Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.

APPENDIX E

Case Studies

CHAPTER 2 CASES
2-11 The Hi-V Company manufactures and cans three orange extracts: juice concentrate,
regular juice, and jam. The products, which are intended for commercial use, are manu-
factured in 5-gallon cans. Jam uses Grade I oranges, and the remaining two products use
Grade II. Table E.1 lists the usages of oranges as well as next year’s demand. A market
survey shows that the demand for regular juice is at least twice as high as that for the
concentrate.
In the past, Hi-V bought Grade I and Grade II oranges separately at the respective
prices of 25 cents and 20 cents per pound. This year, an unexpected frost forced growers
to harvest and sell the crop early without sorting them into Grade I and Grade II. It is
estimated that 30% of the 3,000,000-lb crop falls into Grade I and only 60% into Grade II.
For this reason, the crop is being offered at the uniform discount price of
19 cents per pound. Hi-C estimates that it will cost the company about 2.15 cents per
pound to sort the oranges into Grade I and Grade II. The below-standard oranges
(10% of the crop) will be discarded.
For the purpose of cost allocation, the accounting department uses the following
argument to estimate the cost per pound of Grade I and Grade II oranges. Because
10% of the purchased crop will fall below the Grade II standard, the effective average
cost per pound can be computed as (19 +.9 2.15) = 23.5 cents. Given that the ratio of
Grade I to Grade II in the purchased lot is 1 to 2, the corresponding average cost per
pound based on the old prices is (20 * 2 +3 25 * 1) = 21.67 cents. Thus, the increase in
the average price (= 23.5 cents - 21.67 cents = 1.83 cents) should be reallocated to the
two grades by a 1:2 ratio, yielding a Grade I cost per pound of 20 + 1.83 A 13 B = 21.22
cents and a Grade II cost of 25 + 1.83 A 23 B = 25.61 cents. Using this information, the
accounting department compiles the profitability sheet for the three in Table E.2.
Establish a production plan for the Hi-C Company.

1
Motivated by “Red Brand Canners,” Stanford Business Cases 1965, Graduate School of Business, Stanford
University.

E.1
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.2

E.2 Appendix E Case Studies

TABLE E.1

Pounds of oranges Maximum demand

Product Orange grade per 5-gal can (cans)

Jam I 5 10,000
Concentrate II 30 12,000
Juice II 15 40,000

TABLE E.2

Product (5-gal can)

Jam Concentrate Juice

Sales price $15.50 $30.25 $20.75

Variable costs 9.85 21.05 13.28
Allocated fixed overhead 1.05 2.15 1.96
Total cost $10.90 $23.20 $15.24
Net profit 4.60 7.05 5.51

2-22 A steel company operates a foundry and two mills. The foundry casts three types of
steel rolls that are machined in its machine shop before being shipped to the mills.
Machined rolls are used by the mills to manufacture various products.
At the beginning of each quarter, the mills prepare their monthly needs of rolls
and submit them to the foundry. The foundry manager then draws a production plan
that is essentially constrained by the machining capacity of the shop. Shortages are
covered by direct purchase at a premium price from outside sources. A comparison
between the cost per roll when acquired from the foundry and its outside purchase
price is given in Table E.3. However, management points out that such shortage is
not frequent and can be estimated to occur about 5% of the time.
The processing times on the four different machines in the machine shop are given
in Table E.4. The demand for rolls by the two mills over the next 3 months is given in
Table E.5.
Devise a production schedule for the machine shop.

TABLE E.3

Internal cost External purchase price

Roll type Weight (lb) ($ per roll) ($ per roll)

1 800 90 108
2 1200 130 145
3 1650 180 194

2
Based on S. Jain, K. Stott, and E. Vasold, “Orderbook Balancing Using a Combination of Linear
Programming and Heuristic Techniques,” Interfaces, Vol. 9, No. 1, pp. 55–67, 1978.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.3

Chapter 2 Cases E.3

TABLE E.4

Processing time per roll

Number of Available hour per
Machine type Roll 1 Roll 2 Roll 3 machines machine per month

1 1 5 7 10 320
2 0 4 6 8 310
3 6 3 0 9 300
4 3 6 9 5 310

TABLE E.5

Demand in rolls

Mill 1 Mill 2

Month Roll 1 Roll 2 Roll 3 Roll 1 Roll 2 Roll 3

1 500 200 400 200 100 0

2 0 300 500 300 200 200
3 100 0 300 0 400 200

2-3. ArkTec assembles PC computers for private clients. The orders for the next four
quarters are 400, 700, 500, and 200, respectively. ArkTec has the option to produce more
than is demanded for the quarter, in which case a holding cost of $100 per computer per
quarter is incurred. Increasing production from one quarter to the next requires hiring
additional employees, which increases the production cost per computer in that quarter
by $60. Also, decreasing production from one quarter to the next would require laying
off employees, which results in increasing the production cost per computer in that
quarter by $50.
How should ArkTec schedule the assembly of the computers to satisfy the demand
for the four quarters?
2-4. The Beaver Furniture Company manufactures and assembles chairs, tables, and book-
shelves. The plant produces semifinished products that are assembled in the company’s
assembling facility. The (unassembled) monthly production capacity of the plant
includes 3000 chairs, 1000 tables, and 580 bookshelves. The assembling facility employs
150 workers in two 8-hour shifts a day, 5 days a week. The average assembly times
per chair, table, and bookshelf are 20, 40, and 15 minutes, respectively.
The size of the labor force in the assembly facility fluctuates because of
the annual leaves taken by the employees. Pending requests for leaves include
20 workers for May, 25 for June, and 45 for July. Sales of the three products for the
months of May, June, and July are forecast by the marketing department as given
in Table E.6. The production cost and selling price for the three products are in
Table E.7. If a unit is not sold in the month in which it is produced, it is held over
for possible sale in a later month. The storage cost is about 2% of the unit
production cost.
Should Beaver approve the proposed annual leaves?
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.4

E.4 Appendix E Case Studies

TABLE E.6

Sales forecast units

End-of-April
Product May June July inventory

Chair 2800 2300 3350 30

Table 500 800 1400 100
Bookshelf 320 300 600 50

TABLE E.7

Product Unit cost ($) Unit price ($)

Chair 150 250

Table 400 750
Bookshelf 60 120

CHAPTER 3 CASES
3-1. A small canning company produces five types of canned goods that are extracted from
three types of fresh fruit. The manufacturing process uses two production departments
that were originally designed with surplus capacities to accommodate possible future
expansion. In fact, the company operates currently on a one-shift basis and can easily
expand to two or three shifts to meet increase in demand. The real restriction for the
time being appears to be the limited availability of fresh fruit. Because of the limited
refrigeration capacity on the company’s premises, fresh fruit must be brought in daily.
A young operations researcher has just joined the company. After analyzing the
production situation, the analyst decides to formulate a master LP model for the plant.
The model involves five decision variables (for the five products) and three constraints
(for the raw materials). With three constraints and five variables, LP theory says that
the optimum solution cannot include more than three products. “Aha,” the analyst says,
“the company is not operating optimally!”
The analyst schedules a meeting with the plant manager to discuss the details of
the LP model. The manager, who seems to follow the modeling concept well, agrees
with the analyst that the model is a close representation of reality. The analyst then goes
on to explain that, according to LP theory, the optimal number of products should not
exceed three because the model has only three constraints. As such, it may be worth-
while to consider discontinuing the two nonprofitable products.
The manager listens attentively, then tells the analyst that the company is committed
to producing all five products because of the competitive nature of the market and
that in no way can the company discontinue any of the products. The operations
researcher responds that the only way to remedy the situation is to add at least two
more constraints, in which case the optimal LP model is likely to include all five products.
At this point the manager gets confused, because the idea of having to add more
restrictions to be able to produce more products does not suggest optimality. “That is
what the LP theory says,” is the analyst’s answer.
What is your opinion of this “paradox”?
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.5

Chapter 3 Cases E.5

3-2.3 An LTL trucking company, specializing in LTL shipments, operates a number of termi-
nals that are strategically located across the United States. When the loads arrive at a
terminal, they are sorted either for delivery to local customers or for transfer to other
terminals. The terminal docks are staffed by bid and casual workers. Bid workers are
union employees who are guaranteed a 40-hour workweek. A bid employee assigned
to one of the standard three shifts of the day is expected to work the same shift for five
consecutive days, but may start on any day of the week. Casual employees are hired
temporarily for any number of hours to account for peak loads that may exceed the
work capacity of available bid workers. Union contract restricts casual employees to
less than 40 hours per week.
Loads arrive at the terminal at all hours of the day and, for all practical purposes,
their level varies continuously with the time of the day. A study of historical data shows
that the load level takes on a repetitive weekly pattern that peaks during the weekend
(Friday through Sunday). The company’s policy specifies that a load must be processed
within 16 hours of its arrival at the terminal.
Develop a model to determine the weekly assignment of bid workers.
3-3.4 The Elk Hills oil field has a majority ownership (80%) by the U.S. Federal Government.
The Department of Energy (DOE) is authorized by law to sell the government’s share
of the oil produced to the highest qualified bidders. At the same time, the law limits the
quantity of oil delivered to any one bidder. The oil field has six delivery points with dif-
ferent production capacities (bbl/day). The amounts of daily production (in bbl/day) at
each of the delivery points are presented daily as line items, and a bidder may submit
bids on any number of line items. DOE collects the bids and evaluates them, starting
with line item 1 and terminating with line item 6, awarding delivery to the highest
bidder but taking into account the caps set by law on the quantity of oil any one bidder
can receive. To be specific, Table E.8 provides a summary of bonus prices bid on a
certain day. A bonus is an increment over the highest price offered for similar grade oil
produced in the delivery point area. No bidder can receive more than 20% of the total
daily production of 180,000 bbl from all delivery points.

TABLE E.8

Bonus in $/bbl bid by bidder

Production
Line item 1 2 3 4 5 6 7 8 (1000 bbl/day)

1 1.10 .99 1.20 1.10 .95 1.00 1.05 1.02 20

2 1.05 1.02 1.12 1.08 1.09 1.06 1.11 1.07 30
3 1.00 .95 .97 .94 .93 1.01 1.02 .98 25
4 1.30 1.25 1.31 1.27 1.28 1.26 1.32 1.32 40
5 1.09 1.12 1.15 1.07 1.08 1.11 1.05 1.10 35
6 .89 .87 .90 .86 .85 .91 .88 .91 30

3
Based on a study conducted by the author for a national LTL trucking company.
4
Based on B. Jackson and J. Brown, “Using LP for Crude Oil Sales at Elk Hills: A Case Study,” Interfaces, Vol.
10, pp. 65–69, No. 3, 1980.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.6

E.6 Appendix E Case Studies

DOE uses a ranking scheme for awarding the bids. Starting with line item 1,
bidder 3 has the highest bid (bonus = $1.20) and hence is awarded the maximum
amount allowed by both line item 1 production and the 20% limit imposed by law
( = .2 * 180,000 = 36,000 bbl). From the data in the table, all line 1 item production
(20,000 bbl) is allocated to bidder 3. Moving to line item 2, bidder 3 again offers the
highest bonus but can only be awarded a maximum of 16,000 bbl because of the
20% limit. The remaining quantity is assigned to the bidder with the next-best bonus
( = $1.11), thus allocating 14,000 bbl ( = 30,000 – 16,000) to bidder 7. The process is
repeated until line item 6 is awarded.
Does the proposed scheme guarantee maximum daily revenue for the govern-
ment? Can the government do better by changing the 20% limit either up or down?

CHAPTER 4 CASES
4-1.5 MANCO produces three products P1, P2, and P3. The production process uses raw
materials R1 and R2, which are processed on facilities F1 and F2. Table E.9 provides
the pertinent data of the problem. The minimum daily demand for P2 is 70 units, and
the maximum demand for P3 is 240 units. The unit revenue contributions of P1, P2,
and P3 are $300, $200, and $500, respectively.
MANCO management is exploring means to improve the financial situation of the
company. Discuss the feasibility of the following proposals:
1. The per-unit revenue of P3 can be increased by 20%, but this will reduce the market
demand to 210 units instead of the present 240 units.
2. Raw material R2 appears to be a critical factor in limiting current production.
Additional units can be secured from a different supplier whose price per pound
is $3 higher than that of the present supplier.
3. The capacities of F1 and F2 can be increased by up to 40 minutes a day, each for
an additional cost of $35 per day.
4. The chief buyer of product P2 is requesting that its daily supply be increased from
the present 70 units to 100 units.
5. The per unit processing time of P1 on F2 can be reduced from 3 to 2 minutes at an
additional cost of $4 per day.

TABLE E.9

Usage per unit

Maximum
Resource Units P1 P2 P3 daily capacity

F1 Minutes 1 2 1 430
F2 Minutes 3 0 2 460
R1 lb 1 4 0 420
R2 lb 1 1 1 300

5
Based on D. Sheran, “Post-Optimal Analysis in Linear Programming—The Right Example,” IIE
Transactions, Vol. 16, No. 1, pp. 99–102, March 1984.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.7

Chapter 5 Cases E.7

4-2. The Reddy Mikks Company is preparing a future expansion plan. A study of the market
indicates that the company can increase its sales by about 25%. The following proposals
are being studied for the development of an action plan. (Refer to Example 3.3-1 for
the details of the model and its solution.)
Proposal 1. Because a 25% increase roughly equals a $5250 increase in revenue, and
the worths per additional ton of M1 and M2 are $750 and $500, respectively, the
desired increase in production can be achieved by making a combined increase of
$250 , ($750 +2 $500) = 8.4 tons in each of M1 and M2.
Proposal 2. Increase the amounts of raw materials M1 and M2 by 6 tons and 1 ton,
respectively. These increments equal 25% of the current levels of M1 and M2 (= 24
and 6 tons, respectively). Because these two resources are scarce at the current opti-
mum solution, a 25% increase in their availability produces an equivalent increase
in the levels of production of interior and exterior paints, as desired.
What is your opinion of these proposals? Would you suggest a different approach
for solving the problem?

CHAPTER 5 CASES
5-1.6 ABC Cola operates a plant in the northern section of the island nation of Tawanda. The
plant produces soft drinks in three types of packages that include returnable glass bot-
tles, aluminum cans, and nonreturnable plastic bottles. Returnable (empty) bottles are
shipped to the distribution warehouses for reuse in the plant.
Because of the continued growth in demand, ABC wants to build another plant.
The demand for the soft drinks (expressed in cases) over the next 5 years is given in
Table E.10. The planned production capacities for the existing plant extrapolated over
the same 5-year horizon are given in Table E.11. The company owns six distribution
warehouses: N1 and N2 are located in the north, C1 and C2 in the central section, and
S1 and S2 in the south. The share of sales by each warehouse within its zone is given
in Table E.12. Approximately 60% of the sales occur in the north, 15% in the central
section, and 25% in the south.
The company wants to construct the new plant either in the central section or in
the south. The transportation cost per case of returnable bottles is given in Table E.13.
It is estimated that the transportation costs per case of cans and per case of nonreturn-
ables are, respectively, 60% and 70% of that of the returnable bottles.
Should the new plant be located in the central or the southern section of the
country?

TABLE E.10

Year

Package 1 2 3 4 5

Returnables 2400 2450 2600 2800 3100

Cans 1750 2000 2300 2650 3050
Nonreturnables 490 550 600 650 720

6
Based on T. Cheng and C. Chiu, “A Case Study of Production Expansion Planning in a Soft-Drink
Manufacturing Company,” Omega, Vol. 16, No. 6, pp. 521–532, 1988.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.8

E.8 Appendix E Case Studies

TABLE E.11

Year

Package 1 2 3 4 5

Returnables 1800 1400 1900 2050 2150

Cans 1250 1350 1400 1500 1800
Nonreturnables 350 380 400 400 450

TABLE E.12

Warehouse Share percentage

N1 85
N2 15
C1 60
C2 40
S1 80
S2 20

TABLE E.13

Transportation cost per case ($)

Warehouse Existing plant Central plant South plant

N1 0.80 1.30 1.90

N2 1.20 1.90 2.90
C1 1.50 1.05 1.20
C2 1.60 0.80 1.60
S1 1.90 1.50 0.90
S2 2.10 1.70 0.80

5-2.7 The construction of Brisbane International Airport requires the pipeline movement of
about 1,355,000 m3 of sand dredged from five clusters at a nearby bay to nine sites at
the airport location. The sand is used to help stabilize the swampy grounds at the pro-
posed construction area. Some of the sites to which the sand is moved are dedicated to
building roads both within and on the perimeter of the airport. Excess sand from a site
will be moved by trucks to other outlying areas around the airport, where a perimeter
road will be built. The distances (in 100 m) between the source clusters and the sites are
summarized in Table E.14. The table also shows the supply and demand quantities in
100 m3 at the different locations.
(a) The project management has estimated a [volume (m3) × distance (100 m)] sand
movement of 2,495,000 units at the cost of $.65 per unit. Is the estimate given by
the project management for sand movement on target?

7
Based on C. Perry and M. Ilief, “Earth Moving on Construction Projects,” Interfaces, Vol. 13, No. 1,
pp. 79–84, 1983.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.9

Chapter 5 Cases E.9

TABLE E.14

1 2 3 4 5 6 7 8 9 Supply

1 22 26 12 10 18 18 11 8.5 20 960

2 20 28 14 12 20 20 13 10 22 201

3 16 20 26 20 1.5 28 6 22 18 71

4 20 22 26 22 6 q 2 21 18 24

5 22 26 10 4 16 q 24 14 21 99

Demand 62 217 444 315 50 7 20 90 150

(b) The project management has realized that sand movement to certain sites cannot be
carried out until some of the roads are built. In particular, the perimeter road (desti-
nation 9) must be built before movement to certain sites can be done. In Table E.15,
the blocked routes that require the completion of the perimeter road are marked
with x. In view of these restrictions, how should the sand movement be made?
5-3. Ten years ago, a wholesale dealer started a business distributing pharmaceuticals from a
central warehouse (CW). Orders were delivered to customers by vans. The warehouse
has since been expanded in response to growing demand. Additionally, two new ware-
houses (W1 and W2) have been constructed. The central warehouse, traditionally
well stocked, occasionally supplies the new warehouses with some short items. The
occasional supply of short items has grown into a large-scale operation in which the
two new warehouses receive for redistribution about one-third of their stock directly
from the central warehouse. Table E.16 gives the number of orders shipped out by each
of the three warehouses to customer locations C1 to C6. A customer location is a town
with several pharmacies.
The dealer’s delivery schedule has evolved over the years to its present status. In
essence, the schedule was devised in a rather decentralized fashion, with each ware-
house determining its delivery zone based on “self-fulfilling” criteria. Indeed,
in some instances, warehouse managers competed for new customers mainly to

TABLE E.15

1 2 3 4 5 6 7 8 9

1 x x x

2 x x x

3 x x

4 x x

5 x x x x
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.10

E.10 Appendix E Case Studies

TABLE E.16

Route

From To Number of orders

CW W1 2000
CW W2 1500
CW C1 4800
CW C2 3000
CW C3 1200
W1 C1 1000
W1 C3 1100
W1 C4 1500
W1 C5 1800
W2 C2 1900
W2 C5 600
W2 C6 2200

increase their “sphere of influence.” For instance, the managers of the central ware-
house boast that their delivery zone includes not only regular customers but the other
two warehouses as well. It is not unusual, then, that several warehouses deliver supplies
to different pharmacies within the same town (customer location). The distances in
miles traveled by vans between locations are given in Table E.17. A vanload usually
hauls 100 orders.
Evaluate the present distribution policy of the dealer.
5-4. Kee Wee Airlines flies eight two-way flights between Waco and Macon according to
the schedule in Table E.18. A crew can return to its home base (Waco or Macon) on the
same day, provided there is at least a 90-minute layover in the other city. Otherwise, the
crew can return the next day. It is desired to pair the crews with the flights originating
from the two cities to minimize the total layover time of all the crews.

TABLE E.17

CW W1 W2 C1 C2 C3 C4 C5 C6

CW 0 5 45 50 30 30 60 75 80

W1 5 0 80 38 70 30 8 10 60

W2 45 80 0 85 35 60 55 7 90

C1 50 38 85 0 20 40 25 30 70

C2 30 70 35 20 0 40 90 15 10

C3 30 30 60 40 40 0 10 6 90

C4 60 8 55 25 90 10 0 80 40

C5 75 10 7 30 15 6 80 0 15

C6 80 60 90 70 10 90 40 15 0
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.11

Chapter 6 Cases E.11

TABLE E.18

Flight From Waco To Macon Flight From Macon To Waco

W1 6:00 8:30 M1 7:30 9:30

W2 8:15 10:45 M2 9:15 11:15
W3 13:30 16:00 M3 16:30 18:30
W4 15:00 17:30 M4 20:00 22:00

CHAPTER 6 CASES
6-1. An outdoors person who lives in San Francisco (SF) wishes to spend a 15-day vacation
visiting four national parks: Yosemite (YO), Yellowstone (YE), Grand Teton (GT), and
Mount Rushmore (MR). The tour, which starts and ends in San Francisco, visits the
parks in the order SF : YO : YE : GT : MR : SF and includes a 2-day stay at
each park. Travel from one park location to another is either by air or car. Each leg of
the trip takes 1/2 day if traveled by air. Travel by car takes 1/2 day from SF to YO, 3 days
from YO to YE, 1 day from YE to GT, 2 days from GT to MR, and 3 days from MR back
to SF. The trade-off is that car travel generally costs less but takes longer. Considering
that the individual must return to work in 15 days, the objective is to make the tour as
inexpensively as possible within the 15-day limit. Table E.19 provides the one-way cost
of traveling by car and air. Determine the mode of travel on each leg of the tour.
6-2.8 A benefactor has donated books to the Springdale Public Library. The books come in
four heights: 12, 10, 8, and 6 inches. The head librarian estimates that 12 feet of shelving
will be needed for the 12-inch books, 18 feet for the 10-inch ones, 9 feet for the 8-inch
books, and 10 feet for the 6-inch ones. The construction cost of a shelf includes both a
fixed cost and a variable cost per foot length, as Table E.20 shows.
Given that smaller books can be stored on larger shelves, how should the shelves
be designed?
6-3. A shipping company wants to deliver five cargo shipments from ports A, B, and C
to ports D and E. The delivery dates for the five shipments are given in Table E.21.
Table E.22 gives trip times (in days) between ports (the return trip is assumed to take
less time). The company wants to determine the minimum number of ships needed to
carry out the given shipping schedule.

TABLE E.19

Air travel cost ($) to Car travel cost ($) to

From SF YO YE GT MR SF YO YE GT MR

SF — 150 350 380 450 — 130 175 200 230

YO 150 — 400 290 340 130 — 200 145 180
YE 350 400 — 150 320 175 200 — 70 150
GT 380 290 150 — 300 200 145 70 — 100
MR 450 340 320 300 — 230 180 150 100 —

8
Based on A. Ravindran, “On Compact Storage in Libraries,” Opsearch, Vol. 8, No. 3, pp. 245–252, 1971.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.12

E.12 Appendix E Case Studies

TABLE E.20

Shelf height Fixed cost Variable cost

(in.) ($) ($/ft length)

12 25 5.50
10 25 4.50
8 22 3.50
6 22 2.50

TABLE E.21

Shipment Shipping route Delivery date

1 A to D 10
2 A to E 15
3 B to D 4
4 B to E 5
5 C to E 18

6-4.9 Several individuals set up unregulated brokerage firms overseas that traded in highly
speculative stocks. The brokers operated under a loose financial system that allowed
extensive interbrokerage transactions, including buying, selling, borrowing, and lending.
For the group of brokers as a whole, the main source of income was the commission
they received from sales to outside clients.
Eventually, the risky trading in speculative stocks became unmanageable, and
all the brokers declared bankruptcy. The financial situation at that time was that all
brokers owed money to outside clients, and the interbroker financial entanglements
were so complex that almost every broker owed money to every other broker in the
group.

TABLE E.22

A B C D E

A 3 4

B 3 2

C 3 5

D 2 2 2

E 3 1 4

9
Based on H. Taha, “Operations Research Analysis of a Stock Market Problem,” Computers and Operations
Research, Vol. 18, No. 7, pp. 597–602, 1991.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.13

Chapter 7 Comprehensive Problems E.13

The brokers whose assets could pay for their debts were declared solvent.
The remaining brokers were referred to a legal body whose purpose was to resolve
the debt situation in the best interest of outside clients. Because the assets and
receivables of the nonsolvent brokers were less than their payables, all debts were
prorated. The final effect was a complete liquidation of all the assets of the
nonsolvent brokers.
In resolving the financial entanglements within the group of nonsolvent brokers,
it was decided that the transactions would be executed only to satisfy certain legal
requirements because, in effect, none of the brokers would be keeping any of the funds
owed by others. The legal body requested that the number of interbroker transactions
be reduced to an absolute minimum. This meant that if A owed B an amount X, and B
owed A an amount Y, the two “loop” transactions were reduced to one whose amount
was |X - Y|. This amount would go from A to B if X 7 Y and from B to A if Y 7 X.
If X = Y, the transactions were completely eliminated. The idea was to be extended to
all loop transactions involving any number of brokers.
How would you handle this situation? Specifically, you are required to answer two
questions.
1. How should the debts be prorated?
2. How should the number of interbroker transactions be reduced to a minimum?

CHAPTER 7 COMPREHENSIVE PROBLEMS

7-1. Suppose that you are given the points

A = (6, 4, 6, - 2), B = (4, 12, - 4, 8), C = (-4, 0, 8, 4)

Develop a systematic procedure that will allow determining whether or not each of
the following points can be expressed as a convex combination of A, B, and C:
(a) (3, 5, 4, 2)
(b) (5, 8, 4, 9)
7-2. Consider the following LP:

Maximize z = 3x1 + 2x2

Subject to

x1 + 2x2 … 6

2x1 + x2 … 8

- x1 + x2 … 1
x1, x2 Ú 0

Determine the optimum simplex tableau (use TORA for convenience), and then directly
use the information in the optimum simplex tableau to determine the second-best
extreme-point solution (relative to the “absolute” optimum) for the problem. Verify the
answer by solving the problem graphically. (Hint: Consult the extreme points that are
adjacent to the optimum solution.)
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.14

E.14 Appendix E Case Studies

7-3. Interval programming, Consider the following LP:

Maximize z = {CX|L … AX … U, X Ú 0}

where L and U are constant column vectors. Define the slack vector such that
AX + Y = U. Show that this LP is equivalent to

Maximize z = {CX|AX + Y = U, 0 … Y … U - L, X Ú 0}

Use the proposed procedure to solve the following LP:

Minimize z = 5x1 - 4x2 + 6x3

subject to

20 … x1 + 7x2 + 3x3 … 46
10 … 3x1 - x2 + x3 … 20
18 … 2x1 + 3x2 - x3 … 35
x1, x2, x3 Ú 0

7-4. The optimum solution of the LP in Problem 7-2 is given as x1 = 10 4

3 , x2 = 3 , and
38 10
z = 3 . Plot the change in optimum z with u, given that x1 = 3 + u, where u is unre-
stricted in sign. Note that x1 = 10
3 + u tracks x1 above and below its optimal value.
7-5. Consider the following minimization LP:

Minimize z = (10t - 4)x1 + (4t - 8)x2

subject to

2x1 + 2x2 + x3 = 8
4x1 + 2x2 + x4 = 6 - 2t
x1, x2, x3, x4 Ú 0

where - q 6 t 6 q. The parametric analysis of the problem yields the following results:

-q 6 t … -5: Optimal basis is B = (P1, P4)

-5 … t … -1: Optimal basis is B = (P1, P2)
-1 … t … 2: Optimal basis is B = (P2, P3)

Determine all the critical values of t that may exist for t Ú 2.

7-6. Suppose that the optimum linear program is represented as

Maximize z = c0 - a (zj - cj)xj

jeNB

subject to
xi = x *i - a aijxj, i = 1, 2, . . . , m
jeNB

all xi and xj Ú 0
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.15

Chapter 8 Cases E.15

where NB is the set of nonbasic variables. Suppose that for a current basic variable
xi = xi we impose the restriction xi Ú di, where di is the smallest integer greater than
xi. Estimate an upper bound on the optimum value of z after the constraint is added
to the problem. Repeat the same procedure assuming that the imposed restriction is
xi … ei, where ei is the largest integer smaller than xi.

CHAPTER 8 CASES
8-1.10 The Warehouzer Company manages three sites of forest land for timber production
and reforestation with the respective areas of 100,000, 180,000, and 200,000 acres. The
main timber products include three categories: pulpwood, plywood, and sawlogs. Several
reforestation alternatives are available for each site, each with its cost, number of
rotation years (i.e., number of years from seedling size till harvesting), return from rent,
and production output. Table E.23 summarizes this information.
To guarantee sustained future production, each acre of reforestation in each
alternative requires that as many acres as years in rotation be assigned to that
alternative. The rent column represents the stumpage value per acre.
The goals of Warehouzer are as follows:
1. Annual outputs of pulpwood, plywood, and sawlogs are 200,000, 150,000, and
350,000 cubic meters, respectively.

TABLE E.23

Annual $/acre Annual m3/acre

Rotation
Site Alternative Cost Rent year Pulpwood Plywood Sawlogs

1 A1 1000 160 20 12 0 0
A2 800 117 25 10 0 0
A3 1500 140 40 5 6 0
A4 1200 195 15 4 7 0
A5 1300 182 40 3 0 7
A6 1200 180 40 2 0 6
A7 1500 135 50 3 0 5
2 A1 1000 102 20 9 0 0
A2 800 55 25 8 0 0
A3 1500 95 40 2 5 0
A4 1200 120 15 3 4 0
A5 1300 100 40 2 0 5
A6 1200 90 40 2 0 4
3 A1 1000 60 20 7 0 0
A2 800 48 25 6 4 0
A3 1500 60 40 2 0 4
A4 1200 65 15 2 0 3
A5 1300 35 40 1 0 5

10
Based on K. Rustagi, Forest Management Planning for Timber Production: A Goal Programming
Approach, Bulletin No. 89, Yale University, New Haven, 1976.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.16

E.16 Appendix E Case Studies

2. Annual reforestation budget is $2.5 million.

3. Annual return from land rent is $100 per acre.
How much land at each site should be assigned to each alternative?
8-2. A charity organization runs a children’s shelter. The organization relies on volunteer
service from 8:00 A.M. until 2:00 P.M. Volunteers may begin work at the start of any hour
between 8:00 A.M. and 11:00 A.M. A volunteer works a maximum of 6 hours and a minimum
of 2 hours, and no volunteers work during lunch hour between 12:00 noon and 1:00 P.M.
The charity has estimated its goal of needed volunteers throughout the day (from 8:00 A.M.
to 2:00 P.M., and excluding the lunch hour between 12:00 noon and 1:00 P.M.) as 15, 16, 18, 20,
and 16, respectively. The objective is to decide on the number of volunteers that should start
at each hour (8:00, 9:00, 10:00, 11:00, and 1:00) such that the given goals are met as much as
possible. Formulate and solve the problem as a goal programming model.

CHAPTER 9 CASES
9-1. A development company owns 90 acres of land in a growing metropolitan area, where it
intends to construct office buildings and a shopping center. The developed property is
rented for 7 years and then sold. The sale price for each building is estimated at 10 times
its operating net income in the last year of rental. The company estimates that the
project will include a 4.5-million-square-foot shopping center. The master plan calls
for constructing three high-rise and four garden office buildings.
The company is faced with a scheduling problem. If a building is completed too early,
it may stay vacant; if it is completed too late, potential tenants may be lost to other projects.
The demand for office space over the next 7 years based on appropriate market studies is
given in Table E.24. Table E.25 lists the proposed capacities of the seven buildings.

TABLE E.24

Demand (thousands of ft2)

Year High-rise space Garden space

1 200 100
2 220 110
3 242 121
4 266 133
5 293 146
6 322 161
7 354 177

TABLE E.25

Capacity High-rise Capacity

Garden (ft2) buildings (ft2)

1 60,000 1 350,000
2 60,000 2 450,000
3 75,000 3 350,000
4 75,000 — —
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.17

Chapter 9 Cases E.17

TABLE E.26

U of A Scores for gymnast

Event 1 2 3 4 5 6

Vault 6 9 8 8 4 10
Bars 7 9 7 8 9 5
Beam 9 8 10 9 9 8
Floor 6 6 5 9 10 9

The gross rental income is estimated at $25 per square foot. The operating expenses
are $5.75 and $9.75 per square foot for the garden and high-rise buildings, respectively.
The associated construction costs are $70 and $105 per square foot, respectively. Both
the construction cost and the rental income are estimated to increase at roughly the
inflation rate of 4%.
How should the company schedule the construction of the seven buildings?
9-2.11 In a National Collegiate Athletic Association women’s gymnastic meet, competition
includes four events: vault, uneven bars, balance beam, and floor exercises. Each team
may enter the competition with six gymnasts per event. A gymnast is evaluated on
a scale of 1 to 10. Past statistics for the U of A team produce the scores in Table E.26.
The total score for a team is determined by summarizing the top five individual
scores for each event. An entrant may participate as a specialist in one event or an
“all-rounder” in all four events but not both. A specialist is allowed to compete in
at most three events, and at least four of the team participants must be all-rounders.
Set up an ILP model that can be used to select the competing team, and find the
optimum solution.
9-3.12 In 1990, approximately 180,000 telemarketing centers employing 2 million individuals
were in operation in the United States. In the year 2000, more than 700,000 companies
employed approximately 8 million people to telemarket their products. The questions
of how many telemarketing centers to employ and where to locate them are of
paramount importance.
The ABC company is in the process of deciding on the number of telemarketing
centers to employ and their locations. A center may be located in one of several
candidate areas selected by the company and may serve (partially or completely)
one or more geographical areas. A geographical area is usually identified by one or
more (telephone) area codes. ABC’s telemarketing concentrates on eight area codes:
501, 918, 316, 417, 314, 816, 502, and 606. Table E.27 provides the candidates’ locations,
their served areas, and the cost of establishing the center. The communication costs
per hour between the centers and the area codes are given in Table E.28.
ABC would like to select three or four centers. Where should they be located?
9-4.13 An electric utility company serving a wide rural area wants to decide on the number
and location of Customer-Service Linemen (CSL) centers that will provide responsive

11
Based on P. Ellis and R. Corn, “Using Bivalent Integer Programming to Select Teams for Intercollegiate
Women’s Gymnastic Competition,” Interfaces, Vol. 14, No. 3, pp. 41–46, 1984.
12
Based on T. Spencer, A. Brigandi, D. Dargon, and M. Sheehan, “AT&T’s Telemarketing Site Selection System
Offers Customer Support,” Interfaces, Vol. 20, No. 1, pp. 83–96, 1990.
13
Based on T. Erkut, Myrdon, and K. Strangway, “Transatlanta Redesigns its Service Delivery Network,”
Interfaces, Vol. 30, No. 2, pp. 54–69, 2000.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.18

E.18 Appendix E Case Studies

TABLE E.27

Center location Served area codes Cost ($)

Dallas, TX 501, 918, 316, 417 500,000

Atlanta, GA 314, 816, 502, 606 800,000
Louisville, KY 918, 316, 417, 314, 816 400,000
Denver, CO 501, 502, 606 900,000
Little Rock, AR 417, 314, 816, 502 300,000
Memphis, TN 606, 501, 316, 417 450,000
St. Louis, MO 816, 502, 606, 314 550,000

TABLE E.28

Area code

To 501 918 316 417 314 816 502 606

From

Dallas, TX ($) 14 35 29 32 25 13 14 20
Atlanta, GA ($) 18 18 22 18 26 23 12 15
Louisville, KY ($) 22 25 12 19 30 17 26 25
Denver, CO ($) 24 30 19 14 12 16 18 30
Little Rock, AR ($) 19 20 23 16 23 11 28 12
Memphis, TN ($) 23 21 17 21 20 23 20 10
St. Louis, MO ($) 17 18 12 10 19 22 16 22

service regarding repairs and connections. The company groups its customer base in
five clusters according to Table E.29. The company has selected five potential location
for its CSL centers. Table E.30 summarizes the average travel distance in miles from
the CSLs to the different clusters. The average speed of the service truck is approxi-
mately 45 miles per hour.

TABLE E.29

Cluster 1 2 3 4 5

Number of customers 400 500 300 600 700

TABLE E.30

CSL Center

Cluster 1 2 3 4 5

1 40 100 20 50 30
2 120 90 80 30 70
3 40 50 90 80 40
4 80 70 110 60 120
5 90 100 40 110 90
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.19

Chapter 9 Cases E.19

The company would like to keep the response time to a customer request to
around 90 minutes. How many CSL centers should be in operation?
9-5.14 In the automobile industry, prototype vehicles are used to test new designs. The building
of these vehicles represents a major investment that may exceed $250,000 per prototype.
Separate tests are carried out by different groups, each concentrating on checking certain
attributes of the new design. For example, possible attributes of a transit vehicle could
include body style, engine size, roof height, transmission type, rear closure, gross vehicle
weight, and wheel base. To examine a worst-case scenario for high-altitude drivability
requires a prototype with highest gross vehicle weight, automatic transmission, and smallest
engine. Attributes such as roof height and wheelbase are not important for this type of test.
At the outset, prototypes can be built to meet the individual attributes specified
by the tester. For example, if test 1 involves attribute A, and test 2 requires attribute
B, two different prototypes can be built: one for A and the second for B. Alternatively,
two identical units of prototype (A, B) can be used for the two tests. The advantage is
that the production of two identical prototypes is considerably less expensive than
building two distinct units. In this case (A, B) is said to be a shared prototype.
In a general situation, let Ai, i = 1, 2, Á , n , be the set of configurations for
attribute i. For example, if attribute 1 is transmission, then A1 = {standard, automatic} .
A prototype is built using one configuration from each attribute. Thus, if the number
of configurations in attribute i is mi , then the maximum possible number of configura-
n
tions is q i = 1mi . New designs may thus involve thousands of prototypes, and the idea
is to make a judicious selection of shared prototypes that meet testers’ specifications.
Let Bj, j = 1, 2, Á , t, represent the set of configurations requested by tester j. For
example, B1 = {V8 engine, standard}. The elements of a set B must thus be a buildable
prototype or a subset of it.
Based on the given information, how should buildable prototypes be selected
to meet the requirements of the testers? Apply the developed model to the situation
in Table E.31. The associated testers’ requirements are given in Table E.32. [Note:

TABLE E.31

Attribute Configurations

Engine {4 cyl, 6 cyl, 8 cyl}

Transmission {automatic, standard}
Body style {4-door, coupe, wagon}

TABLE E.32

Tester Desired configurations

1 {4 cyl, standard}
2 {8 cyl, coupe}
3 {6 cyl, wagon}
4 {8 cyl}
5 {standard, wagon}
6 {6 cyl, automatic}

14
Based on K. Chelst, J. Sidelko, A. Przebienda, J. Lockledge, and D. Mihailidis, “Rightsizing and Management
of Prototype Vehicle Testing at Ford Motor Company,” Interfaces, Vol. 31, No. 1, pp. 91–107, 2001.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.20

E.20 Appendix E Case Studies

TABLE E.33

From city To city Flight number

C1 C2 F1
C1 C3 F2
C1 C5 F3
C1 C8 F4
C2 C4 F5
C2 C7 F6
C2 C8 F7
C3 C1 F8
C3 C2 F9
C3 C6 F10
C4 C1 F11
C4 C8 F12
C5 C2 F13
C5 C7 F14
C6 C4 F15
C6 C1 F16
C7 C4 F17
C8 C3 F18

The given situation is oversimplified and obviously can be solved by inspection. In

a real situation with thousands of buildable prototypes and tens of requested tests,
the solution will not be as obvious.]
9-6.15 American Express Airlines operates between 8 cities (C1 through C8) with 18 flights
(F1 through F18) and 10 flight crews (R1 through R10). Crews normally start from a
given base and return to the same base after completing their assignments. Table E.33
provides the daily flight schedules for the airline. The scheduling department is in charge
of developing crew pairings that take into account legalities as well as crew preferences.
A feasible pairing defines the routes (i.e., flights) a crew can service during the planning
period. Table E.34 provides the feasible pairings for the 10 crews. The cost of assigning
a crew to a pairing is proportional to the number of flight legs a pairing covers.

TABLE E.34

Crew Feasible pairings

1 (C3, C6, C4, C8, C3), (C3, C2, C8, C3)

2 (C1, C5, C7, C4, C1)
3 (C4, C8, C3, C2, C4), (C4, C1, C5, C7, C4)
4 (C1, C8, C3, C6, C4, C1), (C1, C5, C2, C1)
5 (C2, C4, C8, C3, C2), (C2, C4, C8, C3, C1, C2), (C2, C7, C4, C1, C2)
6 (C8, C3, C1, C8)
7 (C5, C2, C8, C3, C1, C5), (C5, C7, C4, C8, C3, C1, C5)
8 (C6, C1, C3, C6), (C6, C4, C1, C5, C7, C4, C8, C3, C6), (C6, C4, C8, C3, C6)
9 (C7, C4, C2, C7), (C7, C4, C8, C3, C2, C7)
10 (C1, C3, C6, C1), (C1, C2, C8, C3, C1), (C1, C8, C3, C1)

15
Based on G. Yu, M. Arguello, G. Song, S. McCowan, and A. White, “A New Era for Recovery at Continental
Airlines,” Interfaces, Vol. 3, No. 1, pp. 5–22, 2003.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.21

Chapter 9 Cases E.21

Because the pairings are developed to satisfy the crew preferences as well as
Federal Aviation Administration (FAA) regulations, the pairings proposed by the
scheduling department may not produce a feasible solution that covers all the flights
and engages all the crews. The objective then is to assign crew pairings in a manner that
will eliminate infeasibilities as much as possible. The assumption is that if the developed
solution is infeasible, then the scheduling department should either propose additional
feasible pairings or seek the service of reserve crews.
Develop a model that can be used to evaluate the pairings proposed by the
scheduling department and interpret the solution.
9-7.16 The core manufacturing flow for microelectronic parts starts with wafers (CD-like
round thin pieces of silicon) on which thousands of circuits are etched. A completed
wafer is then cut into small rectangular parts, called devices, placed on a substrate, and
packaged to create a module. After testing, the devices are found to fall in different
categories, each with distinct circuits. Given that N is the number of devices cut from
a wafer with n categories, the number of units binned into category j is estimated at rjN,
where r1 + r2 + . . . + rn = 1, rj Ú 0, for all j. Produced devices may be used inter-
changeably in the production of a module, so that one unit of device i or device j may
be used to produce one unit of module k. Interchangeability of devices is a function of
the specification of the module. Given the binning ratios rj, how many wafers should
be produced to satisfy a specific demand for the modules? How should the produced
devices be allocated to the modules?
Test the developed model for a specific situation with 5 devices and 3 modules
using the data in Tables E.35 and E.36.
9-8.17 In days past, banks used to clear checks against a customer’s account in the random order
in which the checks were received. Nowadays, and with the advent of modern data pro-
cessing capabilities, some banks are legally allowed to sequence the daily debiting process
in a manner that garners higher return fees for insufficient funds. For example, suppose
that a bank account has a balance of $1000 and that in a specific day four successive
checks are received in the amounts $100, $100, $100, and $1000. If these checks are debited
in their order of receipt, only one check ($1000) should be returned for not-sufficient
funds (NSF). In this case, the customer is responsible for one NSF charge (of about $20).

TABLE E.35

Device 1 2 3 4 5

Binning ratio .21 .19 .1 .3 .2

Initial inventory 10 4 8 0 3

TABLE E.36

Module Demand (units) Interchangeable devices

1 20 2, 3, or 5
2 30 1, 3, or 5
3 45 4, or 5

16
Based on P. Lyon, R. Milne, R. Orzell, and R. Rice, “Matching Assets with Demand in Supply-Chain
Management at IBM Microelectronics,” Interfaces, Vol. 31, No. 1, pp. 108–124, 2001.
17
Based on A. Apte, U. Apte, R. Beatty, I. Sarkar, and J. Semple, “The Impact of Check Sequencing on NSF
(Not-Sufficient Funds) Fees,” Interfaces, Vol. 34, No. 2, pp. 97–105, 2004.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.22

E.22 Appendix E Case Studies

TABLE E.37

Minimum quantity required by bidders 1-8

of line items 1-6 (in 1000 barrels)

Line item 1 2 3 4 5 6 7 8

1 10 13 10 25 18 14 20 14
2 14 11 11 20 17 16 18 14
3 20 15 20 10 16 16 17 17
4 11 17 15 6 20 16 17 8
5 15 29 18 12 14 10 15 10
6 18 20 19 22 18 8 19 15

If, on the other hand, the checks are debited in the order $1000, $100, $100, and $100,
three NSF checks will result, and the bank collects three NSF charges. From this example,
it appears that a bank can maximize its NSF charges by using a high-low sequence that
debits the higher checks first. This is not true, in general. For example, consider the high-
low sequence of $900, $675, $525, $200, $100, $75, and $25 against an account balance of
$1200. In this case, the checks $900, $200, and $100 are cleared, and the remaining four
checks carry NSF charges. Actually, the bank could collect one extra NSF if it skipped the
$900 check and cleared the $675 and $525 checks (= $1200) first.
(a) Develop a model that will allow banks to process the daily checks in a manner
that guarantees the collection of maximum NSF charges, and apply the model
to the given data.
(b) Ethically, one should expect banks to offer the best service to customers by
minimizing the NSF charges. How should the checks be processed in this case?
9-9. Consider Case 3-3 (Chapter 3). For some bidders, an awarded quantity is acceptable
only if it satisfies the specific minimum requirement given in Table E.37. A successful
bidder must receive at least the minimum requirement (but within the 20% limit
specified by law). Else no award is made to the bidder.
How can this task be accomplished?
9-10.18 A construction company has been awarded contracts for 8 projects located in different
geographical locations around the United States. Each project is administered by one of
the company’s 5 managers. The managers are stationed in different home bases around
the country, and their travel times to different project locations vary. High-cost projects
are administratively more demanding. To be equitable, the company assigns managers
to the projects depending on both the size of the project and also the proximity of the
manager’s home base to the location of the project. Table E.38 gives the estimated costs
of the projects (in millions of dollars). The travel times are given in Table E.39. How
should the managers be assigned to the projects? [Hint: Base assignments on project

TABLE E.38

Project 1 2 3 4 5 6 7 8
Cost ($106) 10 2 24 5 15 12 7 9

18
Based on L. LeBlanc, D. Randels, Jr., and T. K. Swann, “Heery International’s Spreadsheet Optimizations
Model for Assigning Managers to Construction Projects,” Interfaces, Vol. 30, No. 6, pp. 95–106, 2000.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.23

Chapter 13 Cases E.23

TABLE E.39

Travel time to project locations

in hours by manager

Project a b .c d e

1 2 5 3 1 6
2 5 4 2 5 3
3 4 1 3 2 2
4 5 3 6 3 4
5 1 4 5 6 1
6 2 4 6 2 3
7 6 7 2 3 3
8 4 2 1 5 4

intensity, defined here as (travel time in hours + 1) * 6 * log (project cost in million $) + 1.
The expression is a well-known measure of project intensity in construction.]

CHAPTER 12 CASE
12-1. A company reviews the status of heavy equipment at the end of each year, and a decision
is made either to keep the equipment an extra year or to replace it. However, equipment
that has been in service for 3 years must be replaced. The company wishes to develop
a replacement policy for its fleet over the next 10 years. Table E.40 provides the
pertinent data. The equipment is new at the start of year 1.

CHAPTER 13 CASES
13-1. The distribution center of the retailer Walmark Stores engages on a daily basis
in buying many staple, nonfashionable inventory items. Steady demand for the
various items comes from the numerous stores Walmark owns. In the past,

TABLE E.40

Maintenance cost ($) Salvage value ($)

Purchase
Year price ($) 0 1 2 1 2 3

1 10,000 200 500 600 9,000 7,000 5,000

2 12,000 250 600 680 11,000 9,500 8,000
3 13,000 280 550 600 12,000 11,000 10,000
4 13,500 320 650 700 12,000 11,500 11,000
5 13,800 350 590 630 12,000 11,800 11,200
6 14,200 390 620 700 12,500 12,000 11,200
7 14,800 410 600 620 13,500 12,900 11,900
8 15,200 430 670 700 14,000 13,200 12,000
9 15,500 450 700 730 15,500 14,500 13,800
10 16,000 500 710 720 15,800 15,000 14,500
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.24

E.24 Appendix E Case Studies

decisions regarding how much and when to order were relegated to the buyers,
whose main purpose was to acquire the items in sufficiently large quantities to
guarantee the low purchase prices. This policy was carried out without conscious
concern about the inventory status of the items. Indeed, decisions regarding how
much to buy were based on the annual dollar usage of the item at the distribution
center level. For example, if an item was purchased for $25 a unit and consumed
at the rate of 10,000 units a year, then its annual dollar usage is estimated at
$250,000. The main guideline the buyers used was that the higher the annual dollar
usage of an item, the higher should be its stock level in the distribution center. This
guideline translated into expressing the amount of inventory that must be kept on
hand at the distribution center as the period between replenishments. For example,
a buyer might purchase a prespecified amount of an item every 3 months.
To exercise better inventory control, Walmark decided to enlist the help of an opera-
tions research consultant. After studying the situation, the consultant concluded that the
consumption rate of most items in the distribution center was, for all practical purposes,
constant and that Walmark operated under the general policy of not allowing shortages.
The study further indicated that the inventory-holding cost for all the items under consid-
eration was a constant percentage of the unit purchase price. Furthermore, the fixed cost
a buyer incurred with each purchase was the same regardless of the item involved. Armed
with this information, the consultant was able to develop a single curve for any single
item that related the annual dollar usage to the average time between replenishments.
This curve was then used to decide on which items currently were overstocked or
understocked. How did the analyst do it?
13-2. A company manufactures a final product that requires the use of a single component.
The company purchases the component from an outside supplier. The demand rate
for the final product is constant at about 20 units per week. Each unit of the final
product uses 2 units of the purchased component. Table E.41 inventory data are
available. Unfilled demand of the final product is backlogged and costs about
$8 per lost unit per week. Shortage in the purchased component is not expected
to occur. Devise an ordering policy for the purchase of the component and the
production of the final product.
13-3. A company deals with a seasonal item, for which the monthly demand fluctuates
appreciably. Table E.42 provides demand data (in number of units). Because of the
fluctuations in demand, the inventory control manager has chosen a policy that orders
the item quarterly on January 1, April 1, July 1, and October 1. The order size covers
the demand for each quarter. The lead time between placing an order and receiving it
is 3 months. Estimates for the current year’s demand are taken equal to the demand
for year 5, plus an additional 10% safety factor.
A new staff member believes that a better policy can be determined by using the
economic order quantity based on the average monthly demand for the year. Fluctuations
in demand can be “smoothed” out by placing orders to cover the demands for consecutive
months, with the size of each order approximately equal to the economic lot size. Unlike

TABLE E.41

Component Product

Setup cost per order ($) 80 100

Unit holding cost per week ($) 2 5
Lead time (week) 2 3
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.25

Chapter 15 Cases E.25

TABLE E.42

Year

Months 1 2 3 4 5

January 10 11 10 12 11
February 50 52 60 50 55
March 8 10 9 15 10
April 99 100 105 110 120
May 120 100 110 115 110
June 100 105 103 90 100
July 130 129 125 130 130
August 70 80 75 75 78
September 50 52 55 54 51
October 120 130 140 160 180
November 210 230 250 280 300
December 40 46 42 41 43

the manager, the new staff member believes that the estimates for next year’s demand
should be based on the average of years 4 and 5.
The company bases its inventory computations on a holding cost of $.50 per unit
inventory per month. A setup cost of $55 is incurred when a new order is placed.
Suggest an inventory policy for the company.

CHAPTER 15 CASES
15-1.19 A shop manager is considering three alternatives to an existing milling machine.
(a) Retrofit the existing mill with a power feed (PF).
(b) Buy a new mill with a computer-aided design (CAD) feature.
(c) Replace the mill with a machining center (MC).
The three alternatives are evaluated based on two criteria: monetary and performance.
Table E.43 provides the pertinent data. The manager surmises that the monetary criterion

TABLE E.43

Criterion PF CAD MC

Monetary
Initial cost ($) 12,000 25,000 120,000
Maintenance cost ($) 2000 4000 15,000
Training cost ($) 3000 8000 20,000

Performance
Production rate (units/day) 8 14 40
Setup time (min) 30 20 3
Scrap (lb/day) 440 165 44

19
Based on S. Weber, “A Modified Analytic Hierarchy Process for Automated Manufacturing Decisions,”
Interfaces, Vol. 23, No. 4, pp. 75–84, 1993.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.26

E.26 Appendix E Case Studies

is 1 1冫2 times as important as the performance criterion. Additionally, the production rate
is twice as important as the setup time and 3 times as important as the scrap. The setup
time is regarded as 4 times as important as the scrap. As for the monetary criterion, the
manager estimates that the maintenance and training costs are of equal importance, and
the initial cost is twice as important as either of these two costs.
Analyze the situation, and make an appropriate recommendation.
15-2.20 A company operates a catalog sales operation encompassing more than 200,000 items
stocked in many regional warehouses. In the past, the company considered it essential
to keep accurate records of the actual inventory in each warehouse. As a result, a full
inventory count was ordered every year—an intense and unwelcome activity that was
done grudgingly by all warehouses. The company followed each count by an audit that
sampled about 100 items per warehouse to check the quality of the logistical operation
in each region. The result of the audit indicated that, on the average, only 64% of the
items in each warehouse matched the actual inventory, which was unacceptable. To
remedy the situation, the company ordered more frequent counts of the expensive and
fast-moving items. A system analyst was assigned the task of setting up procedures for
targeting these items.
Instead of responding directly to the company’s request for identifying the target
items, the system analyst decided to identify the cause of the problem. The analyst
ended up changing the goal of the study from “How can we increase the frequency of
inventory counts?” to “How can we increase the accuracy of inventory counts?” The
study led to the following analysis: Given that the proportion of accurately counted
items in a warehouse is p, it is reasonable to assume that there is a 95% chance that an
item that was counted correctly in the first place will again be recounted correctly in
a subsequent recount. For the proportion 1 - p that was not counted correctly in the
first round, the chance of a correct recount is 80%. Using this information, the analyst
developed a decision tree to graph a break-even chart that compared the count accuracy
in the first and second rounds. The end result was that the warehouses that had an
accuracy level above the break-even threshold were not required to recount inventory.
The surprising result of the proposed solution was a zealous effort on the part of each
warehouse to get the count right the first time around, with a resounding across-
the-board improvement in count accuracy in all the warehouses.
How did the analyst convince management of the viability of the proposed
threshold for recounting?
15-3.21 In the airline industry, working hours are ruled by agreements with the unions. In
particular, the maximum length of tour of duty may be limited to 16 hours for Boeing-747
flights and 14 hours for Boeing-707. If these limits are exceeded because of unexpected
delays, the crew must be replaced by a fresh one. The airlines maintain reserve crews
for such eventualities. The average annual cost of a reserve crew member is estimated at
$30,000. Conversely, an overnight delay resulting from the unavailability of a reserve crew
could cost as much as $50,000 for each delay. A crew member is on call 12 consecutive
hours a day for 4 days of the week and may not be called on during the remaining 3 days of
the week. A B-747 flight may also be served by two B-707 crews. Table E.44 summarizes
the callout probabilities for reserve crews based on 3-year historical data. As an illustration,

20
Based on I. Millet, “A Novena to Saint Anthony, or How to Find Inventory by Not Looking,” Interfaces,
Vol. 24, No. 2, pp. 69–75, 1994.
21
Based on A. Gaballa, “Planning Callout Reserves for Aircraft Delays,” Interfaces, Vol. 9, No. 2, Part 2,
pp. 78–86, 1979.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.27

Chapter 15 Cases E.27

TABLE E.44

Callout probability

Trip category Trip (hr) B-747 B-707

1 14.0 .014 .072

2 13.0 .0 .019
3 12.5 .0 .006
4 12.0 .016 .006
5 11.5 .003 .003
6 11.0 .002 .003

the data indicate that for 14-hour trips, the probability of a callout is .014 for B-747 and .072
for B707. Table E.45 provides a typical peak-day schedule. The present policy for reserve
crews calls for using two (seven-member) crews between 5:00 and 11:00, four between
11:00 and 17:00, and two between 17:00 and 23:00.
Evaluate the effectiveness of the present reserve crew policy. Specifically, is the
present reserve crew size too large, too small, or just right?
15-4.22 During the well-publicized 1982 trial of John Hinkley, accused of attempting to assassi-
nate U.S. President Ronald Reagan, the defense attorney wanted to introduce Hinkley’s
CAT scan results as evidence that his client was mentally ill. Hinkley’s CAT scan did
show brain atrophy. Expert testimony during the trial stipulated that 30% of individuals
diagnosed with schizophrenia had brain atrophy, as opposed to only 2% of those who
were not schizophrenic. Statistics show that approximately 1.5% of the U.S. population
suffer from schizophrenia.
Analyze the situation from the standpoint of the impact of introducing CAT scan
results as evidence on the outcome of the trial.
15-5.23 An instructor wants to estimate the probability that students in his junior–senior class
have ever cheated in a test during their tenure at the university. To obtain unbiased

TABLE E.45

Time of day Aircraft Trip category

8:00 707 3
9:00 707 6
707 2
10:00 707 3
11:00 707 2
707 4
15:00 747 6
16:00 747 4
19:00 747 1

22
Based on A. Barnett, I. Greenberg, and R. Machol, “Hinckley and the Chemical Bath,” Interfaces, Vol. 14,
No. 4, pp. 48–52, 1984.
23
Based on R. Sheaffer, J. Witmer, A. Watkins, and M. Gnanadesikan, Activity-Based Statistics, Springer-
Verlag, New York, 1996, p. 133.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.28

E.28 Appendix E Case Studies

(truthful) answers from the students, each student is asked to toss a coin privately to
answer a decoy question if the outcome is Heads or a real question if the outcome is
Tails. The real question is “Did you ever cheat in a test?” and the decoy question is
“Are you a graduating senior?” Each student answers “yes” or “no” on a sheet of paper,
and the sheets are then collected and tallied by the instructor. Privacy is guaranteed
because no one but the individual student knows which question was answered. Of the
35 students participating in the experiment, 20 are graduating seniors. The tallied results
of the experiment show 18 yes and 17 no answers. Use this information to estimate the
probability that a student in the designated class has ever cheated in a test.

CHAPTER 16 CASES
16-1.24 A telephone company operates telephone centers that provide residential services to
customers in their respective domains. There are more than 60 telephone models to
choose from. Currently, each phone center holds from 15 to 75 days of stock. The
management considers such stock levels to be excessive because they are replenished
on a daily basis from a central warehouse. At the same time, the management wants to
ensure that sufficient stock is maintained at the telephone centers to provide a service
level of 95% for the customers. The team studying the problem started by collecting
pertinent data. The team’s objective was to establish an optimal stock level for each
telephone model. Table E.46 shows the number of sets issued in a day of the green,
desktop, rotary-dial model (Green 500). Similar tables were developed for all the
models.
The desired cost parameters needed to determine the optimal stock level for each
telephone model are difficult to estimate, so traditional inventory models cannot be
applied. Based on the observation that both regression and time series analyses failed
to detect appreciable trends in demand, the team has decided to use a more basic
approach for determining appropriate stock levels for the different phone models.
Suggest a method for determining adequate stock levels for the different models.
State all the assumptions made to reach a decision.
16-2.25 The inventory manager of a small retail store places orders for items to take advantage
of special offers or to combine orders received from one supplier. The result is that both
the order quantity and the cycle length (interval between successive orders) become
essentially random. Moreover, because the manager’s policy is driven mostly by
noninventory considerations, the order quantity and cycle length can be considered
independent, in the sense that shorter cycle lengths do not necessarily mean smaller
order quantities and vice versa.

TABLE E.46

Sets issued 0 1 2 3 4
Frequency 189 89 20 4 1

24
Based on R. Cohen, and F. Dunford, “Forecasting for Inventory Control: An Example of When ‘Simple’
Means ‘Better,’” Interfaces, Vol. 16, No. 6, pp. 95–99, 1986.
25
Based on A. Holt, “Multi-Item Inventory Control for Fluctuating Reorder Intervals,” Interfaces, Vol. 16,
No. 3, pp. 60–67, 1986.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.29

Chapter 16 Cases E.29

TABLE E.47

Order quantity (units)

Cycle length
(months) Item 1 Item 2 Item 3

2.3 10 8 1
2.6 4 6 0
4 1 4 2
2.0 8 6 2
1.2 7 0 2
1.4 0 10 1
1.7 1 2 0
1.3 0 5 2
1.1 9 4 3
1.8 4 6 2
1.6 2 0 0
.5 5 3 1
2.1 10 7 2
2.3 4 12 4
2.4 8 9 3
2.1 10 8 5
2.2 9 13 2
1.8 12 8 4
.7 6 4 2
2.1 5 4 0

Table E.47 provides typical data for three items that were ordered simultaneously.
The data show that both the order quantity and the cycle length are random. Moreover,
a cursory look at the entries of the table reveals the lack of correlation between the
order quantity and the cycle length.
A goodness-of-fit analysis of the complete set of data (see Chapter 12) reveals that
the distribution of the demand rates (order quantity divided by cycle length) for the
three items follows a Weibull distribution, f(r), of the form

2r -r2/a
f1r2 = e ,r Ú 0
a

where r is the demand rate for the item. Similarly, the analysis shows that the distribu-
tion of the reciprocal of the cycle length, s(x), is exponential of the form

s(x) = be-b(x - a), x Ú a

where a is the minimum value assumed by x.

The determination of the optimal order quantity is based on the maximization
of the expected profit per month, which is defined as

e
1
Expected profit = u(q, r, t)f (r)drfg (t)dt
L t L

ex b f(r)dr fs(x)dx
1
= uaq, r,
L L x
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.30

E.30 Appendix E Case Studies

where t and g(t) are the cycle length and its density function. The profit function
u(q, r, t) is based on p, the net unit profit for the item, h, the holding cost per unit
per month, and K, the fixed order cost.
(a) Use the data for the three items to determine the probability density function
for each demand rate.
(b) Use the data for the cycle length to determine s(x).
(c) Develop the mathematical expression for u(q, r, t).
Determine the optimal order quantity for the three items, given the following cost data:
p1 = $100, p2 = $150, p3 = $125, h1 = $2 , h2 = $1.20 , h3 = $1.65 , and K = $30.

CHAPTER 18 CASES
18-1.26 Foote (1976). The Bank of Elkins currently operates a traditional drive-in station
and two “robo” lanes that connect to the inside of the bank through a pneumatic
cartridge. The bank would like to expand the existing facilities so that an arriving car
would complete its business in no more than 4 minutes, on the average. This time limit
was based on psychological studies that show that customers base their impatience
on the movement of the minute hand between two marks, which on most watches
represents five minutes. To collect the necessary data, the team observed the opera-
tion of the existing tellers. After studying the system for a while, a member of the
team noticed that there was a marked difference between the time a customer spent
in the drive-in lane and the time the teller spent carrying out the necessary bank
transactions. In fact, the time a car spent in the system consisted of (1) realizing the
car in front had moved, (2) moving to the teller window, (3) giving the teller instruc-
tions, (4) teller taking action, and (5) moving out. During the first, second, and fifth
components of this time period, the teller was involuntarily idle. Indeed, during each
cycle, the teller was busy serving the customer only 40% of the time. Based on this
information, the team discovered that there was room for reducing the operating
cost of the present system.
What was the team’s suggestion for improving the existing drive-in operation?
Discuss all the implications of the suggestion.
18-2. A state-run child abuse center operates from 9:00 A.M. to 9:00 P.M. daily. Calls report-
ing cases of child abuse arrive in a completely random fashion, as should be expected.
Table E.48 gives the number of calls recorded on an hourly basis over a period of
7 days. The table does not include lost calls resulting from the caller receiving a
busy signal. Each received call lasts randomly for up to 12 minutes with an average
of 7 minutes. Past records show that the center has been experiencing a 15% annual
rate of increase in telephone calls.
The center would like to determine the number of telephone lines that must
be installed to provide adequate service now and in the future. In particular,
special attention is given to reducing the adverse effect of a caller’s receiving a busy
signal.
18-3. A manufacturing company employs three trucks to transport materials among six
departments. Truck users have been demanding that a fourth truck be added to the fleet
to alleviate the problem of excessive delays. The trucks do not have a home station from

26
Based on B. Foote, “A Queuing Case Study in Drive-In Banking,” Interfaces, Vol. 6, No. 4, pp. 31–37, 1976.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.31

Chapter 18 Cases E.31

TABLE E.48

Total number of calls for day

Starting
hour 1 2 3 4 5 6 7

9:00 4 6 8 4 5 3 4
10:00 6 5 5 3 6 4 7
11:00 3 9 6 8 4 7 5
12:00 8 11 10 5 15 12 9
13:00 10 9 8 7 10 16 6
14:00 8 6 10 12 12 11 10
15:00 10 9 12 4 10 6 8
16:00 8 6 9 14 12 10 7
17:00 5 10 10 8 10 10 9
18:00 5 4 6 5 6 7 5
19:00 3 4 6 2 3 4 5
20:00 4 3 6 2 2 3 4
21:00 1 2 2 3 3 5 3

which they can be called. Instead, management considers it more efficient to keep
the trucks in continuous motion about the factory. A department requesting the use of
a truck must await its arrival in the vicinity. If the truck is available, it will respond to the
call. Otherwise, the department must await the appearance of another truck. Table E.49
gives the frequency of the number of calls per hour. The service time for each depart-
ment (in minutes) is approximately the same. Table E.50 summarizes a typical service
time histogram for one of the departments.
Analyze the effectiveness of the present operation.
18-4. A young industrial engineer, Jon Micks, was recently hired by Metalco. The company
owns a 30-machine shop and has hired 6 repairpersons to take care of repairs. The
shop operates for one shift that starts at 8:00 A.M. and ends at 4:00 P.M. Jon’s first
assignment was to study the effectiveness of the repair service in the shop. To that

TABLE E.49

Calls/hr Frequency

0 30
1 90
2 99
3 102
4 120
5 100
6 60
7 47
8 30
9 20
10 12
11 10
12 4
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.32

E.32 Appendix E Case Studies

TABLE E.50

Service time, t Frequency

0 … t 6 10 61
10 … t 6 20 34
20 … t 6 30 15
30 … t 6 40 5
40 … t 6 50 8
50 … t 6 60 4
60 … t 6 70 4
70 … t 6 80 3
80 … t 6 90 2
90 … t 6 100 2

end, he collected the data in Table E.51 from the repair log for three randomly select-
ed machines. Additionally, by checking the repair records for five randomly selected
days, Jon was able to compile the data in Table E.52 representing the number of
broken machines (including those being repaired) at the beginning of every hour
of the work day.
Jon has a meeting with his supervisor, Becky Steele, regarding the data he
has collected. He states that he is confident that the breakdown/repair process
in the shop is totally random and that it is safe to assume that the situation can

TABLE E.51

Machine 5 Machine 18 Machine 23

Failure hour Repair hour Failure hour Repair hour Failure hour Repair hour

8:05 8:15 8:01 8:09 8:45 8:58

10:02 10:14 9:10 9:18 9:55 10:06
10:59 11:09 11:03 11:16 10:58 11:08
12:22 12:35 12:58 13:06 12:21 12:32
14:12 14:22 13:49 13:58 12:59 13:07
15:09 15:21 14:30 14:43 14:32 14:43
15:33 15:42 14:57 15:09 15:09 15:17
15:48 15:59 15:32 15:42 15:50 16:00

TABLE E.52

Total number of broken machines at the hour of

Date 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00

October, 2 6 6 9 6 8 8 7 7
October, 29 9 8 5 9 5 5 6 8
November, 4 6 6 5 7 7 8 6 5
December, 1 9 5 9 7 5 7 5 5
January, 19 6 5 8 5 9 8 8 6
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.33

Chapter 22 Cases E.33

be described as a Poisson queue. Becky confirms that her long experience in

the shop indicates that the situation is indeed totally random. Based on this
observation, she examines Jon’s data, and after making some computations, she
announces to Jon that there is something wrong with the data. How did Becky
reach that conclusion?
18-5 The Yellow Cab Company owns four taxis. The taxi service operates for 10 hours
daily. Calls arrive at the dispatching office according to a Poisson distribution
with a mean of 20 calls per hour. The length of the ride is known to be exponential
with mean 11.5 minutes. Because of the high demand for cabs, Yellow limits the
waiting list at the dispatching office to 16 customers. Once the limit is reached,
future customers are advised to seek service elsewhere because of the expected
long wait.
The company manager, Kyle Yellowstone, is afraid that he may be losing too much
business and thus would like to consider increasing the size of his fleet. Yellowstone
estimates that the average income per ride is about $5. He also estimates that a new
cab can be purchased for $18,000. A new cab is kept in service for 5 years and then
sold for $3500. The annual cost of maintaining and operating a taxi is $20,000 a year.
Can Mr. Yellowstone justify increasing the size of his fleet, and if so, by how many?
For the analysis, assume a 10% annual interest rate.

CHAPTER 22 CASES
22-1.27 The department of industrial engineering at U of A has 3 faculty members and
offers a total 5 courses in a two-semester academic year. The department has 2 graduate
students who can teach courses C1 and C3, but only as a last resort if the regular
faculty cannot teach these classes. A student may not teach more than one course
per semester. Tables E.53 and E.54 specify each professor’s preferences for teaching
certain courses and the number of sections per semester that must be taught of
each course.
Develop a model that can be used to assign faculty (and graduate students, if
necessary) to the designated classes.

TABLE E.53
Number of sections
per semester
Number of sections per
Course academic year Fall Spring

C1 2 1 1
C2 2 1 or 2 1
C3 2 1 or 2 1 or 2
C4 1 1 0
C5 1 0 1

27
Based on J. Dyer and J. Mulvey, “An Integrated Information/Optimization for Academic Planning,”
Management Science, Vol. 22, No. 12, pp. 582–600, 1976.
Z05_TAHA5937_09_SE_APPE.QXD 7/24/10 4:25 AM Page E.34

E.34 Appendix E Case Studies

TABLE E.54

Teaching load
per semester
Teaching load per Order of preference
Professor academic year Fall Spring for courses

P1 1 0 or 1 0 or 1 C1 Ɑ C2 Ɑ C5
P2 3 1 or 2 1 or 2 C1 Ɑ C3 Ɑ C2 Ɑ C4
P3 2 0, 1, or 2 0 or 1 C5 Ɑ C4 Ɑ C3 Ɑ C1

CHAPTER 23 CASE
23-1. A published argument advocates that the recent rise in the mean score of the Scholastic
Aptitude Test (SAT) for high school students in the United States be attributed to
demographic reasons rather than to improvement in teaching methods. Specifically,
the argument states that the decrease in the number of children per family has created
environments in which kids are interacting more frequently with adults (namely,
their parents), which increases their intellectual skills. Conversely, children of large
families are not as “privileged” intellectually because of the immature influence of
their siblings.
What is your opinion regarding the development of a predictive regression
equation for the SAT scores based on this argument?

CHAPTER 24 CASE
24-1. UPPS uses trucks to deliver orders to customers. The company wants to develop
a replacement policy for its fleet over the next 5 years. The annual operating cost of
a new truck is normally distributed with mean $300 and standard deviation $50. The
mean and standard deviation of the operating cost increases by 10% a year thereafter.
The current price of a new truck is $20,000 and is expected to increase by 12% a year.
Because of the extensive use of the truck, there is a chance that it might break down
irreparably at any time. The trade-in value of a truck depends on whether it is broken or
in working order. At the start of year 6, the truck is salvaged, and its salvage value again
depends on its condition (broken or in working order). Table E.55 provides the data of
the situation as a function of the age of the truck.
If the truck is in working condition, its trade-in value after 1 year of operation
is 70% of the purchase price and decreases by 15% a year thereafter. The trade-in
value of the truck is halved if it is broken. The salvage value of the truck at the start
of year 6 is $200 if it is in working condition and $50 if it is broken. Develop the
optimal replacement policy for the truck.

TABLE E.55

Truck age (year) 0 1 2 3 4 5 6

Probability of breakdown .01 .05 .10 .16 .25 .40 .60