— 21 16 + 5 4–5 4 12 16 12 16 0
TLj 0 5 12 12 16 21 5–6 5 16 21 16 21 0

LF – EF or LS – ES
12 – 6 or 12 – 7

16 – 3 or 12 – 0

LF – dur
ES + dur
min(TLj − Dij)

Difined as TEi

TEi

TLj
16 – 4
21 – 5
5–5

6 Solve by the Intuitive Method

A B C D
1 2 3 5 6
0 10 10 10 10 20 21 10 31 31 10 41
0 10 11 21 21 21 31 31 41
21

E 21 21 F
4
10 11 21 21 9 30
10 21 22 31

7 Solve by the Computer Simulated Method

act i j dur ES EF LS LF TF
A 1 2 5 0 5 0 5 0
B 2 3 6 5 11 5 11 0
C 2 4 5 5 10 7 12 2
D 3 5 4 11 15 11 15 0
E 4 5 3 10 13 12 15 2
F 5 6 5 15 20 15 20 0
 Part

Advanced Topics
6

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
This page intentionally left blank
 Chapter

33
Specifying CPM

Specification of a process such as CPM is usually the owner’s responsibility,

usually delegated to either the design professional or the construction
manager. If CPM is not a contract requirement, but a contractor elects to
use it, selection/specification of the CPM approach is the contractor’s
option. Many contractors have selected the option of having CPM con-
sulting assistance. In that event, they need a means to describe the scope
of the services required (ie a specification.)
The specifying of the ADM process produced predictable results. Today,
the response is, “we are using Primavera and WYSIWYG applies” (ie What
You See Is What You Get.) This chapter explores optimum ways to spec-
ify the PDM scheduling process. More importantly, it explains why spe-
cific language should be used since it is paramount that the engineer
should understand the specification and not merely enforce terms copied
from a guideline specification.

33.1. Attorney’s Viewpoint to Writing a CPM Specification

From an owner’s viewpoint, two objectives should be accomplished in a
specification requiring a submission. The first is that the submission
assists the contractor to accomplish the end product desired by the owner.
The second, as important, is that the submission be made in a format
which may be readily reviewed by the owner.
The traditional ADM system accomplished both objectives. It forced the
contractor to logically address the planning needs of a project and thus to
schedule the work in a fashion most likely to achieve completion on time.
The simplicity of the system made the job of the reviewer easy. Logic was
easily followed in pure logic drawings, in barcharts, in logic notated

525

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
526 Advanced Topics

barcharts and in time scaled logic diagrams. Logic was also easily fol-
lowed in tabular printouts as the logic is encoded in I-J activity notation.
However, the PDM system has the advantage that, if used properly,
relationships difficult to depict in ADM, and difficult to update during
the course of the project, may be better depicted using the powerful non-
traditional lead/lag relationships available. The PDM system may assist
the contractor to make a model closer to reality than may be made with
ADM. And this means that, used properly, the schedule calculated is
even more likely to assist the contractor to achieve timely completion.
The downsides are that the additional power can be misused by the con-
tractor to avoid the need to logically address the planning needs of a proj-
ect. The contractor may choose to merely guess when various activities
should be performed, line such up in a barchart format and link such
together with any type of logic restraint relationships that holds the bar-
chart together for the initial submission printout.
And the initial submission looks like a million dollars. In multicolor
graphic format or tabular format it is definitely the output of a sophisti-
cated computer program – and must be correct. But the reviewer must
beware because the schedule submittal may be a charade. Logic rela-
tionships are often not shown. If logic connectors are shown, they do not
readily indicate the type of relationship or existence of lags, and are often
placed so closely together as to make visual review impossible. Additional
powerful features in both ADM and PDM systems, such as the ability to
assign artificial constraints to an activity without clear notation on the
output, further reduces the need of the contractor to sit down and clearly
think through the logic of the project.
Therefore, the job of the specifier is to permit the contractor to utilize
the more powerful features of PDM and modern software programs to
better model the real world, but at the same time continue to force the
contractor to perform the basic planning required for the preparation
of a CPM using the older ADM system.

33.2. Pure Logic Drawing

The first step in preparation of such a specification is to require the pure
logic diagram which is the hallmark of proper planning, and which has
been discussed throughout this text. While it is desired that the pure logic
drawing be prepared prior to entry into the computer, such may be diffi-
cult to enforce. A requirement that the pure logic diagram be hand drawn
is an easy fix, but it is the function and not the form that counts. Logic rela-
tionship lines should be clearly and easily followed, possibly of a minimum
length and separation from other lines, and notated if other than the tra-
ditional Finish-to-Start with no lag. A further suggestion is to require
 Specifying CPM 527

notation if the purpose of the logic relationship is to allocate resources

(“crew logic”) as opposed to a physical requirement. (see Figure 22.6.2)
The size of the diagram should be chosen to allow for easy review by the
owner or its engineer. Since it is desired that this drawing be prepared as
a planning tool, calculated schedule information (e.g. early and late dates)
should not appear thereon. For such information, this drawing may be aug-
mented by a separate barchart or time scaled logic diagram.
To permit proper use but minimize abuse of the powerful features of
PDM, non-traditional relationships should be required to be highlighted
and the need therefore, on an individual basis, explained in an attached
narrative. Hand drawn (or similar) fragnets expressing the detail sum-
marized by the non-traditional relationships should be required as part
of these narratives. Similarly, artificial constraints should be explained
on an individual basis. If multiple calendars are used, they should be
clearly marked and a copy of the features of each calendar attached along
with an explanation of its features and need therefore. Many specifica-
tions require a copy of the calendar – page upon page – when what is really
desired is a tabulation of the basis (such as M-Tu-W-Th-F) and exceptions
(holidays and other non-work days, and special weekends on which work
is to be performed.) The algorithm of the software chosen by the con-
tractor should be explained. This requirement may be skipped if the
owner specifies the software to be used, but where options within the soft-
ware exist (such as continuous or interruptible activities,) the options
taken should be noted and explained.
In summary, if the CPM is to be a tool which is to assist the entire proj-
ect team run the project in the most expeditious fashion, limits on the con-
tractor’s use of the most modern technology should not be mandated.
However, the rigors required by the old ADM system should not be relaxed
and each extension from that system should be explained.

33.3. Content of the Logic Network

So much for the technical issues of the CPM submittal. The content of
the CPM submittal must also be considered. It has already been noted
as useful and is here again suggested that each activity be keyed to at
least one contract drawing or specification number that best illustrates
the scope of the activity. If the engineer or owner desires to be able to
summarize the status of work by location or function, it is important to
specify such a code be assigned to each activity—but at the same time
remembering that such coding requirements be subordinate to the axiom
that each activity be a “set of instructions . . . to one responsible fore-
man . . .” Thus if a wall that would be poured monolithically runs
between two area codes, it should still be reported as one activity—the
528 Advanced Topics

accuracy of the coding structure and not the activity content must be
degraded as necessary.
A good specification should also require that all assumptions relating
to durations of activities and between activities be recorded. This may
require duplicate recording of resources – once at the nominal tolerance
for supporting the duration and once at the close tolerance for cost engi-
neering and estimating purposes. (And a third record at an exact toler-
ance if the system is to be used for accounting purposes.) Recording the
assumptions behind lags is difficult in most commercial software sys-
tems as code, note and log fields are by activity and not by restraint.
However, a specification can and should require a separate listing, per-
haps as an Excel spreadsheet file, providing this information.
Equally important is what the specification should not include. The
CPM specification is a requirement for a shop drawing. It does not, and is
not meant to specify how the contractor is to perform its scope of work or
even how to schedule its work. The Spearin doctrine (United States v.
Spearin (1918), 248 US 132, 63 L Ed 166, 39 S Ct 59) provides that if an
owner tells a contractor how to perform its scope of work, and the contractor
carefully follows those instructions, that the owner will be responsible for
any failures without regard to any other language in the contract that says
otherwise. Thus, if a contractor was by contract to guarantee that a base-
ment wall be watertight, and then by that same contract erect the wall,
apply waterproofing and carefully backfill the wall all in accordance with
detailed instructions by the specification, stringently enforced by an inspec-
tor, the owner will have no recourse under the guarantee if the wall should
leak (MacKnight Flintic Stone Company v. Mayor, Alderman and
Commonalty of the City of New York (1899), 160 N.Y. 73, 54 N.E. 661.) The
engineer must be similarly careful with coordinating the CPM specifica-
tion with what is desired of the contractor.
For example, some specifications in current use require that the con-
tractor’s CPM show that the contractor intends to complete the project
exactly upon and not earlier than the stipulated project completion date.
This is very much akin to a specification requiring epoxy coated rebar, but
also requiring the contractor to submit shop drawings showing stainless
steel will be used. The specification does not require the contractor to
actually remain on the site after all work is complete, and it is doubtful
that any judge or factfinder would believe that the owner understood this
to be the contract’s or contractor’s intent. As noted in Chapter 22, adher-
ence to such a specification may greatly increase the probability that the
project will be late as the foremen, subcontractors, vendors and inspectors
all work towards published schedule dates that are later than the prudent
contractor would plan and each uses some of the contingency reserved for
extraordinary events. When these events occur late in the project, it is too
late to recover.
 Specifying CPM 529

Another example where a CPM submittal may be misused is where an

engineer prohibits the contractor from starting work on an activity prior
to its published early start date per the CPM. The mathematics of CPM
actually indicate that the activity merely has a probability (usually less
than 50%) of starting on or before the date printed as the early start and
not that the contractor should not attempt to beat this date. Other than
the normal requirements of notice to the engineer (so that necessary
inspection staff can be allocated,) the contractor should be permitted to
work wherever he pleases. The engineer should, however, document and
place the contractor on notice if resources are being diverted away from
critical path work to work off of the critical path.

33.4. Updates and Revisions

Many specifications get confused when discussing updates and revisions.
Instead of discussing the technical and content aspects of the CPM Update
or CPM Revision submittal, the specification attempts to duplicate or
override the contract provisions dealing with payment and claims. In
many cases the result is a hodge-podge of partially concurring and par-
tially conflicting instructions on these two important areas of the contract.
If progress payments require the submittal of a CPM update, as well as
several other submittals of documentation, such fact belongs in the
Progress Payment section of the contract documents. The format and con-
tent of the CPM update submittal are the only subjects that should be dis-
cussed in the specification section for the CPM. Similarly, if the process
by which a contractor should submit a claim requires a CPM analysis, as
well as several other submittals of documentation, such fact belongs in the
Disputes section of the contract documents with the specification on CPM
merely noting the required format and content of such analysis.
However, a word of warning is called for here. Should a contract clause
state that resolution of a contractor’s claim will be based upon the read-
ing of “the entrails of a goat” or other obtuse “analysis” favoring the owner,
a court may honor such clause, or it may choose to ignore the clause or it
may choose to punish the party promoting the clause.
The specification should require that updates and revisions be kept sep-
arate. The inclusion of Claims Digger in the latest Primavera software, or
similar software, makes for easy enforcement of such a restriction. Updates
should only include information on progress during the past reporting
period and the specification should require special notation where work
completed several reporting periods ago is suddenly discovered and is
being reported with this update. Where the contractor is required to or
desires to submit a revision, such should be for the same DataDate as the
last update, and a new update should first be performed before the revi-
sion if necessary.
530 Advanced Topics

A revision should be required if the contractor chooses to perform the

scope of the project to a significantly different “plan of execution” than orig-
inally envisioned. However, the mere fact that the contractor is perform-
ing work out-of-sequence is not, of itself, indicative of a decision to alter
the “plan of execution.” In many, if not most such cases, the contractor may
claim the alternate plan was forced upon it due to events outside of its
control, including those claimed caused by the engineer or owner. Thus,
as part of the narrative to a revision, it is important to require the con-
tractor to elucidate the reasons for such a change to its “plan of execution.”
Neither the engineer nor owner has to agree with the given reason, but
it is important to get such stated at the time of the event.
It is not too much to demand, as part of the CPM specification, that if
the lead project manager or superintendent is replaced for any reason, the
replacement be required to certify in writing completion of a full review
and revision as deemed desired of the remainder of the As-Planned logic
network and schedule (or most recent revision thereof) of a minimum of
eight, sixteen or even twenty-four contact hours depending upon the size
and complexity of the project. There may be other major events within the
responsibility of the contractor that similarly call for a revision. However,
in all likelihood, the submission of a revision will be concurrent with
claims by the contractor to a change order.
A revision should be required if the contractor claims that a change in
condition or change order will significantly impact its “plan of execution.”
If responsibility for the change is arguably that of the owner, the change
order section of the contract documents should require the proposed revi-
sion be submitted with the change order request for payment. The cost
of preparing the revision is part of the cost of the change order and should
be chargeable as such. (Clauses stating otherwise may be deemed
inequitable and subject to being deemed unenforceable by a court.)
The preparation of a full revision will also assist the contractor to deter-
mine the incidental and consequential impact of a change order upon all
of its subcontractors, and allow for a complete price tag to be assigned the
change. While the owner may not wish to remind the contractor to add
yet more costs to the change order, this process is more likely to lead to
complete closure of the issue. Obviously, if there are multiple proposed
change orders for the same time period pending, this process may be
compromised and call for yet more professionalism by the Engineer.

33.5. Standard References

The easiest way to specify the use of CPM would be to invoke a standard
reference, such as one authored by the American National Standards
Institute (ANSI). However, ANSI has not issued a CPM standard as of the
 Specifying CPM 531

date of this publication. Over a period of many years, two ANSI sub com-
mittees addressed the question of a CPM standard and some draft mate-
rial was released. The British counterpart of ANSI has published a
standard on networks. The usefulness of this published material is lim-
ited as a reference for a specification because it is so even handed that all
options are left open.
One book traditionally used to describe or specify CPM used by general
contractors was published by the Associated General Contractors of
America (AGC) as CPM in Construction: A Manual for General
Contractors. This book, first published in 1965, was written by Glenn
L. White. It provides a clear description of CPM theory and technique but
stops short of describing methods and philosophies of application of the
technique. However, the advent of PDM and the specifying of CPM by
owners desiring an owner oriented, rather than AGC oriented viewpoint,
has led to the supplanting of the book by newer specifications tailored to
the needs of such owners.
A standard which has evolved over the past 40 years is the U.S. Army
Corps of Engineers reference regulation ER-1-1-11, “Network Analysis
System,” which was used into the late 1990s as the Corps reference stan-
dard when it specified the use of network scheduling. ER-1-1-11 describes
network theory and technique; it leaves the method of application to the
particular project specification.
Another standard which has evolved over the years is the Network
Analysis Schedules (NAS) section of the General Conditions prepared
by the U.S. Navy Naval Facilities Command (NAVFAC.) The first joint
effort of the authors of this text was in the total rewrite of this specifi-
cation in 1987. Harkening back to an era where sole source specifica-
tions were discouraged, a good portion of the specification was devoted
to having the contractor submit its files in a format readable by
Primavera without requiring the contractor to use that one vendor.
(Figures 33.5.1 and 2)

2.1.4 Software: It is intended that the information submitted in computer readable

form may, but need not, be used by the Government to verify Contractor submissions.
The Government intends, but is not committed, to use PRIMAVERA PROJECT
MANAGEMENT AND CONTROL software to implement this review, if performed,
and said software may be used to prepare all reports (hard copy and computer
readable.) However, other software systems may also have this capability, and if
desired, Contractor may use other software packages, or perform the mathematical
analysis required by hand calculation, provided that the submissions are made in the
format required by this specification.

Figure 33.5.1 Excerpt from 1987 specification.

532 Advanced Topics

FILENAME: PROJ02
ALL RECORDS or LINES
COL 1- 5: key “ACTVA”
COL 12-15: key Activity Number of activity 0001 to 9999
COL 16: key “O” (letter “O”)
COL 17-20: key number of work days duration of activity (right justified, integers only)
COL 25-71: key description of activity
COL 72-73: imposed constraint code
COL 74-80: imposed constraint date

Figure 33.5.2 Excerpt from 1987 specification for format of submitted ASCII files in lieu
of Primavera.

This standard, too, has evolved over the years and is now a standard
used by several governmental entities and known as the “Unified
Facilities Guide Specification” which can be customized by the user.
The base document is 27 pages long and is included as Appendix B.

33.6. Sanctions
The discussion of specifications is a proper place to identify the actions
that will be taken if the CPM schedule and methods are not properly
applied. The most common sanction is a refusal to make progress pay-
ments unless the CPM schedule has been submitted and approved or to
limit progress payments to the first 3 months or some other reasonable
time frame.
An example taken from the Dade County specifications for its metro
system spelled out the following sanctions:
FAILURE TO SUBMIT NETWORK ANALYSIS: Failure of the Contractor
to submit the network analysis or any required revisions thereto within
the time limits stated, shall be sufficient cause for certification that the
Contractor is not performing the Work required by this Section, or that
the Contractor’s personnel directly responsible for planning, scheduling,
and maintaining progress of the Work are not performing their work in
a proper and skillful manner, or both. The Engineer may withhold
approval of the Contractor’s invoices for progress payment until such
delinquent submittal is made.

Dade County in its general contracts has one of the strongest sections
on sanctions used to date by anyone or any organization:

A. The Contractor shall prosecute the Work in accordance with the latest
approved network analysis. In the event that the progress of items along
the critical path is delayed, the Contractor shall revise his planning to
include additional forces, equipment, shifts or hours as necessary to meet
 Specifying CPM 533

the time or times of completion specified in this Contract. Additional costs

resulting therefrom will be borne by the Contractor. The Contractor shall
make such changes when his progress at any check period does not meet
at least one of the following two tests:
1. The percentage of dollar value of completed work with respect to the
total amount of the Contract is within ten percentage points of the per-
centage of the contract time elapsed, or
2. The percentage of dollar value of completed work is within ten per-
centage points of the dollar value which should have been performed accord-
ing to the Contractor’s own network analysis previously approved by the
Engineer.
B. Failure of the Contractor to comply with the requirements under this
provision will be grounds for determination that the Contractor is not pros-
ecuting the work with such diligence as will ensure completion within the
time of completion specified in this Contract. Upon such determination,
MDC may terminate the Contractor’s right to proceed with the Work, or
any separate part thereof . . .

Of course, the provision of onerous sanctions does not mean that such
can or will be used. In one instance when performing work for the U.S.
Navy on a project, the contractor simply refused to provide a CPM sub-
mittal in accordance with the contract documents. The commanding offi-
cer complained that the contractor made it known that if payment was
stopped, work would stop, and all understood the urgent need for the proj-
ect was such that there was little real threat of termination. What was rec-
ommended by the consultant was increasing the retainage from 10 percent
to 20 percent until such time as the CPM was properly submitted and
employing and back charging for a third-party consultant to warn of poten-
tial delay issues. A nuclear bomb is not always effective; sometimes you
just have to send in the Marines.

33.7. Sample CPM Specification

A sample CPM specification is provided in Appendix A for the purposes
of assisting the engineer in preparation of the specification for the next
project. As with any canned guide specification, it is necessary to care-
fully review the contents and modify the contents to fit the specific proj-
ect and mesh with other provisions of the contract documents. For
example, for a highway project, the language at Sections 1.3.3.1 and
2.1.13.8, relating to building system and codes for floor, are certainly out
of place. Similarly, the provisions relating to cost loading are designed
to mesh with specific language in a separate section of progress pay-
ments. Finally, the laws of the jurisdiction where the specification may
be enforced must be reviewed.
534 Advanced Topics

33.8. Summary
CPM as a theory needs a specification to bring it into contractual reality.
The availability of an acceptable reference standard can make this easier.
Currently, there are no ANSI or ASTM standard references to fill this
role. The 1965 AGC book can fill the role evenhandedly, but the 1976 AGC
effort does so less evenhandedly.
The balance of the chapter provides examples of various modes of sched-
uling specifications that could be used with minimal changes to a project.
 Chapter

34
CPM in Claims and Litigation

34.1. Introduction
An important function of scheduling in the construction industry, for
both the owner and those doing the construction, is to evaluate claims
based on failure to meet schedules. CPM can affect claims in two ways.

1. It establishes a realistic schedule through CPM planning, which can

furnish a legal basis for the enforcement of damages.
2. Perhaps even more important, it can be used to evaluate actual
claims through the reconstruction of a project’s history or the use of
an existing CPM plan to indicate the effects of changes on the orig-
inal schedule.

In one instance, a contractor, a consortium, was asked by a bridge

authority to show why it should not be pressed for $550,000 in liquidated
damages. The authority believed that the contractor had done a good
job, but because of the public trust involved, it felt that it needed tan-
gible proof of good performance.
In response, the contractor used a construction as-built CPM plan to
demonstrate the effects of three different unforeseen circumstances:
unusually bad weather, loss of special equipment by fire, and time lost
in doing work claimed as extra. The presentation demonstrated the
combined effect of the three causes (which was less than the serial
effect) and the effects of any one or two of them alone and together.
Thus, if any one or two of the factors had been deemed non-excusable,
the effect of the remaining factor or factors was still quantified. On the
basis of the finite presentation, the bridge commission did not press for
the liquidated damages.

535

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
536 Advanced Topics

In a complex multimillion dollar suit and counter-suit the owner, an

airport authority, used a detailed as-built CPM to realistically evaluate
the overall effects of the changes that both the owner and the contrac-
tor had imposed on the project. The network, set up on a historical
basis, could be run to consider the combined effect of the changes as well
as the separate effects of individual changes.
Information from daily, weekly, and monthly field reports was used
to prepare the historical CPM network. The calculated results were
invaluable to the owner’s engineer for preparing a factual testimony. The
pretrial and trial periods extended over a number of years, and with-
out the historical network, factual testimony would have become almost
impossible.
In negotiating extra work, contractors often neglect the effects a
change order will have on working time, so they either request no time
extension or an extension equaling the total period they estimate the
additional work will require. However, extra work on a project usually
affects float areas, and any time extension granted should be less than
the total incremental time needed to complete the additional work.
At Cape Canaveral, the combined emphasis on time and public pres-
sure to complete projects reversed this situation. Contractors recog-
nized more clearly the time–money relations and usually made
substantial requests for additional time as well as for extra money to
implement changes. The Corps of Engineers and NASA required net-
work analysis for the basic work on most of the major projects under-
taken. Thus, most of the contractors prepared network-oriented fragnets
to demonstrate the effects that additional work would have on sched-
uling. There were abuses, but in the long run, CPM was used fairly by
both parties to evaluate requests for time extensions, and many claims
were settled without the drudgery of formal suits.
Also at Cape Canaveral, a new type of claim evolved: a claim for accel-
eration charges. Contractors would often accept extra work items and
agree to perform them in the originally allotted time span. To balance
the obvious inequity of additional work but no time extensions, a fee for
work acceleration would be charged to compensate for the costs of over-
time and other problems that arose, such as inefficiencies generated by
overstaffing particular areas of work.
The type of contract originally signed impacts whether there is a
potential for easy resolution or settlement of claims should they arise.
Construction management and negotiated contract claims in the private
sector can often be resolved by an objective report based on schedules
and other factual information. Objective evaluation is important not
only in regard to the legalities of the settlement proceedings, but as doc-
umentation for proving to both plaintiff and defendant that a proper set-
tlement has been reached. Claims in the public sector are usually not
 CPM in Claims and Litigation 537

so easily settled, however, and an increasing number of disputes are run-

ning the full course of litigation.

34.2. Evidentiary Use of CPM

During the 1960s, CPM schedulers, technicians, and engineers antic-
ipated that the critical path method would be used as a tool in con-
struction claims and litigation at some time. In fact, as early as
1963–1964, consultants to the litigants on both sides of a case involv-
ing the Atomic Energy Commission used CPM to prepare their posi-
tions, although a case citation is not available, and no wide exposition
of the results was made.
In the 1970s, 1980s, and 1990s, CPM techniques were often used in
presenting and defending delay claims cases. In no case in which OKA
was involved was the use of CPM questioned by opposing counsel or the
court. Some of the cases include the following (dates are approximate):

■ IBM vs. Henry Beck Construction, Federal Court, Florida, 1973.

■ Somers Construction vs. H.H. Robertson, arbitration, Philadelphia,
1973.
■ E.C. Ernst vs. City of Philadelphia, Eastern Federal District Court,
Philadelphia, 1976.
■ Arundel vs. Philadelphia Port Corp., Commonwealth court,
Pennsylvania, 1979.
■ Buckley vs. New York City, New York State Court, 1982.
■ Kidde-Briscoe vs. University of Connecticut, Connecticut State Court,
1980–1982.
■ Keating vs. City of Philadelphia, Eastern District Court, Philadelphia,
1981.
■ Glasgow vs. Commonwealth of Pennsylvania, Commonwealth Board
of Claims, 1982–1983.
■ PT & L Construction vs. NJDOT, New Jersey State Court, 1983.
■ I. DuPont Hospital vs. Gilbane, et al., mini-trial, Delaware, 1985.
■ White Oak Construction vs. Connecticut, arbitration, Hartford, 1987.
■ G.E. Environmental Systems vs. Chevron, arbitration, Philadelphia,
1988.
■ Santa Fe Construction vs. U.S. Navy, Armed Services Board of Claim,
Alexandria, Virginia 1989.
■ Shoemaker-Driscoll vs. Smith Kline Beckman, mini-trial, Philadelphia,
1989.
538 Advanced Topics

■ Mergentime vs. Washington Metropolitan Area Transit Authority, U.S.

District Court, Washington, DC, 1992, Retried 2001.
■ Cris Tech vs. Joint Meeting (EUC), arbitration, Somerset, New Jersey,
1993.
■ Brooks Construction vs. British Petroleum, Commonwealth Court,
Philadelphia, 1995
■ Newmont Gold Co. vs Lurgie, et al., Arbitration under International
rules, NYC, NY, 1998

In many more OKA cases entered and en route to trial, CPM was a
factor in settlement. In the early 1970s, several lawyers researched
the question of CPM as an evidentiary tool. A series of articles and pre-
sentations followed, a number of which used the same thread, start-
ing with the article, “The Use of Critical Path Method Techniques in
Contract Claims,” by Jon M. Wickwire and Richard F. Smith, in the
Public Contract Law Journal of October 1974. Extracts from that
article follow:
Judicial acceptance of CPM analyses as persuasive evidence of delay and
disruption has been slow to develop, primarily due to technical errors in the
analysis submitted or a failure of a presentation to realistically portray the
work as actually done. In spite of the early reluctance to accept CPM pre-
sentations, the current state of the law is that use of CPM schedules to prove
construction contract claims has become the standard, rather than the
exception. Scheduling techniques which cannot display activity interrela-
tionships are not favorably regarded as evidence of delay and disruption.
In Minmar Builders, Inc., GSBCA, 3430, 72-2 BCA ¶ 9599 (1972) the
General Services Administration Board of Contract Appeals commented
upon Minmar Builder’s construction schedules (bar charts) which were
offered to show project completion delay due to government’s failure to
timely issue ceiling change instructions:
Although two of Appellant’s construction schedules were introduced in
evidence, one which had been approved by the government and one which
had not, neither was anything more than a bar chart showing the duration
and projected calendar dates for the performance of the various contrac-
tual tasks. Since no interrelationship was shown as between the tasks the
charts cannot show what project activities were dependent on the prior per-
formance of the plaster and ceiling work, much less whether overall proj-
ect completion was thereby affected. In short, the schedules were not
prepared by the Critical Path Method (CPM) and hence are not probative
as to whether any particular activity or group of activities was on the crit-
ical path or constituted the pacing element for the project.
The greatest difficulty encountered by contractors using CPM techniques
in claim presentation is the requirement for the presentation to be thoroughly
grounded in the project records. The failure of contractors to properly docu-
ment CPM studies has been held controlling in many board decisions . . . .
 CPM in Claims and Litigation 539

Guidelines for the use of CPM presentations were set forth in the General
Services Administration Board of Contract Appeals decision in Joseph E.
Bennett Co. (GSBCA 2362, 72-1 BCA ¶ 9364 (1972)) which . . . affirms the
need to properly update a CPM and support the study with accurate
records. The contractor’s claim in this appeal was founded on a letter from
the contracting officer ordering completion of the work by the contract
completion date. The contractor argued this requirement was an acceler-
ation order, which was denied by the contracting officer because of a lack
of meaningful evidence. The contracting officer rejected the accuracy of
the contractor’s critical path method construction plan on the basis of
errors in the interrelationships of activities.
At the board, the appellant presented a computer analysis of the CPM
used on the project to isolate the delays caused by government activities.
The board held that the usefulness of this analysis was dependent upon
three things: 1) the extent to which the individual delays are established
by substantial evidence-this requirement is concerned with the project
records and evidence available for the appellant to show the underlying
causes of delay; 2) the soundness of the CPM system itself—this requires
the contractor to demonstrate the logic of the CPM and show that its the-
oretical and scheduling analyses are sound; and 3) the nature of and reason
for any changes to the CPM schedule in the process of reducing it to a com-
puter program—this relates to the exactness and accuracy with which the
appellant has reduced the CPM network to a computer analysis and how
effectively this analysis can be used in a claim presentation.
As expected, the appellant in Bennett argued that the CPM was the
proper basis for any analysis of the project since the plan was submitted
by the appellant and approved by the government.
However, the board rejected the appellant’s CPM analysis because it: 1)
contained numerous mathematical errors; 2) failed to consider foreseeable
weather conditions; 3) changed the critical path and float times without
reason; and 4) was prepared without the benefit of any site investigation
and after the project was already completed . . . .
The gradual acceptance of CPM presentations when properly documented
is demonstrated in the case of Continental Consolidated Corp. ENG BCA
2743, 2766, 67-2 BCA ¶ 6624 (1967) . . . .
In this case a claim was submitted for extra costs due to suspension of
work and subsequent acceleration directed by the government. The appel-
lant alleged it was entitled to time extensions due to government delay in
approving shop drawings. The government’s failure to grant time exten-
sions for these delays made the work appear to be behind schedule as of
certain dates when in fact, if proper time extensions had been granted, the
appellant would have been on schedule. As a result, government directives
to work overtime and/or extra shifts would have been unnecessary . . . .
The contract set completion dates for various elements of the work which
in effect required a critical path for each element within an overall work
plan. With the use of the appellant’s CPM analysis, the board was able to
separate out the delay costs due appellant and the additional costs incurred
due to a compensable acceleration order. This evidentiary tool allowed the
540 Advanced Topics

board to identify the periods of delay and actual progress on the job and
thereby determine when an acceleration order was properly issued from
that point in time when such an order was compensable because the con-
tractor was back on schedule.
Thus the boards have recognized the value of a CPM developed contem-
poraneously with the work or subsequent to the work so long as it is based
upon the relevant records available. The records may include daily logs, time
sheets, payroll records, diaries, purchase orders. While the boards have
accepted the CPM as an evidentiary tool, this tool cannot rise above the
basic assumptions and records upon which it is founded. The board can
accept the theoretical value of a CPM presentation, but reject its conclusion
for failure to base the analysis on the actual project records. (See C. H.
Leavell & Co., GSBCA 2901, 70-2 BCA ¶ 8437 (1970); 70-2 BCA ¶ 8528 (1970)
[on reconsideration] where the contractor failed to establish the accuracy of
the input data for its computer analysis of delays due to design deficiencies.)
Where the board has received persuasive evidence that the CPM network
is either logically or factually inaccurate, incomplete or prepared specifi-
cally for the claim, the board will discount its evidentiary value. A CPM
must be linked to the job records, as a CPM analysis is primarily con-
cerned with visually portraying the job records to establish the cause of
delay or disruption.
The extent to which a CPM presentation may be used to document a
claim can be seen in Canon Construction Co. (ASBCA 16142, 72-1 BCA ¶
9404 1972) where the contractor gained total acceptance of its CPM sched-
ule to establish a delay claim. In this opinion, the board recognized the
underlying logic and evidence presented in the appellant’s original CPM
schedule and the value of CPM techniques to prove extended overhead
costs.
In Canon, the contractor was awarded his overhead costs determined by
the difference between the actual date of completion and the date the con-
tractor would have completed the work absent government fault and per-
formance of changed work. But the recovery of extended overhead costs was
held to be limited by either the extended period of performance time or the
aggregate net extent of delays caused by government fault or change work,
whichever was the lesser. Using this formula the board recognized that the
contractor was not entitled to recovery for the group of excusable but non-
compensable delays including weather delays, reasonable suspensions of
work, etc. . . .
The Canon decision is extremely important since it shows that a prop-
erly prepared and presented CPM schedule will be accepted by the board
as the basis for computing project delays. In this regard it is noted that the
board clearly indicated that it was “relying principally on the CPM chart
and only using the witness’ testimony to ascribe an aspect of reasonable-
ness to the chart.”
The Canon decision is also significant since it provided further guidance
as to the application of CPM principles to claims. For example, the board
acknowledged that delays incurred off the critical path would not delay
 CPM in Claims and Litigation 541

ultimate performance. Further, the board found that where the sequence
established by the network was violated, costly start and stop operations
would result and implied that the contractor’s planned network operations
need not be the only way to accomplish the work shown, but must be shown
to be economical in both cost and time. (Reference: Stagg Construction Co.,
GSBCA, 2644, 69-2 BCA ¶ 8241 (1970) [on reconsideration]).

In 1975, coauthors Paul J. Walstad, Jon M. Wickwire, Thomas Asselin,

and Joseph H. Kasimer wrote a book titled Project Scheduling and
Construction Claims, a Practical Handbook, which is published by A.
James Waldron Enterprises. On page 14-1, the authors note:
There was reluctance at first to accept the use of CPM analysis as evidence
of delays and disruption. Of paramount concern were possible technical
errors in the system or a failure of the system or analysis to realistically
portray the work as actually done. See e.g., A. Teichert & Sons, Inc., ASBCA
No. 10265, 68-2 BCA ¶ 7151 (1968) . . . .
This concern no doubt stemmed from early presentations which based
CPM analysis to a great extent on speculation, inferences, or innuendo
rather than hard, documented facts. Thus, even though the CPM has
become recognized as a competent source of evidence . . . its usefulness in
providing a claim has been held dependent upon at least four factors:

1. The soundness of the CPM schedule itself . . . . This requires proof of the
reasonableness and feasibility of the schedule so as to show that on a
theoretical basis the scheduling was sound;
2. The extent to which any individual delays can be established by sub-
stantial evidence. This goes to the basic records and evidence available
to the claimant to show the underlying causes of delay and disruption;
3. The nature of any changes to the CPM schedule made during the
claim analysis process. This relates to the exactness and accuracy with
which the claimant has analyzed the project scheduling in making his
presentation;
4. Proof that the work sequence shown was the only possible or reasonable
sequence by which the work could be completed on time.

In the late 1970s and early 1980s, Engineering News-Record pre-

sented a series of professional seminars on claims and litigation. Paul
J. Walstad, Esq., has been a leader in the formulation and presentation
of a number of these. The comments on evidentiary value of CPM con-
tinue as previously described. By 1980, Walstad had added the follow-
ing in this regard:∗
In Blackhawk Heating & Plumbing Co., Inc., GSBCA No. 2432, 75-1 BCA,
the contractor claimed 403 days as a result of ductwork design deficiencies.
The Board found the deficiencies were the fault of the Government.
However, the Board indicated the main question was whether the ductwork
542 Advanced Topics

delay had extended contract completion; the Government contended a

delay involving electrical fixtures was the critical item.
In support of its position, the Government produced its own CPM
analysis, which had been prepared after the delays had occurred. The
Government CPM showed the ductwork design problems were not on the
critical path; the activities which the contractor had contended were
delayed actually had “float” time remaining even after the delay was con-
sidered, and the critical path ran through the electrical fixture approval,
delivery and installation cycle.
The Board carefully analyzed the Government’s CPM, and found it . . .
established a sound network diagram and computer run showing just how
the project was actually constructed up to the date of substantial comple-
tion on December 7, 1970 . . . .
After reviewing the delay analysis set forth in the Government CPM, the
Board further concluded it had provided “a sound basis upon which to eval-
uate various project delays.” Based upon the finding the electrical fixture
delay was the factor which delayed ultimate completion, the Board then
proceeded to allocate responsibility for the fixture delays. Upon reconsid-
eration, the Board refused to modify its original decision, indicating the
as-built CPM was the best evidence of delay.
The use of CPM as an evidentiary tool in claims and court proceedings is
not confined to administrative boards. In the Brooks Towers Corporation vs.
Hunkin-Conkey Construction Company, 454 F. 2d 1203 (10th Cir. 1972), the
owner claimed delay damages from the contractor. The Tenth Circuit Court
of Appeals affirmed an award in favor of the contractor, and in so doing
placed great weight on the CPM analysis provided by an expert witness:
“The testimony of Richard N. Green, a Construction Consultant, is cor-
roborative of Ratner’s grant of some 185 days extensions and significant
in relation to the ‘clockwork’ scheduling of work components required to
accomplish the original contract completion schedules. Green’s study took
into consideration the plans and specifications, the computerized Critical
Path Scheduling program, all Bulletins, formal Change Orders, related
correspondence, Daily Progress Report and Monthly Pay Requests. He
computed some 394 days involving requests for extensions. He eliminated
those of an ‘overlapping’ nature and those which were not critical. He did
not consider delays resulting from labor disputes or severe weather con-
ditions. He arrived at a total of 180 days extension of time to which the
Contractor was entitled.”
In its decision of July 18, 1983, the General Services Administration
Board of Contract Appeals (GSA BCA) complained about the misuse of
CPM schedules in a claim by Welch Construction, Inc. Welch filed a claim
for damages as a result of owner delay in the modification of a geological
survey center. When presenting its claim, Welch used CPM diagrams that
purported to present as-planned and as-built schedules. In its opinion,
GSA BCA, denying the claim, stated:
Candor compels us to admit that we may not have figured out what it
was that Appellant thought its exhibits would show. If so, Appellant has
only itself to blame . . . [One] of the surest ways of losing a case for lack of
 CPM in Claims and Litigation 543

proof is submitting complex exhibits to a tryer of facts with no attempt to

explain what they show or how they relate to the other evidence in the record.
The Board believed that the schedules used in presenting the claim ignored
both contractual and actual completion dates.

Courtesy of the Public Contract Law Journal.

Used with permission. Courtesy of A. James Waldron.
Engineering News-Record, “Advanced Course on Construction Claims,”
p. 269, May 1 and 2, 1980. Used with permission, McGraw-Hill: New York,
and Construction Education Management Corporation.
This page intentionally left blank
 Chapter

35
Delay Analysis

The principal dimension measured by schedules is delay. In the past,

delays in construction used to be a mutually accepted condition. Courts,
on occasion, even recognized that delay was a normal situation in the
construction process. Today, however, delay is a very problematic area,
because owners have tighter budgets and contractors staying on a job
longer than planned incur real costs.
When delays occur during construction, the parties involved attempt
to shift the costs that result onto each other. If litigation results after nego-
tiations fail, the lawsuits are between two or more losers, all of whom are
attempting to mitigate their losses. There are no winners in delay.
To the private owner, delay can mean a loss of revenues through the
resulting lack of production facilities and rentable space, as well as through
a continuing dependence on present facilities. To the public owner, it can
mean that a building or facility is not available for use at the proper time.
The service revenues lost through delay can never be recovered.
To the contractor, delay means higher overhead costs resulting from
the longer construction period, higher prices for materials because of
inflation, and escalation costs due to labor cost increases. Further, work-
ing capital and bonding capacity are so tied up that other projects cannot
be undertaken.

35.1. Delay versus Disruption

Another reason why courts traditionally did not recognize delays is the
confusion over terminology, distinguishing delay from disruption. During
the course of the project, not every delay to a specific task is going to
result in a delay to the project. In fact, the majority of delays to specific
tasks will not delay the project as a whole.

545

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
546 Advanced Topics

A court facing competing claims of major delays to electrical instal-

lation due to alleged under manning and millwork installation due to
owner indecision might well throw up its hands. Thus, parties to a dis-
pute may paper the file with numerous claims of such instances for the
very purpose of confusing the court.
The introduction of the CPM process, distinguishing those activities
that are “critical” from those having “float,” has provided the courts with
a new means to segregate the complaints into those causing delay from
those merely causing a disruption to the flow of work. As noted previously,
courts have recognized that a certain level of disruption is to be expected
in the construction process, but also understand that unreasonable levels
of disruption can have their own cost and even contribute to responsibil-
ity for a delay to the project by diverting resources from critical activities.

35.2. Responsibility / Types / Force Majeure

The assignment of responsibility for delay after the fact is often diffi-
cult, and courts have often remarked that delay should be anticipated
in any construction project. Traditionally, the courts have protected
owners more than contractors. In recent years, no-damage-for-delay
clauses have often been enforced in many states, with contractors receiv-
ing only time extensions when delays occurred. However, granting time
extensions evades another owner-oriented remedy for problems con-
nected with delay: liquidated damages. Even when courts are inclined
to consider recovery of damages for owner-caused delays, the burden is
on the contractor to prove active interference on the part of the owner
to receive a favorable decision.
There are four general categories of responsibility:

1. Owner (or owner’s agents) is responsible.

2. Contractor or subcontractors are responsible.
3. Neither contractual party is responsible.
4. Both contractual parties are responsible.

When the owner or owner’s agents has caused the delay, the courts
may find that the language of the contract, in the form of the typical no-
damage-for-delay clause, protects the owner from having to pay dam-
ages but requires a compensatory time extension to protect the
contractor from having to pay liquidated damages. If the owner can be
proved guilty of interfering with the contractor’s progress on the proj-
ect or has committed a breach of contract, however, the contractor can
probably recover damages from the owner.
If the contractor or subcontractors cause the delay, the contract lan-
guage does not generally offer the protection against litigation on the
 Delay Analysis 547

part of the owner to recover damages. If the delay is caused by forces

beyond the control of either party to the contract, the finding generally
is that each party must bear the brunt of its own damages. If both par-
ties to the contract contribute to the delay or cause concurrent delays,
the usual finding is that the delays offset one another. An exception
would be instances in which the damages can be clearly and distinctly
separated.
There are three basic types of delay: classic, concurrent, and serial.
Classic delay occurs when a period of idleness and/or uselessness is
imposed on the contracted-for work. In Grand Investment Co. vs. United
States, 102 Ct. Cl. U.S. 40 (1944), the government issued a stop order
by telegraph to the contractor that resulted in a work stoppage of 109
days. The contractor sued for damages caused by the delay, basing the
suit on a claim of breach of contract.
The court allowed, among other things, damage due to the loss of uti-
lization of equipment on the job site, finding inability to use equipment
on the job site, and stating:
When the government in breach of its contract, in effect, condemned a con-
tractor’s valuable and useful machines for a period of idleness and use-
lessness . . . it should make compensation comparable to what would be
required if it took the machines for use for a temporary period.

Johnson vs. Fenestra, 305 F. 2d 179, 181 (3d Cir. 1962), also involved
a classic delay: Workers were idled by the failure of the general con-
tractor to supply materials. That type of delay, to be legally recognized
as such, must be substantial, involve an essential segment of the work
to be done, and remain a problem for an unreasonable amount of time.
Generally, if two parties claim concurrent delays, the court will not
try to unravel the factors involved and will disallow the claims by both
parties. In United States vs. Citizens and Southern National Bank, 367
F. 2d 473 (1966), a subcontractor was able to show delay damages caused
by the general contractor. However, the general contractor, in turn, was
able to demonstrate that portions of the damages were caused by fac-
tors for which he was not responsible. In the absence of clear evidence
separating the two claims, the court rejected both claims, stating:
As the evidence does not provide any reasonable basis for allocating the
additional costs among those contributing factors, we conclude that the
entire claim should have been rejected.

Similarly, in Lichter vs. Mellon-Stuart, 305 F. 216 (3d Cir. 1962), the
court found that the facts supported evidence of delay imposed on a
subcontractor by a general contractor. It also found that the work had
been delayed by a number of other factors including change orders,
delays caused by other trades, and strikes.
548 Advanced Topics

The subcontractor had based its claim for damages solely on the delay
imposed by the general contractor, and both the trial court and the
appeals court rejected the claim on the basis that:
Even if one could find from the evidence that one or more of the interfer-
ing contingencies was a wrongful act on the part of the defendant, no basis
appears for even an educated guess as to the increased costs . . . due to that
particular breach . . . as distinguished from those causes from which defen-
dant is contractually exempt.

It should be noted that in recent decisions, the courts increasingly

have demonstrated a willingness to allocate responsibility for concur-
rent delays.
Serial delay is a linkage of delays (or sometimes of different causes
of a delay). Thus, the effects of one delay might be amplified by a later
delay. For instance, if an owner’s representative delays reviewing shop
drawings and the resulting delay causes the project to drift into a strike
or a period of severe weather resulting in further delays, a court might
find the owner liable for the total serial delay resulting from the initial
incremental delay.
Force Majeure causes include what are known as “acts of God.” The
general contract usually provides a list of such events: fires, strikes,
earthquakes, tornadoes, floods, and so on. Should such an event occur,
the contract provides for a mutual relief from demands for damages
that are due to delay, and the owner is obligated to provide a reasonable
(usually a day-for-day) time extension.
In the case of weather-related delays, usually only the occurrences
shown to be beyond the average weather conditions expected for the area
based on past records can be considered as a reason for time exten-
sions. That can, however, vary with contract language. A number of
states and cities allow a day-for-day time extension (noncompensable)
for all bad weather.
Many contracts have clauses stating the time extensions for delay
caused by acts of God shall be granted only to the portions of the proj-
ects that are specifically affected by such events. Thus, a severe down-
pour after a site has been graded and drained and the building closed
in may cause no actual delay, so that claims for time extensions because
of it would not be accepted even though it would qualify under other
methods of evaluation as a Force Majeure act.

35.3. As-Planned Logic Network

The first step in preparing an analysis to determine responsibility for
delay to a project is to locate or otherwise acquire the as-planned logic
network. This may be, but often is not, the submitted and approved
baseline schedule. The key factor in determining if a “baseline” is the
 Delay Analysis 549

proper starting point is whether it includes information known after the

start of the project. A proper as-planned logic network will not include
any information known after the start of the project, although there may
be some leniency relating to the period of time between submission of
the bid and NTP.
The rationale for this rule is based upon the concept of contract. In
the rush to bid a project, a contractor rarely has sufficient time to care-
fully plan all the details of how it will perform the work and often does
not have the input from all members of the project team that will super-
vise such work. However, the contractor does anticipate that once a bid
is won, the project team will carefully review the project drawings and
specifications and choose one of many possible “plans of execution” to
effectuate performance in what it deems the most expedient and cost-
efficient manner. This thought pattern, however expressed and recorded,
is the basis of the as-planned logic network.
The CPM as initially submitted may be rejected for real or imagined
flaws. If technical flaws are in the submitted CPM, they must be cor-
rected, and it is not uncommon for there to be several submittals and
rejections over a period of months until a proper CPM may be approved.
However, during this period it is important that neither party use the
approval process to modify the initial “plan of execution” to account for
later unanticipated events. From the viewpoint of a proper analysis, the
maneuvering of the parties over acceptance of the CPM submittal should
be ignored. Thus, an acceptance too hastily made should not bind the
owner and an improper resubmittal required by an owner should not
bind the contractor.
In Edwin J. Dobson, Jr. Inc. vs. Rutgers (157 N.J. Super. 357, 384A.
2d 1121 [1978]), the Court found that the schedule was not complete
enough to use to measure delay until the third update. In Dobson vs.
Rutgers, the Court also held that the schedule does not have to be for-
mally accepted by the owner or agent to be accepted as the basis for delay
analysis.
Once a proper as-planned logic network has been chosen, the CPM can
be useful in establishing the facts and also the intentions of the parties
to a contract. The network can be used by the owner to demonstrate
areas of failure on the part of the contractor, and it can be used by the
contractor to demonstrate points of interference on the part of the owner
or owner’s agents.
A project involving regular (usually monthly) reviews or updates of
the CPM plan should provide a good basis, through the CPM reports,
for evaluating the progress of the work done on it. Unfortunately, many
such projects have only a collection of CPM diagrams and computer
runs to show for the reviews. The CPM reports are far more valuable if
each update is accompanied by a comprehensive narrative. The narra-
tives, which should be normal portions of the project documentation, are
550 Advanced Topics

prepared in the normal order of business and, therefore, can be accepted

later at face value, with due weight given to their origins.

35.4. As-Should-Have-Been CPM Network

While it is best to start with an as-planned logic network, there are sit-
uations where a good as-planned network did not exist or the one used
was flawed or inadequate. In this case, an as-should-have-been net-
work can be produced. Obviously, what is desired here is to recreate the
“plan of execution” envisioned by the project team at the time the work
was starting and not to utilize “Monday morning quarterbacking” to
create a CPM of how the work should have been planned. If the matter
is in dispute, the urge for one party to submit a plan that exaggerates
the impact of faults of the other party and sidesteps its own faults may
be large. Obviously, it is here that the question of the credibility of the
Scheduler will come into play, thus, it is also important to document the
sources of information used to recreate the as-should-have-been to
become the as-planned CPM.
Thus the terminology means “as-should-have-been-submitted” and
not “as-should-have-been-planned.” In the event that the contractor’s
CPM is missing some portion of the scope of work (such as installation
of a pipe below a foundation), this is to be treated similarly as when the
engineer’s drawings are missing an existing condition (such as a buried
pipe) or necessary scope of work requiring a change order. In either
case, the work additional to the contractor’s original plan will be added
to the as-impacted network as described in section 35.9
In some cases, the preparer of the as-should-have-been network has
a bar graph to utilize as a guideline. In other cases other contempora-
neous documents must be used as primary source material, or to bol-
ster and validate interviews with key staff members on how they
planned to perform the project. In one major project, the new Library
of Congress building (James Madison Memorial Library), it was recog-
nized by both the owner, the Architect of the Capitol, and the contrac-
tor, Bateson Construction Co., that there would be delay claims as a
result of certain delay problems in the project. It was mutually agreed
that it would be advantageous to convert the contractual as-planned bar
graph into a CPM network, which would prove more useful in evaluat-
ing the effects of delay impacts.
The contractor’s scheduling consultant, A. James Waldron, converted the
network into a CPM diagram and printout. This was reviewed for the
Architect of the Capitol by O’Brien-Kreitzberg & Associates (OKA) and,
after some adjustments, a mutually agreed upon baseline was stipulated.
The network was useful to both sides in determining the responsibility
 Delay Analysis 551

for delays and the resulting costs. Often, an as-should-have-been network

is more of an uphill situation. If both parties do not agree to a previously
approved as-planned network, whoever produces the as-should-have-
been network must be able to provide a foundation for it and to justify
its use.
In one such application, the New Jersey Department of Transportation
specification had an elaborate narrative description of the sequencing
required for implementing a project. At that time, the state did not use
CPM planning, and the contractor, a major heavy construction con-
tractor, submitted a totally inadequate bar graph that used fewer than
25 activities to describe the work to be accomplished. The contractor also
worked in such a fashion that he produced a large amount of excava-
tion soil, which was to be used on and/or sold to other projects. The eco-
nomic plan made sense, but the logic did not. OKA used experienced
highway engineers to develop an in-depth, as-should-have-been net-
work, which resulted in being 24 sheets long in its logic and made up of
more than 4000 activities. The computer run demonstrated the impro-
priety of the contractor’s initial actions and illustrated a lack of plan-
ning in regard to the project.

35.5. As-Planned Schedule

Emphasis in the prior section has been placed on calling the CPM the
as-planned logic network to distinguish such from the as-planned sched-
ule. The reason for this emphasis on terminology is to stress that it is
the logic that will be the basis for further steps of the analysis and that,
at this point of the analysis, schedule dates are mostly irrelevant. A quick
calculation of the schedule from the as-planned logic network may allow
the team to determine if some of the logic is missing or incorrect. A
schedule showing pouring of concrete in northern latitudes during the
winter may indicate some adjustment is necessary. But the flip side is
not necessarily true, that is, an otherwise good logic network may neg-
lect to include many of the assumptions therein. Bumping the start
date of the CPM by 3 months, rescheduling and reviewing the output
may point out some of this missing logic. So it remains the careful and
detail review of the as-planned logic network to provide assurance that
each activity is preceded by a physical restraint to some physical object,
and is also preceded by a resource restraint for each necessary resource
(crew, equipment, forms, etc.) that will be necessary for a proper analy-
sis. An as-planned schedule may exclude assumptions that may be
ignored so long as the project is going smoothly. An as-planned logic net-
work for analysis of delay requires that these assumptions be stated
within the logic.
552 Advanced Topics

If the contractor manipulates the as planned schedule, it can be dis-

regarded as the basis for comparison with the as-built. In Hensel
Phelps v. U.S. (ASBCA No. 49,270, 99-2), Hensel Phelps, the general
contractor, prepared the CPM. Their mechanical subcontractor esti-
mated 8 work weeks each for the duct line work for 12 process exhaust
fans for a total of 96 weeks. In the CPM schedule, Hensel Phelps
reduced the estimate to 3 weeks each for a total of 36 weeks. The sub-
contractor was not given a copy of the completed CPM as-planned
schedule. The as-planned CPM was rejected as the basis for a delay
claim because the general contractor had manipulated the as-planned
schedule duration estimates.

35.6. As-Built Schedule

When the activities on the as-planned network have been identified,
work can start on an as-built schedule. A copy should be made of the
As-Planned Logic Network and renamed as the As-Built Schedule. The
second schedule should include the same activities as the first, for com-
parison purposes, but be based on actual performance dates. Those
dates are researched from the updates of the original CPM plan, the
progress reports, and any other documentation available. Sparse or
faulty project documentation may make development of an accurate as-
built schedule difficult.
For that reason, CPM updates should plug in actual dates for all activ-
ities as they start and as they are completed. However, it is often the case
that erroneous dates are entered when two activity descriptions are sim-
ilar or when the contractor has performed general condition work not on
the CPM and yet desired to record progress to something. In such cases
it may be necessary to research other contemporaneous project records,
such as daily diaries or job photos, to determine a correct actual start and
finish date. It is important to note where such changes have been made
and to footnote the source of replacement data in such cases.
For a quick review, the as-planned and as-built schedules can be plot-
ted side-by-side to the same time scale for a rough comparison. This review
can assist in highlighting where the two schedules diverge and to help
determine where further research is desired, but it is certainly not suffi-
cient in itself to document a claim as neither the causes for variances nor
the criticality of activities before and after impact are yet determined.
The work involved in preparing the two schedules will vary with the
input information available, its organization, and the information on the
levels of the work provided by the client and/or the client’s attorney. Two
to five people will be needed to work on them over a period of several
months. The work should be under the direction of a CPM scheduling
professional who is qualified to testify in regard to the final products.
 Delay Analysis 553

From Preston-Brady Co. v. U.S. (VABCA Nos. 1892, 1991, 2555 87-1 ),
“a general statement that disruption or impact occurred, absent any
showing through use of updated CPM schedules, Logs or credible and spe-
cific data or testimony, will not suffice to meet that burden.......of prov-
ing the extent of any delay which it claims...This is particularly so where,
as here, the Logs, when contrasted to the as-planned CPM schedule,
show minimal delay to the very trades most directly involved in the
change order at issue.”

35.7. As-Built Logic Network

Emphasis in the previous section was placed on preparing an as-built
schedule and not an as-built logic network. In this phase of the analy-
sis, it is the actual dates of performance that are important and not the
actual reason why one activity is performed before another. Thus, if
work was performed out-of-sequence from the as-planned logic, it is
entirely possible that the as-built schedule will show an activity start-
ing or even finishing before its predecessor.
An as-built logic network would record the actual logic—the “why”
each activity was performed before the next. It may be possible to trace
the actual logic for select portions of the project, but to do so for an
entire project is usually difficult or impossible. Contemporaneous proj-
ect records will rarely provide sufficient detail for such an endeavor. An
example of the detail necessary to prepare a proper as-built logic net-
work might include, as shown in Figure 35.7.1:

Figure 35.7.1 As-Built Logic Network.

554 Advanced Topics

■ As-planned called for forms used for Wall “A” to be used for Wall “B”
■ An RFI is issued relating to a rebar conflict, delaying completion of
Wall “A”
■ Additional forms rented and delivered to site
■ Wall “B” constructed prior to completion of Wall “A”

Since the cause for deviation from plan is often not so straightforward
and the reasons, therefore, rarely are recorded contemporaneously, the
preparation of an as-built logic network for the entire project usually will
involve a large degree of conjecture and subjectivity on the part of the
preparer.

35.8. Causative Factors

Once the as-planned logic network and as-built schedule are completed,
a uniform format for evaluating the causative factors in the delay is now
available. (Even before the completion of the network and schedule, a
separate group under the direction of the scheduling professional can
begin the evaluation.) The identity of most of the causative factors
should be readily apparent, but the specific impact of different factors
may not be obvious.
One of the first areas to be identified is Force Majeure. The most
common being strikes and bad weather. Strikes should be documented
in terms of their length, the remobilization time once they are over, and
the trades and areas of work affected by them. Most contracts provide
for time extensions because of strikes but not for compensation. In the
case of a contractor making a claim, it would be important to be able to
demonstrate that a strike had little or no impact on the critical path of
a project, so that other compensable factors could be shown to be the
cause of the damages being claimed. Conversely, an owner defending
against claims would try to demonstrate that strikes did indeed cause
the delays and other problems were, at worst, concurrent.
Other causative factors include:

1. RFIs, or request for information: claimed by the contractor to the

owner
2. CICs, or change in condition: claimed by the contractor to the owner
3. CORs, or change order requests: presented by the contractor to the
owner
4. PCOs, or proposed change orders: presented by the owner to the
contractor
5. COs, or change orders: signed by the contractor and then the owner
 Delay Analysis 555

6. PROs, or proceed orders: presented by the owner to the contractor

7. CCOs, or constructive change orders: claimed by the contractor to
the owner
8. SWOs, or stop work orders:presented by the owner to the contractor
9. CQCs, or contractor quality control deficiencies noted: presented
by the contractor to the owner
10. ODNs, or owner deficiency notices: claimed by the owner to the
contractor
11. REJs, or rejection of submitted shop drawings: claimed by the owner
to the contractor
12. REWs, or rework: reported by the contractor to distinguish from
baseline productivity

The acronyms may differ from jurisdiction to jurisdiction and the list
could go on to include other causative factors.
Causative factors are evaluated in terms of the specific impact that
they have on the progress of a project. This is done in two ways. First,
a determination is made at what point in the network a particular
causative factor impacted the fieldwork. In addition, when dealing with
changing or modifying the scope of work to be performed, activities that
were preparatory for implementing the change work are identified.
Examples are change order proposals, ordering material, mobilization,
and any other pre-implementation factors. Other examples include dem-
olition of defective work, reordering of material, and remobilization.
A separate evaluation is done for every causative factor in the proj-
ect. In addition to identifying the basic impact each has had on the
plan, the analysis must also identify the times of issue of the individ-
ual causative factor and reason therefore. While a change order issued
in the eleventh month may be tied to the notice-to-proceed when all con-
tract scope should be known and thus override contractor delays per-
haps caused by under-manning, in an individual case, it may merely be
tied to an activity preceding the one impacted when a visual inspection
suggested the change to the owner. In such a case, the delay caused by
the change would not override and be considered partially concurrent
with the previous delays but merely add new delay time to the delays
already encountered.
It is important to try to include all causative factors that may have
impacted the project. During this phase responsibility for the various
causative factors is not assigned. Not only is it important to air what is
initially thought to be your own “dirty laundry,” because the other side
certainly will if you do not, but also in many cases, there may be some
question as to who is the responsible party. For example, a rejected
556 Advanced Topics

shop drawing is typically charged to the contractor who is responsible

for all of its subcontractors and vendors. However, if the engineer’s
rejection is later deemed improper, the delays for resubmittal, review,
and all consequential delays, would be shifted to the owner. Similarly,
poor productivity is usually charged to the contractor, unless such is the
serial effect of shifting work from one season to the next due to causative
factors chargeable to the owner.
Also keep in mind that most causative factors are not going to impact
activities upon the critical path and will not actually cause delay to the
project. However, it will be important to defuse the claims of “but he was
doing that” when presenting the results of the delay analysis.

35.9. As-Impacted Logic Network

Once all of the causative factors have been determined, they should be
applied to another copy of the as-planned logic network. The rules of net-
work development should again be rigorously adhered to, such as not
permitting open ends or entering of actual dates. Where it is known that
a causative event did not occur until a specific date, such may be entered
via a SNET constraint.
The courts are firm that causation must be specifically connected to
the resulting delay: In Titan Mountain States Construction Corp. v.
U.S. (ASBCA Nos.22, 617, 22, 930, 23, 095, 23, 188, 85-1 the Board
found: “A contractor was not entitled to time extensions for delay and
impact allegedly resulting from modifications because his critical path
analysis did not establish a causal relationship between the modifica-
tions and the alleged delays attributable to them.”
In Hoffman Construction Co. of Oregon v. U.S. (40 Fed Cl.184 [1998]):
“proof that the government was the ‘sole proximate cause’ of the delay
entails proof ‘that no concurrent cause would have equally delayed the
contract regardless of the government’s action or inaction’ [Mega Constr.
Co., Inc. v. United states, 29 Fed. Cl. 396, 424 (1993)]. . .”
In Fru-Con Construction Corp. v. U.S. 44 Fed.Cl. (1999) two burdens
of the claimant contractor were described in the decision: (1) “Unless the
Government retains control over the evidence, plaintiff bears the burden
of establishing an excusable delay by a preponderance of the evidence”
and (2) “It is not sufficient to establish that some work was prevented;
the work prevented must be work that will delay the overall completion
of the job.”
In PCL Construction Services v. U.S. (47 Fed. Cl. 745 [2000]), the
court held that the claim must meet three tests: “(1) the extent of the
delay with a reasonable degree of accuracy; (2) that the delay proxi-
mately was caused solely by the government’s actions; and (3) that the
delay caused specific, quantifiable injury to the contractor.”
 Delay Analysis 557

35.10. As-Impacted Schedule

At this time, the as-impacted logic network should be saved to a secure
file name and then (and only then) the scheduling routine applied. The
purpose of this step is to be able to demonstrate that a true evaluation
was performed rather than merely going through the motions to back
into a desired result. At this point, the Scheduler has all of the infor-
mation needed in a format suited to perform an analysis of delay.

35.11. Time Impact Evaluations

When all the causative factors have been identified, a time impact eval-
uation (TIE) is prepared for each factor. The information is assembled
as previously described, and it is prepared in a format so that the
impact of each factor on the as-planned network can be determined and
applied to it.
When the impacts of all the causative factors have been correctly
determined and applied, the result should be an approximation of the
as-built schedule. The as-impacted schedule is then compared with the
as-built one, and any major disparities between them examined to iden-
tify whether or not the TIEs were incorrectly applied or there were
additional causative factors not identified.
The theoretical effects of the impacting factors on the as-planned net-
work must be explainable in terms of the as-built network, otherwise
the proposed analysis is probably incorrect. Some professionals take a
different position, however. One well-known scheduling consultant
expounds the theory of the 500 bolts: If an owner is to provide 500 bolts
and has delivered only 499, in the consultant’s opinion, the activity
involved will be impacted until that last bolt has been delivered. But it
appears more logical to examine the function of the last bolt. For
instance, if the bolt is a spare or there is a readily acceptable substitute
that permits construction to proceed, then it is not, theoretically, proper
to claim that the as-planned network has been impacted by its absence.
Another position often taken by Schedulers who conduct impact analy-
ses on as-planned networks for contractor evaluations, is that all float
belongs to the contractor. This has been a continuing argument in the
profession. In fact, some recent owner’s specifications, in order to coun-
teract such claims, state, “All float belongs to the owner.” Neither posi-
tion is tenable, however.
Float is a shared commodity. Like a natural resource, it must be used
with common sense. The owner should be permitted to use float for
order changes, shop drawing reviews, and other owner-responsible
areas. On the other hand, it is obvious that owners should not use float
to the point that the entire project becomes totally critical. This would
558 Advanced Topics

be an overreach on the part of owners. Conversely, contractors should

be expected to use float only to balance their work forces and to work
efficiently, to complete projects on time and at optimum budgets.
Once all of the TIE information has been imposed on the as-planned
network, a standard CPM calculation is made. The calculation should
correlate, as discussed previously, with the as-built network.

35.12. Zeroing Out

If there are only a few causative factors impacting the as-planned logic
network, it is suggested that the TIEs be selectively zeroed out by cat-
egory. For instance, the Force Majeure changes are zeroed out and a run
is made to determine the overall impact of their absence on the network.
Similarly, contractor-related TIEs are zeroed out, and whatever further
improvement their absence makes in the schedule is noted. Then, the
owner-related TIEs, involving changes and any hold orders, and so on,
are zeroed out, and the final result should bring the network back to its
as-planned status.
Because each category of change is zeroed out step-by-step, the effects
of concurrency can be observed from the results of the three separate
runs. This can provide an arbitrator or a court with the means to allocate
delay damages and impacts caused by the various parties. One of the first
applications of this approach was to a major airport project. The airport
authority had contracted for the installation of a $15 million underground
fueling system. The contractor for the work, who was the low bidder by
several million dollars, prepared a construction CPM plan that was never
accepted by the owner, and all of the milestone dates were completely
missed. The airport authority took under advisement whether to enter
suit for delay damages that were due to losses in interest on money and
in airport operating efficiency, as well as for other direct delay damages.
When the contractor filed a $6 million delay suit against the author-
ity, the authority promptly filed a counterclaim and litigation ensued.
In the absence of a mutually acceptable as-planned CPM, the owner
directed that an as-built CPM be prepared to evaluate the real causes
of the delays. The daily, weekly, and monthly reports, as well as personal
observations by the owner’s field team and the CPM consultant, were
used to develop the comprehensive plan. It contained milestone points
reflecting actual dates of accomplishment for various activities. Between
the milestone points, the estimates for the time that the work should
have taken were inserted, and the CPM team then divided the delay pro-
portionally by its causes. The causes were either by contractor, owner,
combined, or neither.
The first computer run of the network showed the actual dates for
all the events. The next computation established the amount of delay
 Delay Analysis 559

due to the contractor alone. The third established the amount of delay
due to the owner alone. The fourth identified the amount of delay due
to both. But the total actual delay was less than the combined total
when the amounts caused by the owner alone and the contractor alone
were added together.
Using this very specific information, the managing engineer for the
owner was able to facilitate an out-of-court settlement that took more
than a year to negotiate. (Part of the owner’s management team’s will-
ingness to negotiate was because they recognized the very real delays
they caused by a slow shop drawing review. Many of the delays were due
to the high workload of the owner’s engineering department, but many
were caused by the engineers trying to redesign the shop drawing sub-
missions, a common mistake made in the course of reviews.)

35.13. Zeroing to a Collapsed As-Impacted Logic Network

If there are many causative factors or if the determinations of respon-
sibility for such are unclear, zeroing out by category may not be practi-
cal and an alternate means will be required to determine the impact
(if any) of each causative factor. To effectuate this portion of the analy-
sis, it is necessary to prepare a separate spreadsheet or provide addi-
tional activity codes for each activity representing a causative factor. The
additional columns or code fields include:

1. order in which this causative factor is zeroed out,

2. the project completion date prior to this causative factor being zeroed
out,
3. the number of days difference between the project completion date
before and after zeroing out, and
4. the number of days between the activity impacted and project com-
pletion.

The process of zeroing out of causative factors one by one again begins
with scheduling the as-impacted logic network. Next, starting at the last
activity, or project completion, the analyst works backward, tracing the
driving relationships until reaching either a causative factor or the
start of the network. This will constitute the critical path of the TIE.
Often there will be more than one such path either converging to one
causative factor or more than one. Code the one or several causative fac-
tors as “001,” note the finish date for project completion and start date
for each causative factor, calculating and recording the number of days
from the causative factor to project completion.
560 Advanced Topics

Then delete all successors to the one or several causative factors

deemed the roots of the critical path(s.) Next, reschedule the network
and trace the critical paths back to the next root causative factor(s.) Code
as “002,” record the new project completion date, calculate the number
of days between the last recorded completion date and the new one,
note the start date for causative factor(s) and calculate and record the
number of days from the causative factor(s) to project completion.
This process is repeated until reaching the start of the network rather
than a causative factor. This final result should again bring the network
back to its as-planned status. You may note that many of the causative
factors still remain. This matches the reality that a majority of even the
most “serious” causative factors encountered on a project are not the
cause of delay. However inclusion of such in the analysis proves such an
assertion and may defuse quite a bit of argument related thereto.
Where a project has intermediate milestone deadlines that are sub-
ject to actual or liquidated damages, the same approach may be used to
determine time extension entitlement toward individual milestones.
Starting from the milestone and working backward through the driv-
ing relationships leading thereto, the individual root causative factors
can be ascertained until again reaching the starting point of the logic
network. This should ideally be done with a fresh copy of the as-impacted
schedule for each milestone as an individual causative factor may have
a separate impact on various milestones and project completion.
At this point, two more columns may be added to the spreadsheet: one
for the party alleged to be responsible for the causative factor, and the other
to note whether that causative factor was the sole cause for the incre-
mental impact to project completion, or was concurrent with one or more
other causative factors.

35.14. Limitations of the TIE Methodology

Either via grouping by categories or by the more tedious method of zero-
ing out causative factors one by one, the TIE method may be used to
determine entitlement for relief from liability for damages caused by the
delay. The fact that the project completion date of the as-impacted sched-
ule may be later than the actual completion date is irrelevant so long
as the impact to activities immediately successor to causative factors
roughly matches the as-built. In fact, it is more likely than not that the
project completion date of the as-impacted schedule will be later than
the actual completion date since most contractors will make some
attempt to mitigate the impact of various causative factors.
However, these efforts by the contractor to mitigate cannot be used to
reduce the contractor’s entitlement. To some unquantified extent (at least
 Delay Analysis 561

at this point of the analysis), the difference between the calculated

project completion date of the as-impacted schedule and actual project
completion date of the as-built schedule represents the cost of the efforts
of the parties to accelerate work and mitigate the delays incurred by the
project.
On the other hand, the contractor should not be entitled to monetary
compensation for damages that have been avoided by such mitigation.
Thus, while the TIE methodology is conclusive in determining entitle-
ment to avoid the payment of actual or liquidated damages for late com-
pletion of the project or stipulated milestones thereto, it may tend to
overstate entitlement to compensatory damages. To determine the
appropriate number of days for which a contractor may be entitled to
compensatory damages for field and home office overheads associated
with an extended project timeframe, it is necessary to also factor in the
impact of such mitigation. This is discussed in the section on the win-
dows analysis methodology that follows.

35.15. TIE Example of John Doe Project

Take the 34-day CPM plan for the initial portion of the John Doe proj-
ect as a schedule and use it to measure delays or impacts. If, for instance,
the well pump required a 6-week delivery time, the equivalent number
of work days would be 30. The impact area is measured by adding an
activity starting at 0 and going to event 4. The activity would be titled
“late delivery of well pump,” and adding it would produce the result
shown in Figure 35.15.1, the time scale version of the initial part of the
John Doe project. Because the well work was on the critical path, the
delay would force the late start of activity 4–5, install well pump, to
await the delivery of the well pump. In this example, 30 minus 22, or a
delay of 8 working days.
Of course, it is necessary to view the entire contractual universe. For
instance, if there was a 2-week delay in the notice to proceed for rea-
sons other than the pump delivery, then the pump delivery delay would
be better represented by disconnecting the initial, or i end, of the delay
arrow from the 0 event and bringing it into the network as a new start-
ing point with a specified date. Thus, if a 2-week Force Majeure delay
were imposed on the start of the site work, the additional time needed
for delivery of the well pump would become a concurrent delay.
Figure 35.15.2 shows a TIE form describing the delay in the delivery
of the well pump.
Review of the TIE points out a number of additional issues. The
evaluation does not note the original duration for this “stock delivery”
item was 10 days, but does note that the responsibility for the delay
562

Figure 35.15.1 Time scale network showing late delivery of well pump delay.
 Delay Analysis 563

Figure 35.15.2 Time impact evaluation (TIE) describing delay of well pump.

belongs to the A/E for failure to deliver the specification for the pump.
Thus, after leaving a minimum of 5 days for contractor procurement,
this activity would initially have had 7 days of float. The TIE does not
provide when the A/E finally provided the specification (that presum-
ably should have been provided no later than the notice to proceed) and
564 Advanced Topics

thus this must be presumed to have been on day 15 to provide the fol-
lowing sequence:

15 days to provide spec + 5 days to shop among 3 local vendors

+ 10 days to prep and deliver to site = 30 days total

But what if the A/E had taken 22 days to provide the specification but
the contractor had mitigated by ordering the same day the spec was
received with a vendor preselected to expedite the delivery in return for
a slightly higher price? If the TIE had been prepared on that day 22, the
contractor would be entitled to a delay of 8 + 7 = 15 days. Should the
contractor’s entitlement be reduced due to his/her own initiative and
additional expense? What if the A/E had provided the specification
within the 15 days of this example, but the contractor chose to exten-
sively shop for the least expensive vendor, taking 7 days to order the
pump (and prep/delivery taking the normal 10 days)? The delay to the
project would then be 10 days rather than the 8 calculated by the TIE.
Keeping in mind that the contractor initially expected to need 5 days
for shopping, and had an additional 7 days float, will this delay be
charged totally to the owner (who is responsible for the A/E) or split with
the contractor?
Similarly, the TIE notes that the contractor was ready for the pump
on day 22. If the 15 day activity of drilling the well was interrupted by
rain on one day, was it really necessary for the contractor to make up
the lost day by working a Saturday when he/she knew the pump would
be delayed until day 30? The TIE process does not look at these ques-
tions of who knew what when, but rather only at the day the impact is
expected to or did occur.
A second delay, a 2-month delay to delivery of structural steel, further
illustrates this issue, starting with the TIE form in Figure 35.15.3. The
design change was also noted at the commencement of the project as a
result of an RFI generated by the contractor’s efforts to prepare a CPM.
The vendor, in pricing the change, agreed that it would deliver the steel
no later than day 118, but ran into its own production problems and was
not able to provide delivery until day 123.
When the two problems are imposed on the overall network, the
critical path goes through procurement of the structural steel, as
shown in Figure 35.15.4. The owner, knowing he/she would be held
responsible for the delay to the steel, had no reason to rush out the
specifications on the pump. The contractor had no reason to rush the
procurement process, other than to mitigate the disruption to opera-
tions. The well pump was delivered late to the project; however, there
is no impact on the overall project because the late steel delivery takes
precedence.
 Delay Analysis 565

Figure 35.15.3 TIE for 2-month delay in delivery of structural steel.

As may be seen from this example, the TIE methodology alone may
provide some difficulty in sorting out responsibility for multiple delays
and in making adjustments for acceleration by the injured party to mit-
igate the impact of such delays. However, the use of the TIE is defini-
tive in determining the total potential impact of any delaying causative
566

Figure 35.15.4 Time scale network showing steel delay.

 Delay Analysis 567

event and calculating the time extension to which the injured party is
entitled. Thus, to determine the cumulative effect of all delays, all TIEs
should be developed and impacted against the network simultaneously.

35.16. Windows Analysis

Just as the use of CPM allows project personnel to better understand a proj-
ect by breaking large scopes of work down into small activities, the Windows
analysis allows a better understanding of large and overlapping delays. The
purpose of this analysis is to measure the actual impact of various causative
factors upon the progress of the work, as opposed to measuring the theo-
retical impact to the as-planned logic network and “plan of execution” by
which the contractor was entitled to not only a lack of impediment, but an
express obligation of assistance under the general precepts of contract law.
The period of the windows run from the start of the project to the first
significant causative factor, and then from that point to the next signifi-
cant causative factor and on to project completion. It may be possible to
run a window to each causative factor, however, the resultant analysis
may become a day-by-day account of the project if each claimed RFI, CIC,
SWO, or other claimed causative event are included in the analysis. Thus,
just as a Scheduler needs to use some judgment in splitting, combining,
and otherwise defining the scope of individual activities, so must some
judgment be used in setting the timeframe of individual windows.
In applying this judgment, the Scheduler may look to the likelihood
that a specific causative event will have an impact upon the schedule.
This task is greatly simplified if the TIE analysis has been subjected to
a zeroing out analysis with the commencement of the various root
causative factors being the start and end of each window. Further lati-
tude may be provided if a string of short duration windows all appear to
be caused by factors chargeable to one party. A more detailed breakdown
may be necessary if causative factors of differing or questioned respon-
sibility overlap. In a complex delay claim situation, however, with dozens
or hundreds of causative factors, it may be necessary to simply pick a
standard timeframe, such as 1 week or 1 month. Although careful tai-
loring of the window periods will be more accurate, a question of dimin-
ishing returns for the effort required must also be considered.
At this point yet another copy must be made of the as-planned logic
network. If the scheduling software permits, it may be useful to import
to a custom code field the actual dates recorded in the as-built sched-
ule. To this copy should be added only those causative factors that start
within the first window. However, the durations of the causative factors
will have to be reviewed since in this forward-looking analysis, only the
anticipated duration, rather than the actual duration, should be used.
568 Advanced Topics

The activities of the window are then statused to the end of the window
using the actual start and finish dates from the as-built schedule. A spe-
cial problem exists for activities started but not finished during this
update in determining the remaining duration as of the new data date.
If the window conveniently ends on or about the same date as one of
the project updates, it may be possible to extract the remaining duration
reported in that contemporaneously recorded document. However, often
the two data dates will differ or misreporting during the progress update
may make such information less than accurate. (For example, a con-
tractor forgets to report progress on some activities and report comple-
tion of such in a following update.)
The solution that best alleviates this issue is to set the remaining
duration as the lesser of the actual finish minus the data date or origi-
nal duration. Although this computed remaining duration may not be
precisely what the field personnel may have anticipated on the data
date, it will be a close approximation. Obviously, if either contempora-
neous documentation or common sense dictates that reasonable field
personnel would anticipate a larger duration (up to or even greater
than the actual duration experienced), that estimate of remaining dura-
tion should be used, although it should be footnoted appropriately.
An important technical point in entering causative factors is the dura-
tion that will be assigned to such. For purposes of the TIE performed
“after the fact,” the entire and known duration of the causative event
may be added. In many cases, rather than tediously calculating these
durations, they may be entered by means of the expected finish con-
straint, allowing the computer to calculate the number of days from the
known start to finish. If a causative factor, such as a stop work order
pending resolution of a request for information, lingered from week to
week for several months, this is the same as being told “replacement
steel will be delivered in 12 weeks.”
However, when performing a windows-based analysis, the “what was
known and when” issues come into play. In the first case, the duration
should be 1 week since resolution is always expected by next week. In
the second case, the duration should be the full 12 weeks. This may cause
the causative factor of a 1-month window (to match an update cycle) to
calculate an impact greater than the duration of the window.
At this point, the window file should be rescheduled with the new data
date being the end of the window. The impact, if any, from the causative
factors added and from the performance of work upon the baseline net-
work logic, will be calculated. This incremental delay to project com-
pletion should be recorded. Working backward via driving relationships
from project completion will determine if one (or more) of the causative
factors added is the root cause for this incremental delay, or if it is
attributed solely to poor production during this timeframe.
 Delay Analysis 569

35.17. Zeroing Out within the Windows Analysis

If there are numerous causative factors in a specific window of an analy-
sis and there is the possibility that differing parties may be responsible
for overlapping delays, either smaller windows may be utilized for this
timeframe or a zeroing out analysis, as described, may be used.

35.18. Windows Example of John Doe Project

To illustrate an example of the windows methodology, let us modify our
previous example. Instead of the steel problem being determined around
the time of the notice-to-proceed, it was not discovered until day 43
when the building layout was underway and the steel was already par-
tially fabricated. Again, the fabricator agreed as part of the change
order price adjustment to deliver the revised steel by day 118, but did
not make actual delivery until day 123.
The two delays to the project are now distinct, as may be seen in
Figure 35.18.1. The earlier delay of the well pump added the first 8
days to the project. This will be properly charged to the owner. Delivery
of the pump marks the end of the first window. The second window
begins when the steel issue is discovered on day 43. Since all work
during this period went according to schedule, no additional time is
added to the project or charged to either party.

Figure 35.18.1 Overview of relative float created in site work because of late deliveries.
570 Advanced Topics

The third window begins with discovery of the steel issue and con-
tinues until the steel is delivered. The steel design issue adds only an
additional 27 days and not the full 35 days calculated by the TIE since
the first 8 days are now attributed solely to the well pump issue.
Responsibility for this delay is further split to 22 days chargeable to the
owner (and the A/E) and 5 days to the contractor (and the fabricator).

35.19. Summary
The use of CPM in claims and legal cases increased dramatically in the
last three decades as parties to construction contracts have come to
increasingly rely on litigation to settle disputes. The as-planned net-
work, preferably approved by the owner, the contracting officer, or the
construction manager, is key in the claim evaluation process. The best
approach to such evaluation is the time impact evaluation (TIE), which
applies all the delay factors to the as-planned schedule to determine
how they impacted it. If there was no as-planned network or it was inad-
equate, an as-should-have-been network can be substituted based upon
what may be ascertained as the contractor’s original “plan of execution.”
A detailed, as-built network, compressed rather than impacted, can
be used to evaluate a situation if a good as-planned network is not avail-
able, but this approach is highly subjective and subject to challenge. The
as-built network can also be compared with the impacted, as-planned
network, or the impacted, as-should-have-been network, to validate the
evaluation of what impacts the delay factors had. Examples of the
impact approach were given. The John Doe network updates are shown
as the basis for a contractor’s claim and an owner’s defense.
The principal reference document for most of the case references in
this chapter is the omnibus work “Construction Scheduling: Preparation,
Liability, and Claims,” second edition, authors Jon M. Wickwire, Thomas
J. Driscoll, Stephen B. Hurlbut, and Scott B. Hillman, 1103 pages pub-
lished in 2003 by Aspen Publishers and its 271 page supplement pub-
lished in 2005.
 Chapter

36
Disruption Analyses

We emphasized that the term “delay” has a special meaning to a

Scheduler. Activities are not subject to “delay” but rather to “disruption.”
If the disruption is to an activity that is on the current critical path of
the project at the time the disruption occurs, then the project is subjected
to a delay. However, project correspondence and conversation among
field personnel do not normally make such a distinction; even a Scheduler
may say “the concrete pour was delayed by inclement weather.”
The fact is that most of the “delays” on a project, even the ones most
subject to dispute, do not delay the project. Some of the disruptions
directly impact an activity that is currently upon the critical path. Some
disruptions have an indirect impact, such as an emergency extra work
order that diverts resources from an activity on the critical path. The
vast majority of disruptions do not directly or indirectly cause delay.
However, they do have a cost to the project.
We noted that the plan of the as-planned logic network must be that
of person (or persons) directing the work to be performed. We noted
that this plan may be one of many possible plans but it is the one chosen,
based upon both the assumptions recorded and unreported by the
Scheduler, to be the most efficient in terms of time and cost. The proj-
ect manager may choose to work the project from the north to the south
or vice versa. It may be presumed that the choice was not random and
a disruption forcing a reversal will have some cost, even if only 15¢.

36.1. Traditional Methodologies

Traditional methodologies for measuring the cost impact of disruptions
were much like those for measurement of delay prior to the advent of
CPM. Often, a blunderbuss of alleged causative factors (deemed the

571

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
572 Advanced Topics

responsibility of the “other” side) leading to disruptions was fired against

a chart indicating planned versus actual costs. The defense was often
the firing of a similar blunderbuss filled with the alleged causative
factors deemed the responsibility of the initial claimant. Substitute
an as-plan versus as-built bar chart for the cost curves and we have a
pre-CPM delay analysis.
A more refined methodology is to measure the difference between
planned and actual productivity and costs during a period of calm (called
the “measured mile”) and compare it to the same difference during a
period of disruption. Obviously, the circumstantial evidence of this analy-
sis is greater, but still lacks a firm cause and effect relationship.

36.2. CPM Out-of-Sequence Methodology

Use of the as-planned logic network adds a great deal of objectivity and
reproducibility to the disruption analysis. Obviously, project records of
actual costs are never complete enough for the lawyers and forensic
consultants, and even if they were, there is always the question of the
accuracy of the initial estimates (cost and time) for the activities dis-
rupted. However, like the value of CPM in breaking down large “ballpark”
estimates of time into estimates of discrete activities—some high, some
low— that cumulatively are more accurate, a similar value is imparted
by the CPM relating to the estimated costs of discrete activities and the
costs of disruption.
If masonry on the upper level of a two-story structure is stopped while
the owner determines if he/she desires a larger or smaller window open-
ing, it is clear that there will be some additional cost in remobilization
and ramping up to speed after a restart; and if the scaffolding is left
standing, there will be costs associated with the rental of the scaffold.
If the scaffolding is removed to permit other trades to have access to the
interior of the building (thus working out-of-sequence from the as-
planned logic), there will be the additional costs of removal and re-
erection. There may also be additional costs of less-than-complete access
by the other trades working around stacks of block, which is, hopefully
(but is not guaranteed), less than the double handling of the mason
completely clearing the area. Although the quantifying of such costs
may still be an estimate, the use of the measured mile approach at the
activity or task level is less subject to variation and dispute than for
entire areas or time periods of the project.
A project manager who is involved in the preparation of the original
as-planned logic network will truly attempt to meet that schedule. One
of the finest compliments given to a Scheduler is when, after all the work
of preparing the CPM is completed, the project manager says “that is
exactly the way I intend to build the project. What did I need you for?”
 Disruption Analyses 573

If a project manager encounters a disruption that can freeze the proj-

ect in its tracks, he/she may attempt to work around it. However, the
project manager is certainly hoping to minimize the distortion to the
“most efficient plan” and intends to return to the plan as soon as prac-
ticable. Only on the worst of projects, where the project manager is con-
stantly running into roadblocks to the plan and even to the workarounds
put in place, may the plan be abandoned and resources assigned wher-
ever there appears to be a task to perform without interruption.
Thus, it is possible to track the disruptions to the project by analysis
of progress of work performed out-of-sequence from time period to time
period. Building upon the Windows methodology discussed previously,
the Scheduler should note the instances of new and continuing work per-
formed out-of-sequence for each statusing or update. As noted in the
Windows methodology, the DataDates for such updates may be a func-
tion of major causative factors or may be periodic if the number of
causative factors makes such an exercise unwieldy.
Primavera’s P3 Project Planner software’s schedule diagnostic is an
excellent tool for such an analysis. Each time an activity is started or fin-
ished out-of-sequence, it is reported. P3’s diagnostic distinguishes between
eight types of work being performed out-of-sequence (Figure 36.2.1).
The diagnostic allows the Scheduler to distinguish between work per-
formed out-of-sequence that may either indicate a crew getting a head
start on the next activity or the start of a disruption (“Activity started,
predecessor has not finished”) from that more clearly indicating a lin-
gering obstruction (“Activity finished, predecessor has not finished”). Not
included in the Primavera diagnostic but of potential use might be a code

1. “Activity started before its predecessor finished.” – used to note actual dates
reported this period may indicate problem within this period
2. “Activity started, predecessor has not finished.” – used for FS and FF relationships,
with or without lag
3. “Activity started before its predecessor’s lag would allow.” – used for FS and SS
relationships with lag
4. “Activity started, predecessor has not started.” – used for SS relationships, with
or without lag
5. “Activity finished, predecessor has not finished.” – used for FF relationships, with
or without lag
6. “Activity finished before its predecessor’s lag would allow.” – used for FF
relationships with lag
7. “Activity started too early to allow it to finish on or after its predecessor’s
finish.” – used for FF relationships without lag
8. “Activity started too early to allow it to finish after the expiration of its predecessor’s
lag.” – used for FF relationships with lag

Figure 36.2.1 Types of work being performed out-of-sequence.

574 Advanced Topics

for “Activity finished, predecessor not started” to distinguish between

disruptions that skip to the next activity when progress on a started
activity is obstructed from those when an activity in a planned sequence
is skipped over entirely. Similarly, the code “Activity finished, prede-
cessor has not finished” is not issued by the P3 diagnostic for breaching
a FS relationship, although it might be useful for this application.
At this point, the reason for each instance of out-of-sequence per-
formance may now be noted. Reasons can range from “superinten-
dent’s choice” or “equipment failure” to “change in condition discovered”
or “stop work order issued” to “unresolved RFI” or “C.O. pending” to
“too much mud, sent crew elsewhere.” Determining the party respon-
sible for each cause would be next. If a log record (of the reason why a
planned activity was not started or was started and was then stopped)
was not kept, a detail review of the daily diaries of the project may be
required to determine the cause of the disruption. This process is greatly
aided if the previously prepared list of causative factors includes all
and not only the most noteworthy incidents that have occurred.
Having determined a disruption and its impact to a specific activity
or flow of activities, a cost can be assigned to the specific disruption. This
may involve some level of subjectivity, but the level of disagreements
should be small if each side renders an opinion in good faith for these
small amounts. After all, remobilization of a drill rig for one piling ini-
tially skipped because of the discovery of an undocumented pipe is
unlikely to cost either $100,000 or zero. In fact, if there are but a few
such disruptions, a contractor can expect to be laughed out of court as
some reasonable level of disruption is expected in any endeavor.
However, if the level of disruption tips that of reasonableness, the total
of the impact of “1000 bee stings” begins to look like real money. Where
the totals of the disruption analysis are similar to that of traditional
methodologies, a very strong case can be made for compensation.

36.3. Summary
The as-planned logic represents the project manager’s “plan of execu-
tion” and, presumably, the most expedient and cost effective means to
perform the scope of work of the project. When the project manager is
hindered from performing work according to this plan, then additional
costs can be expected, even if the disruptions incurred do not impact the
current critical path of the project. Review of selected updates to deter-
mine which activities were performed out-of-sequence and why can be
used to prepare or defend a claim of disruption.
 Chapter

37
Advanced Topics:
Resource Leveling
and Smoothing

Another extension of the basic CPM methodology is scheduling based

upon resource availability rather than strictly on estimates of durations
of individual activities and specified logic between activities. To deter-
mine the optimal time when each activity should be performed requires
the assignment of resources to each activity and knowledge of the avail-
ability of such resources during each time period of the project.

37.1. Resource Leveling and Smoothing

Activity and resource types

Resource-based scheduling encompasses a large field of possible algo-
rithms yielding a variety of results. In one scenario, durations are not
estimated by the project manager or the scheduler, but the units of
resources required (labor hours) and availability (craftsmen). This is the
default system used by Microsoft Project. Obviously, this mimics the
manual process of determining duration in many cases but is not accept-
able in others.
One example of inaccuracy is the old scheduler’s analogy that if one
woman bears a child in 9 months, can 9 women collaborate to bear a
child in 1 month? Other examples involving a possible division of a
large task into smaller portions that can be performed concurrently rec-
ognize the need to add additional resources for the subtasks of splitting
and recombining the various parts. A traditional scheduler’s trick for util-
ity work is to split a long utility line into two, making a final connection

575

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
576 Advanced Topics

at the middle. Recognized is that there is always an extra cost for the
custom connecting piece and effort of installing.
A less intrusive consideration of resources is to use the project man-
ager’s determination of duration but to recognize that an additional
logic restraint to the start of any activity is the availability of resources.
Thus, an “improved” algorithm could determine availability before
scheduling the activity. However, if two or more activities could start on
a specific date, and only sufficient resources are available for one, which
one will start and which will be deferred? Choice of “who goes first” is
neither intuitive nor subject to mathematical solution.

37.2. Limitations of Algorithms

Nondeterministic, Polynomial-Time-Complete, or NP Complete prob-
lems, mathematically have no known optimal solution. However, numer-
ous “leveling” routines are available that can provide a workable, if not
optimal, solution. The danger lies in allocating the scarce resource to the
wrong activity, resulting in a longer duration than absolutely necessary
for the project. While the misassignment of one resource unit among
several activities may have a limited impact, misassignment of multiple
resource units among several activities can result in resource leveling
algorithms that stretch a project duration to double the optimal solution.
Thus, a scheduler using such a leveling routine should always make sev-
eral runs using differing rules for prioritization of activities to level.
A subspecies of resource leveling is resource smoothing, or setting
limitations on deployment of resources. The objective is not to delay
activities because of a limit in the number of resources, but rather to
defer work (hopefully, on activities having sufficient float) so that fluc-
tuations in resource use (hiring and firing of crafts) can be minimized.
Most project managers and schedulers would agree that a desirable
resource use curve is a slow build-up of crafts, a steady number work-
ing during the majority of the project, and a tapering off at the end
until only a punchlist crew remains. Correcting fluctuations creating
peaks and troughs in resource usage can be accomplished manually by
the scheduling team by use of various constraints and “crew logic”
restraints or by smoothing software routines. Again, discretionary deci-
sions as to which activities should be deferred for smoothing, should be
carefully reviewed by the project manager. After all, it is he or she who
is being paid to make these decisions.

37.3. Driving Resources

A number of considerations must be evaluated in choosing a sched-
ule where activity durations are driven by resource, logic is aug-
mented by desired or maximum levels of resource availability, or logic
 Advanced Topics: Resource Leveling and Smoothing 577

is augmented by the desire to smooth the fluctuation of resource

usage. The first question is determining what resources are to be con-
sidered driving.
A driving resource is one that determines the duration of the activ-
ity. If two units of a driving resource can complete a task in six days,
then three can complete in four days. The relationship may be linear,
as in this example, or nonlinear as discussed below. Categories include
craftsmen (total or by union craft), equipment, supervisory limitations,
access, and any other physical restriction on performance.
Determining desired and maximum levels of availability must next
be made. For example, on a project where both high steel rigging and
work below the rigging could be performed concurrently except for safety
considerations, a “safety access” resource could be assigned each activ-
ity. In this case, a total of one unit “available” can be used to permit a
software leveling routine to determine which should “go first.”
The next question is how such resources might impact the perform-
ance of an activity. If an activity requires multiple limited resources, the
most limited of such will usually determine the duration. However, if
two or more resources are required for an activity, but only one is
required at any one time, each portion of the activity being performed
independently of the other, a different algorithm is required to determine
the activity duration.
An example is the activity “Write Software,” where two individuals
are responsible for different sections. The vacation schedules and other
responsibilities of each individual would not impact the other from com-
pleting his or her portion of the work. Obviously, this type of “activity”
should, in fact, be broken into two concurrent activities with a possible
third “coordination” activity at the end. However, for reporting pur-
poses, managers often prefer activities involving concurrent responsi-
bility, and successful software vendors will accommodate such practices
regardless of correctness.

37.4. Resource Calendars

The next question is specifying the limits of resource availability. A sim-
plistic approach is to specify a set limit for the number of resource units
available even though availability may vary over time. More student
interns may be available in the summer than in the winter. Without the
need for formal multi-project scheduling, the expected completion of
one project might release resources for another.
Taking this to a higher level of detail, resources can have their own
calendar of availability. Thus, an activity that may be required to be per-
formed on weekdays, but require one resource be available only the first
10 days of the month and another resource never available on Fridays.
578 Advanced Topics

When multiple calendars are invoked, the question becomes how to

reconcile or combine them. Microsoft Project combines various calendars
excluding all nonwork periods for both the activity and the resource.
Primavera’s resource calendar overrides the activity calendar. Thus, an
activity requiring the independent use of multiple resources could have
one of the resources working on Saturdays, whereas the others may be
limited to the standard five-day workweek. (A better solution from a
theory viewpoint would be to split the activity into two or more con-
current tasks, each having its own activity calendar, converging to a
“coordination” activity or milestone at the end.)
A problem with this approach is that the resource calendar, being
derived from an activity calendar, is limited to use on activities with that
activity calendar. Thus, a resource calendar for the resource previously
theorized as being available only on the first 10 days of the month could
not be applied to two activities; one on a 5-day workweek and the other
on a 6-day workweek.
After determining resource availability, resource usage must be
addressed. One area of divergence may be the assumption of linear, or
constant, usage of a resource versus an expectation of ramp-up, pro-
duction, and taper-off usage. Various software’s treatment of resource
usage curves could cause additional divergence in calculated results.

37.5. Practical Solutions

Finally, great care must be exercised in setting the standards of prior-
itization. Given the choice of “which one goes first,” a project manager
will use a number of factors, including intuition, or other factors that
he cannot express. One project manager client would use three com-
peting factors, none of which is normally encoded in the network. For
rigging process plant equipment onto a previously prepared pad, he
would: (1) rig first the equipment that required the longest reach of the
crane, (2) rig first the equipment that was the heaviest, and (3) rig first
the equipment that had platforms or other equipment rigged above. If
there was a conflict among these directives, he would use his intuition.
However, he was directing the operation of only one or two crane crews
at a time.
If a contractor must choose between 5 of 20 possible “which ones go
first,” or if the project includes hundreds of such decisions, the method-
ology must be determined early in the scheduling process so that the nec-
essary coding fields can be included and information acquired and input.
In the example just given, despite the use of three conflicting sets of rules,
the scheduler rarely had to inquire of the project manager’s “intuition.”
Other technical rules may also be used to determine prioritization. It
would seem obvious that activities having the least amount of total
 Advanced Topics: Resource Leveling and Smoothing 579

float should go first. However, this can occasionally backfire. Primavera’s

default requires that the network first be scheduled using traditional
means, then level on the next available activity with the earliest late
start, then total float, then the activity number. Other activity attrib-
utes that can result in an earlier completion of a leveled project include
prioritizing free float and independent float and those activities that use
the scarcest resources.
However, as noted previously, no matter how carefully a project man-
ager or scheduler chooses the order of priority, there is always the
chance that a better solution could be found. (One example is a small,
10-activity network that takes longer when leveled with three crews
than when leveled with two crews.) The practical solution would be to
merge leveling routines with generalized evaluation review technique
(GERT) software.
After running a large network with many activities competing for
multiple resources of 500 or 1000 iterations, a software routine could
choose a prioritization that results in the shortest project duration.
However, you must realize that the 1001st iteration could have deter-
mined a better solution and that as actual progress is made on the proj-
ect, the optimal solution changes.

37.6. Summary of Resource Leveling

The impact of limited resources and limitations on the deployment of
resources are inherent in every schedule a project manager prepares.
As projects become larger, automation of the structuring of these impacts
is often desired. Software solutions should be used with care, however,
because they can yield unintended results. Also, theoretical limitations
of such algorithms can result in project durations much larger than an
optimal solution. The final caveat is to the reviewer of tabular or graphic
output: Basic CPM calculations that have been modified by such algo-
rithms might not necessarily be noted or apparent.
This page intentionally left blank
 Chapter

Advanced Topics:
38
PERT, SPERT, and GERT

As previously noted, the quality of a CPM network and schedule derived

depend on the care taken in choosing appropriate restraints (or prede-
cessors and successors) and in estimating the duration of individual
tasks. Typically, the level of detail for individual activities is such that
the project manager or scheduler is comfortable estimating a duration
with some degree of accuracy. Sometimes, however, the scope of work
to be performed for the individual task is fuzzy or outside factors do not
permit a reasonable degree of comfort in specifying a set duration.
If the researchers were very lucky, they might locate the right alloy
in an optimistic period of time. If they were unlucky, they might take a
pessimistic period of time. Based on the law of averages and the expe-
rience of the research team, they could specify a most likely period of
time required.
These three estimates, the Optimistic, Most-Likely, and Pessimistic
durations, created the basis for statistically determining the range of
durations that could be experienced. Unfortunately, in 1958, computers
were not fast enough nor had sufficient memory to perform the true sta-
tistical analysis that represented the mathematical model for which
these estimates were collected. Instead, a rough average was made
using the formula

O + 4M + P
6
to reduce the information to a level similar to that used for CPM analy-
sis and subsequent calculations. As computers became more powerful, var-
ious programs were developed, mostly in academia, for demonstrating the

581

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
582 Advanced Topics

power of full implementation of a Statistical or multiple Simulation PERT

under various acronyms, such as SPERT.

38.1. PERT
For example, a research and development (R&D) project, the Navy’s
development of the Polaris missile, contained a parallel development to
the critical path method of schedule analysis—the Navy’s Performance
Evaluation and Review Technique, or PERT. Unlike the CPM system in
which the scope of individual activities could be reasonably quantified
leading to an estimate in labor hours and finally in working days, the
Polaris program had much wider guesstimates. Considering time con-
straints, the researchers could not test all possible alloys for a rocket
nozzle, but must continue testing until a suitable (if not optimal) alloy
and configuration could be found.

38.2. SPERT
Here, allowing the duration to vary among Optimistic, Most-Likely, and
Pessimistic via a random number generator and running the resultant
network for a requisite number of times (such as 100 iterations), the
reviewer could be confident of the estimated duration of the project, notwith-
standing the large variances estimated for the individual activities.
The original “fudge” of using the formula

O + 4M + P
6
led many practitioners to a serious misunderstanding of the mathe-
matics behind PERT. One such example might be the use of statistical
analyses used to quantify the risk inherent in a bid estimate for the
purpose of quantifying the risk of meeting a schedule.
Suppose, similar to PERT, an optimistic, most likely, and pessimistic
estimated cost is assigned to each element of a bid estimate. The aver-
age of these three estimates would be

O+ M+P
3

while the mean would be

O + 4M + P
6
 Advanced Topics: PERT, SPERT, and GERT 583

If the optimistic and pessimistic costs were truly the highest and
lowest that might ever be experienced, the variance would be
P−O
6
while if such estimates were assumed to be at the 5 percent and 95 per-
cent probabilities of best/worst cases, it is suggested the formula would be
P−O
3.2
(See J.J. Moder and E.G. Rodgers, “Judgement estimates of the moments
of PERT type distributions, Management Science, Vol. 15, No. 2, October
1968.) The sum of all line item mean costs becomes the most likely total
cost for the project. Calculating the square root of the sum of the vari-
ances for each line item on the bid estimate would then yield the stan-
dard deviation from that most likely cost.
Any text on statistics will include a table of the cumulative normal
distribution function, which will provide a correlation of the number of
standard deviations from the mean, known as the Z value, to the prob-
ability of such occurrence. For example, the Z value relating to a 90 per-
cent chance that the actual cost of the project will be below a certain
amount is 1.3, while the Z value for an 80 percent chance is only 0.8.
The chances for the project to complete at or below the total estimated
cost is 50 percent, correlating to a Z value of 0.0.
If one adds to the total most likely cost this Z value times the stan-
dard deviation for total cost, the result is the maximum cost of the proj-
ect for the probability stated. Thus, if the total estimated cost is $100,000
and the standard deviation is $20,000, there is a 90 percent chance that
the actual cost for the project will be less than:

Cost Estimated + σ Cost Estimated × Z = Cost Actual = $100000 + $20000 × 1.3

= $ 126000

Likewise, the probability that the project will come in for under
$90,000 is determined by flipping the equation:

CA – CE / σ CE = Z = ($90000 – $100000) / $20000 = –0.5

which correlates to a 31 percent probability.

While this is all quite fascinating and possibly very useful to an esti-
mating department, it works only because accounting and estimating
line items can be totaled in a hierarchical manner. Unfortunately, this
is not true in logic network theory. The existence of the merge bias, as
584 Advanced Topics

noted in Chapter 22, means that the probability of meeting any date will
be lower than calculated in this fashion.
However, because the early practitioners of PERT and CPM/PERT
allowed the “fudge” of using the formula, others tried to build on this
false analogy. Many respected textbooks attempt to use this Z value
method to determine the probability that a project will finish within 1
week, or 1 month, from the date calculated by the CPM algorithm. The
first inkling that this method may not work in this situation is that the
Z value for 50 percent probability is zero. Thus, the assumption is made
that all projects have a 50 percent chance of finishing exactly on time,
just as there is a 50 percent chance that a properly made estimate will
be neither high nor low. The reality is much more bleak.
The use of a Monte Carlo type of analysis, requiring hundreds or thou-
sands of iterations, will generate a much better estimate of the probabil-
ity of completion by any specified date or, conversely, calculate the date
by which there is a specified probability of completion. Thus, if the owner
truly desires the project to be complete by a certain date, the specifica-
tion could provide that a Monte Carlo analysis of the network will pro-
vide an 80 percent or even 90 percent probability of completion by the
stipulated date (Figures 38.2.1 through 38.2.4). Or the engineer can use

Figure 38.2.1CPM calculates completion on 02APR01, MCA SPERT calculates 85%

chance of completion by 11MAY01
Figure 38.2.2 Pertmaster’s MCA software calculates the Activity #30 has only a 52%
chance of being critical.

Figure 38.2.3 Comparison of CPM calculation (second bar) to that of SPERT (first bar) –
default distribution.

585
586 Advanced Topics

Figure 38.2.4 Pertmaster calculates 91 of 154 activities of John Doe project will never
become critical, 19 activities have a 51% to 60% chance of becoming critical, only 3 have
a 81% to 90% of becoming critical, none are guaranteed to be critical.

the old rule of thumb, that for a project that needs to be finished in 12
months, the CPM should show completion in 11 months.

38.3. GERT
At the same time that SPERT studies were being performed on varia-
tions in estimating activity durations, other researchers were studying
alternate forms of logic connection prohibited in the CPM model. For
example, if an activity had two possible successors but only one could
be performed at a time, there was no proper means to convey this in a
CPM model. (One possible work-around to this is to assign the two
activities a common “access” resource, limit such resource to one unit,
and resource level.)
Another problem is that an activity, such as a submittal or a field
test, is not always approved or passed. When such is rejected, addi-
tional work must be performed and then it must be resubmitted or
retested. This type of loop is not permitted in CPM. However, it was
recognized that this type of problem was a more generalized version
of the special type of problem solved by the CPM or PERT algorithm.
Thus, programs to handle the generalized version became known as
GERT programs.
 Advanced Topics: PERT, SPERT, and GERT 587

38.4. Computers Add Power

In the past decade, computers have increased in power, both in speed
and memory, to an astonishing degree. Problems previously not
attempted because they required too much time or memory can now be
easily solved. Extensions of SPERT and GERT, implicit in the original
mathematical concepts of CPM and PERT, can now be performed at little
cost as add-on products to existing CPM software. It is predicted that
within the next few years, CPM products will include such extensions as
a matter of course.
One such product is Primavera’s Monte Carlo software. This program
includes the capability to provide ranges for estimated durations and
decision points which randomly choose a subsequent path, some of which
may even loop back to the decision point. This software can be used to
validate traditionally prepared network or to expand a traditional net-
work to include indeterminate duration activities (such as excavate
extra rock, duration between 0 and 60 days) to major expansions that
include GERT-style loops for resubmittals and retesting.
An example of the validation function is determining the fluctuation
of completion dates of a project based on a reasonable variation of indi-
vidual activity durations. Because each activity duration is, in fact, an
estimate, they are subject to variation. Monte Carlo, as a default,
assumes that each activity duration estimate can be as much as 15%
overstated or 20% understated. Thus, a 10-day activity could be com-
plete in as little as 8.5 days or take as long as 12 days. The user can over-
ride such defaults for the entire project or for individual activities.
Using a random number generator, the program sets the duration of
each activity within the −15%/+20% window and computes the CPM
analysis. The results are stored, the random numbers reassigned, and
the CPM recalculated. After the user-set number of iterations (500 iter-
ations are suggested), the results are tallied and displayed in both tab-
ular and graphical format.
The default graphical format depicts the range of completion dates and
their likelihood. Typically, the likelihood of completing by the completion
date calculated by simple CPM analysis is less than 50%. More important
is the degree of delay or overrun that might be encountered at the 95%
confidence level. Restated, assuming that individual activity durations are
overstated as much as 15% or understated as much as 20%, what is the
latest date the project will finish 95% of the time? What is the likelihood
that the project will be complete a month late? Or a month early?
The tabular format provides the ability to determine what factors are
most likely to impact the project. Using traditional CPM analysis, we
focus upon the critical path. However, if actual durations of activities
are expected to vary from the estimates given in the CPM, what addi-
tional activities should we be concerned with?
588 Advanced Topics

The default tabular report indicates how the critical path can shift
based on the stated variation or possible error in the estimated dura-
tion of individual activities. Often, it is determined that an activity not
on the initial critical path becomes critical for a significant percent of
the simulations.
This often occurs in a construction project CPM where activity dura-
tions are small but procurement and fabrication durations are large. The
relatively large number of small duration activities tend to have their
fluctuation cancel each other. The small number of large duration activ-
ities in a procurement chain (submit, approve, fabricate, and deliver)
tend not to have their variations cancel. Thus, such procurement activ-
ities, although showing a comfortable amount of float in the CPM print-
out, often become critical in actual experience.
Knowing which of these procurement activities have a higher likeli-
hood of becoming problems alerts the project team to be extra vigilant
in tracking such. An expanded critical activities list may be prepared to
display all activities on the critical path, or likely to become critical.

38.5. Summary
Just as the introduction of the first computers led to the introduction of
CPM modeling and analysis, the more powerful computers of today will
permit the more powerful schedule analysis tools of SPERT and GERT
to augment the basic strength of CPM.
 Chapter

39
Conclusion

The Gantt chart or barchart method of scheduling, involving a “graph-

ical interface” or depiction of work, improved upon prior methods such
as todo lists with specified deadlines. Undoubtedly it was considered
complex by its first users. The barchart’s most onerous requirement
was to demand the user to explain how he may meet the various interim
deadlines of a todo list. Similarly, the Critical Path Method system of
scheduling was derided for many years by project managers used to
using barcharts. To many, CPM merely added a burden—the burden of
explaining how the various bars on the barchart, or tasks on the todo
list, interrelated with each other.
The stark simplicity of the ADM or traditional variant of CPM
required this detail to be provided with exacting specificity. Sometimes
the level of detail required would get in the way of practical usage, such
as ease of updating. Sometimes the level of detail, though desired by
upper management of the project team, would be more than desired in
management level reports. Therefore, various means of summarization,
hammocking, filtering and sorting were added to the simple CPM. With
review by non-schedulers came the need for dates to be correct and not
approximate (although the input is still based on early estimates) and
so multiple calendars were added.
Many of the new tools also had a downside—they could be used to cir-
cumvent the rigor required by the ADM system in preparing a network.
The lead/lag capabilities of PDM are chief amongst these new tools.
They provide a great deal of additional power in modeling the real world.
However they must be used with care if they are not to provide an unat-
tainable fantasy world.
The full power and implications of CPM and PERT have always been con-
strained by the limits of the mathematics, software and hardware. Only

589

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
590 Advanced Topics

recently have these limitations been overcome and programs which can
solve the full set of theorized problems been made commercially available.
Undoubtedly these new tools will be considered complex by new users.
Undoubtedly SPERT and GERT systems, such as Primavera’s Monte Carlo
and Pertmaster’s Project Risk and MCA product will be derided for many
years by project managers used to using basic CPM. Undoubtedly these new
tools will be misused, intentionally in some cases, by some who do not care
to expend the additional effort these tools require.
The tools of CPM, whether by ADM, PERT, PDM or RDCPMTM, and
extensions of SPERT and GERT are powerful tools, but they are only
tools. Like your power saw, if these tools are used without the proper
knowledge and respect, they can cause more harm than good. A Master
Carpenter knows how to use a hand saw, but also knows how to prop-
erly use a power saw to accomplish feats the average handyman could
not envision. It is the purpose of this text to advance and promote the
proper use of the tools of project controls into the 21st Century.
The closing words of the 5th Edition, copyright 1999, were that “in ten
years, a CPM without GERT extensions and a SPERT style review will
be considered as naive as a project managed by a todo list prepared on
the back of an envelope by the project manager.” Some readers of the
text suggested the authors lived in an ivory tower and that the real world
was not ready for statistical risk analysis and higher mathematics. Yet
already in 2005, Primavera has teamed up with Pertmaster to provide
a more powerful SPERT/GERT product than its own Monte Carlo soft-
ware. OpenPlan incorporates a SPERT system as part of its basic CPM
product. Even Microsoft Project incorporates the ability to enter the
Optimistic, Most Likely and Pessimistic durations necessary for such an
exercise. Perhaps we are not quite as far progressed as envisioned in
1999, but we are on our way.
 Appendix

A
Sample CPM Specification
as a Guideline
for Preparing your
own Specification

SECTION 01310
CPM SCHEDULES AND REPORTS

PART 1. GENERAL
1.1. RELATED DOCUMENTS

1.1.1. Drawings and General Provisions of Contract, including General and

Supplementary Conditions and other Division-1 Specification Sections, apply to
this Section.

1.2. SUMMARY

1.2.1. General

1.2.1.1. This Section specifies administrative and procedural requirements for

preparation and reporting of Contractor’s preferred sequence of construction of
the Work and other possible sequences of construction of the Work, for moni-
toring and reporting of actual performance of the Work, for incorporation of
changes and unexpected events for determination of possible impact to the
timely completion of the Work and for determination of means and method nec-
essary for time completion of performance of the Work.

1.3. PROJECT SCHEDULE

1.3.1. Detailed Construction Schedule

1.3.1.1. The General Contractor shall develop and maintain the overall Detailed
Construction Schedule, (referred to hereafter as Schedule or Construction

591

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
592 Appendix A

Schedule). The Schedule shall be computer generated in precedence format

using the Critical Path Method (CPM). The Contractor shall perform its obli-
gations generally in accordance with the sequence and time frame provided
by the Schedule. The Contractor shall update the Schedule and shall modify
and change the Schedule as may be required as provided by this Section. The
Contractor shall not modify, change, or update the Construction Schedule or
any activities therein without the prior written approval of the Program
Manager.

1.3.2. Purpose of the Schedule

1.3.2.1. Provide additional assurance by the Contractor of its adequate plan-

ning, scheduling, and reporting during the execution of the construction and
related activities so they may be prosecuted in an orderly and expeditious
manner, within the Contract time and the milestones stipulated herein.
1.3.2.2. Provide additional assurance by the Contractor of the coordination of
the work of the Contractor and the various Subcontractors and suppliers at
all tiers.
1.3.2.3. Assist the Program Manager in monitoring the progress of the work.
1.3.2.4. Assist the Program Manager in evaluation of the Contractor’s monthly
progress payments requests.
1.3.2.5. Assist the Program Manager in evaluating the potential impact of pro-
posed changes to the Contract.
1.3.2.6. Assist and be utilized by the Contractor in the coordination of its forces,
subcontractors, and vendors.
1.3.2.7. Assist in detecting problems for the purpose of taking timely corrective
action and to provide a mechanism or tool for determining and monitoring such
corrective actions.

1.3.3. General Requirements of Submitted Schedules

1.3.3.1. The Work shall proceed at a rate as will meet the specified Milestone
Dates, Substantial Completion (if applicable), and Final Completion dates within
the Contract Time. By execution of the Contract, the Contractor represents
that he has analyzed the Work, the materials, and methods involved, the sys-
tems of the building, availability of qualified mechanics and labor, restrictions
of the site, constraints imposed, their own work load and capacity to perform
the Work, and agrees that the specified times are reasonable considering the
existing conditions prevailing in the locality of the Work, including weather
conditions, and other factors, with reasonable allowance for variations from
average or ideal conditions.
1.3.3.2. The Schedule shall clearly identify the activities illustrating accom-
plishment of the time(s) for completion of the Project set forth in the Contract.
If the Schedule indicates earlier completion time(s) than that set forth in
Contract, the float between the Schedule and the Contract dates shall be con-
sidered to be part of the total float available.
 Sample CPM Specification as a Guideline 593

1.3.3.3. In developing the Schedule, the Contractor shall be responsible for

assuring that Subcontractor work at all tiers, as well as Contractor’s own work,
is included in the Schedule.
1.3.3.4. The Schedule as developed shall show the sequence and interdepend-
ence of activities required for complete performance of the work. The Contractor
shall be responsible for assuring all work sequences are logical and the Schedule
shows a coordinated plan of the work.
1.3.3.5. Failure by the Contractor to include any element of work required for
performance of the Contract or failure to properly sequence the work shall not
excuse the Contractor from completing all work within the Contract Time.

1.3.4. Use of Float and Reasonable Limitations Upon Resources

1.3.4.1. Float time is not for the exclusive use or benefit of either the Contractor
or the Owner. The Contractor shall limit its use of logic restraints based upon
use of a resource such that it equally reserves float for the limited resources of
the Program Manager, the Owner, other entities under the Owner’s control, and
other entities beyond the control of either the Contractor or Owner, and that the
Contractor reserves float for the various types of unexpected events which may
be anticipated on a construction project of this magnitude.
1.3.4.2. The Contractor acknowledges and agrees that actual delays to specific
activities that do not exceed available total float time of such activities will not
have any effect upon Contract completion times and Contractor will take all
actions necessary to maintain the overall schedule.

1.3.5. Requirement for Additional Resources

1.3.5.1. The Contractor shall provide adequate resources, including but not
limited to manpower and construction equipment, to perform its obligations in
a timely manner. The Contractor shall be required to provide additional
resources for additional work or events which may be anticipated on a con-
struction project of this magnitude.
1.3.5.2. If the Contractor and Program Manager agree to a Change Order, such
agreement shall be construed as stating that the Contractor is capable of supply-
ing additional resources as may be required to effectuate such Change Order with-
out the need to reduce the resources available for other work on the project, without
disruption of other work on the project and without additional cost for provision
of additional resources other than as may be included in the agreed Change Order.

1.3.6. Entitlement to Extension of Time and Acceleration

1.3.6.1. Entitlement to extensions of time for performance as described in the

Contract Documents will be granted only to the extent that time adjustments
for the activity or activities affected by any condition or event which entitles the
Contractor to a time extension exceed the total float along the current critical
path of activities affected.
1.3.6.2. If the Program Manager does not provide an extension of time at the
request of the Contractor, the Contractor shall in a timely manner provide a
594 Appendix A

Recovery Schedule and itemized estimate of costs to effectuate such or shall be

deemed to waive its claim for additional compensation therefore.

PART 2. PRODUCTS
2.1. TECHNICAL REQUIREMENTS OF SUBMITTED SCHEDULES

2.1.1. Contractor shall plan, schedule, execute, and report on the Work using
the Critical Path Method (CPM). The principles used herein shall be as set
forth in the text CPM in Construction Management, 6th edition, McGraw Hill,
except that in case of conflict, the provisions of these Contract Documents shall
govern.
2.1.2. The Contractor shall employ the services of at least one fully qualified
scheduler for the duration of the Contract. Said scheduler(s) shall have a min-
imum of 3 years experience in CPM scheduling on projects of similar size and
scope. Said scheduler(s) shall cooperate with the Program Manager and shall
be on the project site full time for the purpose of continuously monitoring, mod-
ifying, or updating the Contractor’s detailed construction schedule.
2.1.3. The Schedule shall be developed utilizing the Precedence Diagramming
Method. Contractor shall use Primavera P3 scheduling software, Version 3.1
for Windows. Contractor shall use such software with scheduling options set
for retained logic, calculate start-to-start lag from actual start, schedule dura-
tions as interruptible, show open ends as critical, and calculate the total float
as the most critical.

2.2. RESTRICTIONS SUBJECT TO DISCRETION OF ENGINEER

2.2.4. Milestone Dates, including Notice to Proceed, Substantial Completion,

and Final Completion must be adhered to and shall be clearly identified on the
Schedule. Milestone Dates may not be changed without the written consent of
the Program Manager. Final Completion shall be a mandatory finish con-
strained date.
2.2.5. The Schedule shall be developed utilizing activities of specified duration
of whole days between one (1) and fifteen (15) working days. Working days are
defined as on a five day per week calendar, less recognized holidays as provided
by the Program Manager. Milestones or other zero duration activities shall not
be permitted except to indicate milestones set forth in this Specification.
Submissions including use of Expected Finish constraints to calculate dura-
tions shall not be permitted.
2.2.6. Logic relationships between activities shall be limited to finish to start
type relationships. The use of durations between activities, or lags, shall not
be permitted.
2.2.7. Exceptions to the requirements above for the purpose of improving the
Program Manager’s ability to monitor the Schedule and permitting the use of
durations larger than fifteen (15) working days, multiple calendars, milestones
or other zero duration activity, calculated durations, logic relationships other
than finish to start or durations between activities (lags) may be permitted on
a case by case basis at the sole discretion of the Program Manager.
 Sample CPM Specification as a Guideline 595

2.3. REQUIRED LEVEL OF DETAIL AND DURATIONS

2.3.1. The level of detail of the Schedule shall be a function of the complexity of
the work involved. The level of detail and total number of activities shall be sub-
ject to approval by the Program Manager. No construction activity shall have
duration of longer than fifteen (15) work days without prior acceptance of the
Program Manager. Non-construction activities (such as procurement and fab-
rication) may have duration in excess of fifteen (15) work days.
2.3.2. Normal weather conditions shall be considered and included in the plan-
ning and scheduling of all work influenced by high or low ambient tempera-
tures and/or precipitation to ensure completion of all work within the Contract
Time. Normal weather conditions shall be determined by an assessment of
average historical climatic conditions based upon the preceding ten (10) year
records published for the locality by the National Ocean and Atmospheric
Administration (NOAA).
2.3.3. Activity descriptions shall be clear and concise. The beginning and end of
each activity shall be readily verifiable. All activity starts and finishes, with the
exception of Milestones, must be tied into the schedule by logical restraints.
2.3.4. Proposed duration assigned to each activity shall be the Contractor’s best
estimate of time required to complete the activity considering the scope and
resources planned for the activity.
2.3.5. Responsibility for each activity shall be identified with a single perform-
ing organization.
2.3.6. For all major equipment and materials fabricated or supplied for this
project, the Schedule shall show a sequence of activities including:
2.3.6.1. Preparation of shop drawings and sample submissions.
2.3.6.2. Review of shop drawings and samples.
2.3.6.3. Shop fabrication and delivery.
2.3.6.4. Erection or installation.
2.3.6.5. Testing of equipment and materials.

2.4. REQUIRED MINIMUM CODING OF ACTIVITIES, RESOURCES AND COSTS

2.4.1. Each activity shall be identified with codes including as a minimum:

2.4.1.1. The party responsible for performance of the Work,
2.4.1.2. Where work is to be subcontracted, the subcontractor to be responsi-
ble for the Work,
2.4.1.3. The size, craft and composition of the labor crew performing the Work,
2.4.1.4. The total man-hours estimated for performing the Work,
2.4.1.5. The major equipment to be used for performance of the Work,
2.4.1.6. The number of hours in a work day for this activity including all shifts,
2.4.1.7. The contract drawing number or specification section where the Work
is best portrayed,
2.4.1.8. Building, Floor and Location of the Work, and
596 Appendix A

2.4.1.9. The primary Schedule of Values and CSI classification associated with
the Work.
2.4.2. Contractor shall assign resource loading, including but not limited to
manpower and major construction equipment, for each activity of the Schedule.
2.4.2.1. Manpower and major construction equipment resources must be listed
in the Resource Library of the Primavera software. The Contractor shall set the
Normal level of resource usage as not less than that required to perform the proj-
ect in a timely manner. The Contractor shall set the Maximum level of resource
usage as not less than 150% of the Normal level of resource usage.
2.4.2.2. The cumulative assigned labor man-hours for each activity must equal
the total man-hours assigned in the activity code tabulation.
2.4.3. Each activity shall be cost-loaded and the assigned dollar value (cost-
loading) of each activity shall cumulatively equal the Total Contract Amount.
Separate activities shall be included and cost loaded for costs associated with
mobilization, demobilization, bond and insurance. The cost for demobilization
shall be at least seventy-five percent (75%) of that for mobilization. Costs for
General Conditions, overhead. and profit shall be prorated throughout all activ-
ities other than those for mobilization, demobilization, bond, and insurance.
2.4.3.1. The assigned dollar value (cost-loading) of each activity shall be coded
against one or more cost category and cost account number. Cost account numbers
shall be the same as line items in the Schedule of Values. Cost categories shall
include Labor, Equipment, Materials and Subcontracted. The assigned dollar
value (cost-loading) for each cost account for each activity shall cumulatively equal
the Total Contract Amount assigned for each respective line item in the Schedule
of Values.
2.4.3.2. The assigned dollar value (cost-loading) for the labor category shall
cumulatively approximate the total labor man-hours times the average cost per
man-hour plus a reasonable percent for overhead and profit.
2.4.3.3. The assigned dollar value (cost-loading) for the equipment shall cumu-
latively approximate the total reasonable rental value for such equipment plus
a reasonable percent for overhead and profit.
2.4.3.4. The assigned dollar value (cost-loading) for the material category shall
cumulatively approximate the total reasonable cost for materials plus a rea-
sonable percent for overhead and profit.
2.4.3.5. The assigned dollar value (cost-loading) for the subcontracted category
shall cumulatively approximate the total labor manhours for subcontracted work
times the average cost per man-hour plus a reasonable percent for overhead and
profit. If requested by the Program Manager, the Contractor shall furnish either
the total subcontracted cost of selected subcontractor(s) or a breakdown by such
subcontractor(s) of labor, equipment, and material cost.

2.5. REQUIRED NARRATIVE

2.5.1. Contractor shall prepare and provide a written narrative to further explain
the plan as set forth in its CPM logic network and schedule. The narrative shall
include a general summary of the Contractor’s proposed plan to execute the
 Sample CPM Specification as a Guideline 597

works of the project. The narrative shall include an explanation of the format
of activity descriptions including standard abbreviations used. The narrative
shall include a general summary of the Contractor’s plan for manning of the proj-
ect including anticipated total manpower per month for direct hire and for sub-
contractor forces. The narrative shall include a list of the major items of
construction equipment intended for use on this Contract’s operations includ-
ing types, number of units, unit capacities, and the proposed time each piece of
equipment will be on the job, keyed to the activities on which the equipment will
be used.
2.5.2. If requested by the Program Manager, the Contractor shall include in its
written narrative the Contractor’s determination of duration for critical, near-
critical and other specified activities. Such explanation shall include the number
of crews, crew composition, number of shifts per day, number of hours in a shift
and the number of work days per week.

PART 3. EXECUTION
3.1 CONTRACTOR’S REPRESENTATIVE DESIGNATION

3.1.1. Within (3) working days after date of Notice to Proceed, Contractor shall
designate its authorized project scheduler (“Scheduler”).
3.1.2. Contractor’s scheduler shall have complete authority to act on behalf of
the Contractor in fulfilling the Construction Schedule requirements of the
Contract and such authority shall not be interrupted throughout the duration
of the Contract unless approved in writing by the Program Manager.

3.2 INITIAL SCHEDULE CONFERENCE

3.2.1. The Program Manager will schedule and conduct an initial schedule con-
ference within five (5) working days of Notice to Proceed shall be prepared to
review and discuss the schedule and sequence of operations including cost,
manpower, and equipment loading methodology. The conference shall be
attended by:
3.2.1.1. Contractor’s Project Manager, Superintendent, and Scheduler
3.2.1.2. Other Contractor key personnel, defined as any individual reporting to
the Contractor’s Project Manager or Superintendent, and being in responsible
charge of more than 20% of field efforts as defined by cost.
3.2.1.3. The Project Manager, Superintendent or person in responsible charge
of each Subcontractor expected to perform more than 10% of field efforts as
defined by cost.
3.2.1.4. A representative in responsible charge of the fabrication and delivery
of materials for this project for each major supplier including each supplier of
more than 10% of the total contract value as defined by cost.
3.2.1.5. [intentionally blank or for known specialty subcontractor or supplier]
3.2.1.6. [intentionally blank or for third party (e.g. utility, owner’s operations
manager, etc.) but either 1) placing obtaining as contractor’s obligation or 2)
relieving Owner if such individual cannot or will not attend]
598 Appendix A

3.2.1.7. Program Manager or its representative(s).

3.2.2. Procedures will be reviewed by the Program Manager for the following:
3.2.2.1. Development of preliminary Schedule by Contractor.
3.2.2.2. Procedures for updating and revisions.
3.2.2.3. Cost and resource loading of activities. Format and use of schedule of
values.
3.2.2.4. Method of generating earned value reports, establishing target sched-
ules, and evaluating cost, resource, and schedule performance.
3.2.2.5. Data exchange and Communications.
3.2.2.6. Procedures for assessing schedule impacts, schedule delays, and time
extensions.
3.2.2.7. Development of recovery schedules.

3.3. PRELIMINARY SCHEDULE

3.3.1. Within ten (10) days after Notice to Proceed, the Contractor shall submit
for the Program Manager’s review and acceptance a Preliminary Schedule.
3.3.2. The Preliminary Schedule shall cover the following project phases and
activities:
3.3.2.1. Proposed Procurement Activities to be accomplished during the first
ninety (90) days of the Contract. Procurement activities shall include mobi-
lization, key shop drawing and sample submittals, reviews, and the fabrication
and delivery of key and long-lead procurement elements. Indicate planned sub-
mittal dates and delivery dates for fabrication and delivery activities.
3.3.2.2. Proposed Construction Activities to be accomplished during the first
ninety (90) days of the Contract
3.3.2.3. Summary Activities not included above which are necessary to prop-
erly indicate the approach to scheduling the remaining work areas or phases
of the work. The work for each phase or area must be represented by at least
one summary activity such that they cumulatively indicate the entire
Construction Schedule. Summary Activities may exceed the duration limita-
tions listed above and may be connected by Start-to-Start and Finish-to-Finish
logic relationships as well as Finish-to-Start logic relationships and such logic
relationships may include durations between activities (lag.).
3.3.2.4. The Preliminary Schedule shall otherwise conform with the require-
ments outlined in the “Technical Requirements for Contractor-submitted
Schedules” in this specification section.
3.3.2.5. Contractor shall assign an approximate cost to the Proposed
Construction Activities and to Summary Activities. If requested by the Program
Manager, the Contractor shall furnish a written narrative supporting such
approximate costs.
3.3.3. The submission shall consist of:

3.3.3.1. A Primavera Backup Disk.

 Sample CPM Specification as a Guideline 599

3.3.3.2. A pure logic diagram of the entire Preliminary Schedule, which may be
either hand or machine drawn. If machine drawn, activities shall be separated
by a minimum of one inch horizontally and one half inch vertically. The diagram
shall be drawn or plotted on D or E size media.
3.3.3.3. Machine plotted time scaled logic diagrams of:
3.3.3.3.1. The entire Preliminary Schedule.
3.3.3.3.2. The first 90 days of the Preliminary Schedule.
3.3.3.3.3. Time scaled diagrams shall be plotted on D or E sized media.
3.3.3.4. Tabular listings of:
3.3.3.4.1. All Procurement Activities grouped by Submissions, Review and
Fabrications, then sorted by Early Start, then Total Float.
3.3.3.4.2. All Construction Activities to be performed in the first ninety (90)
days, sorted by Early Start, then Total Float.
3.3.3.4.3. All Summary Activities sorted by Early Start, then Total Float.
3.3.3.4.4. Tabulations shall include Activity ID, Description, Original Duration,
Remaining Duration, Percent Complete, Cost Percent Complete, Activity Codes,
Early Dates, Late Dates, Total Float, Predecessors, Successors, and Total
Assigned Cost.
3.3.3.5. A hand drafted or machine plotted diagram of anticipated manpower,
including subcontracted manpower versus time.
3.3.3.6. Five (5) copies of all diagrams and tabulations shall be required.
3.3.4. Within seven (7) working days after receipt by the Program Manager of
the Preliminary Schedule or changes thereto, the Program Manager will notify
the Contractor of any concerns the Program Manager may have in regard to the
Preliminary Schedule.
3.3.5. The Contractor shall provide a response to the concerns of the Program
Manager, to the satisfaction of the Program Manager, within five (5) working
days for the first response and three (3) working days for subsequent responses
as may be required.
3.3.6. If the Preliminary Schedule is not approved within thirty (30) working
days after Notice to Proceed, the Program Manager may assess the Contractor
for liquidated damages in the amount of $#,### per day to reimburse the Owner
for the additional risk of late completion and the increased monitoring and
administration associated with attempts to control and mitigate such risk.
3.3.7. The Preliminary Schedule shall be updated on a monthly basis while the
Baseline Schedule is being developed. The monthly updating of the Preliminary
Schedule shall be consistent with the procedures and requirements described
in the “Schedule Updating” section of this specification section.

3.4. BASELINE SCHEDULE

3.4.1. Within thirty (30) working days following Notice to Proceed, Contractor
shall submit to the Program Manager a detailed Baseline Schedule in precedence
format for the Contractor’s construction work scope.
600 Appendix A

3.4.2. The Baseline Schedule shall conform with the requirements outlined in
the “Technical Requirements for Contractor-submitted Schedules” in this spec-
ification section.
3.4.3. The Baseline Schedule shall be prepared in accordance with the com-
ments and concerns raised by the Program Manager relating to the Preliminary
Schedule. If such concerns have not been fully addressed for the Preliminary
Schedule prior to the deadline of section 3.4.1 above, the Contractor shall nev-
ertheless submit its work in progress as of that date.
3.4.4. The submission shall consist of:
3.4.4.1. A Primavera Backup Disk.
3.4.4.2. A pure logic diagram of the entire Baseline Schedule, which may be
either hand or machine drawn. If machine drawn, activities shall be separated
by a minimum of one inch horizontally and one half inch vertically. Diagrams
shall be drafted or plotted on D or E size media.
3.4.4.3. Machine plotted time scaled logic diagrams of:
3.4.4.3.1. The entire Baseline Schedule.
3.4.4.3.2. Activities on the critical path and those having ten (10) days or less
float relative to the critical path.
3.4.4.3.3. Diagrams shall have the critical path highlighted and activities for
which responsibility is other than the Contractor highlighted in a different color
and pattern.
3.4.4.3.4. Diagrams shall be plotted on D or E size media.
3.4.4.4. Tabular listings of:
3.4.4.4.1. All Procurement Activities organized by Submissions, Review and
Fabrications, then sorted by Early Start, then Total Float.
3.4.4.4.2. All Construction Activities sorted by Early Start, then Total Float.
3.4.4.4.3. Tabulations above shall include Activity ID, Description, Original
Duration, Remaining Duration, Percent Complete, Cost Percent Complete,
Activity Codes, Early Dates, Late Dates, Total Float, Predecessors, Successors,
and Assigned Cost. The column for Assigned Cost shall be totaled.
3.4.4.4.4. All activities in activity identification number order, listing the activ-
ity identification number, activity title, successor identification number, suc-
cessor title, logic relationship type, lag, activity calendar number and reason for
the logic relationship. The reason for the logic relationship shall be given as
“Physical,” “Resource,” or “Other.” If the reason given is “Resource” an additional
column shall list what resource. If the reason given is “Other” such shall be sup-
ported by a separate narrative. This tabulation shall be provided in both print
and on diskette in an Excel or compatible format.
3.4.4.5. A machine plotted diagram of manpower versus time for:
3.4.4.5.1. Total manpower on the project.
3.4.4.5.2. Manpower for each subcontractor performing ten percent (10%) or
more of the total labor on the project.
3.4.4.5.3. Manpower for each craft directly employed by the Contractor.
 Sample CPM Specification as a Guideline 601

3.4.5. The Baseline Schedule shall be reviewed in the following manner:

3.4.5.1. Within fifteen (15) working days after receipt by the Program Manager
of the Baseline Schedule or seven (7) working days after receipt by the Program
Manager of changes thereto, the Program Manager shall notify the Contractor of
any concerns the Program Manager may have in regard to the Baseline Schedule.
3.4.5.2. The Contractor shall provide a response to the concerns of the Program
Manager, to the satisfaction of the Program Manager, within five (5) working
days for the first response and three (3) working days for subsequent responses
as may be required.
3.4.5.3. If requested by the Program Manager after receipt of the Contractor’s
response, the Contractor shall meet within three (3) working days of such
request for a joint review, correction, or adjustment of the Contractor’s pro-
posed Baseline Schedule. Within three (3) working days after such joint review
between the Contractor and Program Manager, the Contractor shall revise the
Baseline Schedule in accordance with agreements reached during the joint
review and re-submit it to the Program Manager.
3.4.5.4. If the Baseline Schedule is not approved within sixty (60) working days
after Notice to Proceed, the Program Manager may assess the Contractor for liq-
uidated damages in the amount of $#,### per day to reimburse the Owner for
the additional risk of late completion and the increased monitoring and admin-
istration associated with attempts to control and mitigate such risk. This assess-
ment of liquidated damages shall be in addition to that provided in section 3.3.6
if applicable.
3.4.6. Acceptance of Contractor’s Baseline Schedule:

3.4.6.1. Upon the submittal by the Contractor and review and approval by the
Program Manager of the Baseline Schedule, the Contractor shall submit a copy
of the Approved Baseline Schedule (“ABS”) signed on the face by the Contractor
and by each Subcontractor performing greater than 10% by cost of scope of the
project, indicating the Contractor’s (and Subcontractor’s) acceptance and
approval of the Baseline Schedule.
3.4.6.2. The Primavera files for Accepted Baseline Schedule shall be saved as
Project ID Number/Version “BASELINE” and not further modified. A copy of this
file shall be renamed as Project ID Number/Version “UPDATE <date>” and
used for subsequent updates and shall reference this file as its target. A copy of
the “UPDATE” file shall be renamed as Project ID Number/Version “INTERIM
<date>” and shall reference the most recent “UPDATE” file.
3.4.6.3. Acceptance by the Program Manager of the Contractor’s accepted
Baseline Schedule shall be a condition precedent to the making of any progress
payments under the Contract after the first seventy-five (75) working days of
the Contract at the discretion of the Program Manager. In such event, the
Program Manager, in its sole discretion, may also choose to make partial
progress payments requiring additional retainage or may choose to make par-
tial progress payments based solely upon certified payrolls and vendor invoices.
3.4.6.4. Upon acceptance of the Baseline Schedule by the Program Manager,
the cost-loaded values of the Baseline Schedule shall be used as a basis for
602 Appendix A

determining progress payments. Monthly progress payments shall be based

upon information developed at the monthly Schedule Update.
3.4.6.5. Acceptance by the Program Manager of the Contractor’s Construction
Schedule does not relieve the Contractor of any of Contractor’s responsibility
whatsoever for the accuracy or feasibility of the Construction Schedule, or of the
Contractor’s ability to meet the Contract completion date or Milestone Dates,
nor does such acceptance acknowledge or admit the reasonableness of the activ-
ities, logic, duration, manpower, cost, or equipment loading of the Contractor’s
Construction Schedule.
3.4.6.6. In the event the Contractor fails to define any element of work, activ-
ity, or logic and the Program Manager review does not detect this omission or
error, such omission or error, when discovered by the Contractor or Program
Manager, shall be corrected by the Contractor at the next monthly Schedule
Update (discussed hereinafter) and shall not affect the Contract Time

3.5. USE OF SCHEDULE FOR PROJECT COORDINATION

3.5.1. Weekly Progress Meetings

3.5.1.1. Once each week at the progress meeting, the progress achieved by the
Contractor during previous work week will be assessed. The Contractor shall
update the most recent “INTERIM” Schedule with the Actual Start date of
activities started in the past week, the Remaining Duration of those activities
in progress and the Actual Finish date of activities completed during the past
week. The Contractor shall submit a progress schedule in Primavera’s tabu-
lar/bar-chart format indicating the information used to perform this update for
the previous week and the activities scheduled for the succeeding three (3)
weeks. A bar chart directly derived from the most recent “INTERIM” Schedule
shall be used to generate the four (4) week window. All activities shown in this
short interval schedule will be identified by the same activity numbers and
descriptions as shown in the Construction Schedule.
3.5.1.2. The Contractor shall mark on this computer generated bar-chart sched-
ule the choice and timing of those activities it intends to actually perform during
the upcoming three weeks. The Contractor may add further details to monitor
this short interval Schedule.
3.5.1.3. A copy of this short interval schedule shall be submitted to the Program
Manager.

3.5.2. Minor Revisions to the Schedule for Unanticipated Events

3.5.2.1. lf the Contractor and Program Manager agree to a Change Order (“CO”),
such agreement shall include an Impact Analysis and agreement on the accept-
ance for such impact (in part or whole by each party) and costs for mitigation
thereof. In the event that such agreement is not part of the agreement for the
price of the Change Order, the Contractor shall treat such Change Order as a
directive for purposes of the schedule.
3.5.2.2. If the Contractor believes that a submitted Request For Information
(“RFI”), claimed Change In Conditions (“CIC”), request to delay or defer work
 Sample CPM Specification as a Guideline 603

pending a Proposed Change Order (“PCO”), directive to perform a Change Order

(“CO”) or claimed Constructive Change Order (“CCO”) may impact its work, the
Contractor shall perform an Impact Analysis upon a copy of the most recent
“UPDATE” file and submit such to the Program Manager as soon as practica-
ble after determination of such belief. The failure to so promptly notify the
Program Manager shall be deemed a waiver of any compensation or extension
of time due to such cause. Where the parties are in disagreement over the
responsibility of the delaying event, the Contractor shall use a description for
such which is responsibility neutral.
3.5.2.3. As part of its Impact Analysis, the Contractor shall suggest means of
mitigation including but limited to use of greater resources, modification or
deletion from the logic network of selected restraints and selective overtime. If
the Contractor believes that its efforts to mitigate such impact will entitled it
to additional compensation, the Contractor shall submit an estimate of the
unmitigated and mitigated impact and cost consequences of each. The failure
to provide such a submittal in a timely manner shall be deemed a waiver of any
additional compensation.

3.5.3. Monthly Schedule Update Meetings

3.5.3.1. On a monthly basis, the Contractor shall meet with the Program
Manager for the purpose of updating the Schedule. The Contractor shall submit
its assessment of the Actual Start date of activities started since the last update,
Remaining Duration of those activities in progress, Actual Finish date of activ-
ities completed and Cost Percent Complete of activities in progress or complete.
The Program Manager will either assent to the Contractor’s assessments or
direct the Contractor to use other dates or Cost Percent Complete. The Program
Manager may request the Contractor to provide additional assurance of a
Remaining Duration of work in progress. If the Contractor does not agree to the
direction of the Program Manager, it shall note such in the log field of the activ-
ity and in comments to the Minutes of the meeting.
3.5.3.2. The information shall be entered to a copy of the most recent “UPDATE”
file which will then be saved and not further modified.
3.5.3.3. Monthly Update Reports submission shall consist of:
3.5.3.3.1. A Primavera Backup Disk.
3.5.3.3.2. [Intentionally blank.]
3.5.3.3.3. Machine plotted time scaled logic diagrams of:
3.5.3.3.3.1. The entire Baseline Schedule including indication of progress to date.
3.5.3.3.3.2. Activities not yet completed on the critical path and those having
ten (10) days or less float relative to the critical path.
3.5.3.3.3.3. Diagrams shall have the critical path highlighted and activities for
which responsibility is other than the Contractor highlighted in a different color
and pattern.
3.5.3.3.3.4. Diagrams shall be plotted on D or E size media.
3.5.3.3.4. Tabular listings of:
604 Appendix A

3.5.3.3.4.1. All Procurement Activities organized by Submissions, Review, and

Fabrications, then sorted by Early Start, then Total Float.
3.5.3.3.4.2. All Construction Activities sorted by Early Start, then Total Float.
3.5.3.3.4.3. Tabulations above shall include Activity ID, Description, Original
Duration, Remaining Duration, Percent Complete, Cost Percent Complete,
Activity Codes, Early Dates or Actual Dates, Late Dates (if applicable), Total
Float, Predecessors, Successors, Assigned Cost, and Earned Cost. The columns
for Assigned Cost and Earned Cost shall be totaled.
3.5.3.3.5. A machine plotted diagram of manpower versus time for work on the
remainder of the project including:
3.5.3.3.5.1. Total manpower on the project.
3.5.3.3.5.2. Manpower for each subcontractor performing ten percent (10%) or
more of the total labor on the project.
3.5.3.3.5.3. Manpower for each craft directly employed by the Contractor.
3.5.3.4. If there have been any Minor Revisions to the Schedule for
Unanticipated Events during the past reporting period, such shall be incorpo-
rated into a copy of most recent “UPDATE” file and rescheduled. This file will
be renamed as Project ID Number/Version “REVISED <date>”. The file will
then be saved as the new “UPDATE” file and not further modified.
3.5.3.5. Minor Revision Reports submission shall consist of:
3.5.3.5.1. All reports required for an Update.
3.5.3.5.2. Those portions of the pure logic diagram required for the Baseline
Schedule submission which have been modified with the modifications high-
lighted.
3.5.3.6. After updating and (if required) revising the schedule, it shall be copied
to the next “UPDATE” file and to a new “INTERIM” file.

3.6. MAJOR REVISIONS TO THE SCHEDULE

3.6.1. In the event that, pursuant to a Change Order, a Revised Baseline

Schedule is adopted for the work remaining on the project, such revised base-
line Schedule shall be used as the target for further update to the project.

3.7. RECOVERY SCHEDULE

3.7.1. In the event that the Contractor determines that it can no longer perform
according to the schedule, the Contractor shall prepare and submit a Recovery
Schedule.
3.7.2. In the event that the Most Recent Update indicates that the project is more
than twenty (20) days behind schedule, or that a major subcontractor perform-
ing more than ten percent (10%) of the labor on the site leaves for any reason
without completion of its work, or that a specialty subcontractor employing pro-
prietary means and methods leaves the site for any reason without completion
of its work, or the Contractor becomes aware of an anticipated delay of specially
ordered materials or equipment calculated to delay the project more than twenty
(20) days behind schedule or the Contractor anticipates for any reason that the
 Sample CPM Specification as a Guideline 605

project is likely to be delayed more than twenty (20) days behind schedule, and
upon notice of such to and subsequent request of the Program Manager, the
Contractor shall prepare and submit a Recovery Schedule.
3.7.3. The Recovery Schedule submittal may include, without limitation:
3.7.3.1. Revisions to the Original Durations of Activities not yet started, which
are to individually be supported with a narrative of the actual productivity to
date or increased resources or hours per day to effectuate such.
3.7.3.2. Revisions to the Calendar, including indicating work on Saturdays,
Sundays or holidays, subject to approval by the Program Manager.
3.7.3.3. Splitting of activities to indicate more precise coordination, which are
to be individually supported with a narrative of how a portion of the previously
indicated activity may now suffice for a successor activity.
3.7.3.4. Revisions to Logic Relationships, deleting restraints based upon lim-
ited resources, which are to individually be supported with a narrative indicating
the ability and willingness to engage additional resources.
3.7.4. The Recovery Schedule shall be prepared to indicate, where practicable,
recovery within one month or within ten percent (10%) of the remaining dura-
tion until the mandated deadlines threatened.
3.7.5. Recovery Schedule Reports submission shall consist of:

3.7.5.1. All reports required for an Update.

3.7.5.2. The pure logic diagram required for the Baseline Schedule submission,
highlighted, where practicable, to indicate where the Recovery Schedule differs
from the Baseline Schedule.
3.7.6. Where the Recovery Schedule has been ordered by the Program Manager,
it shall be submitted within five (5) working days. The Contractor and all par-
ties under its control called to the Initial Schedule Conference shall be pre-
pared to attend, upon fortyeight (48) hours notice, a Recovery Schedule Meeting
which may be called by the Program Manager within the next three (3) to seven
(7) working days. The Program Manager may also request the Contractor’s
Surety to attend the Recovery Schedule Meeting.
3.7.6.1. If a Recovery Schedule Meeting is called, the parties attending shall pro-
vide additional assurances to, or revise the proposed Recovery Schedule to the
satisfaction of the Program Manager.
3.7.6.2. Once approved by the Program Manager, the Recovery Schedule shall
be treated as a Minor Revision to the Schedule or a Major Revision to the
Schedule as may be directed by the program Manager.
3.7.6.3. Once approved by the Program Manager, failure by the Contractor to
strictly follow the Recovery Schedule until back on schedule shall be deemed a
Material Breach of the Contract
This page intentionally left blank
 Appendix

B
Unified
Facilities Guide
Specification

SECTION TABLE OF CONTENTS

DIVISION 01—GENERAL REQUIREMENTS
SECTION 01321N
NETWORK ANALYSIS SCHEDULES (NAS)
04/02

PART 1. GENERAL
1.1. DESCRIPTION
1.2. SUBMITTALS
1.3. SCHEDULE ACCEPTANCE
1.3.1. Schedule Acceptance Prior to Start of Work
1.3.2. Acceptance
1.4. SOFTWARE
1.4.1. Computer Hardware
1.4.2. Software Training
1.5. QUALIFICATIONS
1.6. NETWORK SYSTEM FORMAT
1.6.1. Diagrams
1.6.2. Quantity and Numbering of Activities
1.6.2.1. HVAC TAB Activities
1.6.2.2. Procurement Activities
1.6.2.3. Government Activities
1.6.2.4. Construction Activities
1.6.2.5. Anticipated Weather Delays
1.6.2.6. Activity Properties
1.6.3. Mathematical Analysis

607

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
608 Appendix B

1.6.4. Additional Requirements

1.6.5. Required Reports
1.7. SUBMISSION AND ACCEPTANCE
1.7.1. Preliminary Meeting
1.7.2. Schedule Development Session:
1.7.3. Preliminary Network Analysis Schedule
1.7.4. Network Analysis Schedule
1.7.5. Review and Evaluation
1.7.6. Accepted Network Analysis Schedule
1.7.7. Monthly Network Analysis Updates
1.7.8. Summary Network
1.7.9. As-Built Schedule
1.8. CONTRACT MODIFICATION
1.8.1. Time Impact Analysis:
1.8.2. No Reservation-Of-Rights
1.9. CHANGES TO THE NETWORK ANALYSIS SCHEDULE
1.10. FLOAT
1.10.1. Definitions of Float or Slack
1.10.2. Ownership of Float
1.10.3. Negative Float
1.11. TIME EXTENSIONS
1.12. MONTHLY COORDINATION MEETING
1.13. BIWEEKLY WORK SCHEDULE
1.14. WEEKLY COORDINATION MEETING
1.15. CORRESPONDENCE AND TEST REPORTS
PART 2. PRODUCTS
PART 3. EXECUTION

SECTION 01321N
NETWORK ANALYSIS SCHEDULES (NAS)
04/02
Note: This guide specification covers the preparation and use of a contrac-
tor prepared Network Analysis Schedules. This section will be used on most
projects in lieu of Section 01320, “Construction Progress Documentation.”
Section 01320 shall be used only when a hand-drawn bar chart is required
for management and oversight of a project. As prescribed in FAR 36.515, the
Contracting Officer may insert the clause “Schedules for Construction
Contracts” (FAR 52.236-15) in solicitations and contracts when a fixed-price
construction contract is contemplated, the contract amount is expected to
exceed the simplified acquisition threshold, and the period of actual work per-
formance exceeds 60 days. This clause may be inserted in such contracts
when work performance is expected to last less than 60 days and an unusual
situation exists that warrants impositions of the requirements. This clause
should not be used in the same contract with clauses covering other man-
agement approaches for ensuring that a contractor makes adequate progress.
 Unified Facilities Guide Specification 609

Coordination is required with other Division 1 specifications when Network

Analysis Schedules is not specified.
Comments and suggestion on this specification are welcome and should be
directed to the technical proponent of the specification. A listing of the tech-
nical proponents, including their organization designation and telephone
number, is on the Internet.
Recommended changes to a UFGS should be submitted as a Criteria
Change Request (CCR).
Use of electronic communication is encouraged.
Brackets are used in the text to indicate designer choices or locations where
text must be supplied by the designer.

Note: This guide specification requires project costs to be loaded into the
schedule and assigned to activities. When using this section, delete the
requirement for “Schedule of Prices” in Section 01200 “Price and Payment
Procedures.”

PART 1. GENERAL
1.1. DESCRIPTION

Prepare a progress chart pursuant to the clause entitled “FAR 52.236-15,

Schedules for Construction Contracts” of the Contract Clauses that shall con-
sist of a network analysis system. The network analysis system shall consist
of the network analysis schedule (diagram), mathematical analysis, and asso-
ciated reports. The scheduling of construction shall be the responsibility of
the Contractor. Submission of progress and revision data will be used to
measure work progress, aid to evaluate time extensions, and provide basis of
all progress payments. The Critical Path Method (CPM) of network calcula-
tion shall be used to generate the project schedule and will utilize the
Precedence Diagram technique to satisfy both time and cost applications. All
progress payment amounts will be derived from and tied to the cost-loaded
schedule activities.
1.2. SUBMITTALS

Note: The “G” in submittal tags following each submittal item indicates
Government acceptance and should be retained. Add “G” in submittal tags fol-
lowing any added submittals that are determined to require Government
acceptance. Submittal items not designated with a “G” will be approved by
the QC organization.

Submit the following in accordance with Section 01330, “Submittal

Procedures.”
SD-01 Preconstruction Submittals
[Qualifications; G]
Standard Activity Coding Dictionary
[Schedule Development Session scheduler/planner; G]
610 Appendix B

[Preliminary Network Analysis Schedule; G]

Network Analysis Schedule; G
Accepted Network Analysis Schedule; G
Summary Network
As-Built Schedule
SD-07 Certificates
Monthly Network Analysis Updates; G
SD-11 Closeout Submittals

1.3. SCHEDULE ACCEPTANCE

Review comments made by the Government on the Contractor’s construction

schedule will not relieve the Contractor from compliance with requirements of
the Contract Documents. The Contractor is responsible for scheduling, sequenc-
ing, and prosecuting the Work to comply with the requirements of the Contract
Documents. Government acceptance extends only to the activities of the con-
tractor’s schedule that the Government has been assigned responsibility for
and agrees it is responsible. The Government will also review for contract
imposed schedule constraints and conformance, and cost loading of the CPM
activities. Comments offered on other parts of the schedule which the Contractor
is assigned responsibility are offered as a courtesy and are not conditions of gov-
ernment acceptance; but are for the general conformance with established indus-
try schedule concepts.
1.3.1. Schedule Acceptance Prior to Start of Work

Note: Prior to including or editing this paragraph, contact the ROICC Field
Office to determine if the contractor will be allowed to start work prior to
acceptance of the project schedule.

The Accepted Network described in the paragraph entitled “Accepted Network

Analysis Schedule” must be submitted and accepted by the government before
the contractor will be allowed to start work.
1.3.2. Acceptance

a. When the Accepted Network Analysis Schedule is submitted and accepted

by the Contracting Officer, it will be considered the “Baseline CPM
Schedule.” The Baseline CPM Schedule will then be used by the Contractor
for planning, organizing, and directing the work; reporting progress; and
requesting payment for work accomplished. The schedule will be updated
monthly by the Contractor and submitted monthly with the progress pay
request to reflect the current status of the work. [For payment requests
made after the period covered by the Preliminary Schedule.] The submit-
tal and acceptance of the Accepted Network Analysis Schedule and accu-
rate updated schedules accompanying the pay requests are both conditions
precedent to processing pay requests. Only bonds will be paid prior to
acceptance of the Accepted Network Analysis Schedule.
 Unified Facilities Guide Specification 611

b. Submittal of the Network, and subsequent schedule updates, will be under-

stood to be the Contractor’s representation that the submitted schedule
meets all of the requirements of the Contract Documents, accurately
reflects the work accomplished, and that Work will be executed in the
sequence indicated on the submitted schedule.

1.4. SOFTWARE

Note: Check with the ROICC Field Office for local personal computer (PC)
equipment capacity and edit as appropriate. The Contractor’s software may
require more computer capacity than the ROICC Field Office has available,
in which case, subject to the written approval of the Contracting Officer, the
contract may include the requirement for the contractor to provide hardware
and software necessary to allow the government to monitor work progress and
process payments. At the end of the contract term, this equipment software
may be specified to remain the property of the contractor or become govern-
ment property as determined to be most cost effective by the Contracting
Officer. Should this equipment be specified to become government property,
all property control regulations must be followed.

Note: Include the bracketed sentences requiring Primavera software for

LANTDIV, PACDIV and SOUTHDIV projects. Consult with the EFD/EFA 05
and/or Field Office to determine which software will be used. As a general
guide; for projects less than $5 Mil use SureTrak and for projects $5 Mil and
greater use P3.

[The scheduling software that will be utilized by the government on this proj-
ect is [SureTrak by Primavera Systems, Inc.] [Primavera Project Planner (P3)
by Primavera Systems, Inc. If the contractor chooses to use an equally capable
program, the contractor shall convert all data into Primavera Machine Readable
Format (Lotus, D-Base, Excel, etc.) prior to submission of all schedule inputs,
included but not limited to the initial schedule, monthly updates, and changes
to the schedule. It is the responsibility of the Contractor to ensure all data ele-
ments and logic required by this specification are kept intact during the con-
version to Primavera. If scheduling software other than Primavera is being
used, provide]] [Provide] a licensed copy of the Contractor’s scheduling soft-
ware and data. The software will be the most current version available and will
be compatible with all MS-Windows operating systems (e.g., Win NT, Win 95,
etc.). The scheduling software package shall contain all user manuals normally
provided by the software distributor. If the Contractor upgrades their software
during the course of the contract, the upgrade shall also be provided to the
Contracting Officer. The software will remain the property of the government.

1.4.1. Computer Hardware

[The network analysis software shall be capable of running on a [Government

owned] [Contractor provided] personal computer.] [Provide and maintain a
[_____] personal computer (PC) capable of running the network analysis soft-
ware specified herein.] All necessary software and hardware will be provided to
612 Appendix B

make the system a complete and useable package.] [Provide a [_____] [printer]
[plotter] with necessary cables. The contractor PC will remain the property of
the [Contractor] [Government].
1.4.2. Software Training

Note: Select and edit this paragraph when training is needed. Coordinate
with the ROICC Field Office.

[If software other than Primavera is used by the Contractor, provide] [Provide]
schedule software training for [two] [_____] Government personnel. A firm
accredited by the scheduling software manufacturer, as their authorized trainer
shall conduct the training. The training shall last a minimum of 24 hours per
individual. Provide course material the training firm normally distributes at
their software classes. Provide all necessary materials and equipment to con-
duct the training. The Contractor shall provide training within 10 working days
after notification to the Contractor, by the Contracting Officer. Unless agreed
to by the Contracting Officer, the training site shall be at the Contracting Office.

1.5. QUALIFICATIONS

Note: Before editing the following paragraph, coordinate with the ROICC
Field Office.

The Contractor shall designate a [full time] [part time] Scheduler that will be
responsible for the development, preparation, and maintenance of an accurate,
computerized Network Analysis Schedule. [Part time is defined as the sched-
uler performing on-site coordination, attending project meetings, and updates
for [_____] hours per work week.] The Scheduler shall have previously developed,
created and maintained at least [2] [_____] previous computerized schedules of
similar size and complexity of this contract. A resume outlining the qualifica-
tions of the scheduler shall be submitted for acceptance to the Contracting
Officer. If at a later date, the Contracting Officer considers the Contractor’s
Scheduler to be incompetent or objectionable, the Contractor will propose a new
Scheduler, meeting the qualification requirements. Payments will not be
processed until an acceptable Scheduler is provided.

1.6. NETWORK SYSTEM FORMAT

The system shall consist of time scaled logic diagrams accompanying mathe-
matical analyses and specified reports.
1.6.1. Diagrams

Show the order and interdependence of activities and the sequence in which the
work is to be accomplished as planned. The basic concept of a network analysis
diagram will be followed to show how the start of a given activity is dependent
on the completion of preceding activities and how its completion restricts or
restrains the start of following activities. Diagrams shall be [organized by [Work
Phase] [Area Code] and] sorted by Early Start Date and will show a continuous
flow from left to right with no logic (relationship lines) from right to left. With the
 Unified Facilities Guide Specification 613

exception of the Project Start and Project Completion milestone activities, no

activities will be open-ended, each activity will have predecessor and successor
ties. The diagram shall clearly show the activities of the critical path. No onsite
construction activity shall have duration in excess of 20 working days. Once an
activity exists on the schedule it may not be deleted and must remain in the logic.
No more than [20] [_____] percent of the activities may be critical or near criti-
cal. Critical will be defined as having zero days of Total Float. “Near critical” will
be defined as having Total Float in the range of [1 to 14] [[_____] to [_____]] days.
Show the following information on the diagrams for each activity:
a. Activity/Event Number
b. Activity Description
c. Original Duration in work days
d. Actual Duration in Work Days
e. Early Start Date
f. Early Finish Date
g. Total Float [or Slack]
h. Responsibility Code
Provide network diagrams on ANSI E sheets. Updated diagrams shall show the
date of the latest revision.

1.6.2. Quantity and Numbering of Activities

Note: A good knowledge of construction is required when determining the

numbers of activities for a network analysis. Factors such as the nature of
the work, geographical location, completion time, complexity (“the complex-
ity of a project is related to the number of specification sections, the number
of buildings, special phasing requirements and special quality control require-
ments”), cost of maintaining each activity throughout the life of the contract
and level of use by field management personnel must be considered. As a gen-
eral rule, use the following guidance:

Project Number of Construction

Construction Cost Activities Recommended

Up to $1,000,000 150 ± activities

$1,000,000 to $2,000,000 150 to 200 activities
$2,000,000 to $5,000,000 200 to 1000 activities
$5,000,000 to $10,000,000 1000 to 2000 activities
Over $10,000,000 2500 ± activities

*Important
■ When selecting the number of activities, please keep in mind the cost added
to the contract. An activity needs to be maintained throughout the life of
the contract and the use of too many activities will unnecessarily increase
the total contract cost.
614 Appendix B

■ The guidance provided above is meant as GUIDANCE. Use your best

judgement for selecting number of activities. Some contracts may require
less number of activities than recommended amounts. (Ex: A contract to
stripe a 500-mile stretch of highway may have a project cost of $6,000,000
but it should not require between 1000 to 2000 activities).

Numbering shall be assigned so that, in general, predecessor activity num-

bers are smaller numerically than the successor activity numbers. Skip num-
bering shall be used on the network to allow insertion of additional activities
for contract modifications and logic changes. The minimum number of con-
struction activities in the final network diagram shall be [_____]. Types of activ-
ities included in the schedule are specified below.

1.6.2.1. HVAC TAB Activities

Note: This paragraph will be used only when HVAC Testing, Adjusting and
Balancing work is specified in the contract specifications.

Requirements for the activities related to HVAC TAB work, Section entitled,
“HVAC Testing/Adjusting/Balancing,” are specified in Section entitled, “Price and
Payment Procedures.”

1.6.2.2. Procurement Activities

Tasks related to the procurement of material or equipment shall be included

as separate activities in the project schedule. Examples of procurement activ-
ities include, but are not limited to: Material/equipment submittal preparation,
submittal and approval of material/equipment; delivery of O&M manuals;
material/equipment fabrication and delivery, delivery of extra parts, extra stock,
special tools, notification of Government Furnished Material/Equipment deliv-
ery requirement, etc. As a minimum, separate procurement activities will be pro-
vided for every specification section. If the Contractor intends on using
Just-In-Time (JIT) delivery methods, the schedule will show each JIT delivery
with relationship tie to the Construction Activity specifically for the JIT deliv-
ery. Material and equipment for which payment will be requested in advance
of installation shall be cost-loaded with the procurement costs. All activities
within a procurement process/cycle will have a unique identifier in the activity
code to show their relationships and will extend to the related construction
activities (i.e., Work Category).
If the Government’s action on any submittal is “Disapproved” or “Revise and
Resubmit,” a new series of Procurement Activities will be inserted into the
schedule. Predecessor for the new submittal preparation activity will be the orig-
inal approval activity and the successor of the new approval activity will be the
fabrication/deliver activity for the equipment or material.

1.6.2.3. Government Activities

Government and other agency activities that could impact progress shall be
clearly identified. Government activities include, but are not limited to;
Government approved submittal reviews, Government conducted inspec-
 Unified Facilities Guide Specification 615

tions/tests, utility outages, Notice(s) to Proceed and delivery of Government

Furnished Material/Equipment. Show activities indicating Government fur-
nished materials and equipment utilizing delivery dates indicated in “FAR
52.245-2, Government Property (Fixed-Price Contracts).” Government activities
will be driven by calendars that reflect Saturdays, Sundays and all Federal
Holidays as non-work days.
1.6.2.4. Construction Activities

Construction activities shall include, but are not limited to: Tasks related to
mobilization/demobilization; the installation of temporary or permanent work
by tradesmen; testing and inspections of installed work by technicians, inspec-
tors or engineers; start-up and testing of equipment; commissioning of building
and related systems; scheduling of specified manufacture’s representatives;
final clean-up; training to be provided; and administrative tasks necessary to
start, proceed with, accomplish or finalize the contract. Contractor activities will
be driven by calendars that reflect Saturdays, Sundays and all Federal Holidays
as non-work days.
1.6.2.5. Anticipated Weather Delays

Schedule activity duration(s) shall be formulated with allowance for normal

adverse weather conditions. Any activity duration which could be impacted by
normally anticipated adverse weather (precipitation, high or low temperature,
wind, etc.), due to the time period which the Contractor has scheduled the work,
shall include an adjustment to include the anticipated weather delay. The
Contractor shall anticipate delay by comparing the contractually imposed envi-
ronmental restrictions in the Contract Documents to the National Oceanic and
Atmospheric Association’s (NOAA) historical monthly averages for the NOAA
location [at (Enter NOAA Station here)] [closest to the project site]. The number
of anticipated adverse weather delays allocated to an activity will be reflected
in the activity’s calendar. A lost workday, due to weather conditions, is defined
as a day in which the contractor’s workforce cannot work 50 percent or more of
the day. The Contractor shall immediately notify the Contracting Officer when
a lost day has occurred due to weather and will record on the Daily Reports, the
occurrence of adverse weather and resultant impact to the normally scheduled
work. If the number of actual adverse weather delay days exceeds the number
of days anticipated, the Contracting Officer will convert any qualifying delays
to calendar days, giving full consideration for equivalent fair weather work
days and issue a modification in accordance with the contract clauses.
1.6.2.6. Activity Properties

Schedule activities will have the following properties:

a. Standard Activity Coding Dictionary: The Contractor shall submit a coding
scheme for Schedule Activity Numbers that shall be used throughout the
project. The coding scheme submitted shall list the values for each activ-
ity code category and translate those values into project specific designa-
tions. Code length shall not exceed [10] [_____] characters. Once accepted,
the coding scheme will be used for the duration of the project.
616 Appendix B

b. Activity Description: Each activity shall have a narrative description con-

sisting of a Verb or work function (e.g.; form, pour, excavate), an Object (e.g.;
slab, footing, underfloor plumbing), and Area (e.g., 3rd floor, northeast
quadrant, basement).

Note: Include the following paragraph when the project includes the
requirement for Phased Construction.

c. Work Phase: If phasing is specified in the contract, all activities shall be

identified in the project schedule by the phase of work in which the activ-
ity occurs. Activities shall not be contained in more than one Work Phase.]
d. Work Category: All Activities shall be identified in the project schedule
according to the work category which best describes the activity. Examples
of work categories are procurement, government, and construction activ-
ities that are all related to a single Definable Feature of Work. Activities
shall not be contained in more than one Work Category.
e. Area Code: All activities shall be identified in the project schedule by the
Area Code in which the activity occurs. Activities shall not be contained
in more than one Area Code. Area is defined as a distinct separation in con-
struction, such as a story of construction, separate structure, usage or
function difference, utility distribution systems, etc.
f. Responsibility Code: All activities in the project schedule shall be identi-
fied with the party responsible to perform the task. Responsibility includes,
but is not limited to; the prime contractor, subcontracting firm, or
Government agency performing a given task. Activities shall not belong to
more than one responsible party. The responsible party for each activity
shall be identified by a responsibility code. For example, a responsibility
code value, “ELEC”, may be identified as “Electrical Subcontractor.”
g. CSI Code: All activities in the project schedule shall be identified with its
respective 5-digit Specification Section number. Activities shall not belong
to more than one Section number. If an activity does not have an applica-
ble CSI Code, [such as “Mobilize”], the code will be “00000.”
h. Drawing Code: All activities in the project schedule shall be identified
with its respective project drawing code. The drawing code is the Sheet
Number on the primary project drawing which indicates the work to be per-
formed. Activities shall not belong to more than one Drawing Code.
Examples of Drawing Codes are “C-10,” “C.10” or “C10.” The code system
will allow organizing all activities by drawing code in alpha and numeric
order. If an activity does not have an applicable Drawing code, (such as
“Mobilize”), the code will be “00000.”
i. Modification Code: The Modification Code shall identify activities that
are modified or added by contract modification. Activities shall not belong
to more than one Modification Code. The Government will assign the mod-
ification number, which will be shown on the SF 30. Use a shortened ver-
sion of the modification number for the code (e.g.; A00010 = 010).
j. Request for Equitable Adjustment (REA) or Claim Code: Activities that are
modified or added, as a result of a Contractor’s REA or Claim shall be
 Unified Facilities Guide Specification 617

identified by a code generated by the Contractor. Activities shall not belong

to more than one REA or Claim Code.
k. The Three Phases of Control (Preparatory, Initial, and Follow-up): For
each Definable Feature of Work identified in the Contractor’s Quality
Control Plan, include an activity for the Preparatory Phase. The Initial
Phase and Follow-up Phase will be represented by the Construction
Activities in the schedule.
l. Project Milestone Dates: Dates shall be shown on the diagram for the start
of the project, any contract required interim start and completion dates,
contract completion date and other significant milestones.
m.Scheduled Project Duration: The schedule duration shall extend from
notice-to-proceed to the contract completion date.
n. Project Start Date Milestones: The schedule shall start no earlier than the
contract award date and the project duration (Day 1) will start on the
Notice-to-Proceed (NTP) date. The Contractor shall include as the first
activity in the schedule, an activity named “Contract Award” and another
activity on the NTP date named “Start Project.” Both activities will be
zero duration, with constrained start dates equal to the contract award and
NTP dates.
o. Constraint of Last Activity Milestone: The Contractor shall include as the
last activity in the project schedule, an activity named “End Project.” The
“End Project” activity shall be zero duration with a mandatory finish con-
straint equal to the contract completion date for the project. Calculation
of project updates shall be such that if the finish of the last activity falls
after the contract completion date, then the float calculation shall reflect
negative float on the critical path.
p. Early Project Completion: In the event the Contractor’s project schedule
shows completion of the project prior to the contract completion date, the
Contractor shall include an activity named “Contractor Early Completion.”
The activity shall be a zero duration milestone with an unconstrained date
representing the Contractor’s Early Completion date.
q. Substantial Completion: If the contractor elects to include an activity for
Substantial Completion, then it is agreed that Substantial Completion
will be the point in time that the Government considers the project is com-
plete and ready for its intended use. The activity will be named
“Substantial Completion.” The activity shall be a zero duration milestone
with an unconstrained date representing the Contractor’s Substantial
Completion date.

Note: Include the following three paragraphs when the project includes the
requirement for Phased Construction.

r. Phase Start Milestone: The Contractor shall include as the first activity
for a project phase, an activity named “Start Phase X,” where “X” identi-
fies the phase of work. The “Start Phase X” activity shall be zero duration
with an unconstrained start date equal to the date of the Phase NTP. This
unconstrained start date is not a release from contractually required start
618 Appendix B

dates, but is left unconstrained to allow the schedule logic to calculate

without hindrance.
s. End Phase Milestone: The Contractor shall include as the last activity in
a project phase, an activity named “End Phase X” where “X” identifies the
phase of work. The “End Phase X” activity shall be zero duration with an
unconstrained late finish date equal to the contract phase completion date.
This unconstrained completion date is not a release from contractually
required finish dates, but is left unconstrained to allow the schedule logic
to calculate without hindrance.
t. Early Phase Completion: If the contractor expects to finish prior to the con-
tract phase completion date, the milestone will show an early finish date
equal to the Contractor’s early finish date. The name of the activity will
be “Early Phase Completion” and will be zero duration with an uncon-
strained date representing the contractor’s early phase completion date.]

Note: Include Summary Activities if Summary Networks will be requested

or if repetitive groups of activities will be used in a project (e.g. similar hous-
ing units being built several times over). Also include if Summaries will assist
in keeping Customer or Higher Management appraised of progress.

u. Summary (a.k.a., Banding or Hammock) Activities: The Contractor shall

include special activities that are a summary of a chain of activities. The
start of the activity will be the start date of the first activity in the chain
and the finish date will be the finish date of the last activity in the chain.
Generalized work sequences, Categories of Work and all Phase of Work
activity chains will be summarized.]
v. Activity/Event Constraints: Date/time constraint(s), other than those
required by the contract, will not be allowed unless accepted by the
Contracting Officer.
w. Leads and Lags: Leads or lags will not be used when the creation of an
activity will perform the same function (e.g., concrete cure time). Lag dura-
tions contained in the project schedule shall not have a negative value. The
use of any lead or lag will be explained in the Narrative Report.
x. Default Progress Data Disallowed: Actual Start and Finish dates shall
not be automatically updated by default mechanisms that may be included
in the CPM scheduling software system. Actual Start and Actual Finish
dates on the CPM schedule shall match the dates provided from Contractor
Quality Control and Production Reports. These reports will be the sole basis
for updating the schedule. Work activities will be updated by actual work
progression rather than being cash flow driven. The updating of the per-
cent complete and the remaining duration of any activity shall be inde-
pendent functions; program features that calculate one of these parameters
from the other shall be disabled. Out-of-Sequence progress (if applicable)
shall be handled through Retained Logic, not the Default Option of
Progress Override. Actual labor and equipment hours used on activities will
be derived from the daily reports.

1.6.3. Mathematical Analysis

 Unified Facilities Guide Specification 619

The network diagram mathematical analysis shall include a tabulation of each

activity shown on the detailed network diagrams. Provide the following infor-
mation as a minimum for each activity:
a. Activity/Event number
b. Activity/Event description
c. Estimated duration of activities (by work days)
d. Earliest start date (by calendar date)
e. Earliest finish date (by calendar date)
f. Actual start date (by calendar date)
g. Actual finish date (by calendar date)
h. Latest start date (by calendar date)
i. Latest finish date (by calendar date)
j. Total float or slack
k. Material/Equipment costs will be assigned to their respective
Procurement Activities (i.e., the delivery activity). Costs for installation
of the material/equipment (labor, construction equipment, and temporary
materials) will be assigned to their respective Construction Activities.
The value of inspection/testing activities will not be less than [10] [_____]
percent of the total costs for Procurement and Construction Activities.
Evenly disperse overhead and profit to each activity over the duration of
the project.
l. Responsibility code (including prime contractor, subcontractors, suppli-
ers, Government, or other party responsible for accomplishment of an
activity.)
m. Area Code
n. Manpower required (crew size)
o. Percentage of activity duration completed
p. Contractor’s earnings based on accepted work-in-place.
The program or means used in making the mathematical computation shall
be capable of compiling the total value of completed and partially completed
activities. The program shall also be capable of accepting revised completion
dates as modified by approved time extensions and recompilation of tabulation
dates/costs and float accordingly. The total of all cost loaded activities; includ-
ing costs for material and equipment delivered for installation on the project,
and manpower and construction equipment loaded construction activities, shall
total to 100 percent of the value of the contract.

1.6.4. Additional Requirements

Note: The information required by the following paragraphs are optional and
typically not needed for routine work. Include on projects with critical com-
pletion dates. Manpower and equipment loading schedules are of primary
620 Appendix B

importance to the Contractor in deciding the most efficient use of personnel

resources and optimizing equipment usage and is the basis of activity dura-
tion estimates. Since these decisions are the responsibility of the Contractor’s
management process, the information value to the Contracting Officer is in
assuring that the planned manpower and equipment are being supplied
throughout the course of the project.

In addition to the tabulation of activities, in the Paragraph entitled

“Mathematical Analysis,” include the following data:
a. On-site manpower loading schedule: Each construction activity shall have
an estimate of the number of workers per day by trade, man-hours per day
by trade and total expected hours used by trade during the execution of
the activity. If no workers are required for an activity, then the activity shall
be identified as using zero workers per day.
b. Equipment loading schedule: Each construction activity shall have an
estimate of the equipment used per day, number of units per day and total
expected hours for each piece of equipment used during the duration of the
activity. Include a description of the major items of construction equipment
planned for each construction activity on the project. The description shall
include the year, make, model, and capacity. If no equipment is required
for an activity, then the activity shall be identified as using zero equipment
per day.

1.6.5. Required Reports

Note: Consult with the ROICC Field Office to identify which of the follow-
ing reports are preferred. Always include Earned Value Report and Log
Report.

The following reports will be made available in the schedule submittals and in
each updated schedule submission provided on disk by the Contractor:
a. By the preceding event number from lowest to highest and then in the order
of the following activity number (Activity Identification Report) showing
the current status of all activities.
b. By the amount of total float, from lowest to highest and then in order of
[activity number] [early start date] (Total Float or Slack Report) showing
all incomplete activities.
c. By latest allowable start dates and then in order of activity numbers (Late
Start Report).
d. Earned Value Report listing all activities having a budget amount and
cost. A compilation of total earnings on the project from the notice to pro-
ceed to the most recent monthly progress payment request and the dif-
ference between the previous request amount and the current payment
request amount. Sort report first by resource and then by activity.
 Unified Facilities Guide Specification 621

e. By earliest allowable start dates and then in order of activity number

(Early Start Report).
f. By tasks scheduled to start and finish by the end of the next pay period
(30-Day Look Ahead).
g. With each updated schedule submission, provide a computer generated Log
Report using a recognized schedule comparision software listing all changes
made between the previous schedule and current updated schedule. Identify
the name of the previous schedule and name of the current schedule being
compared. This report will as a minimum show changes for: Added & Deleted
Activities, Original Durations, Remaining Durations, Activity Percent
Complete, Total Float (or Slack), Free Float, Calendars, Descriptions,
Constraints (added, deleted or changed), Actual Starts/Finishes, Added/
Deleted Resources, Resource Quantities, Costs, Resource Percents,
Added/Deleted Relations, Changed Relation Lags, Changed Driving
Relations, and Changed Critical Status.
h. By the activity number from lowest to highest, showing preceding and
succeeding activity numbers for each activity (Predecessor/Successor
Report), and showing the current status of each activity.

Note: Include the following two paragraphs if the requirements of the para-
graph entitled “Additional Requirements” are specified.
i. Manpower staffing report and histogram: With each update schedule, a
planned early and planned late versus actual labor resource histogram
will be provided. This histogram shall be based upon and shall be in
agreement with, the number of shifts and crew sizes by craft, in the
Accepted Network Analysis Schedule (planned) and the Monthly Network
Update (actual). Included in the report will be a tabular report that will
list each trade to the activities that were worked on during the con-
struction period.
j. Equipment usage report and histogram: With each update schedule, a
planned early and planned late versus actual equipment resource his-
togram will be provided. This histogram shall be based upon and shall be
in agreement with the equipment allocation accepted on the Accepted
Network Analysis Schedule (planned) and the Monthly Network Update
(actual). Included in the report will be a tabular report that will list equip-
ment (by make and model) to the activities that were worked on during the
construction period.]

1.7. SUBMISSION AND ACCEPTANCE

1.7.1. Preliminary Meeting

At the Pre-Construction Conference, the Contracting Officer, Contractor and

major subcontractors shall participate in a preliminary meeting to discuss the
proposed schedule and requirements of this section prior to submission of the
network. The definition of a “major subcontractor” is one that exceeds [5] [_____]
percent of the contract value.
622 Appendix B

1.7.2. Schedule Development Session:

Note: Contact the ROICC Field Office before including this paragraph in the
specifications. If included, editing of the paragraph will be coordinated with
the Representative. This paragraph will typically be used only on large, com-
plex or schedule sensitive projects.

Upon completion of the 90 day (Preliminary) Network Analysis Schedule, and]

Prior to the submission of the Network Analysis Schedule, the Contractor shall
conduct a Schedule Development Session. The Schedule Development Session
shall include procurement of on site services of an expert scheduler/planner for
not less than a [5] [_____] day period. The Contractor’s choice of Schedule
Development Session scheduler/planner is subject to the acceptance of the
Contracting Officer. The scheduler/planner shall facilitate the session on site [and
shall be fluent in the English language]. The scheduler/planner shall have at least
[10] [_____] years experience developing construction project schedules with
scheduling software programs that the contractor intends to use. Unless agreed
to by the Contracting Officer, the session shall be conducted at the Office of the
Contracting Officer. The Contractor is responsible for providing the necessary
equipment for the session which, as a minimum, includes a personal computer
(PC), a computer display projector to facilitate group viewing, and a printing
device. During the session the facilitator [shall provide all necessary training
to participants and] shall lead the development of the project’s schedule. As a min-
imum, the scheduler/planner shall facilitate development of activity coding and
work breakdown structures; establishment of procurement, government, and con-
struction activities; activity relationship; resourcing; budgeted costs; and reports
to be used during the project. Members of the Contracting Officer’s staff will attend
the session as well as [members of the designer of record,] [customer who will
occupy the facility,] [major subcontractors (those which exceed [5] [_____] percent
of the contract value), and] the Contractor’s home and field project management
staff. [Past experience has revealed that these services do not exist in [Indicate proj-
ect location] which has resulted in the Contractor forming agreements with
Scheduling Firms [in the United States] to meet the terms of the specification
requirement.] All costs associated with the Schedule Development Session are
to be borne by the Construction Contractor.]
1.7.3. Preliminary Network Analysis Schedule

Note: This paragraph should only be used on complex contracts. Do not use
this paragraph on contracts that require an Accepted Network Analysis
Schedule to be submitted and accepted by the Government prior to beginning
work.

Submit a preliminary network defining the planned operations during the first
[90] [_____] calendar days after contract award within [20] [_____] days after con-
tract award. The general approach for the balance of the project shall be indi-
cated. Cost of activities expected to be completed or partially completed before
submission and acceptance of the Accepted Network Analysis Schedule should
be included. Submit three copies of both the preliminary network diagrams and
required reports listed in paragraph entitled “Required Reports.” In accordance
 Unified Facilities Guide Specification 623

with paragraph entitled “Monthly Reports,” the preliminary network may be

used for requesting progress payments for a period not to exceed 90 calendar
days after receipt of “Contract Award.” Submittal and acceptance of the
Preliminary Network is condition precedent to the processing of the Contractor’s
pay requests on this schedule. Payment requests after the first [90] [_____] cal-
endar day period shall be based upon the Accepted Network Analysis Schedule.
The activities and relationships of the preliminary schedule shall coincide and
mesh with the activities of the Network Analysis Schedule. As part of this sub-
mittal, provide the Project Name format (and Project Group Name if used) that
will be used by the Contractor to identify initial schedule submittals, updates,
fragnets, changes, etc. Include [1] [_____] copy of the Preliminary Network
Analysis Schedule on 3.5” disk(s).
1.7.4. Network Analysis Schedule

Note: In the first sentence, remove the language in the brackets if the sched-
ule is to be submitted prior to allowing the contractor to commence work. If
a Preliminary Schedule is required remove the language for Project Name
format.

Submit the complete network system, consisting of the network mathematical

analysis and network diagrams [, within 40 [_____] calendar days after contract
award]. Submit [three] [_____] copies of the diagrams described in the paragraph
entitled “Diagrams,” the required reports listed in the paragraph entitled “Required
Reports,” [and] the analysis described in the paragraph entitled “Mathematical
Analysis” [and information required by the paragraph entitled “Additional
Requirements”]. [As part of this submittal, provide the Project Name format (and
Project Group Name if used) that will be used by the Contractor to identify initial
schedule submittals, updates, fragnets, changes, etc.] Include [1 copy] [_____ copies]
of the Network Analysis Schedule on 3.5” disk(s) formatted to hold 1.44 MB of data.
1.7.5. Review and Evaluation

After the Government’s review, the Contractor shall meet with the Contracting
Officer to discuss the review and evaluation of the NAS submittal. Revisions nec-
essary as a result of this review shall be resubmitted for acceptance within 10
calendar days after the meeting.
1.7.6. Accepted Network Analysis Schedule

Once review comments are resolved and the network has been accepted by the
Contracting Officer, the Contractor shall within 5 calendar days furnish:
a. [Two] [_____] copies of the network diagrams
b. [Two] [_____] copies of the required reports listed in paragraph entitled
“Required Reports”
c. [Two] [_____] copies of the “Mathematical Analysis.”
d. [Two] [_____] copies of the Cash Flow Report indicating the cash flow based
upon both the early and late start schedules.
624 Appendix B

e. [Two] [_____] copies of each major subcontractor’s statement certifying

their concurrence with the Contractor’s Accepted Network Analysis
Schedule. Each certifying statement will be made on the subcontractor’s
letterhead.
f. [Two] [_____] sets of data disks containing the project schedule shall be pro-
vided for the initial submission and every periodic project update. Data shall
be submitted on 3.5: disk(s), formatted to hold 1.44 MB of data. A permanent
exterior label shall be affixed to each disk submitted. The label shall indicate
the type of schedule (Preliminary, NAS Submittal, Accepted, Update,
Recovery, or Change), full contract number, Project Name used to identify
project in scheduling software, contract name & location, data status date,
diskette number with total number of diskettes in set, software name and ver-
sion used to run the schedule, and the name and telephone number of person
responsible for the schedule.
For major revisions, updates or changes to the network diagrams, once
accepted by the Contracting Officer, the Contractor shall submit these same dia-
grame and reports.

1.7.7. Monthly Network Analysis Updates

At monthly intervals the Contractor, Government representatives and major

subcontractors will meet to jointly update the project schedule and agree on per-
centage of payment for each activity progressed during the update period. The
purpose of the meeting is to determine progress payment amounts for each
activity, allow all parties to evaluate project status at the data date, provide a
complete and accurate update of procurement and construction progress, create
an historical record of the project and establish prediction of completion date(s)
based upon current status. The Contractor is responsible to gather all support-
ing documentation, propose the update data for the schedule and record the
meeting minutes. All progress payment amounts will be derived from and tied
to the cost-loaded schedule activities. Submit at monthly intervals a report of the
actual construction progress by updating the required reports, the time scaled
logic diagram, and mathematical analysis. Meeting to update the schedule and
the submission of an error free, acceptable updated schedule to the Government
is a condition precedent to the processing of the Contractor’s pay request. As a
minimum, the following actions will be accomplished during the meeting:
a. Identify activities started and completed during the previous period and
enter the Actual Start and Actual Finish dates.
b. Show estimated duration (in workdays) to complete each activity started
but not completed (remaining duration).
c. Indicate percentage of cost payable for each activity.
d. Reflect changes in the network diagram. All changes (i.e., duration changes,
logic changes, new logic, conformed change orders, new activities, changes
due to Conformed Modifications, changes in work sequence, etc.) shall be
recorded and a note added to the activity log field. The log shall include as a
minimum, the date and reason for the change, and description of the change.
 Unified Facilities Guide Specification 625

e. Submit [two] [_____] copies of a Narrative Report describing: 1) Progress

made in each area of the project; 2) Changes in the following; activities,
original durations, logic interdependencies, milestones, planned sequence
of operations, critical path, and resource and loading; 3) Pending items and
status thereof, including permits, change orders, and time extensions; 4)
Status of Contract Completion Date and interim milestones; 5) Current
and anticipated delays (describe cause of the delay and corrective
action(s)); and 6) Description of current and future schedule problem
areas. Each entry in the narrative report will cite the respective Activity
ID and Activity Description.
f. Submit [two] [_____] copies of the required reports listed in paragraph
entitled “Required Reports.”
g. Submit [two] [_____] copies of the Update Meeting minutes.

1.7.8. Summary Network

Note: Before specifying Summary Networks, verify with the ROICC Field
Office that the Summary will be useful on the project being designed. Choose
type of summary to be provided.

A summary network shall have the same network form as the Accepted Network
Analysis Schedule. The summary network will contain a minimal number of
activities that represent the general approach of work sequence. The Summary
will be a time-scaled logical sequence of [Work Phases] [Work Category] [Area
Code]. The Contractor shall submit a summary network diagram immediately
after acceptance of the Accepted Network Analysis Schedule. A summary net-
work update shall be submitted every [6] [_____] months during the contract
duration and immediately following acceptance of each major schedule change.
Submit the following:
a. [Two] [_____] copies of the summary network diagram.
b. [Two] [_____] copies of the Activity Identification Report.
c. [Two] [_____] copies of the Total Float (or slack) Report.
d. [Two] [_____] copies of the Earned Value Report indicating the actual cash
flow for the current updated (not summary) network based upon both the
early and late start schedules.

1.7.9. As-Built Schedule

Note: Before specifying As-Built Schedules, verify with the ROICC Field
Office that the schedule will be required.

As a condition precedent to the release of retention, the last update of the sched-
ule submitted shall be identified by the Contractor as the “As-Built Schedule.”
The As Built shall reflect the exact manner in which the project was actually
constructed (including actual start and finish dates, activities, sequences, and
logic) and shall be certified by the Contractor’s Project Manager and
Construction Scheduler as being a true reflection of the way the project was actu-
626 Appendix B

ally constructed. If more than one person filled the position(s) during the course
of the project, each person will provide certification for the period of time they
were responsible.

1.8. CONTRACT MODIFICATION

When a contract modification to the work is required, submit proposed revisions

to the network with a fragnet and a cost proposal for each proposed change. All
modifications shall be incorporated into the network analysis system as separately
identifiable activities broken down and inserted appropriately on the first update
following issuance of a directive to proceed with the change. Submit [one copy]
[_____ copies] of the Total Float Report, Log Report and a copy of the proposed
Time Impact Analysis on disk, with the cost proposal. Unless the Contracting
Officer requests otherwise, only conformed contract modification fragnets will be
added into the subsequent monthly updates. All revisions to the current base-
line schedule activities that are necessary to further refine the schedule so that
the changed work activities can be logically tied to the schedule shall be made.
Financial data shall not be incorporated into the schedule until the contract
modification is signed by the Contracting Officer.
1.8.1. Time Impact Analysis:

Time Impact Analysis shall be used by the Contracting Officer in determining

if a time extension or reduction to the contract milestone date(s) is justified. The
Contractor shall provide a Time Impact Analysis to the Contracting Officer for
any proposed contract change or as support for a Value Engineering Proposal,
Claim or Request for Equitable Adjustment by the Contractor.
a. The Contractor shall submit a Time Impact Analysis (TIA) illustrating
the influence of each change or delay on the Contract Completion Date or
milestones. Unless the Contracting Officer requests an interim update to
the schedule, the current monthly updated schedule accepted by the gov-
ernment shall be used to display the impacts of the change. Unless
requested by the Contracting Officer, no other non-conformed changes will
be incorporated into the schedule being used to justify the change impact.
b. Each TIA shall include a Fragmentary Network (fragnet) demonstrating
how the Contractor proposes to incorporate the impact into the Project
Schedule. A fragnet is defined as the sequence of new activities and/or
activity revisions, logic relationships and resource changes that are pro-
posed to be added to the existing schedule to demonstrate the influence of
impacts to the schedule. The fragnet shall identify the predecessors to the
new activities and demonstrate the impacts to successor activities. Include
a narrative report describing the effects of new activities and relation-
ships to interim and contract completion dates, with each TIA.
c. Following the Contractor’s receipt of a contract modification on a Standard
Form 30 signed by the Government; all changes in the fragnet used to
determine impacts, shall be incorporated into the schedule. Changes will
occur during the next monthly schedule update meeting.

1.8.2. No Reservation-Of-Rights
 Unified Facilities Guide Specification 627

All direct costs, indirect costs, and time extensions will be negotiated and made
full, equitable and final at the time of modification issuance.

1.9. CHANGES TO THE NETWORK ANALYSIS SCHEDULE

If changes in the method of operating and scheduling are desired, the

Contracting Officer shall be notified in writing stating the reasons for the
change. If the Contracting Officer considers these changes to be of a major
nature, the Contractor may be required to revise and submit for acceptance,
without additional cost to the Government, the network diagrams and required
sorts. A change may be considered of a major nature if the estimated time
required or actually used for an activity or the network logic is varied from the
original plan to a degree that there is a reasonable doubt as to the effect on the
contract completion date(s) [or phase completion dates]. Changes that affect
activities with adequate float time shall be considered a major change when their
cumulative effect could extend the contract completion date.
1.10. FLOAT

Use of float suppression techniques, such as; preferential sequencing (arrang-

ing critical path through activities more susceptible to government caused
delay), special lead/lag logic restraints, zero total or free float constraints,
extended activity times, or imposing constraint dates other than as required by
the contract, shall be cause for rejection of the project schedule or its updates.
The use of Resource Leveling (or similar software features) used for the purpose
of artificially adjusting activity durations to consume float and influence the crit-
ical path is expressly prohibited.
1.10.1. Definitions of Float or Slack

Free Float is the length of time the start of an activity can be delayed without
delaying the start of a successor activity. Total Float is the length of time along
a given network path that the actual start and finish of activity(s) can be delayed
without delaying the project completion date. Project Float is the length of time
between the Contractor’s Early Completion (or Substantial Completion) and the
Contract Completion Date.
1.10.2. Ownership of Float

Float available in the schedule, at any time shall not be considered for the
exclusive use of either the Government or the Contractor. During the course of
contract execution, any float generated due to the efficiencies of either party is
not for the sole use of the party generating the float; rather it is a shared com-
modity to be reasonably used by either party. Efficiencies gained as a result of
favorable weather within a calendar month, where the number of days of nor-
mally anticipated weather is less than expected, will also contribute to the
reserve of float. A schedule showing work completing in less time than the
Contract time, and accepted by the Government, will be considered to have
Project Float. Project Float will be a resource available to both the Government
and the Contractor. No time extensions will be granted nor delay damages paid
unless a delay occurs which impacts the Project’s critical path, consumes all
628 Appendix B

available float or contingency time, and extends the work beyond the Contract
Completion Date.
1.10.3. Negative Float

Negative float will not be a basis for requesting time extensions. Any extension of
time will be addressed in accordance with the Paragraph “Time Extensions.”
Scheduled completion date(s) that extend beyond the contract [or phase] completion
date(s) (evidenced by negative float) may be used in computations for assessment
of payment withholdings. The use of this computation is not to be construed as a
means of acceleration.

1.11. TIME EXTENSIONS

Extension of time for performance required under the clauses entitled

“Changes,” “Differing Site Conditions,” “Default (Fixed-Price Construction)”
or “Suspension of Work” will be granted only to the extent that equitable
time adjustments for the activity or activities affected exceed the total float
or slack along the network paths involved at the time Notice to Proceed was
issued for the change. The Contractor acknowledges and agrees that delays
in activities which, according to the network analysis schedule, does not in
fact actually affect any milestone completion dates or the contract completion
date shown on the CPM network at the time of delay, will not be a basis for
a contract extension. Submit time extension requests with a Time Impact
Analysis and three copies of the Total Float (or Slack) Report, Narrative
Report and Log Report.

1.12. MONTHLY COORDINATION MEETING

Note: Consult with the local ROICC Field Office on whether to use this
paragraph. Include this paragraph for larger or more complex projects.

In conjunction with receipt of the Monthly Network Update submission, a coor-

dination meeting will be held each month [on site] [in the Contracting Officer’s
conference room] to discuss the report. The Contractor shall make a presenta-
tion of the previously submitted and current Monthly Network Update to the
Contracting Officer so as to provide an overview of the project’s schedule and
provide an opportunity to discuss items of coordination.

1.13. BIWEEKLY WORK SCHEDULE

Note: Consult with the local ROICC Field Office on whether to use this
paragraph. Include this paragraph for larger or more complex projects.

To provide a more detailed day-to-day planning of upcoming work, the Contractor

shall prepare and issue detailed work plans that coordinate with and supple-
ment the above defined network analysis. The work plans shall be keyed to the
CPM activity numbers and shall be submitted each week and shall show the
projects activities that will occur during the following two-week interval.
 Unified Facilities Guide Specification 629

Additionally, the critical path activities are to be identified on the Biweekly

Work Plan. The detail work plans are to be bar chart type schedules prepared
by the Contractor in sufficient detail to define the work to be accomplished, the
crews, construction tools and equipment to be used during the current and next
two-week interval. The bar charts shall be formatted to allow reproduction on
8 1/2 by 11 sheets. Three copies of the bar chart schedules shall be delivered to
the Contracting Officer not less than 3 work hours prior to the start of the
weekly coordination meeting.
1.14. WEEKLY COORDINATION MEETING

Note: Consult with the local ROICC Field Office on whether to use this
paragraph. Include this paragraph for larger or more complex projects.

In conjunction with the receipt of the Bi-Weekly Work Schedule, a coordination

meeting will be held each week [on site] [in the Contracting Officer’s conference
room] to discuss the work schedule. The Contractor shall make a presentation
of the previously submitted and current Bi-Weekly Work Schedule to the
Contracting Officer so as to provide an overview of the project’s schedule and
provide an opportunity to discuss items of coordination. Consideration of mate-
rials, crews, and equipment shall be addressed to ascertain their respective
availability. The meeting shall identify actions necessary to provide adherence
to the Bi-Weekly Work Schedule and the overall network for the project defined
above. The Contractor will take meeting minutes. All meeting minute entries
will be keyed to the schedule activity number(s) being addressed. Within one
day of the meeting, the Contractor will provide a draft copy of the meeting min-
utes to the Contracting Officer for review and comment. Final copies of the
minutes containing the comments provided by the Contracting Officer, will be
issued within 3 days of the meeting.
1.15. CORRESPONDENCE AND TEST REPORTS

All correspondence (e.g., letters, Requests for Information (RFIs), e-mails, meet-
ing minutes, Production and QC Daily Reports, material delivery tickets, pho-
tographs, etc.) shall reference the Schedule Activity Number(s) that are being
addressed. All test reports (e.g., concrete, soil compaction, weld, pressure, etc.)
shall reference the Schedule Activity Number (s) that are being addressed.
This page intentionally left blank
 Appendix

C
Notation for RDCPM

Event ID# Internal event ID# Event (milestone)

& description description
Activity ID# Logic type Activity ID# Logic type
Activity description Logic description Activity description Logic description
ID# ID#
OD cal CRSZ CRTY Lead/lag cal OD cal CRSZ CRTY Lead/lag cal
Other codes Other codes Other codes Other codes
Logic type Activity ID# Logic type
Logic description Activity description Logic description
ID#
Lead/lag cal OD cal CRSZ CRTY Lead/lag cal
Other codes Other codes Other codes
Logic type Activity ID# Logic type
Logic description Activity description Logic description
ID#
Lead/lag cal OD cal CRSZ CRTY Lead/lag cal
Other codes Other codes Other codes

Figure C.1 Elements of a RDCPM Logic Network.

631

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
632 Appendix C

1441i at 2500 SY Building

of 5000 SY watertight
1440 Crew 2440
Main bldg-install roofing roofers Garage–install roofing
1440i 2440i
10 1 6 RF FS 0 1 51 6 RF
A12 S23 RF1 A15 S23
Physical 1450
-- Main bldg–studs & drywall
1450i
FS 0 1 10 1 4 CP
-- S06 S07
Physical 1460
Penetrations at 50% Main bldg–roof HVAC
1460i
PS 50% 41 4 SM
S23 M14
Figure C.2 Segment of a RDCPM Logic Network. Activity 1440, starting after event
1440i, uses 6 roofers to install the roof on the main building, as shown on contract draw-
ings A12 and S23, and has three successors. Activity 2440, starting after event 2440i, is
the next activity to use the roofing crew for the garage roof, as shown on drawings A15
and S23. Activity 1430, after event 1430i which represents the milestone of “building
watertight,” uses 4 carpenters to install studs and drywall, as shown on drawings S06
and S07. Event 1441i, in the middle of activity 1440, represents 50% of performance by
area, allowing the start of activity 1460, HVAC penetrations, as shown on drawings S23
and M14, by a four person sheet metal crew.

Act A EF or data date Act B

FS 7

EF or actual finish Act C

ES 7
Figure C.3 Two variations of Finish-to-Start in RDCPM.
Activity B may start 7 days after Activity A duration reduced
to zero. Activity C may start 7 days after Activity A has
reported end date.
 Notation for RDCPM 633

Act A Act B

ES or data date Act C

SS 7

Act A Act B

ES or actual start Act C

BS 7

Act A Act B
10

7 days performed Act C

PS 7

Act A Act B
10
3 days remaining Act C
RS 3
Figure C.4 Four variations of Start-to-Start in RDCPM.

Act A Act B

EF or data date
FF 7
Act C
12

Act A Act B

EF or actual finish
EF 7
Act C
12

Act A Act B

7 days remaining of C
FR 7
Act C
12

Act A Act B

5 days performed of C
FP 5
Act C
12
Figure C.5 Four variations of Finish-to-
Finish in RDCPM.
634 Appendix C

Act A Act B

ESA or data date to EFB

SF 7
Act C

Act A Act B

ESA or actual start to EFB

BF 7
Act C

Act A Act B
10
6d performed of A to 3d remaining of C
PR 6 3
Act C
12

Act A Act B
10
4d remaining of A to 3d remaining of C
RR 4 3
Act C
12

Act A Act B
10
4d remaining of A to 9d performed of C
RP 4 9
Act C
12

Act A Act B
10
6d performed of A to 9 days performed of C
PP 6 9
Act C
12
Figure C.6 Six variations of Start-to-Finish in RDCPM.
 Glossary

activity The work item that is the basic component of the project schedule.
activity code Alpha or numeric code added to the activity description to
facilitate sorting by specific categories.
activity list A tabulation of the project activities sorted by activity time
and/or code.
activity number In the ADM variant of CPM, this is a unique two-alphanumeric
pair, known as i-j number, assigned to identify the activity and its logic to the
computer. In the PDM variant of CPM, this is a unique one-alphanumeric
identifier assigned to identify the activity but not its logic to the computer.
activity-on-arrow (AOA) See arrow diagramming method (ADM)
activity-on-node (AON) See precedence diagramming method (PDM)
activity times Time information generated through the CPM calculation that
identifies the start and finish times for each activity in the network.
arrow The graphical representation of an activity in the CPM network. One
arrow represents one activity. The arrow is not a vector quantity and is not
necessarily drawn to scale.
arrow diagram CPM network. See also network and arrow diagramming
method.
arrow diagramming method (ADM) A network in which the scheduled
sequence of activities is represented by arrows where the tail of the arrow
represents the start of the activity and the head represents the finish of the
activity.
backward pass The calculation of late finish and late start times working
backward from the last activity finish time.
bar chart Chart with activities plotted to time scale. Also called Gantt chart.
baseline schedule The approved project schedule.
calendar Assigned calendar (typically 5-day, 6-day, or 7-day) used to convert
CPM calculated times to calendar dates.
constraint An artificial limitation based upon information not recorded in the
logic network that affects when an activity can be scheduled.
cost The monetary price of an activity.
crashing The technique of reducing overall schedule time frame by either re-
sequencing activities and/or reducing activity times by expenditures such as
overtime.

635

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
636 Glossary

critical activity An activity on the critical path.

critical path The longest connected route through the CPM network.
critical path method (CPM) An approach to developing a schedule that starts
with a project plan sequenced from beginning to end, followed by time driven
forward and backward passes to establish (1) the critical path and (2) float or
scheduling flexibility on all other paths.
data date Starting calendar date for a network calculation.
dummy activity Colloquial term to describe a logic restraint. In ADM, a zero-
duration activity used to achieve unique numbering (ij) for an activity, or to
record complex logic between activities. In PDM, a zero-duration activity used
to achieve a hammock.
duration The number of work days (periods) to complete the activity.
early event time The earliest time an event can be started.
early finish The earliest time an activity can be completed.
early start The earliest time an activity can be started (equal to early event
time).
edit A computer sort by i-j, total float, code, or early or late dates.
event A point in time representing the intersection of two or more arrows. The
event has no time duration. It can be a milestone. Junction between two or
more activities in a logical network.
event number Unique random number assigned to an event in an ADM
network. Its purpose is to identify the event to the computer. See i-j number.
event times Time information generated through CPM calculation that
identifies the start and finish times for each event in a network.
expected time The mean activity duration for a PERT activity.
fast track A project where implementation starts before design is finalized.
finish-to-finish In PDM, predecessor activity must finish before the successor
can complete.
finish-to-start In PDM, the predecessor must complete before the successor can
begin.
float The flexibility that an activity has vs. the critical path. See total float and
slack.
forward pass The calculation of the early start and early finish dates for all
network activities.
fragnet A part or fragment of a network. See subnetwork.
free float Activity float that identifies the scheduling flexibility that will not
delay the early start of any succeeding activities if used.
hammock activity A summary activity representing a continuous segment of
a logic network.
 Glossary 637

horizontal event numbering Assigning event numbers in horizontal order.

imposed date A fixed date, usually in the form of a “start no earlier than”
(SNET) date, or “finish no later than” (FNLT) date.
independent float Activity float that identifies the scheduling flexibility that
will neither reduce the total float (delay the early start) of any succeeding
activities if used nor reduce the total float (push back the late finish) of any
preceding activities if planned to be used.
input The data that must be introduced into the computer before a computation
is started.
lag Duration between activities traditionally imposed on following or successor
activities.
late event time The latest time an event can be reached without lengthening
the project.
late finish time The latest time an activity can be completed without
lengthening the project.
late start time The latest time an activity can start without lengthening the
project.
lead An acceleration of the successor activity allowing it to start before
completion of its predecessor activity.
logic diagram See network.
logic loop A circular connection of illogical arrows that cannot be computed.
logical relationship Dependency between two schedule activities; four possible
types are finish-to-start, finish-to-finish, start-to-start, and start-to-finish.
logic restraint A connection used as a logical connector that does not represent
actual work items. Usually represented by a dashed line. Sometimes called a
“dummy,” because it does not represent work; it is an indispensable part of a
CPM network.
master schedule The merging of several distinct project schedules sharing a
common site, resources, or program with additional logic between such projects
that may be optionally considered when scheduling an individual project.
matrix Grid system used in the graphical solution of mathematical problems.
Once used in manual CPM solution but outmoded by later techniques.
milestone Significant event in the project.
milestone schedule A summary level project schedule that includes milestone
events.
Monte Carlo analysis Analytical approach that computes the project schedule
on an iterative basis to predict the probability meeting a range of completion
dates.
most likely time Activity duration estimate (PERT terminology). Same as
expected duration.
638 Glossary

near-critical activity An activity (or network path) that has low total float.
network Connected sequence of arrows representing the project. This is the
basis of CPM and PERT. The network must have one start point and one
terminal point. See also arrow diagramming method.
network path A continuous series of logically connected activities in a network.
node A junction point connecting two or more activities. Same as an event
in ADM.
open end A schedule activity either without any predecessor activities or
without any successor activities.
optimistic time In PERT, the fastest time an activity can reasonably expected
to be completed.
original duration The duration initially assigned to a baseline schedule activity.
output Usually refers to the result of the CPM computer calculation.
path convergence Two or more network paths merging into the same event
or node.
path divergence Two or more network paths emerging from an event or node.
pessimistic time In PERT, the slowest time an activity can be reasonably
expected to be completed.
PERT Originally, program evaluation research task, now performance
evaluation and review technique. Event oriented. Time or duration between
events has three estimates: optimistic; pessimistic; and most likely.
plan The sequence in which a project is to be done. It is independent of the
schedule.
precedence diagramming method (PDM) A variant of the CPM network in
which the scheduled sequence of activities are represented by nodes, incorporating
the scope and duration of the activities, separated by logic restraints that may
also convey durations between activities.
preconstruction CPM Plan and schedule for the concept and design phase
preceding the award of contract.
preliminary CPM plan CPM analysis of the construction phase made before the
award of contract to determine a reasonable construction period.
project The overall work being planned. It must have one start point and one
finish point.
project management information systems (PMIS) A system organizing, usually
at the summary level, all projects for a major entity or governmental unit.
project schedule Planned dates for performing project schedule activities and
meeting project schedule milestones usually derived from a project schedule
network.
project time Time dimension in which the project is being planned. It must be
consistent and is a net value (less holidays).
 Glossary 639

remaining duration The time remaining in a schedule activity in progress.

resource leveling Schedule network analysis in which the activity start and
finish are driven by resource constraints as well as logic.
schedule Planned dates for performing scheduled activities and meeting
schedule milestones usually derived from a planned logical network.
s. curve Graphic display of cumulative project factors, such as cost and
manpower, which is derived from the typical flat rate of progress at the project
start and conclusion, with higher rate of progress in between.
slack In PERT, the scheduling flexibility available for an event, equivalent to
total float in CPM.
sort Same as edit.
stakeholder Persons and/or organizations who are involved in a project schedule
who may either influence the success of the schedule or are impacted by it.
start-to-finish(S-F) Logical relationship where completion of predecessor
activity must precede start of successor.
start-to-start(S-S) Predecessor activity must start before (or at the same time)
successor activity can start.
subnetwork Amplification or overlay of a section of the CPM network to study
a special sequence or establish a difficult time estimate. See fragnet.
summary network Summary of the CPM network for presentation purposes
comprised of hammock activities.
task A component scope of work that is part of an activity.
time impact evaluation Use of a fragnet or subnetwork to evaluate the impact
of an event such as a change of order or unusual occurrence on the baseline
schedule; known as TIE. This is also known as time impact analysis (TIA).
total float Measure of scheduling flexibility available on a network path.
updating Regular, periodic review, analysis, evaluation, and recomputation of
a CPM schedule.
vertical event numbering Assigning event numbers in vertical order.
This page intentionally left blank
 List of Acronyms
and Symbols

ADM arrow diagramming method

AOA activity-on-arrow
AON activity-on-node
a optimistic PERT activity time
b pessimistic PERT activity time
CPM critical path method
EET early event time
EF early finish
ES early start
FF free float
FF finish-to-finish
FNET finish not earlier than constraint (forward pass)
FNLT finish not later than constraint (backward pass)
FS finish-to-start
i starting event, typical activity
IF independent float
j finish event, typical activity
JF junior float
LET late event time
LF late finish
LS late start
m most likely time, PERT
o optimistic PERT activity time
p pessimistic PERT activity time
PERT performance evaluation and review technique
PDM precedence diagramming method

641

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
642 List of Acronyms and Symbols

RDCPM Relationship Diagramming Method (Certified Standard)

RDM Relationship Diagramming Method
SF start-to-finish
SS start-to-start
SNET start not earlier than constraint (forward pass)
SNLT start not later than constraint (backward pass)
Te early event time
TL late event time
TS scheduled event time
 Index

Acceleration, 373, 375, 482, 487, 488, American Society for Testing and
503, 536, 539 (See also Recovery Materials (ASTM), 534
schedule) ANSI (see American National Standards
Activity calendar, 117, 118, 140, 144, 578 Institute)
Activity codes, 22, 70, 133, 207–216, 221, AOA (see Activity on arrow)
223, 240, 392, 407, 413, 415, 434, 559 AON (see Activity on node)
Activity descriptions, 19, 22, 57, 63, 66, Apollo project, 17, 25, 493
93, 105, 113, 147, 210, 212, 220, 311, architectural planning, 5, 6, 8, 46, 69,
348, 393, 418, 432, 448, 449, 470, 313, 322, 324, 325, 329,
486, 552 357, 362
Activity duration, 67, 72, 166, 285, 369, Arrow diagramming method (see ADM)
577, 586–588 Artemis, 27
Activity list, 47, 254, 276 AS (see Actual start)
Activity on arrow, 19, 75, 199 (See also As-built CPM, 498, 517, 535, 542,
ADM) 552–561, 567–572
Activity on node, 19 (See also PDM) As-planned CPM, 349, 530, 542, 548–560,
Activity steps, 133, 216, 220, 463 567, 570–574
Activity title (see Activity description) As-should-have-been CPM, 550, 570
Actual duration (AD), 169, 174, 477 Associated General Contractors of
Actual finish (AF), 129, 130, 138, 166, America (AGC), 531, 534
175, 238, 404, 405, 425, 426, 439, ASTM (see American Society for Testing
461, 463, 467, 476, 568 and Materials)
Actual start (AS), 111, 113, 129, 130, 138, Attributes of an activity, 21, 73–83, 86,
166–168, 175, 238, 241, 291, 404, 108, 111, 131, 184, 246
425, 439, 451, 461, 463–470, actual duration (AD), 169, 174, 477
476–478, 501, 552, 568 actual finish (AF), 129, 130, 138, 166,
ADM, 19–23, 27–31, 75, 81, 88, 111, 115, 175, 238, 404, 405, 425, 426, 439,
131, 132, 143, 150, 153, 161, 163, 461, 463, 467, 476, 568
172–181, 200, 201, 231, 234–239, actual start (AS), 111, 113, 129, 130,
254, 411, 421, 515, 517–521, 589, 590 138, 166–168, 175, 238, 241, 291,
AF (see Actual finish) 404, 425, 439, 451, 461, 463–470,
AGC (see Associated General Contractors 476–478, 501, 552, 568
of America) early finish (EF), 75–88, 94–110, 115,
airport construction, 490 116, 122, 129, 131, 142, 153, 166, 175,
Allegheny County Airport, 490 180, 184, 192, 241, 246, 276, 309–311,
American National Standards Institute 362, 396, 417, 422, 425, 441, 463,
(ANSI), 530, 531, 534 516, 521

643

Copyright © 2006, 1999, 1993, 1984, 1971, 1965 by The McGraw-Hill Companies, Inc.
Click here for terms of use.
644 Index

Attributes of an activity (Cont.): 205, 231, 253, 387, 389, 390, 393–403,
early start (ES), 75–88, 93–111, 427, 429, 437, 438, 448, 453, 503, 538,
115–117, 122, 126, 129, 131, 142, 152, 572 (See also Gantt chart)
158, 166, 171, 180, 184, 192, 208, Baseline schedule, 26, 154, 349, 384, 407,
241, 246, 269, 276, 279, 280, 290, 441, 463, 482, 486, 488, 497, 498,
303, 307, 309–312, 351, 371, 405, 548, 550, 555, 568
417, 421, 425, 428, 432, 439, 449, Bridge construction, 503
451, 463, 466, 469, 516, 521, 529 Broken arrow, 59
finish float, 175, 180 Bus bar technique, 61
free float (FF), 23, 75, 79, 80, 100, 102, Butler Manufacturing plant, 494
109, 110, 123, 129, 131, 163, 166, 168,
176–180, 236, 241, 303, 398, 419, Calendar, 115–118, 190, 207–209,
424, 506, 573 227–229, 243, 274–276, 293, 337,
independent float (IF), 75, 79, 80, 100, 342–345, 453
102, 103 multiple, 28, 63, 76, 102, 106, 111,
junior float (JF), 23, 439 117–120, 139, 141, 178–181, 208, 226,
late finish (LF), 75–88, 94–110, 115, 227, 346, 350, 453, 527, 576–579,
120, 122, 125, 129, 131, 153, 175, 582, 589
180, 241, 246, 275, 309–312, 362, resource, 117, 118, 139, 140, 144, 146,
364, 371, 397, 417, 420, 429, 516, 186, 188, 214, 219, 221, 223, 232,
521, 522 236, 287–305, 332, 349, 575–578
late start (LS), 75–88, 94–111, 115–117, weather, 208, 209, 226, 229, 258, 261,
120–123, 129–132, 153, 180, 184, 343, 344, 350, 353, 432, 434, 535
241, 246, 276, 281, 290, 309, 310, Case histories, 489, 497, 510
318, 319, 351, 397, 417, 516, 521, airport construction, 490
522, 561 Allegheny County Airport, 490
multi-calendar float (MF), 23 Apollo project, 17, 25, 493
original duration (OD), 111–117, 124, bridge construction, 503
139, 162, 166, 169, 199, 240, 393, Butler Manufacturing plant, 494
397, 420, 425, 426, 434, 438, 441, Chicago Courthouse, 489
460–462, 474, 477 Chicago Marina City Tower, 491
percent complete (PCT), 22, 66–68, 74, high-rise construction, 272, 351,
111–113, 130, 138, 162, 166, 168, 171, 492, 510
231, 235, 239, 243, 396, 397, 404, highway construction, 407, 503, 508,
439, 452, 461, 464, 468, 470, 477 509, 551
remaining duration (RD), 22, 68, hospital construction, 45, 323, 337, 349,
111–113, 120, 130, 162, 166, 169, 240, 375, 510
243, 397, 404, 417, 420, 425, 438, housing, 493
440, 461–470, 474, 476–478 Interstate 76, 202 & 422 Interchange,
scheduled finish, 73 508
scheduled start, 73, 77 JFK Airport Program 2000, 497
smoothed start, 77, 80 manufacturing facilities, 494
start float, 180, 451 Maricopa County Prison Project, 512
total float (TF), 74, 75, 79–88, 97–108, NASA, 16, 17, 18, 81, 292, 493,
117, 123, 126, 128, 180, 184, 241, 246, 510, 536
269, 303, 340, 347, 351, 386, 397, New Jersey Turnpike Authority, 496
417, 425, 427, 430, 434, 439, 451, 469 oil refinery, 511
Phoenixville-Mont Claire Bridge, 503
Backward arrow, 59 railroad construction, 252, 256, 261,
Backward pass, 75, 78, 79, 86, 106, 108, 351, 490, 491, 495, 496
123, 132, 153, 241 Saturn program (NASA), 493
Bar chart, 6–9, 26, 29, 30, 36, 73, 76–78, school construction, 511
129, 131, 136, 146, 169, 176, 192, 199, SEPTA RailWorks, 495
 Index 645

Times Tower, 490 weather, 208, 226, 229, 258, 261, 343,
Toronto Transit Let’s Move Program, 500 350, 353, 432, 434, 453, 535,
U.S. Navy housing, 493 539–542, 548, 554, 571
Cash flow, 137, 321, 500 Corps of Engineers, 16, 17, 97, 493, 531,
Categories for cost breakdown, 145, 213, 536
356, 445 Cost
Chaining, 23, 55, 95, 97, 98, 100, 102, cost codes, 137, 214, 215, 223, 357, 358
104, 332, 344, 351, 588 percent complete, 22, 66–68, 74,
Change order, 139, 303, 344, 367, 447, 111–113, 130, 138, 162, 166, 168, 171,
448, 455, 471, 477–482, 486, 487, 231, 235, 239, 243, 396, 397, 404,
500, 512, 530, 536, 542, 547, 550, 439, 452, 461, 464, 468, 470, 477
553–555, 574 report, 18, 214, 215, 401
Chicago Courthouse, 489 vs. time, 368
Chicago Marina City Tower, 491 cost percent complete, 138, 439
Claims and litigation, 31, 537, 541 (See cost/schedule control system (C/SCS), 18
also Legal issues) crashing, 17, 187, 369, 370, 371, 372,
Claims avoidance using CPM, 512 373, 375, 376
Classic delay, 547 Critical activities, 58, 95, 102, 104, 269,
Coded networks, 58 283, 304, 305, 313, 339, 350, 396,
Codes for activities (see Activity codes) 440, 452, 471, 473, 477, 546, 588
Completion backward pass procedure, 78, 79, 86,
percent complete, 22, 66–68, 74, 106, 108, 123, 132, 153
111–113, 130, 138, 162, 166, 168, 171, forward pass procedure, 77, 78, 83, 108,
231, 235, 239, 243, 396, 397, 404, 120, 122, 123, 469
439, 452, 461, 464, 468, 470, 477 i-j list, 55
substantial, 18, 19, 226, 292, 294, 299, Critical path, 16, 19, 26, 58, 95–102,
321, 332, 342, 350, 413, 536, 539, 117–126, 132, 153, 180, 186, 193, 225,
541, 542, 547 269, 276, 279, 283, 290, 304, 309,
Concurrent delay, 547, 561 338, 346–348, 370, 386, 422,
Connections (see Restraints) 427–431, 436, 444, 452, 457, 471,
Constraint, 110, 112, 120–124, 142, 177, 473, 477–481, 485, 489–492, 529,
226, 233, 386, 395, 413, 420, 444, 532, 537–542, 554, 556, 559–561,
450, 462, 486, 532, 556, 568 564, 571, 574, 582, 587
finish-not-earlier-than (FNET) date, 78, Critical path method
110, 120, 122, 123, 233 detail level required, 29, 37, 65, 448,
finish-not-later-than (FNLT) date, 78, 452, 459, 493, 577, 581, 589
110, 120–125, 233, 234, 386, 387, 413 history of, 3, 6, 535
mandatory finish date, 110, 122, 594, Critical Path Method (CPM), 3, 5, 9–11,
617 26, 37, 100, 146, 182, 537–539,
mandatory start date, 110, 122, 123 582, 589
start-not-earlier-than (SNET) date, Criticality, 111, 117, 124, 126, 134, 192,
120–123, 233, 234, 420, 556 429, 434, 477, 552
start-not-later-than (SNLT) date, crossovers, 59
120–123, 233 CSI (see Construction Specifications
start-on date, 110 Institute)
Construction phase, 321, 322, 333
Construction Specifications Institute Data acquisition, 197, 207, 417, 427, 444
(CSI), 145, 213, 356, 445 Delay, 5, 9, 21, 26, 28, 76, 100, 102, 120,
Contingency, 67, 69, 209, 225, 229, 344, 123, 130, 149, 159, 175, 209, 304,
348, 350, 353, 368, 436, 452, 453, 307, 337, 341, 343, 349, 350, 358,
479, 480, 528 441, 448, 462, 469, 471, 478, 482,
calculated, 344 485, 494, 512, 533, 537–542,
experience, 344 545–572, 576, 587
646 Index

Delay (Cont.): remaining duration (RD), 22, 68,

change orders, 139, 344, 477, 500, 512, 111–113, 120, 130, 162, 166, 169, 240,
530, 547, 554, 555 243, 397, 404, 417, 420, 425, 438,
classic delay, 547 440, 461–470, 474, 476–478, 568
concurrent delay, 547, 561 tracking, 68, 111, 171, 186, 406, 477,
force majeure, 546, 548, 554, 558, 561 507, 588
negative float, 122, 124 trend, 22, 189, 298, 299, 513
responsibility for delay, 31, 546
serial delay, 548 Early event time (TE), 74, 75, 77, 78, 79,
strikes as cause for delay, 547, 81, 83, 84, 85, 86, 87, 88, 89, 90, 91,
548, 554 93, 95, 96, 98, 102, 131, 133
weather, 208, 209, 226, 229, 258, 261, Early finish (EF), 75–88, 94–110, 115,
343, 344, 350, 353, 432, 434, 535 116, 122, 129, 131, 142, 153, 166, 175,
Design phase, 321, 322, 324, 325, 327, 180, 184, 192, 241, 246, 276, 309–311,
330, 332 362, 396, 417, 422, 425, 441, 463,
Diagnostic report (Primavera P3), 176, 516, 521
181, 395, 396, 397, 418–426, 450, Early start (ES), 75–88, 93–111, 115–117,
470, 473, 476, 507, 519, 573, 574 122, 126, 129, 131, 142, 152, 158,
calculation options chosen, 396, 400, 166, 171, 180, 184, 192, 208, 241,
411, 414, 416, 417, 421, 476, 503, 246, 269, 276, 279, 280, 290, 303,
527, 531 307, 309–312, 351, 371, 405, 417,
constraints, 395–397, 417, 420, 422, 421, 425, 428, 432, 439, 449, 451,
448, 450, 452, 470, 476, 509, 526, 463, 466, 469, 516, 521, 529
527, 576, 582 EBS (see Enterprise project structure)
extended, software for, 18, 35, 211, 220, EF (see Early finish)
227, 303, 335, 346, 422, 423, 425, End to end time, 148
426, 519, 536, 540, 542, 561 End to start time, 148
loop error report, 181, 418 EnProMaC, 63, 418, 420, 422
open end, 241, 396, 414, 420, 421, 422, Enterprise project structure (EPS), 144,
425, 451, 470, 476, 556 145, 214, 406, 409
out-of-sequence performance, 396, 405, ES (see Early start)
421, 469, 470, 471, 472, 473, 474, Event, 51–59, 62–64, 70, 74–81, 83, 86,
530, 553, 572–574 89–98, 100, 102, 104, 110, 131,
Disruption, 31, 461, 470, 471, 473, 478, 143–153, 172, 233, 259–263, 285,
481, 485, 538, 540, 541, 545, 546, 292, 293, 309, 311, 317, 338, 340, 351,
553, 564, 571–574 370, 375, 392, 397, 414, 420, 451,
Drawing a network, 41, 49–51, 55, 61, 63, 457, 462, 486, 525, 530, 532, 548,
200, 210, 219, 223, 246, 262, 272, 550, 561, 567
297, 313, 350, 392, 418, 434, 439, causative, 485, 554–560, 567–574
445, 478, 526–528, 556–559 early time (TE), 74–98, 131, 133
Drexel University, 31, 508, 510 late time (TL), 74, 78–81, 91, 94, 96,
Driving resources, 577 98, 131
Du Pont, 9, 10, 11, 15, 17, 41, 355, 474, Event numbers, 52–55, 64, 90, 153, 285
479, 537
Duration Fast track scheduling, 341, 353
actual (AD), 169, 174, 477 FF (see Free float)
forecast, 69, 137, 298, 305, 362–364, Filters, 124, 136, 316, 406, 432
366, 368, 379, 441, 442, 483, 500, Finish
502, 512 early finish (EF), 75–88, 94–110, 115,
original duration (OD), 111–117, 124, 116, 122, 129, 131, 142, 153, 166, 175,
139, 162, 166, 169, 199, 240, 393, 180, 184, 192, 241, 246, 276, 309–311,
397, 420, 425, 426, 434, 438, 441, 362, 396, 417, 422, 425, 441, 463,
460–462, 474, 477, 561 516, 521
 Index 647

late finish (LF), 75–88, 94–110, 115, Hierarchical systems, 145, 146, 213,
120, 122, 125, 129, 131, 153, 175, 180, 214, 583
241, 246, 275, 309–312, 362, 364, 371, High-rise construction, 272, 351,
397, 417, 420, 429, 516, 521, 522 492, 510
scheduled, 73 Highway construction, 150, 210, 212, 236,
Finish float, 175, 180 (See also PDM) 237, 290, 343, 407, 409, 413, 442,
Finish-not-later-than (FNLT) date, 78, 490, 495, 503, 508, 533, 551
110, 120–125, 233, 234, 386, 387, 413 Hospital construction, 45, 323, 337, 349,
Finish-to-finish, 129, 163, 168, 170, 173, 375, 510
179, 181, 238, 240, 396, 417, 421, 486 Housing, 493
(See also PDM)
Float IF (see Independent float)
finish, 175, 180 i-j
free float (FF), 23, 75, 79, 80, 100, 102, list, 3, 6, 10, 30, 37, 39, 47, 53, 66, 73,
109, 110, 123, 129, 131, 163, 166, 168, 81, 106, 107, 113, 115, 126, 127, 163,
176–180, 236, 241, 303, 398, 419, 214, 216, 222, 223, 227, 228, 229,
424, 506, 573 231, 243, 245, 253, 254, 269, 271,
independent (IF), 75, 79, 80, 100, 103, 579 275, 285, 310, 313, 319, 342, 474
junior (JF), 23, 439 numbering, 53, 54, 55, 56, 59, 81, 83,
multi-calendar (MF), 23 261, 356, 394
start, 180, 451 sort, 22, 108, 127, 154, 158, 269, 270,
total float (TF), 74, 75, 79–88, 97–108, 272, 288, 307, 318, 365, 407, 427
117, 123, 126, 128, 180, 184, 241, 246, Independent float (IF), 75, 79, 80, 100,
269, 303, 340, 347, 351, 386, 397, 103, 579
417, 425, 427, 430, 434, 439, 451, Interstate 76, 202 & 422 Interchange,
469, 516, 521 508
FNET (see Finish-not-earlier-than) Intuitive manual computation, 88, 89,
FNLT (see Finish-not-later-than) 104, 106, 144, 152, 153, 181, 187,
Fondahl John W., 18, 147, 234, 302, 191, 198, 233, 269, 287, 305, 378,
303, 304 389, 405, 451, 575, 576
Force Majeure, 546, 548, 554, 558, 561
Forecasting JF (see Junior float)
cash flow, 137, 321, 500 JFK Airport Program 2000, 497
Forward pass, 77, 78, 83, 108, 120, 122, Junior float (JF), 23, 439
123, 469
Fragnet, 292 Kelley James E. Jr., 9, 10, 15, 16, 17, 41,
Free float (FF), 23, 75, 79, 80, 100, 102, 86, 170, 376, 474
109, 110, 123, 129, 131, 163, 166, 168, Kelley-Walker method, 10
176–180, 236, 241, 303, 398, 419, Key materials, 307
424, 506, 573
Lag (see Lead/lag)
Gantt Henry L., 6, 171, 173, 243, 417, Late event time (TL), 74, 75, 78, 79, 81,
474, 589 91, 93, 94, 95, 96, 98, 131
GERT, 18, 19, 20, 22, 31, 37, 112, 200, Late finish (LF), 75–88, 94–110, 115, 120,
203, 579, 581, 583, 586–590 122, 125, 129, 131, 153, 175, 180,
Graphical user interface (GUI), 131, 241, 246, 275, 309–312, 362, 364,
184, 189 371, 397, 417, 420, 429, 516,
Grid paper, 49 521, 522
GUI (see Graphical user interface) Late start (LS), 75–88, 94–111, 115–117,
120–123, 129–132, 153, 180, 184,
H.B. Zachry Co., 18, 147 241, 246, 276, 281, 290, 309, 310,
Hammock, 132, 133, 142, 172, 173, 174, 318, 319, 351, 397, 417, 516, 521,
175, 181, 396 522, 561
648 Index

Lead/lag, 18, 30, 43, 59, 68, 89, 149, 150, logic spreader, 45, 317, 340
153, 161, 170, 184, 197, 199, 293, longest path, 100, 153, 422, 452
314, 347, 351, 462, 494, 512, 518, loop, logical (see Logic loop)
526, 530, 589 LS (see Late start)
Legal cases cited, 6, 8, 18, 26, 60, 69, 70,
111, 112, 115, 126, 129, 130, 137, 138, Main chain, 16, 95
139, 140, 147, 167, 170, 171, 175, Management scheduling and control
176, 177, 178, 179, 180, 191, 198, system (MSCS), 27, 163, 164, 168,
224, 225, 226, 288, 290, 342, 363, 186, 227, 287
364, 365, 370, 371, 373, 375, 389, Manipulating the schedule, 349, 450
473, 477, 479, 487, 510, 528, 537, Maricopa County Prison Project, 512
538, 542, 545, 547, 549, 552, 553, Matrix method
556, 572, 578 backward pass, 78, 79, 86, 106, 108,
Legal issues 123, 132, 153
acceleration charges, 373, 375, 482, forward pass, 77, 78, 83, 108, 120, 122,
487, 488, 503, 536, 539, 540, 565 123, 469
as-built CPM, 498, 535, 536, 542, 552, Mauchly Associates, 15, 17, 287
553, 554, 557, 558, 560, 561, 567, Mauchly John W., 9, 15, 17, 287
568, 570, 572 McAuto, 27, 287
as-planned CPM, 349, 542, 548, 549, MF (see Multi-calendar float)
550, 551, 552, 553, 554, 556, 557, Microsoft Project, 71, 72, 117, 186, 188,
558, 560, 567, 570, 571, 572, 574 189, 238, 383, 575, 578, 590
as-should-have-been CPM, 550, 551, 570 Milestones, 21, 52, 74, 77, 83, 93, 112,
causative factors, 485, 554, 555, 556, 116, 131, 132, 143, 213, 233, 337, 349,
557, 558, 559, 560, 567, 568, 569, 396, 432, 440, 451, 560, 561
571, 573, 574 Moder J.J., 18, 147, 583
change orders, 139, 344, 477, 500, 512, Multi-calendar float (MF), 23
530, 547, 554, 555 Multiple calendars, 111, 117, 118, 178,
claims avoidance using CPM, 512 179, 181, 193, 208, 226, 427, 453,
classic delay, 547 527, 578, 589
concurrent delay, 547, 561 Multiple starting/ending activities, 118,
contracts, 8, 18, 276, 342, 462, 493, 497, 120, 420
511, 532, 548, 554, 570 Multiproject scheduling, 25, 294, 296,
evidentiary use of CPM, 538, 539, 540, 297, 305, 383
541, 542 Multisheet networks, 61
force majeure, 546, 548, 554, 558, 561
serial delay, 548 Narrative summary report, 229, 336, 352,
strikes as cause for delay, 547, 548, 554 422, 436, 443, 444, 473, 474, 478,
time impact evaluation (TIE), 557–570 485, 527, 530, 549, 551
Leveled finish, 76 NASA, 16, 17, 18, 81, 292, 493, 510, 536
Leveled start, 76, 80 Negative float, 122, 124
LF (see Late finish) Negative lag, 170, 240, 427
Log (Primavera P3), 211, 222, 225, 227, Network logic, 51, 67, 285, 469
234, 240, 394, 440, 442, 445, 528, 574 Network scheduling, 10, 17, 294, 304, 531
Logic connection, 395, 452, 586 New Jersey Turnpike Authority, 496
Logic diagrams, 48, 215, 526 Nonconstruction CPM project, 457
Logic loop, 22, 212 Non-deterministic polynomial-time
Logic loop detection, 418 complete problem, 576
Logic loop erroneous detection, 419
Logic resource codes, 219 O’Brien James J., preface
Logic restraint, 116, 120, 132, 193, 237, O’Brien-Kreitzberg & Associates (OKA),
396, 450, 483, 518, 526, 576 291, 492, 495, 496, 497, 498, 537,
Logic splitter, 45 538, 550, 551
 Index 649

OD (see Original duration) start-to-start relationships, 153, 162,

Oil refinery, 511 163, 168, 170, 175, 179, 181, 238,
Open Plan, 189, 190, 191, 192, 193, 243 404, 421, 451, 463, 486
Ordinal calendar, 266 Precedence networks, 18, 19,
Organization breakdown structure (OBS), 147, 148
144, 145, 146, 214 Pre-construction phase, 322
Original duration (OD), 111–117, 124, Preferential logic, 259, 297
139, 162, 166, 169, 199, 240, 393, Preliminary schedule, 623
397, 420, 425, 426, 434, 438, 441, Primavera Project Planner P3
460–462, 474, 477, 561 alias code, 216
Overlapping activities, 36 diagnostic tools/reports, 176, 181,
395–397, 418–426, 450, 451,
PDM (see Precedence diagramming 470–476, 507, 519, 573, 574
method) hot keys, 389
finish float, 175, 180 menus, 189, 389, 405, 453
finish-to-finish, 129, 163, 168, 170, 173, Primavera Project Planner P3e/c, 28, 30,
179, 181, 238, 240, 396, 417, 421, 486 31, 35, 36, 67, 70, 113, 114, 140, 144,
start float, 180, 451 181–187, 188, 191, 208, 209, 213,
start-to-finish, 170 216, 221, 222, 227, 238, 242, 383,
start-to-start, 153, 162, 163, 168, 170, 405–407, 409, 414–419, 421, 442,
175, 179, 181, 238, 404, 421, 451, 451, 464–470, 476, 485, 509,
463, 486 511, 517
Percent complete (PCT), 22, 66–68, 74, Primavera Suretrak Project Planner, 28,
111–113, 130, 138, 162, 166, 168, 171, 181–188, 216, 227, 241, 243, 383,
231, 235, 239, 243, 396, 397, 404, 395, 405, 414, 416, 451, 476, 517, 611
439, 452, 461, 464, 468, 470, 477 Procurement, 311–320, 341, 346–348,
cost percent complete, 138, 439 393, 459, 494, 563, 564, 588
Personal computer (PC), 27, 159, 183, free float, 23, 75, 79, 80, 100, 102, 103,
397, 405, 438, 499, 500 123, 177, 303, 506, 579
PERT, 11–22, 31, 37, 70–74, 80–88, 95, junior float, 23, 439
112, 131, 147, 159, 200, 224, 389, 390, Progress override, 22, 111, 180, 396, 414,
391, 394, 472, 581, 582, 583, 584, 451, 469
586, 587, 589, 590 Progress payments, 356, 358, 379, 497,
PERT/Cost, 18 529, 532, 533
Phoenixville-Mont Claire Bridge, 503 Project control system (PCS), 25, 26
Physical logic, 232, 448, 507 Project management system (PMS), 18,
Plot code, 428, 431 287
Plotnick Fredric, preface, 30, 31 PROJECT/2, 27, 287
Precedence diagramming method (PDM),
18–23, 27, 28, 30, 31, 46, 66, 74, 108, Railroad construction, 495
112, 129, 131, 132, 143, 147–156, 159, RD (see Remaining duration)
161–181, 200, 202, 231, 235, 236, RDCPM standard, 20–23, 204, 205, 590,
237, 239, 263, 264, 302, 348, 421, 631–634
423, 425, 463, 464, 465, 518, 525, RDM (see Relationship Diagramming
526, 527, 531, 589, 590 Method)
finish-to-finish relationships, 129, 163, Recovery schedule, 487, 598 (See also
168, 170, 173, 179, 181, 238, 240, Acceleration)
396, 417, 421, 486 Relationship Diagramming Method, 21,
nontraditional relationships in, 166, 102, 164, 236, 237, 438
167, 168 Remaining duration (RD), 22, 68,
overlapping activities, 36 111–113, 120, 130, 162, 166, 169, 240,
start and finish total float, 175, 243, 397, 404, 417, 420, 425, 438,
180, 451 440, 461–470, 474, 476–478, 568
650 Index

Resource allocation and labor planning Scheduled start, 73, 77

(RAMPS), 17, 287 Scheduling
Resource calendar, 117, 118, 144, 578 baseline, 26, 154, 349, 384, 407, 441,
Resource codes for restraints, 219 463, 482, 486, 488, 497, 498, 548,
Resource leveling, 140, 302, 576 550, 555, 568
Resource planning, 17, 287, 292, 294, 298 change orders, 139, 344, 477, 500, 512,
algorithm limitations, 576 530, 547, 554, 555
GERT, 18–20, 22, 31, 37, 112, 200, 203, contingency scheduling, 67, 69, 209,
579, 581, 583, 586, 587, 588, 590 225, 226, 229, 344, 348, 350, 353,
limited resources, 76, 117, 577, 579 368, 436, 452, 453, 479, 480, 528
manpower vs. resource usage, 136, 224, cost report, 18, 214, 215, 401
246, 287, 288 crash events expediting, 17, 187, 338,
materials, 11, 76, 102, 138, 145, 147, 369, 370, 371, 372, 373, 375, 376,
232, 240, 307–311, 313, 314, 320, 344, 378, 379
440, 490, 491, 545, 547 delay, 5, 9, 21, 26, 28, 76, 100, 102, 120,
nondeterministic polynomial-time 123, 130, 149, 159, 175, 209, 304,
complete, 576 307, 337, 341, 343, 349, 350, 358,
resource calendars, 117, 118, 144 441, 448, 462, 469, 471, 478, 482,
resource leveling, 140, 302, 576 485, 494, 512, 533, 537–542,
resource planning and control (RPC), 287 545–572, 576, 587
resource planning and scheduling diagnostic report, 395, 396, 418, 422,
method, 17 425, 426, 450, 470
resource smoothing, 77, 103, 112, frequency of updates, 68
140, 576 lead time, 314
Resource-driven scheduling, 577 manipulating the schedule, 349, 450
Restraints, 19–22, 28, 37, 43, 53, 57, 102, materials procurement milestones, 11,
108, 110, 116, 131, 140, 154, 161–163, 76, 102, 138, 145, 147, 232, 240,
166, 168–181, 189, 231, 232–243, 307–314, 320, 344, 440, 490, 491,
257, 265, 292, 293, 342, 372, 394, 545, 547
396, 404, 405, 414, 420, 425, 426, monitoring project progress, 456
429, 450, 451, 463, 465, 466, 486, monthly updates, 438, 455–474,
499, 508, 526, 527, 556, 572, 574, 475–478, 482–483
576, 577, 581 preliminary schedule, 496, 498
bogus, 464 rate charts, 473
finish-to-finish, 129, 163, 168, 170, 173, resource planning, 17, 287, 292, 294,
179, 181, 238, 240, 396, 417, 421, 486 298
finish-to-start, 170, 178, 200, 234, 235, resource-driven scheduling, 577
237, 238, 240, 265, 414, 421, 425 responsibility for meeting,
physical, 3, 4, 21, 22, 49, 69, 170, 200, 342, 530
212, 219, 221, 223, 231, 232, 234, School construction, 511
236, 239, 240, 243, 262, 327, 332, SEPTA RailWorks, 495
445, 448, 452, 461, 480, 483, 486, Sequencing, 76, 140, 212, 340, 551
487, 507, 527, 551, 577 Serial delay, 548
resource, 232, 487, 551 Smoothed start, 77, 80
start-to-finish, 170 Smyth Richard, 31
start-to-start, 153, 162, 163, 168, 170, SNET (see Start-not-earlier-than)
175, 179, 181, 238, 404, 421, 451, SNLT (see Start-not-later-than)
463, 486 Specifying CPM, 349, 517, 528,
Retained logic, 22, 241, 396, 414, 451, 469 530, 533
sample CPM specification, 519, 533,
S curve, 8, 9 591
Saturn program (NASA), 493 Unified Facilities Guide Specification,
Scheduled finish, 73 519, 532, 607
 Index 651

SPERT, 18, 19, 22, 31, 37, 71, 112, 193, calendar dates, 111, 276, 538
226, 349, 581, 582, 583, 584, 585, cost vs. time, 368
586, 587, 588, 590 crashing/expediting, 17, 187, 369, 370,
Splitter logic, 45 371, 372, 373, 375, 376
Spreader logic, 45, 317, 340 critical activities, 58, 95, 102, 104, 269,
Start 283, 304, 305, 313, 339, 350, 396,
actual (AS), 111, 113, 129, 130, 138, 440, 452, 471, 473, 477, 546, 588
166–168, 175, 238, 241, 291, 404, duration, 12, 72, 80, 139, 304, 552, 582,
425, 439, 451, 461, 463–470, 587, 588, 624
476–478, 501, 552, 568 early event times (TE), 74–98, 102,
early start (ES), 75–88, 93–111, 131, 133
115–117, 122, 126, 129, 131, 142, 152, early finish (EF), 75–88, 94–110, 115,
158, 166, 171, 180, 184, 192, 208, 116, 122, 129, 131, 142, 153, 166, 175,
241, 246, 269, 276, 279, 280, 290, 180, 184, 192, 241, 246, 276, 309–311,
303, 307, 309–312, 351, 371, 405, 362, 396, 417, 422, 425, 441, 463,
417, 421, 425, 428, 432, 439, 449, 516, 521
451, 463, 466, 469, 516, 521, 529 early start (ES), 75–88, 93–111,
float, 180, 451 115–117, 122, 126, 129, 131, 142, 152,
late (LS), 75–88, 94–111, 115–117, 158, 166, 171, 180, 184, 192, 208,
120–123, 129–132, 153, 180, 184, 241, 246, 269, 276, 279, 280, 290,
241, 246, 276, 281, 290, 309, 310, 303, 307, 309–312, 351, 371, 405,
318, 319, 351, 397, 417, 516, 521, 417, 421, 425, 428, 432, 439, 449,
522, 561 451, 463, 466, 469, 516, 521, 529
leveled, 76, 80 end to end time, 148
scheduled, 73, 77 end to start time, 148
smoothed, 77, 80 late event times (TL), 74–81,
Start float, 180, 451 (See also PDM) 91–98, 131
Start-not-earlier-than (SNET) constraint, late finish (LF), 75–88, 94–110, 115,
120–123, 233, 234, 420, 556 120, 122, 125, 129, 131, 153, 175, 180,
Start-not-later-than (SNLT) constraint, 241, 246, 275, 309–312, 362, 364, 371,
120, 122, 123, 233 397, 417, 420, 429, 516, 521, 522
Start-to-finish, 170 late start (LS), 75–88, 94–111, 115–117,
Start-to-start, 153, 162, 163, 168, 170, 120–123, 129–132, 153, 180, 184,
175, 179, 181, 238, 404, 421, 451, 241, 246, 276, 281, 290, 309, 310,
463, 486 (See also PDM) 318, 319, 351, 397, 417, 516, 521,
Steps (see Activity steps) 522, 561
Subsurface conditions, 492 multiple calendars, 111, 117, 118, 178,
Summary networks, 112, 293, 336 179, 181, 193, 208, 226, 427, 453,
527, 578, 589, 594
Taylor Frederick W., 6 percent complete, 22, 66, 67, 68, 74, 75,
TE (see Early event time) 111, 112, 113, 130, 138, 139, 162, 166,
TF (see Total float) 168, 169, 231, 235, 239, 243, 396,
Time 397, 404, 439, 452, 461, 462, 464,
actual duration (AD), 13, 22, 169, 174, 468, 470, 477
477, 567, 568, 587 project time, 95, 281, 294, 332, 339,
activity time computations, 105, 106 343, 357, 362, 561
actual finish (AF), 129, 130, 138, 166, Time estimates, 69, 81, 91, 335, 344,
175, 238, 404, 405, 425, 426, 439, 353, 457
461, 463, 467, 476, 568 Time impact evaluation (TIE), 557, 558,
actual start (AS), 111, 113, 129, 130, 559, 560, 561, 563, 564, 565, 567,
138, 166–168, 175, 238, 241, 291, 568, 570
404, 425, 439, 451, 461, 463–470, Time scaled logic diagram, 434
476–478, 501, 552, 568 Times Tower, 490
652 Index

Time-scaled logic network, 176 progress payments, 356, 358, 379, 455,
TL (see Late event time) 457, 497, 529, 532, 533
To do list, 6
Toronto Transit Let’s Move Program, 500 Walker Morgan, 9, 10, 16, 17, 170, 474
Total float (TF), 74, 75, 79–88, 97–108, Walstad Paul J., 541
117, 123, 126, 128, 180, 184, 241, 246, WBS (see Work breakdown structure)
269, 303, 340, 347, 351, 386, 397, Weather, 208, 209, 226, 229, 258, 261,
417, 425, 427, 430, 434, 439, 451, 343, 344, 350, 353, 432, 434, 453,
469, 516, 521, 579 535, 539, 540, 542, 548, 554, 571
Trend analysis, 69, 137, 298, 299, 305, calendar, 115–118, 190, 207, 208, 229,
362, 363, 364, 366, 368, 379, 441, 243, 274–276, 293, 337, 343, 432,
442, 483, 500, 502, 512 434, 453
Trend duration (TD), 22 contingency, 69, 209, 229, 344, 353, 453
Turnaround application, 11, 67, 210, 292, delay, 540
297, 298, 299, 468, 469, 472, 475, 511 White Glenn L., 531, 537
Wickwire Jon M., 26, 538, 541, 570
U.S. Navy housing, 493 Windows methodology, 569
UNIVAC computer, 9, 10 Work breakdown structure (WBS), 142,
Updating and reviewing 143, 144, 145, 146, 214,
baseline schedule, 407, 441, 548 392, 499, 501
cost reports, 18, 214, 215, 401 Working schedule, 297
diagnostic report, 395, 396, 418, 422,
425, 426, 450, 470 Zachry Co., H.B., 18, 147
frequency of updates, 68, 365, 396 Zeroing out, 558–560, 567, 569