24-09-2021

Introduction to Operations Research

Decision Models in Management

SOM 608

Prof. Pankaj Dutta

SJMSOM, IITBombay

2021 1

OR : Research on Operations

Text References:
⚫ Hiller, F. S. and Lieberman, G. J., Introduction to Operations Research,
McGraw-Hill, Ninth Edition, 2010.
⚫ Hiller & Hiller. Introduction to Management Science: A Modeling and Case
Studies Approach with Spreadsheets, 6th edition, 2019.
⚫ Render, B., Stair, R. M, Quantitative Analysis for Management, Prentice
Hall of India Private Limited, Tenth Edition, 2009.
⚫ Taha, H. A., Operation Research: An Introduction, Pearson Education,
Eighth Edition, 2007.
⚫ Sharma, S. D., Operations Research, Kedar Nath Ram Nath & Co,
Fourteenth Edition, 2005.

2021 2

1
 24-09-2021

Origins of Operations Research

(World War II)

The modern field of operations research emerged during World War

II. Scientists in the UK (Patrick Blackett) and in the US (George
Dantzig) looked for ways to make better decisions in such areas as
logistics and training schedules.
After the war it began to be applied to similar problems in industry.

2021 3

What is Operation Research?

⚫ OR is concerned with the application of scientific tools and

techniques to decision-making problems involving operations of
an integrated system so as to provide optimum solutions to a
problem.
⚫ The methods of OR are very often used in Management Science,
Industrial Engineering, Mathematics, Economics, etc. to analyze
complex real-world systems, typically with the goal of improving
or optimizing performance.

⚫ It provides an understanding that gives the managers’ new

insights and capabilities to determine better solutions in their
decision-making problems.

2021 4

2
 24-09-2021

Managerial Approach to Decision Making

Manager analyzes
situation (alternatives)

Makes decision to
resolve conflict

Decisions are
implemented

2021 5

THE MODELING PROCESS

The Role of Managerial Judgment in the Modeling Process:

as applied to the first two stages of decision making.

Analysis
Model Results
Interpretation
Abstraction

Symbolic
World

Real
World

Management Intuition
Decisions
Situation

2021 6

3
 24-09-2021

Five Stages of applying modeling to real world

decision making

Phases of OR Study:

Stage 1: Study the environment to Frame the Managerial

Situation.
Stage 2: Construct a symbolic (quantitative) model.

Stage 3: Analyze the model to generate results.

Stage 4: Interpret and validate model results.

Stage 5: Implement validated knowledge.

2021 7

The general OR model can be depicted as

Maximize/Minimize
Objective Function
Subject to
Constraints

2021 8

4
 24-09-2021

Types of Models:

⚫ Optimizations Models
⚫ Analytic Models
⚫ Decision Analysis Models
⚫ Simulations Models
⚫ Knowledge-Based Models

2021 9

Applications of Operations Research:

Private Sector

Service Industrial
⚫ Portfolio management ⚫ Food and chemical blending
⚫ Location of retail facilities ⚫ Optimal inventory policies
⚫ Insurance and risk ⚫ Capital budgeting
management ⚫ Distribution of products
⚫ Airplane scheduling ⚫ New product introduction
⚫ Advertising media mix ⚫ Assignment of facilities
⚫ Public Transportation ⚫ Queuing analysis of facilities

2021 10

5
 24-09-2021

Applications of Operations Research:

Public Sector
Social Health
⚫ Air-water Pollution control ⚫ Hospital admissions
⚫ Natural resources planning ⚫ Dietary planning
and allocation ⚫ Hospital utilization and
⚫ Emergency response system scheduling
⚫ Political campaign strategies ⚫ Blood inventory policies
⚫ Educational planning and ⚫ Disease control
school bus scheduling Aerospace-Military
⚫ Solid waste disposal ⚫ Search and rescue efforts
⚫ Inventory, distribution,
maintenance of equipment
⚫ Missile defense and
allocation

2021 11

Application: Federal Express (FedEx)

Federal Express (FedEx) is the world’s largest express transportation company. Every working day, it
delivers more than 6.5 million documents, packages, and other items throughout the United States
and more than 220 countries and territories around the world. In some cases, these shipments can be
guaranteed overnight delivery by 10.30 a.m. the next morning.
The logistical challenges involved in providing this service are staggering. These millions of daily
shipments must be individually sorted and routed to the correct general location (Usually by aircraft)
and then delivered to the exact location (usually by motorized vehicle) in an amazingly short period of
time. How is all this possible?
Operations Research (OR) is the technological engine that drives this company. Ever since its
founding in 1973, OR has helped make it major business decisions, including equipment investment,
route structures, scheduling, finances, and location of facilities. After OR has credited with literally
saving the company during its early years, it became the custom to have OR represented at the
weekly senior management meetings and, indeed, several of the senior corporate vice-presidents
have come up from the outstanding FedEx OR group.
FedEx has come to be acknowledged as a world class company. It routinely ranks among the top
companies on fortune magazine’s annual listing of the “Worlds Most Admired Companies.” It also was
the first winner (1991) of the prestigious prize now known as the INFORMS Price, which is awarded
annually for the effective and repeated integration of OR in to organizational decision making in
pioneering, varied novel, and lasting ways.
Source: R.O. Mason et al., “Absolutely, Positively Operations Research: The Federal Express Story,”
Interfaces, 27(2): 17-36, March-April 1997

2021 12

6
 24-09-2021

2021 13

Decision Models in Management

Introduction to Linear Programming

Sessions 1-2

2021 14

7
 24-09-2021

Distributing Goods through a Distribution Network

⚫ The Problem. The Distribution Unlimited Co. will be producing the same new
product at two different factories, and then the product must be shipped to two
warehouses, where either factory can supply either warehouse. The distribution
network available for shipping this product is shown in DN Fig. , where F1 and F2
are the two factories, W1 and W2 are the two warehouses, DC is a distribution
center. The amounts to be shipped from F1 and F2 are shown to their left, and the
amounts to be received at W1and W2 are shown to their right. Each arrow
represents a feasible shipping lane. Thus, F1can ship directly to W1 and has three
possible routes (F1->DC-> W2, F1->F2->DC->W2, and F1->W1->W2) for
shipping toW2. Factory F2 has just one route to W2 (F2->DC->W2) and one to W1
(F2->DC->W2->W1).The cost per unit shipped through each shipping lane is
shown next to the arrow. Also shown next to F1->F2 and DC->W2 are the
maximum amounts that can be shipped through these lanes. The other lanes have
sufficient shipping capacity to handle everything these factories can send.
The decision to be made concerns how much to ship through each
shipping lane. The objective is to minimize the total shipping cost.

2021 15

Distributing Goods through a Distribution Network

2021 16

8
 24-09-2021

Distributing Goods through a Distribution Network

Minimize Z = 2XF1-F2 + 4XF1-DC + 9XF1-W1 + 3XF2-DC +XDC-W2 +3XW1-W2

+2XW2-W1
Subject to;
⚫ XF1-W1 + XF1-DC + XF1-F2 =50

⚫ XF1-F2 + 40 = XF2-DC
⚫ XF1-DC + XF2-DC = XDC-W2
⚫ XF1-W1 + XW2-W1 = 30 + XW1-W2
⚫ XDC-W2 + XW1-W2 = 60 + XW2-W1
⚫ Upper bound constraints
⚫ XF1-F2 <= 10, XDC-W2 <= 80
⚫ Non-negativity constraints

2021 17

Distributing Goods through a Distribution Network

Minimize Z = 2XF1-F2 + 4XF1-DC + 9XF1-W1 + 3XF2-DC +XDC-W2 +3XW1-W2

+2XW2-W1
Subject to;
⚫ 5-Net flow constrains
⚫ Amount shipped out - amount shipped in = required amount
⚫ Upper bound constraints
⚫ XF1-F2 <= 10, XDC-W2 <= 80
⚫ Non-negativity constraints

⚫ Answer: 0, 40, 10, 40, 80, 0, 20, $49000.

2021 18

9
 24-09-2021

Acme Manufacturing Company

Acme Manufacturing Company has contracted to deliver home windows over

the next 6 months. The demands for each month are 100, 250, 190, 140, 220,
and 110 units, respectively. Production cost per window varies from month to
month depending on the cost of labor, material, and utilities. Acme estimates
the production cost per window over the next 6 months to be $50, $45, $55,
$48, $52, and $50, respectively.
To take advantage of the fluctuations in manufacturing cost, Acme may elect
to produce more than is needed in a given month and hold the excess units for
delivery in later months. This, however, will incur storage costs at the rate of
$8 per window per month assessed on end-of-month inventory.
Develop an optimum production schedule.

2021 19

LP model includes three basic elements

⚫ Decision variables that we seek to determine

⚫ Objective that we aim to optimize as a function of

the decision variables

⚫ Constraints that we need to satisfy (restrictions on

the values that can be assigned to the decision
variables)

2021 20

10
 24-09-2021

Characteristics of LP Problems

⚫ A well-defined single objective must be stated.

⚫ There must be alternative courses of action.

⚫ The total achievement of the objective must be
constrained by scarce resources or other restraints.

⚫ The objective and each of the constraints must be

expressed as linear mathematical functions.

2021 21

Assumptions of LPP

Proportionality: The contribution of each activity to the value of

objective function and to the constraint is proportional to the
level of the activity xj.
This assumption rules out any exponent other than 1 for any variable in
any term of any function
Additivity: Every function in a LP model is the sum of the individual
contributions of the respective activities.
This assumption rules out the possibility of cross-product term in any
function.
Divisibility: Decision variables in a LP model are not restricted to just
integer values. It is assumed that the activities can be run at
fractional levels.
Certainty: The value assigned to each parameter (cj, aij, bi) of a LP
model is assumed to be a known constant.

2021 22

11
 24-09-2021

Steps in Formulating LP Problems

1. Define the decision variables.

2. Define the objective. (min or max)
3. Write the mathematical function for the objective.
4. Identify the number of constraints.
5. Express the constraints in mathematical form.
6. Write <, =, or > for each constraint.
7. Impose the non-negativity criterion for decision variables.

2021 23

The general LPP can be described as

Maximize
Z = CX
Subject to
AX ≤ B,
X ≥ 0.
Minimize
or, Z = CX
Subject to
AX ≥ B,
X ≥ 0.

2021 24

12
 24-09-2021

Common Terminology :

⚫ Resources : Production capacities of plants (m

plants)
⚫ Activities : Production of products (n products)
Z - value of overall measure of performance
Xj - level of activity j (j = 1,2, … , n)
cj - increase in Z that would result from each unit increase in
level of activity j
bi - amount of resource i that is available for allocation to
activities
aij - amount of resource i consumed by each unit of activity j

2021 25

Standard form of LPP

Let there be m resources to be allocated to n activities.
The problem is to find the values of x1, x2, …, xn so as to

Maximize Z = c1x1+ c2x2+ …+ cnxn

subject to
a11x1+ a12x2+ …+ a1nxn ≤ b1
a21x1+ a22x2+ …+ a2nxn ≤ b2
.
.
am1x1+ am2x2+ …+ amnxn ≤ bm

where all x1, x2, …, xn are non-negative.

LPP is a powerful technique for dealing with the problems of allocating

limited resources among competing activities as well as other problems
having a similar mathematical formulation.

2021 26

13
 24-09-2021

Example 1.

A firm manufactures two types of products A and B and sells

them at a profit of Rs. 2 on type A and Rs. 3 on type B.
Each product is processed on two machines G and H. Type A
requires one minute of processing time on G and two minutes
on H; type B requires one minute on G and one minute on H.
The machine G is available for not more than 6 hours 40
minutes while machine H is available for 10 hours during any
working day.

Formulate the problem as a LPP.

2021 27

Example 1.

Machine Product A Product B Capacity of the

(X1) (X2) machines in
minutes
G: 1 1 <= 400

H: 2 1 <= 600

2021 28

14
 24-09-2021

Continued…

Let, x1 = Number of products of type A.

x2 = Number of products of type B.
Maximize Z = 2x1 + 3x2
Subject to
x1 + x2 < 400
2x1 + x2 < 600
and x 1 , x2 > 0

2021 29

Example 2.
A glass company produces high-quality glass products, including
windows and glass doors. It has three plants. Aluminum frames and
hardware are made in Plant 1, wood frames are made in Plant 2 and
Plant 3 produces the glass and assembles the products. Two new
products having large sales potential are launched.

• Product 1: An 8-foot glass door with aluminum framing .

• Product 2: A 4 x 6 foot wood-framed window.

2021 30

15
 24-09-2021

Continued…

Product 1 requires some of the production

capacity in Plants 1 and 3, but none in Plant 2.
Product 2 needs only Plants 2 and 3.

As both the products would be competing

for the same production capacity in Plant 3,
it is not clear which mix of the two products
would be most profitable. The following
data were gathered to study this question.

2021 31

Continued…

Let, x1 = no. of batches of product 1 produced per week

x2 = no. of batches of product 2 produced per week
Z = total profit per week from producing these two
products
Maximize z = 3x1 + 5x2
Subject to
x1 < 4
2x2 < 12
3x1 + 2x2 < 18
x 1, x2 > 0

2021 32

16
 24-09-2021

Example 3.
Evening shift resident doctors in a Govt. hospital work five consecutive
days and have two consecutive days off. Their five days of work can
start on any day of the week and the schedule rotates indefinitely. The
hospital requires the following minimum number of doctors working:

Sun. Mon. Tue. Wed. Thu. Fri. Sat.

35 55 60 50 60 50 45

No more than 40 doctors can start their five working days on the same
day. Formulate the general LP model to minimize the number of
doctors employed by the hospital.

2021 33

Continued…

Let xj (j = 1 to 7) be the number of doctors who start their duty

on jth day of the week.
Min. z = x1+ x2 + x3 + x4 + x5 + x6 + x7
s.t. x1 + x4 + x5 + x6 + x7 >= 35
x 1+ x 2 + x5 + x6 + x7 >= 55
x1+ x2 + x3 + x6 + x7 >= 60
x1+ x2 + x3 + x4 + x7 >= 50
x1+ x2 + x3 + x4 + x5 >= 60
x 2 + x3 + x4 + x 5 + x6 >= 50
x3 + x4 + x5 + x6 + x7 >= 45
x1, x2 , x3 , x4 , x5 , x6 , x7 >=0

2021 34

17
 24-09-2021

Example 4. Advanta India Ltd.

The First Indian Multinational Seed Company with global footprint. Advanta
has leadership position in crops like Sunflower, Rice, Corn, Mustard, Cotton,
Vegetables etc. The Company has been established with the prime objective of
conducting research, development, production and marketing of superior high
yielding hybrid seeds of crops of national importance.
Advanta India Ltd. has 100 acre farm in its Sonepat division. In 1996, they
decided to produce vegetables, namely, tomatoes, lettuce & radishes. They can
sell all tomatoes, lettuce, or radishes they can raise. The price they can obtain
is Re. 1.00 per kg for tomatoes, Rs. 0.75 a head for lettuce, and Rs. 2.00 per kg
for radishes. The average yield per-acre is 2,000 kg of tomatoes, 3000 heads of
lettuces, and 1000 kgs of radishes. Fetilizer is available at Re. 0.50 per kg and
the amount required per acre is 100 kgs each for tomatoes and lettuce, 50 kgs
for radishes. Labour required for sowing, cultivating and harvesting per acre is
5 man-days for tomatoes and radishes, and 6 man-days for lettuce. A total of
400 man-days of labor are available at Rs. 20 per man-day.
Formulate this problem as a LP model to maximize the farm’s total profit.
2021 35

Continued…

Let the farmer decide x1, x2, x3 acre of his land to grow
tomatoes, lettuce and radishes respectively.

Maximize z = 1850x1 + 2080x2 + 1875x3

Subject to
x1 + x2 + x3 < 100
5x1 + 6x2 + 5x3 < 400
and x 1 , x 2 , x3 > 0

2021 36

18
 24-09-2021

Example 5.

⚫ A businessman has the option of investing his money in

two plans. Plan A guarantees that each rupee invested
will earn seventy paise a year hence, while plan B
guarantees that each rupee invested will earn two
rupees two year hence. In plan B, only investments for
periods that are multiples of two years are allowed.
The problem is how should he invest Rs. 10000 in
order to maximize the earnings at the end of three
years.
Formulate the problem as a LPP.

2021 37

Continued..

Let, a1 and b1 be the initial investment in 1st year in plan A and B ,

respectively. (1)
Return after one year = 1.7 a1.
Now, you can invest this return into a2 and b2 parts for plan A & B, resp.
(2)
Return after two year = 1.7a2 +3b1.

Invest this amount in year three in plan A only. (3)

Return after third year = 1.7 (1.7a2 +3b1 )+ 3 b2

Maximize total return: 1.7 (1.7a2 +3b1 )+ 3 b2

s.t. constraints (1)-(3)

and the non-negativity restriction.

2021 38

19
 24-09-2021

Case Study
Acme Manufacturing Company has contracted to deliver home windows over
the next 6 months. The demands for each month are 100, 250, 190, 140, 220,
and 110 units, respectively. Production cost per window varies from month to
month depending on the cost of labor, material, and utilities. Acme estimates
the production cost per window over the next 6 months to be $50, $45, $55,
$48, $52, and $50, respectively. To take advantage of the fluctuations in
manufacturing cost, Acme may elect to produce more than is needed in a given
month and hold the excess units for delivery in later months. This, however,
will incur storage costs at the rate of $8 per window per month assessed on
end-of-month inventory. Develop a linear program to determine the optimum
production schedule.

2021 39

Continued…

Decision variables
Let for i = 1 to 6
Xi = Number of units produced in month i
Ii = Inventory units left at the end of month i

x1 x2 x3 x4 x5 x6

I=0 I1 I2 I3 I4 I5 I6

100 250 190 140 220 110

2021 40

20
 24-09-2021

Continued..

Objective function:
Min (Total Prod. Cost + end-of-month inventory cost)

Constraints:
Beginning Inv.+ Prod. Amount - Ending Inv. = Demand

2021 41

Continued…

Min. z = 50x1+ 45x2 + 55x3 + 48x4 + 52x5 + 50x6

+ 8(I1+ I2 + I3 + I4 + I5 + I6)
Subject to x1 – I1 = 100 (Month 1)
I1 + x2 – I2 = 250 (Month 2)
I2 + x3 – I3 = 190 (Month 3)
I3 + x4 – I4 = 140 (Month 4)
I4 + x5 – I5 = 220 (Month 5)
I5 + x6 – I6 = 110 (Month 6)
xi, Ii > 0, for all i = 1,2,…, 6.

2021 42

21
 24-09-2021

Graphical LP Solution

2021 43

Graphical LP Solution

⚫ Determination of the feasible solution space

⚫ Determination of the optimal solution from among all the
feasible points in the solution space

2021 44

22
 24-09-2021

Graphical LP Solution

⚫ In this method, we consider the inequations of the constraints

as equations and draw the lines corresponding to these
equations in a two dimensional plane and use the non-
negativity restrictions.
⚫ These lines define the region of permissible values of the
variables.
⚫ Then by trail and error method we find a point in this feasible
region whose co-ordinates will give the optimal value (max or
min).

2021 45

Graphical LP Solution

⚫ Feasible solution: A set of values of the decision variables

x1, x2,…,xn which satisfies all the constraints and the
non-negativity restrictions, is called a feasible solution.
⚫ Feasible region: The collection of all feasible solutions.
⚫ Infeasible solution: A solution for which at least one constraint
is violated.
⚫ Optimal solution: A feasible solution which in addition
optimizes the objective function is called the optimal
solution.

2021 46

23
 24-09-2021

Graphical LP Solution

⚫ This point will be a corner (or extreme) point of the

feasible region.
⚫ Therefore, either the extreme point need to be considered as
candidates for the optimal solution
or
the point can be found by translating the straight line given by
the objective function for some particular value of “z”,
through this region.

2021 47

Corner-point feasible solution: A solution that lies

at a corner of the feasible region

⚫ Constraints that are expressed as linear inequalities

necessarily form a convex set.
⚫ Given a linear objective function, regardless of its
inclination or direction of optimization, the optimal
solution will always include a corner point of a linear
convex set.
⚫ Even if the feasible region is convex, the existence of a
curvilinear constraint allows for the possibility of a non-
corner point optimal solution

2021 48

24
 24-09-2021

Convex Set:

A feasible set is convex if the line segment between

every pair of feasible points falls entirely within the
feasible region.
That is, if x1, x2 be any two points of a convex set X, then the
convex combination
x = α x2 + (1- α) x1; 0 ≤ α ≤ 1
must also be in the set.

2021 49

Example 1. Graphical Solution: Revisit to Example 2

Maximize z = 3x1 + 5x2

Subject to
x1 < 4 I
2x2 < 12 II
3x1 + 2x2 < 18 III
x 1, x2 > 0

2021 50

25
 24-09-2021

Graphical Solution

x2 I
A(0,6) B (2,6), Z* = 36
II

z
Feasible C(4,3)
Region

O(0,0) D(4,0) x1
2021
III 51

Example 2.

Consider the LPP:

Minimize z = 2 x1 + x2
Subject to the constraints
3 x1 + x2  9
x1 + x2  6
x1 , x2  0

2021 52

26
 24-09-2021

(0.9)
Feasible region

(0,6)

(0,0) (3,0) (6,0) 3 9

( , )
z minimum at 2 2
2021 53

Example 3. ( Infinite solution /

Alternative optimal solution)

Consider the LPP:

Minimize Z = x1 + x2
Subject to the constraints
5 x1 + 9 x2  45
x1 + x2  2
x2  4
x1 , x2  0
zmin = 2

2021 54

27
 24-09-2021

Example 4. ( Unbounded Solution)

Consider the LPP:

Maximize Z = 3x1 + 4 x2
Subject to the constraints
x1 − x2  0
− x1 + 3 x2  3
x1 , x2  0

2021 55

Example 5. ( No Feasible Solution)

Consider the LPP:

Maximize Z = 2 x1 − 3x2
Subject to the constraints
x1 + x2  2
2 x1 + 2 x2  8
x1 , x2  0

2021 56

28
 24-09-2021

Example 6. ( Unbounded Region, but optimal

solution exist)

Consider the LPP:

Maximize Z = 2 x1 − x2
Subject to the constraints
x1 − x2  1
x1  3
x1 , x2  0 Z max (3,2) = 4

2021 57

Duality and Sensitivity Analysis

2021 58

29
 24-09-2021

Duality in Linear Programming

⚫ LP Problem consists of two forms

- Primal
- Dual
⚫ Associated with every LPP (Primal Problem), there
is a LPP which is called its Dual Problem.
- One being a maximization problem and the
other a minimization problem.

2021 59

Primal to Dual conversion

Primal Problem

Maximize
Z = CX
Subject to Dual Problem
AX ≤ b,
X ≥ 0. Minimize
W = b'Y
Subject to
A'Y ≥ c',
Y ≥ 0.

2021 60

30
 24-09-2021

Primal to Dual conversion

⚫ Number of variables in the dual is equal to the

number of constraints in the primal and vice-versa.
⚫ The coefficients in the objective function of the
primal problem are the right hand sides of the
functional constraints in the dual problem.

⚫ The right hand sides of the functional constraints in

the primal problem are the coefficients in the
objective function of the dual problem.

2021 61

⚫ The coefficients of a variable in the functional

constraints of the primal problem are the
coefficients in a functional constraint of the dual
problem.
⚫ The less than signs of the primal constraints
become greater than signs in the dual constraints.
⚫ The variables in both the problems are non-
negative.

2021 62

31
 24-09-2021

Primal to Dual conversion

n
Maximize Z = c xj =1
j j

subject to
n

a
j =1
ij x j  bi for i = 1, 2...., m

and x j  0 for j = 1, 2...., n. m

Minimize W = b y
i =1
i i

Dual
Objective subject to
W
m

a
i =1
ij yi  c j for j = 1, 2...., n

Primal
Objective Z
and yi  0 for i = 1, 2...., m.

2021 63

Example 1. Wyndor Glass Co. example

Maximize z = 3x1 + 5x2

Subject to
x1 < 4 I
2x2 < 12 II
3x1 + 2x2 < 18 III

x 1, x2 > 0

2021 64

32
 24-09-2021

Example 1. Dual Problem

Minimize W = 4y1 + 12y2 +18y3

Subject to
y1 +3y3 > 3 I
2y2 +2y3 > 5 II

y1, y2 , y3 > 0

2021 65

Example 2.

Maximize z = 1850x1 + 2080x2 + 1875x3

Subject to
x1 + x2 + x3 < 100
5x1 + 6x2 + 5x3 < 400
and x 1 , x 2 , x3 > 0

2021 66

33
 24-09-2021

Why Duality is important in practice?

⚫ Dual of the dual is primal.

⚫ If x be any feasible solution of the primal and y be
any feasible solution of the dual, then
cx ≤ b'y.
⚫ If x* is an optimal solution for the primal problem
and y* is an optimal solution for the dual problem,
then Dual
Objective
W
cx* = b'y*.

Primal
Objective Z

2021 67

Primal Problem Dual Problem Conclusion

⚫ FS FS Finite optimal for both exists.

⚫ NFS FS Dual objective function is

unbounded.

⚫ FS NFS Primal objective function is

unbounded.

⚫ NFS NFS No solution exists

2021 68

34
 24-09-2021

Economic Interpretation:

⚫ Dual variables: the dual variable yi represents the worth per

unit (shadow price) of resource i.
We know that Z < W Dual
Objective

(Revenue) < (Worth of resources) W

n m
z=  c j x j   bi yi = w
j =1 i =1
Primal
Objective Z

i.e.; total revenue from all the activities is less than the worth of
the resources. Optimality occurs when the two quantities are
equal.

2021 69

⚫ Dual Constraints:
m
We know that a
i =1
ij yi  c j

m (usage of resource i per unit of activity j )

 (worth per unit of
i =1 resource i )
 revenue per unit of activity j

i.e.; the assigned cost of all the resources is needed to produce

one unit of activity j.

2021 70

35
 24-09-2021

Post Optimality and Sensitivity

Analysis

2021 71

⚫ By sensitiveness we mean fulfillment of the

condition of optimality as well as
determining the limits of variations of these
parameters for the solution to remain optimal
feasible.

2021 72

36
 24-09-2021

Wyndor Glass Co.

Product 1 requires some of the production

capacity in Plants 1 and 3, but none in Plant 2.
Product 2 needs only Plants 2 and 3.

As both the products would be competing

for the same production capacity in Plant 3,
it is not clear which mix of the two products
would be most profitable. The following
data were gathered to study this question.

2021 73

Wyndor Glass Co.

Let, x1 = no. of batches of product 1 produced per week

x2 = no. of batches of product 2 produced per week
Z = total profit per week from producing these two products

Maximize z = 3x1 + 5x2

Subject to
x1 < 4
2x2 < 12
3x1 + 2x2 < 18
x 1, x2 > 0

2021 74

37
 24-09-2021

Graphical Solution
x2 I
A(0,6) B (2,6), Z* = 36
II

Feasible C(4,3) z=36

Region
O(0,0) D(4,0) x1
2021
III 75

Shadow Price

⚫ The shadow Price for resource i measures the

marginal value of this resource (i.e.) the rate
at which Z could be increased by slightly
increasing the amount of this resource.

2021 76

38
 24-09-2021

Solution by Simplex Method:

Optimal Table

Basic Coefficient of:

Iteration Variable Eq. Z x1 x2 x3 x4 x5 RHS Ratio
Z (0) 1 0 0 0 3/2 1 36 ---
x3 (1) 0 0 0 1 1/3 1/3 2 ---
2
x2 (2) 0 0 1 0 1/2 0 6 ---
x1 (3) 0 1 0 0 -1/3 1/3 2 ---

2021 77

Shadow Price (We can use Solver)

⚫ Shadow price for resource 1:y1 = 0

⚫ Shadow price for resource 2: y2 =3/2
⚫ Shadow price for resource 3: y3 = 1

 b2 and b3 are sensitive parameters

2021 78

39
 24-09-2021

Objective Function Coefficient

Z= C1x1+ 5x2

⚫ The optimal solution (x1*, x2*) remains

the same as long as the slope of this line
lies between the slopes of the binding
constraints. This gives a range for C1.

2021 79

Graphical Solution
x2 I
A(0,6) B (2,6), Z* = 36
II

Feasible C(4,3) z=36

Region
O(0,0) D(4,0) x1
2021
III 80

40
 24-09-2021

⚫Z= C1x1+ 5x2

Or x2 = (-C1/5)x1 + Z/5 (y=mx+c)

For line, 2x2 < 12; m = 0

For line, 3x1 + 2x2 < 18; m = (-3/2)

⚫ 0<=C1<=7.5

2021 81

⚫ Initially, c1=3, c2 = 5, Z = 36.

⚫ With c2 = 5, the allowable range for c1 is
0<= c1 <= 7.5.
⚫ With c1 = 3, the allowable range for c2 is
c2 >= 2.

2021 82

41
 24-09-2021

Constraint Coefficient

⚫ If a constraint is not a binding constraint any small

change in its aij coefficient does not change the
optimal solution and hence is not sensitive.
⚫ As the aij values are determined by the technology
being used, the uncertainty in these values is
relatively less.

2021 83

42
 9/29/2021

Integer Programming
SOM608

Prof. Pankaj Dutta

SJMSOM, IIT Bombay

Application Areas…
◼ Investment Analysis
◼ Site Selection
◼ Designing Production and Distribution Network
◼ Dispatching Shipments
◼ Scheduling Activities
◼ Portfolio Asset Allocation
◼ Airline Applications…

1
 9/29/2021

Integer Programming
◼ Types of ILP
 Examples
◼ Solutions of an ILP
 Excel Solver
 Graphical concepts
 Branch and Bound Algorithm
 Idea of Cutting Plane Method
◼ Application Areas:
◼ Binary ILP
 Project/Portfolio Selection
 Either-Or Constraints
◼ Fixed Charge Problem as IP
◼ Facility Location Problem as IP
◼ Assignment Problem as IP
3

◼ Linear Programming (LP):

The objective and constraint functions are linear and the
decision variables are continuous.
◼ Integer Linear Programming (ILP):
One or more of the decision variables are restricted to
integer values only and the functions are linear.

Divisibility Assumption of LP:

Recall that in LP the variables are continuous, i.e., they
can assume fractional values. However, in IP, (at least
some of) the variables can assume ONLY discrete or
integer values.

2
 9/29/2021

Integer Programming

◼ Pure Integer Programming

- all decision variables are integer.
◼ Mixed Integer Programming
- some decision variables are integer, others are
continuous.
◼ Binary Integer Programming
– i.e., all the variable values must have integer solution of
0 and 1.

Example -1. Consider the following Integer program (IP) and

its LP relaxation (i.e., an LP obtained by dropping integer
restriction on the variables)

◼ IP
Max Z = 21x1+ 11x2
s.t.; 7x1+ 4x2 < 13
x1, x2 > 0 and integers

◼ LP Relaxation of IP
[integrality of variables dropped]
Max Z = 21x1+ 11x2
s.t.; 7x1+ 4x2 < 13
x 1, x 2 > 0
6

3
 9/29/2021

Example 1 (Continued…)
x2

3
◼ As an LPP, x1= 13/7 , x2=0 Feasible Points

2
◼ Rounding off
 x1=2 , x2=0 → infeasible 1

 x1= 1, x2= 0 →not optimal

(0,0) 1 2 x1

◼ The feasible region is

◼ S = { (0,0), (1,0), (0,1), (1,1), (0,2), (0,3) }
◼ This is not a convex set.
◼ The optimal solution is Z= 33, x1=0, x2=3
7

Attempt to get IP Solution by Rounding LP solution

◼ Rounding off is one way to reach the integer solution,

but it often does not yield the best solution.
◼ An IP solution can never be better than the same LP
solution.
◼ The integer solution is worse in terms of higher profit
or lower cost.

4
 9/29/2021

Example 2. x2

Max Z = 4x1+ x2 5

s.t. 4

2x1+ x2 < 5
3 2x1 +x2 < 5

2x1+ 3x2 = 5
2
x1, x2 > 0; 2x1 +3x2 = 5

1
x1, x2: integers
(0,0) 1 2 3 4 5 x1

Example – 2(Continued…)
x2

3 2x1 +x2 < 5

2
2x1 +3x2 = 5

(0,0) 1 2 3 4 5 x1

10

5
 9/29/2021

Example – 2 (Continued…)
◼ As an LPP, the solution is Z= 10, x1=5/2, x2=0
◼ Rounding off,
 x1= 2, x2= 0
 x1= 3, x2= 0

◼ Neither candidate is a feasible solution to the

given problem.
◼ The optimal solution is x1=1, x2= 1.

11

◼ The LP obtained by omitting all integers or 0-1

constraints on the variables is called the
LP relaxation of the IP.
◼ The feasible region for any IP is contained in
the feasible region for the corresponding LP
relaxation.
◼ With n variables in a BIP problem, there are 2n
solutions to be considered. Each time n is
increased by 1, the number of solutions is
doubled.

12

6
 9/29/2021

◼ Therefore, the difficulty of the problem grows

exponentially.
◼ With n = 10, there are 1024 solutions
◼ With n= 20, there are 10,00,000 solutions
◼ The two primary determinants of
computational difficulty for an IP problem are
1)The number of integer variables and
2)Any special structure in the problem

13

Example 3.
Evening shift resident doctors in a Govt. hospital work five consecutive
days and have two consecutive days off. Their five days of work can
start on any day of the week and the schedule rotates indefinitely. The
hospital requires the following minimum number of doctors working:

Sun. Mon. Tue. Wed. Thu. Fri. Sat.

35 55 60 50 60 50 45

No more than 40 doctors can start their five working days on the same
day. Formulate the general ILP model to minimize the number of
doctors employed by the hospital.

14

7
 9/29/2021

Continued…

Let xj (j = 1 to 7) be the number of doctors who start their duty

on jth day of the week.
Min. z = x1+ x2 + x3 + x4 + x5 + x6 + x7
s.t. x1 + x4 + x5 + x6 + x7 >= 35
x 1+ x 2 + x5 + x6 + x7 >= 55
x 1+ x 2 + x3 + x6 + x7 >= 60
x 1+ x 2 + x 3 + x 4 + x7 >= 50
x 1+ x 2 + x3 + x4 + x5 >= 60
x 2 + x3 + x4 + x 5 + x 6 >= 50
x3 + x4 + x5 + x6 + x7 >= 45
x1, x2 , x3 , x4 , x5 , x6 , x7 >=0 and Integer

15

Branch and Bound Method

◼ The B&B algorithm can be used to solve ILP problems (Pure,
MIP or Binary ILP)
◼ Developed in 1960 by A Land and G Doig
◼ B & B is based on the three key ideas:
 Branching: Divide and Conquer
 Bound (Optimal ZIP* lies between LB and UB)
 Fathoming or Pruning

16

8
 9/29/2021

Branch and Bound Method

◼ Assuming a maximization problem. Set an initial
lower bound z = -  on the optimum objective value
of ILP. Set i=0.
◼ Step-1: (Fathoming/bounding): Select LPi, the next
sub problem to be examined. Solve LPi, and attempt
to fathom it using one of three conditions.
(a) The optimal z-value of LPi cannot yield a better objective
value than the current lower bound.
(b) LPi yields a better feasible integer solution than the current
lower bound.
(c) LPi has no feasible solution.

17

Branch and Bound Method

◼ Two cases will arise
(a) If LPi is fathomed and a better solution is found,
update lower bound. If all sub problems have been
fathomed, stop; the optimum ILP is associated
with the current finite lower bound. If no finite
lower bound exists, the problem has no feasible
solution. Else, set i= i+1 and repeat step-1.
(b) If LPi is not fathomed, go to step 2 for branching.

18

9
 9/29/2021

Branch and Bound Method

◼ Step 2: Branching: Select one of the integer variables
xj, whose optimum value xj* in the LPi solution is not
integer. Eliminate the region
[xj*] < xj < [xj*] + 1
(where [v] defines the largest integer < v) by creating
two LP sub problems that corresponds to
xj < [xj*] and xj > [xj*] + 1
Set i = i + 1 and go to step 1.

19

Example 4.
◼ Max Z = 5x1+4x2
St
x1 + x2<5
10x1+ 6x2<45
X2<=1
x1, x2>0
x1, x2 :

20

10
 9/29/2021

Example 4. 1
( B & B Tree)
LP1
x1=3.75, x2 =1.25, z = 23.75

7 2
LP3 LP2
x1=3, x2 =2, z = 23 x1=4, x2 =0.83, z = 23.33
Lower bound (optimum)

4 3
LP4 LP5
x1=4.5, x2 =0, z = 22.5 No feasible solution

6 5
LP6 LP7
x1=4, x2 =0, z = 20 No feasible solution
Lower bound 21

Example 4. (Consider branching x1)

◼ (LP1) Max Z = 5x1+4x2 x =3.75, x =1.25,

LP1
z = 23.75
1 2

St
x1 + x2<5 LP2: x2<1; LP3: x2>2
10x1+ 6x2<45
x1, x2>0

(3.75,1.25)

x1<3 x1>4
22

11
 9/29/2021

Example 4.

LP1
◼ (LP1) Max Z = 5x1+4x2 x =3.75, x =1.25, z = 23.75
1 2

St
x1 + x2<5 LP2: x1<3; LP3: x1>4
10x1+ 6x2<45
x1, x2>0

(3.75,1.25)

x1<3 x1>4
23

Example 4. (Solve the below two new sub-problems as simple LP and

check their solution, if it is integer or not!. Remember, we need to check
the values of Xi’s. If integer solution achieved, then compare the new Z
with previous Z)

◼ LP2 ◼ LP3
◼ Max Z = 5x1+4x2 ◼ Max Z = 5x1+4x2
St St
x1 + x2<5 x1 + x2<5
10x1+ 6x2<45 10x1+ 6x2<45
x1<3 x1>4
x1, x2>0; x1, x2: integers x1, x2>0; x1, x2: integers

24

12
 9/29/2021

Example 4. 1
( B & B Tree)
LP1
x1=3.75, x2 =1.25, z = 23.75

7 2
LP2 LP3
x1=3, x2 =2, z = 23 x1=4, x2 =0.83, z = 23.33
Lower bound (optimum)

4 3
LP4 LP5
x1=4.5, x2 =0, z = 22.5 No feasible solution

6 5
LP6 LP7
x1=4, x2 =0, z = 20 No feasible solution
Lower bound 25

Example 5 & 6. Solve graphically

◼ Max Z = x1+x2 ◼ Min Z = 5x1+4x2
St St
2x1 + 5x2< 16 3x1 + 2x2> 5
6x1+ 5x2<30 2x1 + 3x2> 7
x1, x2>0 x1, x2>0
x1, x2: integers x1, x2: integers

26

13
 9/29/2021

Binary Integer Programming

◼ The LP relaxation problem is obtained by
replacing xj= 0 or 1
by xj < 1 and xj >0
◼ In Formulation
Many Managerial problems are inherently integer
programming problems. Others are not necessarily
integer but possess special characteristics which can
be modeled accurately only through the use of
integer variables.

27

Illustration of a Portfolio Model

Available Funds = $ 300,000

Alternative funding levels for Expected Profits ($M)
each Project Project A Project B Project C
$ 0 $ 0 $ 0 $ 0
100,000 100 120 10
200,000 250 y4=(1,0) 285 215
300,000 310 335 350
Optimum Portfolio Expected Profits
Project A $100,000 $ 100 M
Project B 200,000 285 M
$300,000 $ 385 M

14
 9/29/2021

Project Selection
◼ Max Z= 100y1+120y2+10y3 + 250y4+285y5+215y6

+310y7+335y8+350y9
◼ Subject to

◼ 1 (y1+y2+y3)+2(y4+y5+y6)+3(y7+y8+y9)<=3 (in lakh)

◼ y1+y4+y7 <=1 (for project A)
◼ y2+ y5+y8 <=1 (for project B)
◼ y3+y6+y9 <=1 (for project C)
◼ sum (yi)<=3; i=1 to 9
◼ yi=0 or 1; i=1 to 9

◼ Answer: y1=1; y5=1

29

Example 7.
◼ John is considering four investments. Investment 1 will yield a
net present value (NPV) of 16 lakhs; investment 2, an NPV of
22 lakhs; investment 3, an NPV of 12 lakhs and investment 4,
an NPV of 8 lakhs. Each investment requires a certain cash
outflow at the present time: investment 1 : 5 lakhs, investment
2 : 7 lakhs, investment 3 : 4 lakhs and investment 4 : 3 lakhs.
At present, 14 lakhs is available for investment. Formulate this
problem to maximize the NPV obtained from these
investments.
◼ If investment 2 is made, investment 1 also should be made.
◼ If investment 2 is made, investment 4 cannot be made.

30

15
 9/29/2021

All xi are binary (0,1)

◼ Max Z =sum (npv*Xi)
◼ Subject to
◼ 5Xi+…+3X4<=14
◼ If investment 2 is made, investment 1 also should be made.
 If x2=1, then x1=1
 But, If x2=0, then x1 can be 0 or 1
◼ X2 <=X1
◼ If investment 2 is made, investment 4 cannot be made.
◼ x2+x4<=1

31

The Fixed Charge Problem

◼ e.g. A set up cost is incurred in initiating the
production run. In such cases, the total cost of
the activity (say, j ) can be represented as
fj(xj) = kj+ cjxj ; xj> 0
0 ; xj= 0
where xj denotes the level of activity j,
kj denotes the set up cost and
cj denotes the cost for each incremental unit.

32

16
 9/29/2021

The Fixed Charge Problem

Min Z= ∑ (cjxj + kjyj) ;

y = 1 ; if xj> 0
0 ; if xj= 0

◼ where xj < Myj ; yj for j = 1, 2…..

M is a large no.
and other resource constraints.

33

The Fixed Charge Problem : Example 8.

◼ Three machines (x1, x2, x3) are available to make
5000 units which are required for the internal use of a
company as components to another product. If the
production is to be done by any one of these
machines, a set up cost is incurred apart from the cost
of making each unit on different machines. The cost
data are as follows. What is the best production
strategy?
Machine Set up cost Cost per unit Max Production
1 8000 5 4000
2 5000 4 3000
3 4000 8 1000

34

17
 9/29/2021

The Fixed Charge Problem

◼ Let x1, x2, x3 represents the quantities to be produced on
machines 1, 2 and 3, resp. and d1, d2, d3 indicate whether a
machine is to be used (1) or not (0).
◼ Min Z = 8000d1+5000d2+4000d3+5x1+4x2+8x3
st
x1 + x2+x3 > 5000
x1 < 4000d1
x2 < 3000d2
x3 < 1000d3
x1, x2, x3 > 0; d1,d2,d3 = 0,1.

35

Facility Location Problem

Example 9. The Standard Cars Limited currently has four plants for manufacturing and assembling cars
and six warehouses where these cars are transported for subsequent delivery to its dealers/showrooms.
The company management has collected the following information:
◼ Annual production capacity of each plant,
◼ Annual demand of cars at various warehouses
◼ Transportation cost
◼ Annual fixed operating costs
◼ The information is summarised in the following table.

Table1. Data set

Annual Fixed
Annual operating
Production Costs( '000
Plant Warehouses capacity Rs)
('000 cars)
1 2 3 4 5 6

1 88 15 29 67 38 120 42 418
2 59 82 37 7 19 54 30 302
3 21 80 21 15 32 30 48 362
4 38 30 18 27 21 94 40 168

Annual 21 12 17 6 8 20
demand

36

18
 9/29/2021

Facility Location Problem

Example 9. The Standard Cars Limited currently has four plants for manufacturing and assembling cars
and six warehouses where these cars are transported for subsequent delivery to its dealers/showrooms.
The company management has collected the following information:
◼ Annual production capacity of each plant,
◼ Annual demand of cars at various warehouses
◼ Transportation cost
◼ Annual fixed operating costs
◼ The information is summarised in the following table.

Table1. Data set

Annual Fixed
Annual operating
Production Costs( '000
Plant Warehouses capacity Rs)
('000 cars)
1 2 3 4 5 6

1 x11 x12 x16 42d1 418d1

2 x21 30 302
3 x31 48 362
4 x41 x46 40 168

Annual 21 12 17 6 8 20
demand

37

Example 9…

To illustrate, plant 3 has a total annual capacity of 48 thousand cars, the demand at warehouse 2 is
12,000 cars per annum; the cost of transporting 1,000 cars from plant 3 to warehouse 5 is Rs 32,000
and plant 4 has an annual fixed operating cost equal to Rs 168,000. Currently, the company is sending
the cars as per the following schedule:

Table2. Old schedule

From To Warehouses
Plant
1 2 3 4 5 6
1 12 3
2 6 8
3 9 9 17
4 20

This schedule involves a total cost of Rs 4,41,000 by way of transportation and Rs 1,250,000 as
annual fixed operating costs. The management is contemplating to change the present schedule so as
to attempt to reduce the total cost to the minimum, even considering the closure of some plants. You
are requested to formulate the problem as ILP.

38

19
 9/29/2021

◼ Objective Function?
◼ Minimize Z= sum (transportation cost)+ sum (production cost)
= 88x11+15x12+…..+21x45+94x46 +418y1+…+168y4
◼ Constraints?

Table1. Data set

Annual Fixed
Annual operating
Production Costs( '000
Plant Warehouses capacity Rs)
('000 cars)
1 2 3 4 5 6

1 88 15 29 67 38 120 42 418
2 59 82 37 7 19 54 30 302
3 21 80 21 15 32 30 48 362
4 38 30 18 27 21 94 40 168

Annual 21 12 17 6 8 20
demand 39

Example.
◼ Formulate the assignment problem as an
Binary-IP form.

40

20
 9/29/2021

Examples
◼ Assigning airline crews to flights in order to minimize costs.
◼ Assigning sales personnel to sales districts in order to
maximize sales.
◼ Assigning rescue units to rescue tasks in order to minimize
total combined time to complete all rescue tasks.
◼ Assigning boats or planes to charter trips in order to
maximize total profit.
◼ Assigning social workers to welfare cases in order to
maximize the number of cases closed within a specified
period.

41

The Assignment Problem

◼ The assignment problem is a special type of integer
linear programming problem where assignees are
being assigned to perform tasks.
◼ Assumptions:
- The number of assignees and the number of tasks are the
same.
- Each assignee is to be assigned to exactly one task.

- Each task is to be performed by exactly one assignee.

- There is a cost cij associated with assignee i (i = 1,2,…,n)
performing task j (j = 1,2,…,n).
- The objective is to determine how all n assignments should
be made to minimize the total cost.
42

21
 9/29/2021

Assignment Problem

43

Assignment as a transportation Problem

◼ A resource must be allocated totally or uniquely to a given
task.
◼ Assignment problem is a special type of transportation
problem where the sources are assignees & the destinations
are tasks and where:
 No. of sources (m) = No. of destinations (n)
 Every Supply si= 1
 Every Demand dj=1

44

22
 9/29/2021

Job Assignment (Example 11)

A Production supervisor is considering how he should assign the four jobs
that are to be performed, to four of the workers. He wants to assign the
jobs to the workers such that the aggregate time to perform the jobs is the
least. Based on previous experience, he has the information on the time
taken by the four workers in performing these jobs, as given in the table
below:

Job
Worker
A B C D
1 45.x11 x12 51 67
2 57 42 63 55
3 49 52 48 64
4 41 45 60 55

45

IP Formulation: Job Assignment

xij: 1 or 0 if the worker is assigned to the job or not
Min Z = 45x11+40x12+51x13+ 67x14
+ 57x21+ 42x22+ 63x23+ 55x24
+ 49x31+52x32+ 48 x33+ 64x34
+ 41x41+45x42+ 60 x43+ 55x44
Subject to
x11+x12+x13+x14 = 1
x21+x22+x23+x24 = 1
x31+x32+x33+x34 = 1
X41+x42+ x43+ x44 = 1

x11+x21+x31 +X41= 1
x12+x22+x32 +x42= 1
x13+x23+x33 + x43 = 1
x14+x24+x34 + x44 = 1

Xij ={0,1}, i = 1,2,3,4 j = 1,2,3,4

46
46

23
 9/29/2021

Job Assignment
Solution
Constraints
x11 x12 x13 x14 x21 x22 x23 x24 x31 x32 x33 x34 x41 x42 x43 x44
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1

Objective 45 40 51 67 57 42 63 55 49 52 48 64 41 45 60 55 Min 184

Solution 0 1 0 0 0 0 0 1 0 0 1 0 1 0 0 0

47

Either-or Constraints
◼ A research and development program might be
approved if available scientific personnel are
used from one of two laboratory facilities but
not from both laboratories.
Either 3x1+ 2x2 < 18
or x1+ 4x2 < 16

48

24
 9/29/2021

Either-or Constraints
◼ Let us add a big M (very large value) to either of the constraints
so that M is large enough that it will not eliminate any feasible
solutions.
Either 3x1+ 2x2 < 18 + M
x1+ 4x2 < 16
or
3x1+ 2x2 < 18
x1+ 4x2 < 16 + M
This formulation is equivalent to the set of constraints
3x1+ 2x2 < 18 + 10000y
x1+ 4x2 < 16 + 10000(1-y);
y must be either 0 or 1.
49

25
 10/16/2021

Transportation, Transshipment
and Assignment Problems

Decision Models in Management

SOM 608

Prof. Pankaj Dutta

SJMSOM, IITBombay

Transportation Problems

1
 10/16/2021

The Transportation Problem

◼ The Transportation Problem – Many of its application involve

determining how to optimally transport goods, optimal
production scheduling, etc.

◼ The general transportation problem is concerned with

distributing any commodity from any group of supply centers,
called sources, to any group of receiving centers, called
destinations, in such a way as to minimize the total distribution
cost.

Procter & Gamble (P &G) - An application Vignette

◼ P &G makes and markets over 300 brands of consumer goods worldwide.
To maintain and accelerate it’s growth, a major OR study was undertaken
to strengthen P&G’s global effectiveness. Prior to the study, the company’s
supply chain consisted of hundred of suppliers, over 50 product categories,
over 60 plants, 15 distribution centers, and over 1000 customers zones.
However, as the company moved toward global brands, management
realized that it needed to consolidate plants to reduce manufacturing
expanses, improve speed market, and reduce capital investment.
Therefore, the study focused on redesigning the company’s production and
distribution system for its North American Operations. The result was a
reduction in the number of North American Plants by almost 20 percent,
savings $200 million in pretax cost per year.
◼ A major part of the study revolved around formulating and solving
transportation problems for individual product categories. For each option
regarding the plants to keep open, and so forth, solving the corresponding
transportation problem for a product category showed what the distribution
cost would be for shipping the product category from those plants to the
distribution centers and customer zones. Sources: Interfaces, vol. 27, 1997.
4

2
 10/16/2021

CASE: SHIPPING WOOD TO MARKET

Alabama Atlantic is a lumber company that has three sources of wood and five markets
to be supplied. The annual availability of wood at sources 1, 2, and 3 is 15, 20, and 15
million board feet, respectively. The amount that can be sold annually at markets 1, 2,
3, 4, and 5 is 11, 12, 9, 10, and 8 million board feet, respectively.

In the past the company has shipped the wood by train. However, because shipping
costs have been increasing, the alternative of using ships to make some of the
deliveries is being investigated. This alternative would require the company to invest in
some ships. Except for these investment costs, the shipping costs in thousands of
dollars per million board feet by rail and by water (when feasible) would be the following
for each route:

Unit Cost by Rail Unit Cost by Ship ($1,000’s)

Source ($1,000’s) Market Market
s
1 2 3 4 5 1 2 3 4 5
1 61 72 45 55 66 31 38 24 — 35
2 69 78 60 49 56 36 43 28 24 31
3 59 66 63 61 47 — 33 36 32 26
5

CASE: SHIPPING WOOD TO MARKET

The capital investment (in thousands of dollars) in ships required for each million board
feet to be transported annually by ship along each route is given as follows:

Investment for Ships ($1,000’s) Market

Sources
1 2 3 4 5
1 275 303 238 — 285
2 293 318 270 250 265
3 — 283 275 268 240

Considering the expected useful life of the ships and the time value of money, the
equivalent uniform annual cost of these investments is one-tenth the amount given in the
table. The objective is to determine the overall shipping plan that minimizes the total
equivalent uniform annual cost (including shipping costs).

3
 10/16/2021

CASE: SHIPPING WOOD TO MARKET

You are the head of the OR team that has been assigned the task of
determining this shipping plan for each of the following three options.

Option 1: Continue shipping exclusively by rail.

Option 2: Switch to shipping exclusively by water (except where only rail is
feasible).
Option 3: Ship by either rail or water, depending on which is less
expensive for the particular route.

Present your results for each option. Compare.

Finally, consider the fact that these results are based on current shipping
and investment costs, so that the decision on the option to adopt now
should take into account management’s projection of how these costs are
likely to change in the future. For each option, describe a scenario of future
cost changes that would justify adopting that option now.

The requirements assumption

◼ Each source has a fixed supply of units, where this entire
supply must be distributed to the destinations.
si: no. of units being supplied by source i, i= 1,2,…,m.
◼ Each destination has a fixed demand for units, where this
entire demand must be received from the sources.
dj: no. of units being received by destination j, j= 1,2,…,n.

The cost assumption

◼ The cost of distributing units from any particular source to any
particular destination is directly proportional to the number of
units distributed.
Cij: unit cost for source i and destination j.

4
 10/16/2021

The feasible solution property

◼ A transportation problem will have feasible solutions if and
only if

Integer solution property

◼ For transportation problems where every si and dj have an
integer value, all the basic variables (allocations) in every basic
feasible solution also have integer values.

The Transportation Problem

◼ Z: The total distribution cost
◼ xij: The number of units to be distributed from source i to
destination j

10

5
 10/16/2021

◼The table of constraint coefficient has the following

special structure

m n

◼As 
S =  di
i =1
i
, the number of basic variables in any
j =1

iteration of a transportation problem is m+n-1.

11

Example – P & T COMPANY

◼ One of the main products of the P & T COMPANY is canned peas.
The peas are prepared at three canneries (near Bellingham,
Washington; Eugene, Oregon; and Albert Lea, Minnesota) and then
shipped by truck to four distributing warehouses in the western
United States (Sacramento, California; Salt Lake City, Utah; Rapid
city, South Dakota; and Albuquerque, New Mexico).
◼ Because shipping cost are a major expense, management is initiating
a study to reduce them as much as possible. For the upcoming
season, an estimate has been made of the output from each cannery,
and each warehouse has been allocated a certain amount from the
total supply of peas. This information (in units of truckloads), along
with the shipping cost per truckload for each cannery-warehouse
combination is given in Table 1. Thus, there are a total of 300
truckloads to be shipped.

12

6
 10/16/2021

P &T Co. – (Continued…)

13

P &T Co. – (Continued…)

Table -1 : Shipping data for P & T Co.

Shipping Cost ($) per Truckload
Warehouse
1 2 3 4 Output
1 464 513 654 867 75
Cannery 2 352 416 690 791 125
3 995 682 388 685 100
Allocation 80 65 70 85

◼ The problem now is to determine which plan for assigning

these shipments to various cannery-warehouse combinations
would minimize the total shipping cost.

14

7
 10/16/2021

LP Formulation: P & T COMPANY

xij: number of units of truckloads to be transport from cannery i
to warehouse j

Min Z = 464x11+ 513x12+654x13+ 867x14

+ 352x21+ 416x22+ 690x23+ 791x24
+ 995x31+ 682x32+ 388x33+ 685x34
Subject to:
x11+x12+x13+x14 = 75
x21+x22+x23+x24 = 125
x31+x32+x33+x34 = 100
X11 +x21 +x31 = 80
x12 +x22 +x32 = 65
x13 +x23 +x33 = 70
x14 +x24 +x34 = 85
Xij > 0 ( i = 1,2,3; j = 1,2,3,4)
15

Example – P & T COMPANY

Solution – P & T
COMPANY
constraints
x11 x12 x13 x14 x21 x22 x23 x24 x31 x32 x33 x34
1 1 1 1 75 = 75
1 1 1 1 125 = 125
1 1 1 1 100 = 100
1 1 1 80 = 80
1 1 1 65 = 65
1 1 1 70 = 70
1 1 1 85 = 85
Objective 464 513 654 867 352 416 690 791 995 682 388 685 Min 152535

Solution 0 20 0 55 80 45 0 0 0 0 70 30

16

8
 10/16/2021

Example-2
◼ A power company has three electric power plants that
supply the needs of four cities. Each power plant can
supply the following numbers of kilowatt-hours of
electricity:
 Plant 1 35 million;
 Plant 2 50 million;
 Plant 3 40 million.
◼ The demands in these cities are as follows:
 City 1 45 million;
 City 2 20 million;
 City 3 30 million;
 City 4 30 million.

17

Example – 2
◼ The costs (in ’00 Rs) of sending 1 million kwh of
electricity from plant to city depend on the distance the
electricity must travel and is given below

◼ Formulate an LP to minimize the cost of meeting each

city’s power demand.
18

9
 10/16/2021

L P Formulation
xij: number of (million) kwh produced at plant i and sent to
city j
Min Z = 8x11+ 6x12+10x13+ 9x14
+ 9x21+ 12x22+ 13x23+ 7x24
+ 14x31+ 9x32+ 16 x33+ 5x34
Subject to:
x11+x12+x13+x14 = 35
x21+x22+x23+x24 = 50
x31+x32+x33+x34 = 40
x11+x21+x31 = 45
x12+x22+x32 = 20
x13+x23+x33 = 30
x14+x24+x34 = 30
Xij > 0 i = 1,2,3 j = 1,2,3,4

19

Example – 2
Solution: Example 2
constraints
x11 x12 x13 x14 x21 x22 x23 x24 x31 x32 x33 x34
1 1 1 1 35 = 35
1 1 1 1 50 = 50
1 1 1 1 40 = 40
1 1 1 45 = 45
1 1 1 20 = 20
1 1 1 30 = 30
1 1 1 30 = 30

Objective
function 8 6 10 9 9 12 13 7 14 9 16 5 Min 1020

Solution 0 10 25 0 45 0 5 0 0 10 0 30

20

10
 10/16/2021

Facility Location Analysis

◼ The transportation method is especially useful in helping a
firm to decide where to locate a new factory or warehouse.
◼ Each alternative location should be analyzed within the
framework of one overall distribution system.

◼ The new location that yields the minimum cost for the
entire system is the one that should be chosen

21

Locating a New Factory for Hardgrave

Machine Company

◼ Hardgrave Machine produces computer components at

three plants and they ship to four warehouses
◼ The plants have not been able to keep up with demand so
the firm wants to build a new plant
◼ Two sites are being considered, Seattle and Birmingham

◼ Data has been collected for each possible location

◼ Which new location will yield the lowest cost for the firm
in combination with the existing plants and warehouses

22

11
 10/16/2021

Locating a New Factory for Hardgrave

Machine Company
◼ Hardgrave’s demand and supply data
MONTHLY
DEMAND PRODUCTION MONTHLY COST TO PRODUCE
WAREHOUSE (UNITS) PLANT SUPPLY ONE UNIT ($)
Detroit 10,000 Cincinnati 15,000 48
Dallas 12,000 Salt Lake 6,000 50
New York 15,000 Pittsburgh 14,000 52
Los Angeles 9,000 35,000
46,000
Supply needed from new plant = 46,000 – 35,000 = 11,000 units per month

ESTIMATED PRODUCTION COST

PER UNIT AT PROPOSED PLANTS

Seattle $53

Birmingham $49

23

Locating a New Factory for Hardgrave

Machine Company

◼ Hardgrave’s shipping costs

TO LOS
FROM DETROIT DALLAS NEW YORK ANGELES
CINCINNATI $25 $55 $40 $60
SALT LAKE 35 30 50 40
PITTSBURGH 36 45 26 66
SEATTLE 60 38 65 27
BIRMINGHAM 35 30 41 50

24

12
 10/16/2021

Locating a New Factory for Hardgrave

Machine Company

◼ Optimal solution for the Birmingham location

TO LOS FACTORY
FROM DETROIT DALLAS NEW YORK ANGELES CAPACITY
73 103 88 108
CINCINNATI 10,000 1,000 4,000 15,000

85 80 100 90
SALT LAKE 1,000 5,000 6,000

88 97 78 118
PITTSBURGH 14,000 14,000

84 79 90 99
BIRMINGHAM 11,000 11,000

WAREHOUSE
10,000 12,000 15,000 9,000 46,000
REQUIREMENT

25

Locating a New Factory for Hardgrave

Machine Company
◼ Excel input screen

26

13
 10/16/2021

Locating a New Factory for Hardgrave

Machine Company
◼ Output from Excel QM analysis

27

Locating a New Factory for Hardgrave

Machine Company
◼ Optimal solution for the Seattle location

TO LOS FACTORY
FROM DETROIT DALLAS NEW YORK ANGELES CAPACITY
73 103 88 108
CINCINNATI 10,000 4,000 1,000 15,000

85 80 100 90
SALT LAKE 6,000 6,000

88 97 78 118
PITTSBURGH 14,000 14,000

113 91 118 80
SEATTLE 2,000 9,000 11,000

WAREHOUSE
10,000 12,000 15,000 9,000 46,000
REQUIREMENT

28

14
 10/16/2021

Locating a New Factory for Hardgrave

Machine Company

◼ By comparing the total system costs of the two

alternatives, Hardgrave can select the lowest cost option
◼ The Birmingham location yields a total system cost of
$3,741,000
◼ The Seattle location yields a total system cost of
$3,704,000
◼ With the lower total system cost, the Seattle location is
favored

29

Transportation model can be applied to production

scheduling and inventory control problems. How?
❑ Consider the case of a manufacturer whose product has a seasonal
demand
❑To meet the demand, the production plan of the manufacturer could be:
❑Completely uniform production
❑Variable production, completely matching the demand
❑Something between the two

❑Hence, optimal production plan requires a trade-off between overtime costs

and inventory holding cost

❑ Mapping to a Transportation Problem:

❑Regular and overtime production capacity as “origins”
❑Demand as “destinations”
❑Total cost is (Production cost + inventory cost)
❑Decision variables-production quantities manufactured by a particular
method, namely regular or overtime for meeting the demand

30

15
 10/16/2021

A manufacturer of a certain component has the following estimates of

the demand for its product:

Period 1 2 3 4
Demand 50 70 85 75

The regular production capacity for each period is 60 units while with overtime
working, an additional of upto 20 units can be produced in each period. The
unit cost of production in regular time is worked out to be Rs 6 whereas it
would cost Rs 10 per unit for the items manufactured during overtime. You
may assume that the cost would be identical in all the periods. Any production
in excess of the requirement in the current period may be held at a cost of Rs
3 per unit per period. No holding cost is involved if a unit is sold in the period
in which it is produced.

(a) Formulate the problem as a transportation problem; and

(b) Draw the optimal production plan for the manufacturer

31

Production planning problem as transportation problem

P1-D1 Cost P2-D2 P3-D3 P4-D4 D5

P1 R 6 6+3 6+3*2=12 6+9=15 0 60
O 10 10+3 16 19 0 20
P2 R Large value M 6 6+3 6+3*2=12 0 60
O 10000 10 10+3 16 0 20
P3 R 10000 10000 6 6+3 0 60
O 10000 10000 10 10+3 0 20
P4 R 10000 10000 10000 6 0 60
O 10000 10000 10000 10 0 20
50 70 85 75 40

Total Cost 320

32

16
 10/16/2021

Production planning problem as transportation problem

Destination
Origin Supply
1 2 3 4 D
R1 6 9 12 15 0 60
O1 10 13 16 19 0 20
R2 M 6 9 12 0 60
O2 M 10 13 16 0 20
R3 M M 6 9 0 60
O3 M M 10 13 0 20
R4 M M M 6 0 60
O4 M M M 10 0 20
Demand 50 70 85 75 40 320

R: Regular Time, O: Overtime, D: Dummy

33

Solution: Optimal

Destination
Origin Supply
1 2 3 4 D
R1 6 (50) 9 (10) 12 15 0 60
O1 10 13 16 19 0 (20) 20
R2 M 6 (60) 9 12 0 60
O2 M 10 13 (5) 16 0 (15) 20
R3 M M 6 (60) 9 0 60
O3 M M 10 (20) 13 0 20
R4 M M M 6 (60) 0 60
O4 M M M 10 (15) 0 (5) 20
Demand 50 70 85 75 40 320

R: Regular Time, O: Overtime, D: Dummy

34

17
 10/16/2021

Solution: Production Scheduling

Produce in Month Units For use in Month

1 2 3 4
1 R 60 50 10
O.T. nil
2 R 60 60
O.T. 5 5
3 R 60 60
O.T. 20 20
4 R 60 60
O.T. 15 15
Demand 50 70 85 75

R: Regular Time, O: Overtime

35

Northern Airplane (Production Scheduling)

Northern Airplane Company produces commercial airplanes. The last stage in production
is to produce the jet engines and install them. The company must meet the delivery
deadline indicated in column 2. An option is to produce some engines one month or more
before they are scheduled for installation and store them. Production and storage costs
vary from month to month.

Maximum Unit Cost of

Production Production ($million)
Unit Cost
Scheduled of Storage(*)
Month Installations ($thousand)
1 10 25 1.08 15
2 15 35 1.11 15
3 25 30 1.10 15
4 20 10 1.13

Question: How many engines should be produced in each of the four months so that
the total of the production and storage costs will be minimized?
(*) storage cost is incurred at the end of the month for just those engines that are being held over to
the next month; 36

18
 10/16/2021

Northern Airplane (Production Scheduling)

• Source i - production of jet engines in month i (i = 1,2 ,3 4);

• Destination j – installation of jet engines in month j (j = 1,2 ,3 4)
• xij = number of engines produce in month i to be installed in
month j
• cij = cost associated with each unit of xij

Cost per unit for production + storage if i < j

??? if i >j

• si = ??

• dj = number of scheduled installations in month j

37

Northern Airplane (Production Scheduling)

Cost per unit distributed

Destinations (installation
in month j)
Source
(production
in month i) 1 2 3 4 SUPLLY
1 1.080 1.095 1.110 1.125 ?
2 ? 1.110 1.125 1.140 ?
3 ? ? 1.100 1.115 ?
4 ? ? ? 1.130 ?
Demand 10 15 25 30

Question: How many engines should be produced in each of the four months so
that the total of the production and storage costs will be minimized?
(*) storage cost is incurred at the end of the month for just those engines that are being held over to
the next month; 38

19
 10/16/2021

Transportation Tableau: Cost per unit distributed

Destinations
(installation
in month j)
Source
(production 5(D)
in month i) 1 2 3 4 (*) SUPLLY
1 1.080 1.095 1.110 1.125 0 25
2 M 1.110 1.125 1.140 0 35
3 M M 1.100 1.115 0 30
4 M M M 1.130 0 10
Demand 10 15 25 20 30

(*) The dummy destination can be seen as a slack variable that represents the
unused production capacity. Cost is zero because it is the cost of distributing to a
fictional destination. Note that it would be inappropriate to assign M since we do
not want to force the corresponding values to be zero. In fact these values need
to sum 30.
39

The Transportation Problem

- Alternative Solution Approaches

◼ Initial basic feasible Solution by North West Corner

Method

◼ Initial basic feasible Solution by Vogel’s Approximation

Method (VAM)

◼ The Transportation Simplex method

40

20
 10/16/2021

Example
◼ SunRay Transport ships truckloads of grain from three silos
to four mills. The supply (in truckloads) and the demand
(also in truckloads) together with the unit transportation
costs per truckload on the different routes are summarized in
the transportation model in Table-2.
◼ The unit transportation costs, cij, (shown in the northeast
corner of each box) are in hundreds of dollars. The model
seeks the minimum-cost shipping schedule xij between silo i
and mill j (i=1,2,3; j=1,2,3,4)

41

Example
Table: Sunray Transportation Model.

Mill
1 2 3 4 Supply
10 2 20 11
1 15
x11 x12 x13 x14
12 7 9 20
Silo 2 25
x21 x22 x23 x24
4 14 16 18
3 10
x31 x32 x33 x34
Demand 5 15 15 15

42

21
 10/16/2021

Initial basic feasible Solution by

North West Corner Method
◼ The method starts at the northwest-corner cell (route) of the tableau
(variable x11).
◼ Step-1: Allocate as much as possible to the selected cell, and adjust
the associated amounts of supply and demand by subtracting the
allocated amount.
◼ Step-2: Cross out the row or column with zero supply or demand to
indicate that no further assignments can be made in that row or
column. If both a row and a column net to zero simultaneously,
cross out one only, and leave a zero supply (demand) in the
uncrossed-out row (column).
◼ Step-3: If exactly one row or column is left uncrossed out, stop.
Otherwise, move to the cell to the right if a column has just been
crossed out or below if a row has been crossed out. Go to Sep 1.

43

Example – 3(North West Corner method)

Table-3: NorthWest Corner IBF Solution

Mill
1 2 3 4 Supply
10 2 20 11
1 5 10 15

12 7 9 20
Silo 2 5 15 5 25

4 14 16 18
3 10 10

Demand 5 15 15 15

44

22
 10/16/2021

Initial basic feasible Solution by

Vogel’s Approximation Method (VAM)
◼ VAM is an improved version of the least-cost method that generally,
but not always, produces better starting solutions.

◼ Step-1: For each row (column), determine a penalty measure by

subtracting the smallest unit cost element in the row (column) from
the next smallest unit cost element in the same row (column).

◼ Step-2: Identify the row or column with the largest penalty. Break
ties arbitrarily. Allocate as much as possible to the variable with the
least unit cost in the selected row or column. Adjust the supply and
demand, and cross out the satisfied row or column. If a row and a
column are satisfied simultaneously, only one of the two is crossed
out, and the remaining row (column) is assigned zero supply
(demand).

45

Initial basic feasible Solution by

Vogel’s Approximation Method (VAM)
◼Step-3:
(a) If exactly one row or column with zero supply or demand
remains un-crossed out, stop.

(b) If one row (column) with positive supply (demand)

remains uncrossed out, determine the basic variables in the
row (column) by the least-cost method. Stop.

(c) If all the uncrossed out rows and columns have

(remaining) zero supply and demand, determine the zero
basic variables by the least-cost method. Stop.

(d) Otherwise, go to step-1.

46

23
 10/16/2021

Example – 3: (Vogel’s Approx. Method)

Table: IBF Solution Vogel's Approximation Method(VAM)

Mill Row Penalty

1 2 3 4 Supply I II III IV
10 2 20 11 8 9 -- --
1 15 15

12 7 9 20 2 2 11 20
Silo 2 15 10 25, 10

4 14 16 18 10 2 2 18
3 5 5 10, 5

Demand 5 15 15 15, 5

Colmn I 6 5 7 7
Penalty II -- 5 7 7
III -- -- 7 2
IV -- -- -- 2

47

Example – 3: (Vogel’s Approx. Method)

Table: IBF Solution Vogel's Approximation Method(VAM)

Mill Row Penalty

1 2 3 4 Supply I II III IV
10 2 20 11 8 9 -- --
1 15 15

12 7 9 20 2 2 11 20
Silo 2 15 10 25, 10

4 14 16 18 10 2 2 18
3 5 5 10, 5

Demand 5 15 15 15, 5

Colmn I 6 5 7 7
Penalty II -- 5 7 7
III -- -- 7 2
IV -- -- -- 2

48

24
 10/16/2021

Example – 4

D1 D2 D3 D4 Supply

O1 1 2 1 4 30

O2 3 3 2 1 50

O3 4 2 5 9 20

Demand 20 40 30 10

49

Example – 4

D1 D2 D3 D4 Supply Row Penalty

0
O1 1 2 1 4 30

10 1
O2 3 3 2 1 50

2
O3 4 2 5 9 20

Demand 20 40 30 10
Colum n Penalty 2 0 1 3

50

25
 10/16/2021

Example – 4

D1 D2 D3 D4 Supply Row Penalty

0 0
O1 1 2 1 4 30

10 1 1
O2 3 3 2 1 50 40

20 2 2
O3 4 2 5 9 20

Demand 20 40 30 10
Colum n Penalty 2 0 1 3
2 0 20 1 --

51

Example – 4

D1 D2 D3 D4 Supply Row Penalty

20 0 0 0
O1 1 2 1 4 30 30

10 1 1 1
O2 3 3 2 1 50 40

20 2 2 --
O3 4 2 5 9 20 0

Demand 20 40 30 10
Colum n Penalty 2 0 1 3
2 20 0 20 1 30 -- 0
2 1 1 --

52

26
 10/16/2021

Example – 4
D1 D2 D3 D4 Supply Row Penalty
20 10 0 0 0 1
O1 1 2 1 4 30 10

10 1 1 1 1
O2 3 3 2 1 50 40

20 2 2 -- --
O3 4 2 5 9 20 0

Demand 20 40 30 10
Colum n Penalty 2 0 1 3
2 0 20 1 -- 0
2 1 1 --
-- 1 1 20 --

53

Example – 4
D1 D2 D3 D4 Supply Row Penalty
20 10 0 0 0 1
O1 1 2 1 4 30 10

20 20 10 1 1 1 1
O2 3 3 2 1 50 40

20 2 2 -- --
O3 4 2 5 9 20 0

Demand 20 40 30 10
Colum n Penalty 2 0 1 3
2 0 20 1 -- 0
2 1 1 --
-- 1 1 20 --

54

27
 10/16/2021

The Transportation Simplex method

◼ Step 1: Construct an initial basic feasible solution by North

West corner rule or by Vogel’s approximation method.
◼ Step 2: Derive ui and vj.
A convenient way to derive is by selecting the row having the
largest number of allocations, setting its ui= 0 and then solving
the set of equations cij= ui+ vj for each (i,j) such that xij is
basic.
Calculate ∆ij = cij–ui–vj corresponding to the unoccupied
cell (i, j).
If ∆ij = cij–ui–vj ≥ 0 for every (i, j) such that xij is non basic,
then the current solution is optimal, so stop.

Otherwise, go to the next step.

55

The Transportation Simplex method

◼ Step 3:
•Determine the entering basic variable:
Select the nonbasic variable xij having the largest negative
value of cij–ui–vj.
•Determine the leaving basic variable:
Identify the chain reaction required to retain feasibility
when the entering basic variable is increased. From the
donor cells, select the basic variable having the smallest
value.
•Determine the new basic feasible solution.
Add the value of the leaving basic variable to the allocation
for each recipient cell. Subtract this value from the
allocation for each donor cell.
•Go to step 2.
56

28
 10/16/2021

Interpretation
◼ ui: Multiple of original row i that has been subtracted from
original row 0 by the simplex method during all iterations
leading to the current simplex table.
◼ vj: Multiple of original row m+j that has been subtracted from
original row 0 by the simplex method during all iterations
leading to the current simplex table.
◼ ui and vj can be interpreted as dual variables. For a non basic
variable, cij–ui–vj denotes the rate at which Z will change as xij
is increased. For any transportation problem with m sources
and n destinations, the number of functional constraints is m+n
but the number of basic variables is m+n-1.

57

Example – 5

D1 D2 D3 Supply

O1 2 7 4 5

O2 3 3 1 8

O3 5 4 7 7

O4 1 6 2 14

Demand 7 9 18
58

29
 10/16/2021

Example – 5 (Check all ∆ij > 0 for optimality)

cij= ui+ vj

D1 D2 D3 Supply ui
5
O1 2 7 4 5 1

8
O2 3 3 1 8 -1

7
O3 5 4 7 7 -2

2 2 10
O4 1 6 2 14 0

Demand 7 9 18
vj 1 6 2
59

Example – 5 (Check all ∆ij > 0 for optimality)

D1 D2 D3 Supply ui
5 0 1
O1 2 7 4 5 1
3 -2
8
O2 3 3 1 8 -1
6 7
7
O3 5 4 7 7 -2

2 2 10
O4 1 6 2 14 0

Demand 7 9 18
vj 1 6 2
60

30
 10/16/2021

Example – 5

D1 D2 D3 Supply
5 0 1
O1 2 7 4 5

3 2 6
O2 3 3 1 8

6 7 7
O3 5 4 7 7

2 0 12
O4 1 6 2 14

Demand 7 9 18
61

◼ The Transportation problem always has a feasible

solution.
◼ The solution of the transportation problem is never
unbounded.
◼ Degeneracy: If the no. of occupied cell is not equals to m+n-1, then
add a small positive quantity ϵ to a cell (may be more than one required
to reach m+n-1) such that this does not result in forming a loop among
some or all of the occupied cells and make them dependent. For a
dependent set of cells unique determination of ui and vj will not be
possible.

Now, do the optimality test as it is. If one ∆ij is negative, then add ϵ to
this cell, because it is the most minimum among the concerned loop.
Do the optimality test again.

62

31
 10/16/2021

Example – 6 (For your information)

D1 D2 D3 Supply

O1 8 7 3 60

O2 3 8 9 70

O3 11 3 5 80

Demand 50 80 80

63

◼ Maximization Transportation Problem:

A transportation tableau may contain unit profits instead of unit costs. Select
the largest element from the profit matrix and subtract all element from it.
Now we can find the solution for the minimization problems. Once this is
done, the value of the objective function is determined with reference to the
original profit matrix.
◼ Unbalanced Transportation Problems:
Add a column/row of slack variables with dummy source/destination
representing the requirement equal to the amount of excess demand/supply
with zero transportation cost.
◼ Unacceptable or Prohibited Routes:
Sometimes in a given transportation problem some route(s) may not be
available. Assign a large cost in that cell.
◼ Unique or Multiple Optimal Solutions:
If some cell( or cells) has ∆ij =0, then multiple optimal solutions are
indicated so that there exit other transportation pattern(s) with same cost.
64

32
 10/16/2021

Schematic presentation of transportation method

Start

Write the problem in tabular form

Is it No Balance the table using

balanced? dummy row/column

Yes
Convert it into a maximization problem:
subtract each element of the profit
Is it a Yes matrix from its highest value
maximization
problem?
No

Find its initial basic feasible solution

using NWC rule, VAM etc.

Is it a Yes Estimate degeneracy by assigning ε

degenerate? to requisite number of cells

No

No Is it a
Generate an improved solution
optimal?

Yes
The problem is solved

Stop 65

Transshipment Problems

66

33
 10/16/2021

Transshipment Problems

◼ The transshipment model recognizes that it

may be cheaper to ship through intermediate or
transient nodes before reaching the final
destination.
◼ Transshipment model can be converted to a
regular transportation model using the idea of
a buffer.

67

Transshipment Problems
◼ Pure supply nodes
◼ Transshipment nodes
◼ Pure demand nodes

◼ Supply at a pure supply node = Original supply

◼ Demand at a pure demand node = Original demand
◼ Supply at a transshipment nodes = Original supply + Buffer
amount
◼ Demand at a transshipment nodes = Original demand + Buffer
amount
◼ The buffer amount should be sufficient enough to allow all the
original supply (or demand) units to pass through any of the
transshipment nodes.
68

34
 10/16/2021

The Transshipment Model Characteristics

▪ Extension of the transportation model.

▪ Intermediate transshipment points are added between the sources and
destinations.
▪ Items may be transported from:
▪ Sources through transshipment points to destinations
▪ One source to another
▪ One transshipment point to another
▪ One destination to another
▪ Directly from sources to destinations
▪ Some combination of these

69

Transshipment Model Example

Problem Definition and Data
Wheat is harvested at firms in Nebrasca and Colorado before being shipped to
the three grain elevators in Kansas city, Omaha and Des Moines, which are
now transshipment points. The amount of wheat harvested at each firm is 300
tons. The wheat is then shipped to the mills in Chicago, St. Louis, and
Cincinnati. The shipping cost are given in the following table and on the network
graph.
▪ Extension of the transportation model in which intermediate transshipment
points are added between sources and destinations.
▪ Data:
Grain Elevator
Farm 3. Kansas City 4. Omaha 5. Des Moines
1. Nebrasca $16 10 12
2. Colorado 15 14 17

70
70

35
 10/16/2021

Transshipment Model Example

Problem Definition and Data

Network of Transshipment Routes

71

Transshipment Model Example

Model Formulation
Minimize Z = 16x13 + 10x14 + 12x15 + 15x23 + 14x24 + 17x25 + 6x36 + 8x37 + 10x38
+ 7x46 + 11x47 + 11x48 + 4x56 + 5x57 + x58
subject to:
x13 + x14 + x15 = 300
x23+ x24 + x25 = 300
x36 + x46 + x56 = 200
x37 + x47 + x57 = 100
x38 + x48 + x58 = 300
x13 + x23 - x36 - x37 - x38 = 0
x14 + x24 - x46 - x47 - x48 = 0
x15 + x25 - x56 - x57 - x58 = 0
xij  0

72
72

36
 10/16/2021

Computer Solution with Excel

73

Transshipment Model Example

Computer Solution with Excel

74

37
 10/16/2021

Transshipment Model Example

Transshipment Network Solution for Wheat Shipping

75

Distributing Goods through a Distribution Network

◼ The Problem. The Distribution Unlimited Co. will be producing the same new
product at two different factories, and then the product must be shipped to two
warehouses, where either factory can supply either warehouse. The distribution
network available for shipping this product is shown in DN Fig. , where F1 and F2
to are the two factories, W1 and W2 are the two warehouses, DC is a distribution
center. The amounts to be shipped from F1 and F2 are shown to their left, and the
amounts to be received at W1and W2 are shown to their right. Each arrow
represents a feasible shipping lane. Thus, F1can ship directly to W1 and has three
possible routes (F1->DC-> W2, F1->F2->DC->W2, and F1->W1->W2) for
shipping toW2. Factory F2 has just one route to W2 (F2->DC->W2) and one to W1
(F2->DC->W2->W1).The cost per unit shipped through each shipping lane is
shown next to the arrow. Also shown next to F1->F2 and DC->W2 are the
maximum amounts that can be shipped through these lanes. The other lanes have
sufficient shipping capacity to handle everything these factories can send.
The decision to be made concerns how much to ship through each shipping lane.
The objective is to minimize the total shipping cost.

76

38
 10/16/2021

Distributing Goods through a Distribution Network

The decision to be made concerns how much to ship through each shipping
lane. The objective is to minimize the total shipping cost.

77

Distributing Goods through a Distribution Network

With seven shipping lanes, we need seven decision variables (X F1-F2 ,XF1-DC
,XF1-W1 ,XF2-DC ,XDC-W2 ,XW1-W2 ,XW2-W1) to represent the amounts shipped
through the respective lanes.

Minimize Z = 2XF1-F2 + 4XF1-DC + 9XF1-W1 + 3XF2-DC +XDC-W2 +3XW1-W2

+2XW2-W1
Subject to;
◼ 5-Net flow constrains
 Amount shipped out - amount shipped in = required amount
◼ Upper bound constraints
 XF1-F2 <= 10, XDC-W2 <= 80
◼ Non-negativity constraints

78

39
 10/16/2021

Formulation as a Linear Programming Problem

Minimize Z = 2XF1-F2 +4XF1-DC +9XF1-W1 + 3XF2-DC + XDC-W2 +3XW1-W2 +2XW2-W1

subject to the following constraints:

1. Net flow constraints:

XF1-F2 + XF1-DC + XF1-W1 = 50 (factory 1)
- XF1-F2 + XF2-DC = 40 (factory 2)
- XF1-DC - XF2-DC + XDC-W2 = 0 (distribution center)
- XF1-W1 + XW1-W2 - XW2-W1 = -30
(warehouse 1)
- XDC-W2 - XW1-W2 + XW2-W1 = -60
(warehouse 2)
2. Upper-bound constraints: XF1-F2 <=10, XDC-W2 <=80
3. Non-negativity constraints:
XF1-F2 >=0, XF1-DC >=0, XF1-W1 >=0, XF2-DC >=0,
XDC-W2 >=0, XW1-W2 >=0, XW2-W1 >=0
79

Distributing Goods through a Distribution Network

Solution:

XF1-F2 =0, XF1-DC =40,

XF1-W1 =10, XF2-DC =40,
XDC-W2 =80, XW1-W2 =0,
XW2-W1 =20
The resulting total shipping cost is $49,000.

80

40
 10/16/2021

Test Problem

Two automobile plants, P1 and P2, are linked to three

dealers, D1, D2, and D3, by way of two transit centers, T1
and T2m according to network shown in the following SC
network. The supply amounts at plants P1 and P2 are 1000
and 1200 cars, and the demand at dealers D1, D2 and D3 are
800,900 and 500 cars. The shipping costs per car in hundreds
of dollars between pairs of nodes are shown on connecting
links (or arcs) of the network.

◼ Develop a corresponding transshipment model.

81

Test Problem
D1 800

3 T1 5
1000 P1
6
4
7
D2 900
2
4

1200 5 3
P2
T2
9

D3 500

Fig. Supply Chain Network

82

41
 10/16/2021

Assignment Problems

83

American League Umpires

Scheduling umpires in professionals baseball is a complex problem that must include a
number of criteria. In assigning officials to games, one objective typically is to minimize total
travel cost while satisfying a set of frequency-oriented constraints such as limiting the
number of times an official or crew is exposed to each team, balancing home and away
game exposures, balancing exposures to teams over the course of a season, and so on.
These constraints complicate the problem to such an extent that except for the most trivial
cases, the use of a computer-based system is essential.
The American League is composed of 14 professional baseball teams organized into three
divisions. The game schedule, constructed each winter prior to the start of the baseball
season, is a different scheduling problem in itself. Consideration must be given to such
factors as the number of games played against other teams, both within and outside a
division, the split between home games and road trips, travel time, and possible conflicts in
cities that have teams in the National League.
The objective of balancing crew assignments relatively evenly and minimizing travel costs
are by nature conflicting, Attempting to balance crew assignments necessitates considerable
airline travel and equipment moves, and hence increased travel costs.

Using an assignment model as part of a micro computer-based decision support system, the
American League was able to reduce travel mileage by about 4% during the first year of use.
This not only saved the league $30, 000 but improved the crew exposure balance.
Source: Interface, 1988
84

42
 10/16/2021

Examples
◼ Assigning airline crews to flights in order to minimize costs.
◼ Assigning sales personnel to sales districts in order to
maximize sales.
◼ Assigning rescue units to rescue tasks in order to minimize
total combined time to complete all rescue tasks.
◼ Assigning boats or planes to charter trips in order to
maximize total profit.
◼ Assigning social workers to welfare cases in order to
maximize the number of cases closed within a specified
period.

85

The Assignment Problem

◼ The assignment problem is a special type of linear
programming problem where assignees are being
assigned to perform tasks.
◼ Assumptions:
- The number of assignees and the number of tasks are the
same.
- Each assignee is to be assigned to exactly one task.

- Each task is to be performed by exactly one assignee.

- There is a cost cij associated with assignee i (i = 1,2,…,n)
performing task j (j = 1,2,…,n).
- The objective is to determine how all n assignments should
be made to minimize the total cost.
86

43
 10/16/2021

Assignment Problem

87

Assignment as a transportation Problem

◼ A resource must be allocated totally or uniquely to a given
task.
◼ Assignment problem is a special type of transportation
problem where the sources are assignees & the destinations
are tasks and where:
 No. of sources (m) = No. of destinations (n)
 Every Supply si= 1
 Every Demand dj=1

88

44
 10/16/2021

Job Assignment
A Production supervisor is considering how he should assign the four jobs
that are to be performed, to four of the workers. He wants to assign the
jobs to the workers such that the aggregate time to perform the jobs is the
least. Based on previous experience, he has the information on the time
taken by the four workers in performing these jobs, as given in the table
below:

Job
Worker
A B C D
1 45 40 51 67
2 57 42 63 55
3 49 52 48 64
4 41 45 60 55

89

L P Formulation: Job Assignment

xij: 1 or 0 if the worker is assigned to the job or not
Min Z = 45x11+40x12+51x13+ 67x14
+ 57x21+ 42x22+ 63x23+ 55x24
+ 49x31+52x32+ 48 x33+ 64x34
+ 41x41+45x42+ 60 x43+ 55x44
Subject to
x11+x12+x13+x14 = 1
x21+x22+x23+x24 = 1
x31+x32+x33+x34 = 1
X41+x42+ x43+ x44 = 1

x11+x21+x31 +X41= 1
x12+x22+x32 +x42= 1
x13+x23+x33 + x43 = 1
x14+x24+x34 + x44 = 1

Xij ={0,1}, i = 1,2,3,4 j = 1,2,3,4

90
90

45
 10/16/2021

Job Assignment
Solution
Constraints
x11 x12 x13 x14 x21 x22 x23 x24 x31 x32 x33 x34 x41 x42 x43 x44
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1
1 1 1 1 1 = 1

Objective 45 40 51 67 57 42 63 55 49 52 48 64 41 45 60 55 Min 184

Solution 0 1 0 0 0 0 0 1 0 0 1 0 1 0 0 0

91

Basic Principle
◼ The optimal assignment is not affected if a constant is
added or subtracted from any row or column of the
standard assignment cost matrix
◼ e.g. if the cost of doing any job on machine 1 is
reduced by k, then the objective function of the
assignment problem becomes
n n
Min z =  (c1 j − k ) x1 j +  cij xij
j i = 2 j =1
n n n
=  cij xij − k  xij
i =1 j =1 j =1

s.t.
n

 xi
j =1
j =1 i

92

46
 10/16/2021

Hungarian Algorithm
◼ Step1 : Find the minimum element in each row of
the n x n cost matrix. Construct a new matrix by
subtracting from each cost the minimum cost in its
row.
For this new matrix, find the minimum cost in
each column. Construct a new matrix by subtracting
from each cost the minimum cost in its column.

93

Hungarian Algorithm
◼ Step 2 : Draw the minimum number of lines
(horizontal and/or vertical) that are needed to cover
all the zeros in the reduced cost matrix.
If n lines are required, an optimal solution is
available among the covered zeros in the matrix. If
fewer than n lines are needed, proceed to Step 3.

94

47
 10/16/2021

Hungarian Algorithm
◼ Step 3: Find the smallest element (say k) in the
reduced cost matrix that is uncovered by the lines
drawn in Step 2.
Now subtract k from each uncovered element of
the reduced cost matrix and add k to each element
that is covered by two lines.
Return to step 2.

These operations are the resultant of:

Subtract k from all the uncrossed rows and add k to all
the crossed columns.
95

Example
Machines
M1 M2 M3 M4

J1 8 26 17 11

J2 13 28 4 26
JOBS

J3 38 19 18 15

J4 19 26 24 10

96

48
 10/16/2021

Example
Machines
M1 M2 M3 M4

J1 0 18 9 3

J2 9 24 0 22
JOBS

J3 23 4 3 0

J4 9 16 14 0

97

Example
Machines
M1 M2 M3 M4

J1 0 14 9 3

J2 9 20 0 22
JOBS

J3 23 0 3 0

J4 9 12 14 0

98

49
 10/16/2021

Example (J1→M1, J2→M3, J3→M2, J4→M4)

Machines
M1 M2 M3 M4

J1 0 14 9 3

J2 9 20 0 22
JOBS

J3 23 0 3 0

J4 9 12 14 0

99

Example 2?
Machines
M1 M2 M3 M4

J1 18 26 17 11

J2 13 28 14 26
JOBS

J3 38 19 18 15

J4 19 26 24 10

100

50
 10/16/2021

Example 2?
Machines
M1 M2 M3 M4

J1 7 11 5 0

J2 0 11 0 13
JOBS

J3 23 0 2 0

J4 9 12 13 0

101

Example 2?
Machines
M1 M2 M3 M4

J1 2 6 0 0

J2 0 11 0 18
JOBS

J3 23 0 2 5

J4 4 7 8 0

102

51
 10/16/2021

◼ Maximization Problem: Select the largest element

from the profit matrix and subtract all element from
it. Now we can find the solution to find the minimum
cost which will provide the maximum profit of the
original problem
◼ Unbalanced problem: Add a fictitious job or facility
whichever is required with zero cost.
◼ Impossible assignment: Put a large cost in that cell.
◼ Negative cost: Add the most negative element to all
the elements.
* An Assignment problem may have more than one solution having
the same minimum cost.
103

Example
Table- 1: Maximization problem

I II

A 3 9

B 6 4

I II

A 6 0

B 3 5

104

52
 10/16/2021

Example
◼ Jobshop needs to assign 4 jobs to 4 workers. The
cost of performing a job is a function of the skills of
the workers. Table-1 summarizes the cost of the
assignments. Worker 1 cannot do job 3 and worker 3
cannot do job 4.

◼ Determine optimal assignment using Hungarian

method.

105

Example
Table- 1: Cost data
Job
J1 J2 J3 J4

1 50 50 − 20

2 70 40 20 30
Worker

3 90 30 50 −

4 70 20 60 70

106

53
 10/16/2021

Example
Table- 1: Cost data
Job
J1 J2 J3 J4

1 50 50 M 20

2 70 40 20 30
Worker

3 90 30 50 M

4 70 20 60 70
107

Start
Schematic presentation of Assignment Problem

Write the problem in tabular form

Is it No Add dummy row(s)/

balanced? column(s)

Yes
Convert it into a maximization problem:
either (i) by changing the signs of the
Is it a Yes elements of the table; or by subtracting
maximization
all the values from the largest value
problem?
No

Obtain reduced cost tables:

(i) Subtract from all entries in each row the least
value in the row
(ii) From this table, subtract from all entries in
each column the least value in the column

Improve the solution. For this:

(i) Select the minimum of the uncovered (by
lines) cell values.
Can all zeroes Yes (ii) Subtract this value from all uncovered cell
be covered by values.
less than n lines? (iii) Add this value to the cells lying on the
intersection of any pair of lines.
(iv) Leave the cell values covered by only
No one line undisturbed.
Make assignments on
one-to-one match basis
considering zeroes in
rows/columns.

Stop 108

54