COMSATS Institute of Information Technology,

Virtual Campus

Operations Research (MTH467)

by

Salman Amin MALIK

VCOMSATS
Learning Management System
To my unknown students
 ii

About the author

Dr. Salman Amin Malik did his MS and PhD (Mathematics) from University of
La Rochelle, La Rochelle, France in 2009 and 2012, respectively. Prior to MS and
PhD, Dr. Malik completed his MPhil and MSc (Mathematics) from University of
the Punjab, Lahore, Pakistan. He has been affiliated with several universities in
Pakistan and abroad. He has the experience of teaching a wide range mathematics
courses at undergraduate and graduate level.

Dr. Malik has published several research articles in international journals and
conferences. His area of research includes the study of fractional differential equa-
tions and their applications to image processing. He is also interested in inverse
source problem related to fractional diffusion equations which has numerous appli-
cations ranging from biomedical imaging to the anomalous diffusion in heterogeneous
media .

About the handouts

For the course Operations Research the text book is Hamdy A. Taha, Operations
Research: An Introduction, Eighth Edition, Pearson Prentice Hall, ISBN: 0-13-
188923·0. Consequently, the most of the examples considered in these notes are
from the above mentioned book and its exercises, but not restricted to that book
only. If you find any typing error in the text kindly report to me by writing an email
to [email protected].
 iii

Course Information
Title and Course Code: Operations Research (MTH467)

Number of Credit Hours: 3 credits

Course Objective: The objective of the course is to introduce the basic termi-
nologies of Operations Research. The mathematical modeling of certain real world
problems and adopt different Operations Research techniques to get the best possi-
ble answer.
There are a variety of real world problems will be discussed in the course and
several algorithms will be discussed. This is an introductory course to Operations
Research.

Prerequisites: A little bit knowledge of linear equations/ inequalities is re-

quired.

Course Learning Outcomes: Upon completion of the course, students will be

able to:

1. To formulate a real-world problem as a mathematical programming model.

2. Understand the theoretical workings of the simplex method for linear pro-
gramming and perform iterations of it by hand.

3. Understand the relationship between a linear program and its dual, includ-
ing strong duality and complementary slackness.

4. Perform sensitivity analysis to determine the direction and magnitude of

change of a model’s optimal solution as the data change.

5. Solve specialized linear programming problems like the transportation and

assignment problems.

6. Understand the applications of basic methods for integer programming Model

a dynamic system as a queuing model and compute important perfor-
mance measures.
Contents

1 Lecture No. 01 2
1.1 History (1939-1945) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Decision variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Objective Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.5 Feasible Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.6 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 8

2 Lecture No. 02 9
2.1 Mathematical Modeling . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 15

3 Lecture No. 03 16
3.1 Graphical Linear Programming (LP) Solution . . . . . . . . . . . . . 16

4 Lecture No. 04 19
4.1 Solution of a Minimization Problem . . . . . . . . . . . . . . . . . . 22
4.2 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 25

5 Lecture No. 05 26
5.1 Urban Renewal Model . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.2 Controlling Air Pollution . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.3 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 31

6 Lecture No. 06 32
6.1 The essence of simplex method . . . . . . . . . . . . . . . . . . . . . 32
6.2 The Simplex Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
6.3 LP Model in Equation Form . . . . . . . . . . . . . . . . . . . . . . . 34
6.4 Algebraic Determination of Corner Points . . . . . . . . . . . . . . . 35
6.5 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 37

7 Lecture No. 07 39
7.1 Algebraic Steps for the Simplex Method . . . . . . . . . . . . . . . . 39
7.2 Entering and Leaving variable selection . . . . . . . . . . . . . . . . . 41
7.3 Summary of the Simplex Method . . . . . . . . . . . . . . . . . . . . 43
7.4 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 44

8 Lecture No. 08 45
8.1 M-Technique/Method . . . . . . . . . . . . . . . . . . . . . . . . . . 45
8.2 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 49
Contents v

9 Lecture No. 09 50
9.1 Two Phase Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
9.2 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 54

10 Lecture No. 10 55
10.1 Special Cases in Simplex Method . . . . . . . . . . . . . . . . . . . . 55
10.2 Degeneracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
10.3 Alternative Optima . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
10.4 Unbounded Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
10.5 Infeasible Solution Space . . . . . . . . . . . . . . . . . . . . . . . . . 59
10.6 Some Practice problems . . . . . . . . . . . . . . . . . . . . . . . . . 60

11 Lecture No. 11 62
11.1 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
11.2 Changes in the Objective Coefficients . . . . . . . . . . . . . . . . . . 64

12 Lecture No. 12 68
12.1 Algebraic Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . 68

13 Lecture No. 13 75
13.1 The dual problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
13.2 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 81

14 Lecture No. 14 82
14.1 Simplex Tableau Computations . . . . . . . . . . . . . . . . . . . . . 82
14.2 Economic Interpretation of Dual Variables . . . . . . . . . . . . . . . 85
14.3 Economic Interpretation of Dual Constraints . . . . . . . . . . . . . . 86
14.4 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 88

15 Lecture No. 15 89
15.1 Dual Simplex Method . . . . . . . . . . . . . . . . . . . . . . . . . . 89
15.2 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 92

16 Lecture No. 16 (Revision) 94

17 Lecture No. 17 95
17.1 Transportation Problem . . . . . . . . . . . . . . . . . . . . . . . . . 95
17.2 Transport model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
17.3 Balancing the Transportation Model . . . . . . . . . . . . . . . . . . 100
17.4 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 102

18 Lecture No. 18 103

18.1 The Transportation Algorithm . . . . . . . . . . . . . . . . . . . . . 103
18.2 Least-Cost Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
18.3 Vogel Approximation Method (VAM) . . . . . . . . . . . . . . . . . . 106
18.4 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 109
Contents vi

19 Lecture No. 19 110

19.1 Method of Multipliers . . . . . . . . . . . . . . . . . . . . . . . . . . 110
19.2 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 114

20 Lecture No. 20 116

20.1 The Assignment Model . . . . . . . . . . . . . . . . . . . . . . . . . . 116
20.2 The Hungarian Method . . . . . . . . . . . . . . . . . . . . . . . . . 116
20.3 The transhipment Model . . . . . . . . . . . . . . . . . . . . . . . . . 119

21 Lecture No. 21 122

21.1 Network Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
21.2 Kruskal’s Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
21.3 Prim’s Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

22 Lecture No. 22 128

22.1 Shortest Route Problem . . . . . . . . . . . . . . . . . . . . . . . . . 128
22.2 Dijkstra’s Algorithm For Shortest Route Problem . . . . . . . . . . . 130
22.3 Floyd’s Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

23 Lecture No. 23 135

23.1 LP Formulation of the Shortest Route Problem . . . . . . . . . . . . 135
23.2 Maximum Flow Problems . . . . . . . . . . . . . . . . . . . . . . . . 136
23.3 Maximal Flow Algorithm . . . . . . . . . . . . . . . . . . . . . . . . 137
23.4 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 139

24 Lecture No. 24 140

24.1 Maximum Flow Problems (Contd) . . . . . . . . . . . . . . . . . . . 140
24.2 Linear Programming Formulation of Maximal Flow Model . . . . . . 142
24.3 CPM (Critical Path Method) and PERT (Program Evaluation and
Review Technique) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
24.4 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 145

25 Lecture No. 25 147

25.1 Critical Path (CPM) Computations . . . . . . . . . . . . . . . . . . . 147
25.2 Linear Programming Formulation of CPM . . . . . . . . . . . . . . . 150

26 Lecture No. 26 152

26.1 Integer Linear Programs . . . . . . . . . . . . . . . . . . . . . . . . . 152
26.2 Applications of ILP . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

27 Lecture No. 27 157

27.1 Branch and Bound Method . . . . . . . . . . . . . . . . . . . . . . . 157
27.2 Summary of the B&B Algorithm . . . . . . . . . . . . . . . . . . . . 161
27.3 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 162
Contents vii

28 Lecture No. 28 163

28.1 Cutting-Plane Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . 163

29 Lecture No. 29 168

29.1 Dynamical Programming . . . . . . . . . . . . . . . . . . . . . . . . . 168
29.2 Some Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . . 173

30 Lecture No. 30 174

30.1 Selected DP Applications . . . . . . . . . . . . . . . . . . . . . . . . 174

31 Lecture No. 31 180

31.1 Introduction to Inventory Models . . . . . . . . . . . . . . . . . . . . 182

32 Lecture No. 32 (Revision) 185

33 Frequently Asked Questions 186

List of Figures

1.1 Some case studies and their impact . . . . . . . . . . . . . . . . . . . 4

2.1 The Reddy Mikks problem . . . . . . . . . . . . . . . . . . . . . . . . 11

2.2 The WYNDOR GLASS CO problem . . . . . . . . . . . . . . . . . . 13
2.3 Data for a linear program problem . . . . . . . . . . . . . . . . . . . 14

3.1 Graphical solution of WYNDOR GLASS CO . . . . . . . . . . . . . 16

3.2 The movement of objective function . . . . . . . . . . . . . . . . . . 17
3.3 No feasible solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.4 Infinite many optimal solutions . . . . . . . . . . . . . . . . . . . . . 18
3.5 Unbounded solution . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

4.1 Corner point feasible solutions . . . . . . . . . . . . . . . . . . . . . . 19

4.2 The Reddy Mikks data . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.3 Graphical solution of Reddy Mikks problem . . . . . . . . . . . . . . 21
4.4 Movement of objective function . . . . . . . . . . . . . . . . . . . . . 21
4.5 Diet problem data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.6 Graphical solution of diet problem . . . . . . . . . . . . . . . . . . . 23
4.7 Radiation therapy data . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.8 Graphical solution of radiation therapy problem . . . . . . . . . . . . 24

5.1 The data for steel mill . . . . . . . . . . . . . . . . . . . . . . . . . . 29

5.2 Reduction table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.3 Cost table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.4 Standard for the steel mill . . . . . . . . . . . . . . . . . . . . . . . . 30

6.1 CPF and their adjacent CPFs . . . . . . . . . . . . . . . . . . . . . . 32

6.2 Graphical solution and CPFs . . . . . . . . . . . . . . . . . . . . . . 32
6.3 The essence of simplex method graphically . . . . . . . . . . . . . . . 33
6.4 Graphical solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
6.5 Basic and nonbasic variables . . . . . . . . . . . . . . . . . . . . . . . 36
6.6 Augmented solution of an LP . . . . . . . . . . . . . . . . . . . . . . 37

7.1 Basic and nonbasic variables for corner points of the figure below . . 39
7.2 Graphical solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
7.3 Graphical interpretation . . . . . . . . . . . . . . . . . . . . . . . . . 41
7.4 Simplex tableau iterations for WYNDOR GLASS problem . . . . . . 44

8.1 Graphical solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

8.2 Simplex tableau iteration by M technique . . . . . . . . . . . . . . . 48

10.1 Graphical inspection of degeneracy . . . . . . . . . . . . . . . . . . . 56

List of Figures ix

10.2 Graphical inspection of alternative optima . . . . . . . . . . . . . . . 58

10.3 Infeasible solution space . . . . . . . . . . . . . . . . . . . . . . . . . 59

11.1 Change in RHS of constraint . . . . . . . . . . . . . . . . . . . . . . 63

11.2 Changes in the objective coefficients . . . . . . . . . . . . . . . . . . 65

15.1 The dual simplex iterations . . . . . . . . . . . . . . . . . . . . . . . 92

17.1 General transportation problem . . . . . . . . . . . . . . . . . . . . . 96

17.2 Mileage chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
17.3 Cost table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
17.4 Transportation tableau . . . . . . . . . . . . . . . . . . . . . . . . . . 97
17.5 Optimal solution of the transportation problem . . . . . . . . . . . . 97
17.6 General transportation tableau . . . . . . . . . . . . . . . . . . . . . 98
17.7 Geographical map . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
17.8 Shipping cost table . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
17.9 Pictorial information of the problem . . . . . . . . . . . . . . . . . . 99
17.10Transportation tableau with dummy source . . . . . . . . . . . . . . 100
17.11Transportation tableau with dummy destination . . . . . . . . . . . . 101
17.12Equivalence of terms . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
17.13The transportation tableau for production inventory problem . . . . 102

18.1 Starting tableau . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

18.2 Starting basic solution by north west corner method . . . . . . . . . 105
18.3 Starting basic solution by least cost method . . . . . . . . . . . . . . 106
18.4 VAM row and column penalizes . . . . . . . . . . . . . . . . . . . . . 107

19.1 ui and vj of method of multipliers . . . . . . . . . . . . . . . . . . . 110

19.2 z-row coefficients of nonbasic variables . . . . . . . . . . . . . . . . . 111
19.3 The z-row . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
19.4 Method of multipliers on the transportation tableau . . . . . . . . . 111
19.5 Selection of leaving variable . . . . . . . . . . . . . . . . . . . . . . . 112
19.6 Iteration 2 calculations . . . . . . . . . . . . . . . . . . . . . . . . . . 113
19.7 Iteration 3 the optimal solution . . . . . . . . . . . . . . . . . . . . . 113
19.8 Transportation problem with dummy source . . . . . . . . . . . . . . 114

20.1 General assignment model . . . . . . . . . . . . . . . . . . . . . . . . 116

20.2 The jobs and relative cost . . . . . . . . . . . . . . . . . . . . . . . . 117
20.3 Row minimum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
20.4 Column minimum . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
20.5 Optimal assignment solution . . . . . . . . . . . . . . . . . . . . . . . 118
20.6 Example with no feasible solution . . . . . . . . . . . . . . . . . . . . 118
20.7 Usual solution by Hungarian method . . . . . . . . . . . . . . . . . . 119
20.8 Optimal feasible solution of the problem . . . . . . . . . . . . . . . . 119
20.9 The transhipment problem . . . . . . . . . . . . . . . . . . . . . . . . 120
List of Figures x

20.10Optimal solution of the transhipment problem . . . . . . . . . . . . . 120

20.11Transhipment problem as transportation problem and transportation
tableau . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

21.1 Spanning trees of a connected graph . . . . . . . . . . . . . . . . . . 123

21.2 Konigsberg bridge problem . . . . . . . . . . . . . . . . . . . . . . . 123
21.3 Spanning tree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
21.4 Kruskal’s algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
21.5 Example minimal spanning tree . . . . . . . . . . . . . . . . . . . . . 125
21.6 Minimal spanning tree by Kruskal’s algorithm . . . . . . . . . . . . . 125
21.7 Prim’s algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
21.8 Minimal spanning tree problem . . . . . . . . . . . . . . . . . . . . . 126
21.9 Minimum spanning tree by Prim’s algorithm . . . . . . . . . . . . . . 126
21.10Two different minimum spanning trees . . . . . . . . . . . . . . . . . 127

22.1 The problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

22.2 Minimal spanning tree . . . . . . . . . . . . . . . . . . . . . . . . . . 128
22.3 RentCar data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
22.4 Network corresponding to the RentCar problem . . . . . . . . . . . . 129
22.5 Three jug puzzle as nework . . . . . . . . . . . . . . . . . . . . . . . 129
22.6 Dijkstra’s algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
22.7 Shortest route problem . . . . . . . . . . . . . . . . . . . . . . . . . . 131
22.8 The shortest route by Dijkstra’s algorithm . . . . . . . . . . . . . . . 131
22.9 Shortest route problem . . . . . . . . . . . . . . . . . . . . . . . . . . 131
22.10Intermediate node . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
22.11Shortest route problem . . . . . . . . . . . . . . . . . . . . . . . . . . 132
22.12Starting matrix D0 and S0 . . . . . . . . . . . . . . . . . . . . . . . . 133

23.1 Typical pipeline network . . . . . . . . . . . . . . . . . . . . . . . . . 136

24.1 Project planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

24.2 The role of dummy activities . . . . . . . . . . . . . . . . . . . . . . 144
24.3 Usage of dummy activity . . . . . . . . . . . . . . . . . . . . . . . . . 144
24.4 Network representation of the project . . . . . . . . . . . . . . . . . . 145

27.1 Graphical solution of the problem . . . . . . . . . . . . . . . . . . . . 157

27.2 Addition of new constraints (graphically) . . . . . . . . . . . . . . . . 158
27.3 Branches correspond to two constraints . . . . . . . . . . . . . . . . . 158
27.4 Branches of LP problem . . . . . . . . . . . . . . . . . . . . . . . . . 161

28.1 Cutting plane algorithm graphically . . . . . . . . . . . . . . . . . . 163

Chapter 1

Lecture No. 01

The contents of the course Operations Research can be divided into three major
parts, described below:
Part I:

• The origin, the nature, the impact of Operations Research

• Operations Research Models, i.e., defining the problem, formulating a mathe-

matical model and getting the solution

• Modeling with Linear Programming, graphical solution, assumptions of linear

programming, selected applications, selected linear programming applications

• The simplex method and sensitivity analysis

• The dual of linear programming model

Part II:

• The transportation problems

• The assignment problems

Part III:

• Integer Programming

• Network Models

• Inventory Models

• Dynamic Programming Models

• Queuing theory
1.1. History (1939-1945) 3

The text book:

Frederick S. Hillier and Gerald J. Lieberman, Introduction to Operations

Research, Seventh Edition, McGraw-Hill Higher Education, ISBN: 0072321695.

Hamdy A. Taha, Operations Research: An Introduction, Eighth Edition, Pear-

son Prentice Hall, ISBN: 0-13-188923·0.
Assessment Plan for the Course:

• Four Assignments 10%.

• Four Quiz 10%.

• Graded Discussion Board 5%.

• First Sessional Exam 10%.

• Second Sessional Exam 15%.

• Final Exam 50%.

1.1 History (1939-1945)

The roots of Operations Research OR can be traced back many decades, when early
attempts were made to use a scientific approach in the management of organizations.
However, the beginning of the activity called operations research has generally been
attributed to the military services early in World War II.

Growth of the subject:

1. Extensive Progress in the subject: Simplex method has been developed by

George Dantzig in 1947.

2. The computer revolution.

As its name implies, operations research involves "research on operations." Thus,

operations research is applied to problems that concern how to conduct and coor-
dinate the operations (i.e., the activities) within an organization.

The nature of the organization is essentially immaterial, and, in fact, OR has

been applied extensively in such diverse areas as manufacturing, transportation,
construction, telecommunications, financial planning, health care, the military, and
1.1. History (1939-1945) 4

public services, to name just a few. Therefore, the breadth of application is unusually
wide.

The research part of the name means that operations research uses an approach
that resembles the way research is conducted in established scientific fields.

Figure 1.1: Some case studies and their impact

Important steps in Operations Research (OR):

1. Define the problem of interest and gather relevant data.

2. Formulate a mathematical model to represent the problem.

3. Develop a computer-based procedure for deriving solutions to the problem

from the model.

4. Test the model and refine it as needed.

1.1. History (1939-1945) 5

5. Prepare for the ongoing application of the model as prescribed by management.

6. Implement.

The problem: Imagine that you have a 5-week business commitment between
Fayetteville (FYV) and Denver (DEN).
You fly out of Fayetteville on Mondays and return on Wednesdays. A regular
round-trip ticket costs 400 dollars, but a 20% discount is granted if the dates of the
ticket span a weekend.
A one way ticket in either direction costs 75% of the regular price.
How should you buy the tickets for the 5-week period?

What should I look for?

1. What are the decision alternatives?

2. Under what restrictions is the decision made?

3. What is an appropriate objective criterion for evaluating the alternatives?

Alternatives: These alternatives are considered:

1. Buy five regular FYV-DEN-FYV for departure on Monday and return on

Wednesday of the same week.

2. Buy one FYV-DEN, four DEN-FYV-DEN that span weekends, and one DEN-
FYV.

3. Buy one FYV-DEN-FYV to cover Monday of the first week and Wednesday
of the last week and four DEN-FYV-DEN to cover the remaining legs. All
tickets in this alternative span at least one weekend.

Restrictions: Leave on Monday and return on Wednesday the same week

Alternative 1 cost = 5 × 400 = 2000$
Alternative 2 cost = .75 × 400 + 4 × (.8 × 400) + .75 × 400 = 1880$
Alternative 3 cost = 5 × (.8 × 400) = 1600$

The Problem: Consider forming a maximum-area rectangle out of a piece of wire

of length L inches. What should be the width and height of the rectangle?

Remark: The number of alternatives is not finite.

Solution: Let W = width of the rectangle in inches and h = height of the rectangle
in inches

Restrictions:
1.2. Decision variables 6

1. Width of rectangle + Height of rectangle = Half the length of the wire

2. Width and height cannot be negative.

That is

1. 2(w + h) = L

2. w ≥ 0, h ≥ 0.

Objective of the problem Maximize z = wh

1.2 Decision variables

If there are n related quantifiable decisions to be made, they are represented as
decision variables say (x1 , x2 , ..., xn ) whose respective values are to be determined.

1.3 Objective Function

The appropriate measure of performance (e.g., profit) is then expressed as a mathe-
matical function of these decision variables for example,(P = a1 x1 +a2 x2 +...+an xn )
where a1 , a2 , ..., an ∈ R.
This function is called the objective function.

1.4 Constraints
Any restrictions on the values that can be assigned to these decision variables are
also expressed mathematically. Such mathematical expressions for the restrictions
often are called constraints.

Example: An OR study done for Monsanto Corp. was concerned with optimizing
production operations in Monsanto’s chemical plants to minimize the cost of meeting
the target for the amount of a certain chemical product (maleic anhydride) to be
produced in a given month. The decisions to be made are the dial setting for each
of the catalytic reactors used to produce this product, where the setting determines
both the amount produced and the cost of operating the reactor. The form of the
resulting mathematical model is as follows:
Choose the decision variables Rij (i = 1, 2, ..., s; j = 1, 2, ..., s) so as to
r X
X s
Minimize cij Rij
i=1 j=1
1.5. Feasible Solution 7

subject to
r X
X s s
X
cij pij Rij ≥ T, cij Rij = 1, for i = 1, 2, ..., r, Rij = 0 or 1
i=1 j=1 j=1

where Rij = 1 if reactor i is operated at setting j and otherwise Rij = 0.

cij = cost for reactor i at the setting j, pij = production of reactor i at j, T =
production target, r = number of reactors, s = number of settings.
For Monsanto’s application, this model has over 1, 000 decision variables Rij
(that is, rs ≥ 1, 000). Its use led to annual savings of approximately 2 million $.

Maximize or minimize Objective Function

subject to
Constraints

1.5 Feasible Solution

A solution of the model is feasible if it satisfies all the constraints.
A common theme in OR is the search for an optimal, or best, solution.

An optimal solution for the original model may be far from ideal for the real
problem, so additional analysis is needed. Therefore, postoptimality analysis
(analysis done after finding an optimal solution) is a very important part of most
OR studies.

In part, postoptimality analysis involves conducting sensitivity analysis to deter-

mine which parameters of the model are most critical (the "sensitive parameters")
in determining the solution.
A common definition of sensitive parameter is the following.
For a mathematical model with specified values for all its parameters, the model’s
sensitive parameters are the parameters whose value cannot be changed without
changing the optimal solution.

Rijkswaterstaat: (B. F. Goeller and the PAWN team: "Planning the Netherland’s
Water Resources, Interfaces, 15(1): 3-33, Jan.-Feb. 1985.)
The Netherlands government agency responsible for water control and public
works, the Rijkswaterstaat, commissioned a major OR study1 to guide the develop-
ment of a new national water management policy. The new policy saved hundreds
of millions of dollars in investment expenditures and reduced agricultural damage
by about 15 million$ per year, while decreasing thermal and algae pollution. Rather
than formulating one mathematical model, this OR study developed a comprehen-
sive, integrated system of 50 models.
Furthermore, for some of the models, both simple and complex versions were
developed. The simple version was used to gain basic insights, including trade-off
1.6. Some Practice Problems 8

analysis. The complex version then was used in the final rounds of the analysis or
whenever greater accuracy or more detailed outputs were desired. The overall OR
study directly involved over 125 person-years of effort (more than one-third in data
gathering), created several dozen computer programs, and structured an enormous
amount of data.

This study did not conclude by recommending just a single solution.

Instead, a number of attractive alternatives were identified, analyzed,
and compared. The final choice was left to the Dutch political process,
culminating with approval by Parliament. Sensitivity analysis played a
major role in this study.

1.6 Some Practice Problems

Question: In the tickets example, identify a fourth feasible alternative.

Question: In the rectangle problem, identify two feasible solutions and determine
which one is better.

Question: Determine the optimal solution of the rectangle problem.

(Hint: Use the constraint to express the objective function in terms of one
variable, then use differential calculus.)

Question: Amy, Jim, John, and Kelly are standing on the east bank of a river and
wish to cross to the west side using a canoe. The canoe can hold at most two people
at a time. Amy, being the most athletic, can row across the river in 1 minute.
Jim, John, and Kelly would take 2, 5, and 10 minutes, respectively. If two people
are in the canoe, the slower person dictates the crossing time. The objective is for
all four people to be on the other side of the river in the shortest time possible.
Identify at least two feasible plans for crossing the river (remember, the canoe
is the only mode of transportation and it cannot be shuttled empty).
Chapter 2

Lecture No. 02

2.1 Mathematical Modeling

Problem definition involves defining the scope of the problem under investigation.
This function should be carried out by the entire OR team.

The aim is to identify three principal elements of the decision problem:

(1) description of the decision alternatives,

(2) determination of the objective of the study, and

(3) specification of the limitations under which the modeled system operates.

Model construction entails an attempt to translate the problem definition into

mathematical relationships.

If the resulting model fits one of the standard mathematical models, such as
linear programming, we can usually reach a solution by using available algorithms.

Alternatively, if the mathematical relationships are too complex to allow the de-
termination of an analytic solution, the OR team may opt to simplify the model and
use a heuristic approach, or they may consider the use of simulation, if appropriate.

Model solution is by far the simplest of all OR phases because it entails the use
of well-defined optimization algorithms.

An important aspect of the model solution phase is sensitivity analysis. It deals

with obtaining additional information about the behavior of the optimum solution
when the model undergoes some parameter changes.

Sensitivity analysis is particularly needed when the parameters of the model

cannot be estimated accurately.

In these cases, it is important to study the behavior of the optimum solution in

the neighborhood of the estimated parameters.

Model validity checks whether or not the proposed model does what it purports
to do-that is, does it predict adequately the behavior of the system under study?
2.1. Mathematical Modeling 10

Initially, the OR team should be convinced that the model’s output does not
include "surprises." In other words, does the solution make sense? Are the results
intuitively acceptable?
On the formal side, a common method for checking the validity of a model is to
compare its output with historical output data. The model is valid if, under similar
input conditions, it reasonably duplicates past performance.
Generally, however, there is no assurance that future performance will continue
to duplicate past behavior. Also, because the model is usually based on careful
examination of past data, the proposed comparison is usually favorable. If the
proposed model represents a new (nonexisting) system, no historical data would be
available.
In such cases, we may use simulation as an independent tool for verifying the
output of the mathematical model.

Implementation of the solution of a validated model involves the translation of

the results into understandable operating instructions to be issued to the people
who will administer the recommended system.

The burden of this task lies primarily with the OR team.

Linear Programming: Linear programming uses a mathematical model to de-
scribe the problem of concern.

The adjective linear means that all the mathematical functions in this model are
required to be linear functions.

The word programming does not refer here to computer programming; rather,
it is essentially a synonym for planning.

Thus, linear programming involves the planning of activities to obtain an op-

timal result, i.e., a result that reaches the specified goal best (according to the
mathematical model) among all feasible alternatives.

Example (Frontier Airlines Purchases Fuel Economically):

The fueling of an aircraft can take place at any of the stopovers along the flight
route.
Fuel price varies among the stopovers, and potential savings can be realized by
loading extra fuel (called tankering) at a cheaper location for use on subsequent
flight legs.
The disadvantage of tankering is the excess burn of gasoline resulting from the
extra weight. Linear Programming is used to determine the optimum amount of
tankering that balances the cost of excess burn against the savings in fuel cost.
The study, carried out in 1981, resulted in net savings of about 350,000$ per
year.
2.1. Mathematical Modeling 11

Example (The Reddy Mikks Company) Reddy Mikks produces both interior and
exterior paints from two raw materials, M 1 and M 2. The following table provides
the basic data of the problem:

Figure 2.1: The Reddy Mikks problem

A market survey indicates that the daily demand for interior paint cannot
exceed that for exterior paint by more than 1 ton. Also, the maximum daily
demand for interior paint is 2 tons.
Reddy Mikks wants to determine the optimum (best) product mix of interior
and exterior paints that maximizes the total daily profit.

1. Decision variables that we seek to determine.

2. Objective that we need optimize (maximize or minimize)

3. Constraints that the solution must satisfy.

Objective function:
x1 = Tons produced daily of exterior paint
x2 = Tons produced daily of interior paint
Total profit from exterior paint = 5xl
Total profit from interior paint = 4x2
Letting z represent the total daily profit (in thousands of dollars), the objective
of the company is
Maximize z = 5x1 + 4x2 .
(Usage of a raw material by both paints)≤ (Maximum raw material availability)

Constraints:
Usage of raw material M 1 by exterior paint = 6xl tons/day
Usage of raw material M 1 by interior paint = 4x2 tons/day
Usage of raw material M l by both paints = 6x1 + 4x2 tons/day
Usage of raw material M 2 by both paints = 1x1 + 2x2 tons/day

6x1 + 4x2 ≤ 24 (Raw material M1)

x1 + 2x2 ≤ 6 (Raw material M2)

A market survey indicates that the daily demand for interior paint cannot
exceed that for exterior paint by more than 1 ton. Also, the maximum daily
demand for interior paint is 2 tons.
2.1. Mathematical Modeling 12

Constraints:
x2 − x1 ≤ 1 (Market limit)
x2 ≤ 2 (Demand limit)
An implicit restriction is that variables x1 and x2 cannot assume negative values.
The nonnegativity restrictions, x1 ≥ 0, x2 ≥ 0, account for this requirement.
The complete Reddy Mikks model is

Maximize z = 5x1 + 4x2

subject to

6x1 + 4x2 ≤ 24
x1 + 2x2 ≤ 6
−x1 + x2 ≤ 1
x2 ≤ 2
x1 , x2 ≥ 0

Any values of x1 and x2 that satisfy all five constraints constitute a feasible solu-
tion. Otherwise, the solution is infeasible.
Determine x1 = 3 tons per day and x2 = 1 tons per day is a feasible solution or
not?

The goal of the problem is to find the best feasible solution, or the opti-
mum, that maximizes the total profit.

Properties of the LP Model:

1. Proportionality: This property requires the contribution of each decision

variable in both the objective function and the constraints to be directly pro-
portional to the value of the variable.

Maximize z = 5x1 + 4x2

2. Additivity: This property requires the total contribution of all the variables
in the objective function and in the constraints to be the direct sum of the
individual contributions of each variable.

3. Certainty: All the objective and constraint coefficients of the LP model are
deterministic. This means that they are known constants-a rare occurrence in
real life.

Example: For the Reddy Mikks model, construct each of the following constraints
and express it with a linear left-hand side and a constant right-hand side:

1. The daily usage of raw material M 2 in tons is at most 6 and at least 3.

2.1. Mathematical Modeling 13

2. The demand for interior paint cannot be less than the demand for exterior
paint.
The complete Reddy Mikks model is

Maximize z = 5x1 + 4x2

subject to

6x1 + 4x2 ≤ 24
x1 + 2x2 ≤ 6
−x1 + x2 ≤ 1
x2 ≤ 2
x1 , x2 ≥ 0

Example: Determine the best feasible solution among the following (feasible and
infeasible) solutions of the Reddy Mikks model:
1. x1 = 1, x2 = 4 and x1 = 2, x2 = 2

2. x1 = 3, x2 = 1.5 and x1 = 2, x2 = 1

3. x1 = 2, x2 = −1
Solution: Put all the values of the variables one by one in the objective function
as well as in constraints. The values of x1 , x2 for which all constraints are satisfied
are known as feasible solution points.
Among feasible solution points, the points or values of x1 , x2 that gives maximum
value of the objective function is the best feasible solution.
Can you justify why we are looking for the maximum objective function value.

Example: The WYNDOR GLASS CO. produces high quality glass products, in-
cluding 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.
Product 1: An 8-foot glass door with aluminum framing
Product 2: A 4 × 6 foot double-hung wood-framed window
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.

Figure 2.2: The WYNDOR GLASS CO problem

2.1. Mathematical Modeling 14

Formulate the linear programming problem.

Solution: Let us define the following:
x1 = number of batches of product 1 produced per week
x2 = number of batches of product 2 produced per week
Z = total profit per week (in thousands of dollars) from producing these two
products
Then the mathematical formulation becomes The complete Reddy Mikks model
is
Maximize z = 3x1 + 5x2
subject to

x1 ≤ 4
2x2 ≤ 12
3x1 + 2x2 ≤ 18
x1 , x2 ≥ 0

General Linear program model:

Maximize Z = c1 x+ c2 x2 + ... + cn xn
subject to the restrictions

a11 x1 + a12 x2 + ... + a1n xn ≤ b1

a21 x1 + a22 x2 + ... + a2n xn ≤ b2
.
.
.
am1 x1 + am2 x2 + ... + amn xn ≤ bm

and x1 ≥ 0, x2 ≥ 0, ..., xn ≥ 0

Figure 2.3: Data for a linear program problem

2.2. Some Practice Problems 15

2.2 Some Practice Problems

Question: For the feasible solution x1 = 2, x2 = 2 of the Reddy Mikks model,
determine the unused amounts of raw materials M l and M 2.

Question: Suppose that Reddy Mikks sells its exterior paint to a single wholesaler
at a quantity discount. The profit per ton is 5000$ if the contractor buys no more
than 2 tons daily and 4500$ otherwise.
Express the objective function mathematically. Is the resulting function linear?

Question: Determine the best feasible solution among the following (feasible and
infeasible) solutions of the WYNDOR GLASS CO. model:

1. x1 = 1, x2 = 4 and x1 = 2, x2 = 2

2. x1 = 3, x2 = 1.5 and x1 = 2, x2 = 1

3. x1 = 2, x2 = −1
Chapter 3

Lecture No. 03

Recall: How to plot linear equation or inequality in two variables. Plot x1 +3x2 = 1
and x1 + 3x2 ≤ 1.

3.1 Graphical Linear Programming (LP) Solution

The graphical procedure includes two steps:
1. Determination of the feasible solution space.

2. Determination of the optimum solution from among all the feasible points in
the solution space.

Example: Solve the WYNDOR GLASS CO. model graphically:

Maximize z = 2x1 + 5x2

subject to the constraints

x1 ≤ 4
2x2 ≤ 12
3x1 + 2x2 ≤ 18
x1 , x2 ≥ 0

Figure 3.1: Graphical solution of WYNDOR GLASS CO

3.1. Graphical Linear Programming (LP) Solution 17

Figure 3.2: The movement of objective function

Remark: The point (2, 6) lies at the intersection of the two lines 2x2 = 12 and
3x1 +2x2 = 18, so that this point can be calculated algebraically as the simultaneous
solution of these two equations.

Conclusions: This solution indicates that the Wyndor Glass Co. should produce
products 1 and 2 at the rate of 2 batches per week and 6 batches per week, respec-
tively, with a resulting total profit of 36,000$ per week.
The feasible region is the collection of all feasible solutions.
An optimal solution is a feasible solution that has the most favorable value of the
objective function.

Remark: The Wyndor Glass Co. problem would have no feasible solutions if the
constraint 3x1 + 5x2 ≥ 50 were added to the problem and as shown in figure below.

Figure 3.3: No feasible solution

3.1. Graphical Linear Programming (LP) Solution 18

Remark: The Wyndor Glass Co. problem would have multiple optimal solutions
if the objective function were changed to z = 3x1 + 2x2 were added to the problem
as show in the figure below.

Figure 3.4: Infinite many optimal solutions

Remark: The Wyndor Glass Co. problem would have no optimal solutions if
the only functional constraint were x1 ≤ 4, because x2 then could be increased
indefinitely in the feasible region without ever reaching the maximum value of z =
3x1 + 5x2 .

Figure 3.5: Unbounded solution

Chapter 4

Lecture No. 04

A corner-point feasible (CPF) solution is a solution that lies at a corner of the

feasible region.

Figure 4.1: Corner point feasible solutions

Relationship between optimal solutions and CPF solutions:

Consider any linear programming problem with feasible solutions and a bounded
feasible region.
The problem must possess CPF solutions and at least one optimal solution.
Furthermore, the best CPF solution must be an optimal solution. Thus, if a
problem has exactly one optimal solution, it must be a CPF solution.
If the problem has multiple optimal solutions, at least two must be CPF solu-
tions.

Example (The Reddy Mikks Company) Reddy Mikks produces both interior and
exterior paints from two raw materials, M 1 and M 2. The following table provides
the basic data of the problem:
 20

Figure 4.2: The Reddy Mikks data

A market survey indicates that the daily demand for interior paint cannot
exceed that for exterior paint by more than 1 ton. Also, the maximum daily
demand for interior paint is 2 tons.

Reddy Mikks wants to determine the optimum (best) product mix of interior
and exterior paints that maximizes the total daily profit.

Example: Solve the Reddy Mikks problem by graphical method (we have seen
that the above problem reduces to the following mathematical model) The complete
Reddy Mikks model is
Maximize z = 5x1 + 4x2
subject to

6x1 + 4x2 ≤ 24
x1 + 2x2 ≤ 6
−x1 + x2 ≤ 1
x2 ≤ 2
x1 , x2 ≥ 0
 21

Figure 4.3: Graphical solution of Reddy Mikks problem

Take z = 10 and z = 15 in Maximize z = 5x1 + 4x2

Figure 4.4: Movement of objective function

The optimum solution is x1 = 3, x2 = 1.5 and the corresponding objective

function value is z = 21.

Remark: Notice that the optimal solution is the intersection of the lines 6x1 +4x2 =
24 and x1 + 2x2 = 6. Solving these two equation will also yields the same solution.
4.1. Solution of a Minimization Problem 22

(Verify)

Example: Determine the feasible space for each of the following independent con-
straints, given x1 , x2 ≥ 0
1. −3x1 + x2 ≤ 6

2. x1 − 2x2 ≥ 5

3. x1 − 2x2 ≤ 0
Example: Identify the direction of increase in z in each of the following cases:
1. Maximize z = x1 − x2

2. Maximize z = −x1 + 2x2

4.1 Solution of a Minimization Problem

: (Diet Problem)
Example: Ozark Farms uses at least 800 lb of special feed daily. The special feed
is a mixture of corn and soybean meal with the following compositions:

Figure 4.5: Diet problem data

The dietary requirements of the special feed are at least 30% protein and at most
5% fiber. Ozark Farms wishes to determine the daily minimum cost feed mix.

Solution: The decision variables are:

x1 = lb of corn in the daily mix
x2 = lb of soybean meal in the daily mix
Objective function is Minimize z = .3x1 + .9x2 .
The dietary requirements of the special feed are at least 30% protein and at most
5% fiber.

Constraints: The dietary requirements of the special feed are at least 30% protein
leads to the following constraint

.09x1 + .60x2 ≥ .3(x1 + x2 ) ⇒ .21x1 − .3x2 ≤ 0

4.1. Solution of a Minimization Problem 23

Similarly, the information of at most 5% fiber leads to the following constraint

.02x1 + .06x2 ≤ .05(x1 + x2 ) ⇒ .03x1 − .01x2 ≥ 0

x1 , x2 ≥ 0

The mathematics formulation is

Minimize z = .3x1 + .9x2

subject to

x1 + x2 ≥ 800
.21x1 − .30x3 ≤ 0
.03x1 − .01x2 ≥ 0
x1 , x2 ≥ 0

Figure 4.6: Graphical solution of diet problem

Remark: The optimum solution is the intersection of the two lines x1 + x2 = 800
and .21x1 −.3x2 = 0, which yields x1 = 470.591b and x2 = 329.41 lb. The associated
minimum cost of the feed mix is z = .3 × 470.59 + .9 × 329.42 = 437.65$ per day.

Design of Radiation Therapy: MARY has just been diagnosed as having a

cancer at a fairly advanced stage. Specifically, she has a large malignant tumor in
the bladder area.
Mary is to receive the most advanced medical care available to give her every
possible chance for survival. This care will include extensive radiation therapy.

Tumor cells are typically microscopically interspersed among healthy cells, the
radiation dosage throughout the tumor region must be large enough to kill the
malignant cells, which are slightly more radiosensitive, yet small enough to spare
the healthy cells.
4.1. Solution of a Minimization Problem 24

At the same time, the aggregate dose to critical tissues must not exceed es-
tablished tolerance levels, in order to prevent complications that can be more
serious than the disease itself.

For the same reason, the total dose to the entire healthy anatomy must be minimized.

Figure 4.7: Radiation therapy data

The two decision variables x1 and x2 represent the dose (in kilorads) at the entry
point for beam 1 and beam 2, respectively.
The mathematical model is

Minimize z = .4x1 + .5x2

subject to

.3x1 + .1x2 ≤ 2.7

.5x1 − .5x2 = 6
.6x1 − .4x2 ≥ 6
x1 , x2 ≥ 0

Figure 4.8: Graphical solution of radiation therapy problem

Remark: Thus, the optimal design is to use a total dose at the entry point of 7.5
kilorads for beam 1 and 4.5 kilorads for beam 2.
4.2. Some Practice Problems 25

4.2 Some Practice Problems

Question: For the diet model, suppose that the daily availability of corn is limited
to 450 lb. Identify the new solution space, and determine the new optimum solution.

Question: John must work at least 20 hours a week to supplement his income while
attending school. He has the opportunity to work in two retail stores. In store 1,
he can work between 5 and 12 hours a week, and in store 2 he is allowed between 6
and 10 hours.
Both stores pay the same hourly wage. In deciding how many hours to work
in each store, John wants to base his decision on work stress. Based on interviews
with present employees, John estimates that, on an ascending scale of 1 to 10, the
stress factors are 8 and 6 at stores 1 and 2, respectively. Because stress mounts by
the hour, he assumes that the total stress for each store at the end of the week is
proportional to the number of hours he works in the store. How many hours should
10hn work in each store?
Chapter 5

Lecture No. 05

5.1 Urban Renewal Model

The city of Erstville is faced with a severe budget shortage. Seeking a long-term
solution, the city council votes to improve the tax base by condemning an inner-city
housing area and replacing it with a modern development.

The project involves two phases:

(1) demolishing substandard houses to provide land for the new development, and

(2) building the new development.

The following is a summary of the situation.

1. As many as 300 substandard houses can be demolished. Each house occupies

a 25-acre lot. The cost of demolishing a condemned house is 2000$.

2. Lot sizes for new single, double, triple, and quadruple family homes (units)
are .18, .28, .4, and .5 acre, respectively. Streets, open space, and utility
easements account for 15% of available acreage.

3. In the new development the triple and quadruple units account for at least
25% of the total. Single units must be at least 20% of all units and double
units at least 10%.

4. The tax levied per unit for single, double, triple, and quadruple units is 1,000$,
1,900$, 2,700$, and 3,400$, respectively.

5. The construction cost per unit for single-, double, triple, and quadruple family
homes is 50,000$, 70,000$, 130,000$, and 160,000$, respectively. Financing
through a local bank can amount to a maximum of 15 million$.

How many units of each type should be constructed to maximize tax

collection?

1. As many as 300 substandard houses can be demolished. Each house occupies

a 25-acre lot. The cost of demolishing a condemned house is 2000$.
5.1. Urban Renewal Model 27

2. Lot sizes for new single, double, triple, and quadruple family homes (units)
are .18, .28, .4, and .5 acre, respectively. Streets, open space, and utility
easements account for 15% of available acreage.

3. In the new development the triple and quadruple units account for at least
25% of the total. Single units must be at least 20% of all units and double
units at least 10%.

4. The tax levied per unit for single, double, triple, and quadruple
units is 1,000$, 1,900$, 2,700$, and 3,400$, respectively.

5. The construction cost per unit for single-, double, triple, and quadruple family
homes is 50,000$, 70,000$, 130,000$, and 160,000$, respectively. Financing
through a local bank can amount to a maximum of 15 million$.

Solution: Mathematical Modeling: The decision variables are x1 = Number

of units of single-family homes
x2 = Number of units of double-family homes
x3 = Number of units of triple-family homes
x4 = Number of units of quadruple-family homes
x5 = Number of old homes to be demolished

The objective is to maximize total tax collection from all four types
of homes

Maximize z = 1000x1 + 1900x2 + 2700x3 + 3400x4

1. As many as 300 substandard houses can be demolished. Each house

occupies a .25-acre lot. The cost of demolishing a condemned house
is 2000$.

2. Lot sizes for new single, double, triple, and quadruple family homes
(units) are .18, .28, .4, and .5 acre, respectively. Streets, open space,
and utility easements account for 15% of available acreage.

3. In the new development the triple and quadruple units account for at least
25% of the total. Single units must be at least 20% of all units and double
units at least 10%.

4. The tax levied per unit for single, double, triple, and quadruple units is 1,000$,
1,900$, 2,700$, and 3,400$, respectively.

5. The construction cost per unit for single-, double, triple, and quadruple family
homes is 50,000$, 70,000$, 130,000$, and 160,000$, respectively. Financing
through a local bank can amount to a maximum of 15 million$.
5.1. Urban Renewal Model 28

Constraints correspond to 1 and 2: The first constraint of the problem deals

with land availability.

Acreage needed for new homes = .18x1 + .28x2 + .4x3 + .5x4

Available acreage = .85(.25)x5 and we have the constraint

.18x1 +.28x2 +.4x3 +.5x5 ≤ .2125x5 or .18x1 +.28x2 +.4x3 +.5x5 −.2125x5 ≤ 0

The number of demolished homes cannot exceed 300, which translates to

x5 ≤ 300

1. As many as 300 substandard houses can be demolished. Each house occupies

a .25-acre lot. The cost of demolishing a condemned house is 2000$.

2. Lot sizes for new single, double, triple, and quadruple family homes (units)
are .18, .28, .4, and .5 acre, respectively. Streets, open space, and utility
easements account for 15% of available acreage.

3. In the new development the triple and quadruple units account for
at least 25% of the total. Single units must be at least 20% of all
units and double units at least 10%.

4. The tax levied per unit for single, double, triple, and quadruple units is 1,000$,
1,900$, 2,700$, and 3,400$, respectively.

5. The construction cost per unit for single-, double, triple, and
quadruple family homes is 50,000$, 70,000$, 130,000$, and 160,000$,
respectively. Financing through a local bank can amount to a max-
imum of 15 million$.

Constraints correspond to 3, 4 and 5: The constraint lead to the following

mathematical inequalities

x1 ≤ .2(x1 + x2 + x3 + x4 )
x2 ≤ .1(x1 + x2 + x3 + x4 )
x3 + x4 ≤ .25(x1 + x2 + x3 + x4 )

The only remaining constraint deals with keeping the demolishition/construction

cost within the allowable budget, that is

Construction and demolition cost ≤ Available Budget

which leads to
50x1 + 70x2 + 130x3 + 160x4 + 2x5 ≤ 15000.
5.2. Controlling Air Pollution 29

Consequently, the mathematical modele is

Maximize z = 1000x1 + 1900x2 + 2700x3 + 3400x4

subject to

.18x1 + .28x2 + .4x3 + .5x5 − .2125x5 ≤ 0

x5 ≤ 300
−.8x1 + .2x2 + .2x3 + .2x4 ≤ 0
.1x1 − .9x2 + .1x3 + .1x4 ≤ 0
.25x1 + .25x2 − .75x3 − .75x4 ≤ 0
50x1 + 70x2 + 130x3 + 160x4 + 2x5 ≤ 15000
x1 , x2 , x3 , x4 , x5 ≥ 0

Solution: The optimum solution is

Total tax collection = z = 343, 965$
Number of single homes = x1 = 35.83 ≈ 36 units
Number of double homes = x2 = 98.53 ≈ 99 units
Number of triple homes = x3 = 44.79 ≈ 45 units
Number of quadruple homes = x4 ≈ 0 units
Number of homes demolished = x5 = 244.49 ≈ 245 units.
In our forthcoming lectures we will learn how to get the optimal solution of LP
problems.

Remark: Linear programming does not guarantee an integer solution.

5.2 Controlling Air Pollution

The NORI & LEETS CO., one of the major producers of steel in its part of the
world, is located in the city of Steeltown and is the only large employer there.
Steeltown has grown and prospered along with the company, which now employs
nearly 50, 000 residents.
The three main types of pollutants in this airshed are particulate matter,
sulfur oxides, and hydrocarbons. The new standards require that the company
reduce its annual emission of these pollutants by the amounts shown in table

Figure 5.1: The data for steel mill

5.2. Controlling Air Pollution 30

The steelworks has two primary sources of pollution, namely, the blast furnaces
for making pig iron and the open-hearth furnaces for changing iron into steel. emis-
sion (in millions of pounds per year) can be eliminated from each type of furnace
by fully using any abatement method to its technological limit.

Figure 5.2: Reduction table

The total annual cost would be roughly that fraction of the corresponding quan-
tity in

Figure 5.3: Cost table

Mathematical Modeling: This problem has six decision variables xj , j =

1, 2, ..., 6, each representing the use of one of the three abatement methods for one
of the two types of furnaces, expressed as a fraction of the abatement capacity (so
xj cannot exceed 1).

Minimize z = 8x1 + 10x2 + 7x3 + 6x4 + 11x5 + 9x6

Technological limit:

xj ≤ 1 for j = 1, 2, ..., 6

Figure 5.4: Standard for the steel mill

Mathematical Modeling:
5.3. Some Practice Problems 31

Emission reduction:
12x1 + 9x2 + 25x3 + 20x4 + 17x5 + 13x6 ≥ 60
35x1 + 42x2 + 18x3 + 31x4 + 56x5 + 49x6 ≥ 150
37x1 + 53x2 + 28x3 + 24x4 + 29x5 + 20x6 ≥ 125
Nonnegativity : xj ≥ 0 j = 1, 2, ..., 6

Minimize z = 8x1 + 10x2 + 7x3 + 6x4 + 11x5 + 9x6

subject to
12x1 + 9x2 + 25x3 + 20x4 + 17x5 + 13x6 ≥ 60
35x1 + 42x2 + 18x3 + 31x4 + 56x5 + 49x6 ≥ 150
37x1 + 53x2 + 28x3 + 24x4 + 29x5 + 20x6 ≥ 125
xj ≥ 0 j = 1, 2, ..., 6
xj ≤ 1 j = 1, 2, ..., 6
xj ≥ 0 j = 1, 2, ..., 6
The OR team used this model1 to find a minimum-cost plan
(x1 , x2 , x3 , x4 , x5 , x6 ) = (1, 0.623, 0.343, 1, 0.048, 1)
, with Z = 32.16 (total annual cost of 32.16 million$).

5.3 Some Practice Problems

Question: The city council of Fayetteville is in the process of approving the con-
struction of a new 200, 000f t2 convention center. Two sites have been proposed,
and both require exercising the "eminent domain" law to acquire the property. The
following table provides data about proposed (contiguous) properties in both sites
together with the acquisition cost.

Partial acquisition of property is allowed. At least 75% of property 4 must be

acquired if site 1 is selected, and at least 50% of property 3 must be acquired if site
2 is selected. Although site 1 property is more expensive (on a per fe basis), the
construction cost is less than at site 2, because the infrastructure at site 1 is in a
much better shape. Construction cost is 25 million$ at site 1 and 27 million$ at site
2. Which site should be selected, and what properties should be acquired?
Chapter 6

Lecture No. 06

6.1 The essence of simplex method

The Simplex method is developed by George Dantzig in 1947.
The Wyndor Glass Co:

Figure 6.1: CPF and their adjacent CPFs

Figure 6.2: Graphical solution and CPFs

Two of its CPF solutions are adjacent if they share one constraint boundary.
Optimality test: Consider any linear programming problem that possesses at
least one optimal solution. If a CPF solution has no adjacent CPF solutions that
are better (as measured by z), then it must be an optimal solution.
6.2. The Simplex Method 33

Figure 6.3: The essence of simplex method graphically

6.2 The Simplex Method

The simplex method does (from a geometric viewpoint) to solve the Wyndor Glass
Co. problem.
Initialization: Choose (0, 0) as the initial CPF solution to examine. (This is a
convenient choice because no calculations are required to identify this CPF solution.)
Optimality Test: Conclude that (0, 0) is not an optimal solution. (Adjacent CPF
solutions are better.)

The Key Solution Concepts:

Concept I: The simplex method focuses solely on CPF solutions. For any problem
with at least one optimal solution, finding one requires only finding a best CPF
solution.
Concept II: The simplex method is an iterative algorithm.
Concept III: Whenever possible, the initialization of the simplex method chooses
the origin (all decision variables equal to zero) to be the initial CPF solution. When
there are too many decision variables to find an initial CPF solution graphically,
this choice eliminates the need to use algebraic procedures to find and solve for an
initial CPF solution.
Concept IV: Given a CPF solution, it is much quicker computationally to gather in-
formation about its adjacent CPF solutions than about other CPF solutions. There-
fore, each time the simplex method performs an iteration to move from the current
CPF solution to a better one, it always chooses a CPF solution that is adjacent
to the current one. No other CPF solutions are considered. Consequently, the en-
tire path followed to eventually reach an optimal solution is along the edges of the
feasible region.
Concept V: After the current CPF solution is identified, the simplex method ex-
amines each of the edges of the feasible region that emanate from this CPF solution.
Each of these edges leads to an adjacent CPF solution at the other end, but the
6.3. LP Model in Equation Form 34

simplex method does not even take the time to solve for the adjacent CPF solution.
Instead, it simply identifies the rate of improvement in z that would be obtained by
moving along the edge.
Concept VI: Solution concept 5 describes how the simplex method examines each
of the edges of the feasible region that emanate from the current CPF solution. This
examination of an edge leads to quickly identifying the rate of improvement in Z
that would be obtained by moving along the edge toward the adjacent CPF solution
at the other end.
Therefore, the optimality test consists simply of checking whether any of the
edges give a positive rate of improvement in Z. If none do, then the current CPF
solution is optimal.

6.3 LP Model in Equation Form

The simplex method computations require the following to restrictions on the con-
straints:

1. All the constraints (with the exception of the nonnegativity of the variables)
are equations with nonnegative right-hand side.

2. All the variables are nonnegative.

Converting Inequalities into equations with nonnegative right hand side:

To convert a ≤-inequality to an equation, a nonnegative slack variable is added
to the left-hand side of the constraint.

Example: In the Reddy-Mikks model, the constraint associated with the use of
raw material M 1 is given as
6x1 + 4x2 ≤ 24.
Defining sl as the slack or unused amount of M 1, the constraint can be converted
to the following equation:

6x1 + 4x2 + s1 = 24, s1 ≥ 0.

To convert a ≥-inequality to an equation, a nonnegative surplus variable is

subtracted from the left-hand side of the constraint.

Example: In the Diet problem, the constraint representing the minimum feed
requirements is
x1 + x2 ≥ 800.
Defining Sl as as the surplus variable, the constraint can be converted to the
following equation
x1 + x2 − S1 ≥ 800, S1 ≥ 0
6.4. Algebraic Determination of Corner Points 35

The only remaining requirement is for the right-hand side of the resulting equa-
tion to be nonnegative. The condition can always be satisfied by multiplying both
sides of the resulting equation by −1, where necessary.

Example: The constraint −x1 + x2 ≤ −3 can be transformed into

x1 − x2 − s1 = 3.

Example: Show how the following objective function can be presented in equation
form:
Minimize = max{|x1 − x2 + 3x3 |, | − x1 + 3x2 − x3 |}.

6.4 Algebraic Determination of Corner Points

In a set of m × n equations (m < n), if we set n − m variables equal to zero and
then solve the m equations for the remaining m variables, the resulting solution, if
unique, is called a basic solution and must correspond to a (feasible or infeasible)
corner point of the solution space.

This means that the maximum number of corner points is

n n!
Cm =
m!(n − m)!

Example: Consider the following LP with two variables:

Maximize z = 2x1 + 3x2

subject to

2x1 + x2 ≤ 4
x1 + 2x2 ≤ 5
x1 , x2 ≥ 0

The standard form of the constraints is

2x1 + x2 + s1 = 4
x1 + 2x2 + s2 = 5
x1 , x2 , s1 , s2 ≥ 0
6.4. Algebraic Determination of Corner Points 36

Figure 6.4: Graphical solution

Maximize z = 2x1 + 3x2

Set x1 = 0, x2 = 0 the basic solution will be s1 = 4, s2 = 5. Set s1 = 0, s2 = 0 the
basic solution will be x1 = 1, x2 = 2.

2x1 + x2 + s1 = 4
x1 + 2x2 + s2 = 5
x1 , x2 , s1 , s2 ≥ 0

To summarize the transition from the graphical to the algebraic solution, the
zero n − m variables are known as nonbasic variables. The remaining m variables
are called basic variables and their solution (obtained by solving the m equations)
is referred to as basic solution.

Figure 6.5: Basic and nonbasic variables

Remark: The simplex method alleviates this computational burden dramatically

by investigating only a fraction of all possible basic feasible solutions (corner points)
of the solution space. In essence, the simplex method utilizes an intelligent search
procedure that locates the optimum corner point in an efficient manner.
Slack variables, Augmented form
6.5. Some Practice Problems 37

Figure 6.6: Augmented solution of an LP

An augmented solution is a solution for the original variables (the decision vari-
ables) that has been augmented by the corresponding values of the slack variables.

Example: Augmenting the solution (3, 2) in the example yields the augmented
solution (3, 2, 1, 8, 5) because the corresponding values of the slack variables are
x3 = 1, x4 = 8, and x5 = 5.
A basic solution is an augmented corner-point solution.
A basic feasible (BF) solution is an augmented CPF solution.
The only difference between basic solutions and corner-point solutions (or be-
tween BF solutions and CPF solutions) is whether the values of the slack variables
are included. Thus, the CPF solution (0, 6) in the example is equivalent to the BF
solution (0, 6, 4, 0, 6) for the problem in augmented form.

A basic solution has the following properties

1. Each variable is designated as either a nonbasic variable or a basic variable.

2. The number of basic variables equals the number of functional constraints

(now equations). Therefore, the number of nonbasic variables equals the total
number of variables minus the number of functional constraints.

3. The nonbasic variables are set equal to zero.

4. The values of the basic variables are obtained as the simultaneous solution of
the system of equations (functional constraints in augmented form). (The set
of basic variables is often referred to as the basis.)

5. If the basic variables satisfy the nonnegativity constraints, the basic solution
is a BF solution.

6.5 Some Practice Problems

Question: Consider the following LP with two variables: Maximize z = 2x1 + 3x2
subject to
x1 + 3x2 ≤ 6
3x1 + 2x2 ≤ 6
x1 , x2 ≥ 0
6.5. Some Practice Problems 38

1. Express the problem in equation form.

2. Determine all the basic solutions of the problem, and classify them as feasible
and infeasible.

3. Use direct substitution in the objective function to determine the optimum

basic feasible solution.

4. Verify graphically that the solution obtained in (3) is the optimum LP solution
hence, conclude that the optimum solution can be determined algebraically by
considering the basic feasible solutions only.

5. Show how the infeasible basic solutions are represented on the graphical solu-
tion space.
Chapter 7

Lecture No. 07

Example: Determine the optimum solution for each of the following LPs by enu-
merating all the basic solutions.

1. Maximize z = 2x1 − 4x2 + 5x3 − 6x4 subject to

x1 + 4x2 − 2x3 + 8x4 ≤ 2

−x1 + 2x2 + 3x3 + 4x4 ≤ 1
x1 , x2 , x3 , x4 ≥ 0

2. Minimize z = x1 + 2x2 − 3x3 − 2x4 subject to

x1 + 2x2 − 3x3 + x4 = 4
x1 + 2x2 + x3 + 2x4 = 4
x1 , x2 , x3 , x4 ≥ 0

Also decide algebraically which of the basic solutions are feasible and infeasible.

7.1 Algebraic Steps for the Simplex Method

Consider the objective function Maximize z = 2x1 + 3x2 . The function shows that
an increase in either x1 or x2 (or both) above their current zero values will improve
the value of z.

The design of the simplex method calls for increasing one variable at a time,
with the selected variable being the one with the largest rate of improvement in z.
Entering Variable: The variable which enters the basic solution.

Leaving Variable: The variable which leaves the basic solution.

Figure 7.1: Basic and nonbasic variables for corner points of the figure below
7.1. Algebraic Steps for the Simplex Method 40

Figure 7.2: Graphical solution

Remark: It is important to note that the simplex method moves alongside the
edges of the solution space, which means that the method cannot cut across the
solution space, going from A to C directly.

Example: The Reddy-Mikks model by the simplex method. The problem is

Maximize z = 5x1 + 4x2 + 0s1 + 0s2 + 0s3 + 0s4

subject to

6x1 + 4x2 + s1 = 24 (Raw material M1)

x1 + 4x2 + s2 = 6 (Raw material M2)
−x1 + x2 + s3 = 1 (Market limit)
x2 + s4 = 2 (Demand limit)

x1 , x2 , s1 , s2 , s3 , s4 ≥ 0.
and we write the objective function as z − 5x1 − 4x2 − 0s1 − 0s2 − 0s3 − 0s4 = 0
The starting simplex tableau for the Reddy Mikks problem is

Basic z x1 x2 s1 s2 s3 s4 Solution
z 1 -5 -4 0 0 0 0 0 z-row

s1 0 6 4 1 0 0 0 24 s1 -row
s2 0 1 2 0 1 0 0 6 s2 -row
s3 0 -1 1 0 0 1 0 1 s3 -row
s4 0 0 1 0 0 0 1 2 s4 -row
7.2. Entering and Leaving variable selection 41

7.2 Entering and Leaving variable selection

The entering variable will correspond to the variable with the most negative
coefficient in the objective equation. This rule is referred to as the optimality
condition.

The mechanics of determining the leaving variable from the simplex tableau calls
for computing the nonnegative ratios of the right-hand side of the equations (Solu-
tion column) to the corresponding constraint coefficients under the entering vari-
able.

The minimum nonnegative ratio automatically identifies the current basic vari-
able s1 as the leaving variable and assigns the entering variable x1 the new value of
4.
Feasibility Condition: The rule associated with the ratio computations is referred
to as the feasibility condition because it guarantees the feasibility of the new solution.

Figure 7.3: Graphical interpretation

Basic z x1 x2 s1 s2 s3 s4 Solution
z 1 -5 -4 0 0 0 0 0 z-row

s1 0 6 4 1 0 0 0 24 s1 -row
s2 0 1 2 0 1 0 0 6 s2 -row
s3 0 -1 1 0 0 1 0 1 s3 -row
s4 0 0 1 0 0 0 1 2 s4 -row

Recall: Gauss Elimination method for solving system of linear equations.

Optimality condition: The entering variable is x2 . Decide about the leaving
variable, s2 will leave the basic solution.
7.2. Entering and Leaving variable selection 42

Basic z x1 x2 s1 s2 s3 s4 Solution
z 1 -5 -4 0 0 0 0 0 z-row

s1 0 6 4 1 0 0 0 24 s1 -row
s2 0 1 2 0 1 0 0 6 s2 -row
s3 0 -1 1 0 0 1 0 1 s3 -row
s4 0 0 1 0 0 0 1 2 s4 -row

Basic z x1 x2 s1 s2 s3 s4 Solution
z 1 0 − 23 5
6
0 0 0 20 z-row
2 1
x1 0 1 3 6
0 0 0 4 x1 -row
4
s2 0 0 3
− 16 1 0 0 2 s2 -row
5 1
s3 0 0 3 6
0 1 0 5 s3 -row
s4 0 0 1 0 0 0 1 2 s4 -row

The optimal tableau for the Reddy Mikks problem is

Basic z x1 x2 s1 s2 s3 s4 Solution
2 1
z 1 0 0 3 2
0 0 21 z-row
1
x1 0 1 0 4
− 12 0 0 3 x1 -row
x2 0 0 1 − 81 3
4
0 0 3
2
x2 -row
3
s3 0 0 0 8
− 54 1 0 5
2
s3 -row
1
s4 0 0 0 8
− 34 0 1 1
2
s4 -row

Recommendations: Produce 3 tons of exterior paint daily Produce 1.5 tons of

interior paint daily Daily profit is 21,000$.
Interpretation:The solution also gives the status of the resources. A resource
is designated as scarce if the activities (variables) of the model use the resource
completely. Otherwise, the resource is abundant.
For raw material M1 the slack variable s1 = 0 means that it is scarce.
For raw material M2 the slack variable s2 = 0 means that it is scarce.
Market limit s3 = 52 means that it is abundant.
Demand limit s4 = 12 means that it is abundant.

Remarks:

1. Sensitivity analysis, which deals with determining the conditions that will keep
the current solution unchanged.

2. Post-optimal analysis, which deals with finding a new optimal solution when
the data of the model are changed.
7.3. Summary of the Simplex Method 43

7.3 Summary of the Simplex Method

1. Optimality condition: The entering variable in a maximization (minimiza-
tion) problem is the nonbasic variable having the most negative (positive)
coefficient in the z-row. Ties are broken arbitrarily. The optimum is reached
at the iteration where all the z-row coefficients of the nonbasic variables are
nonnegative (nonpositive).

2. Feasibility condition: For both the maximization and the minimization

problems, the leaving variable is the basic variable associated with the small-
est nonnegative ratio (with strictly positive denominator). Ties are broken
arbitrarily.

3. Gauss Jordan row operations:

• Pivot row: Replace the leaving variable in the Basic column with the
entering variable. New pivot row = Current pivot row ÷Pivot element
• All other rows, including z,
New row = (Current row) - (pivot column coefficient)× (New pivot row)

Steps of the Simplex Method:

1. Determine a starting basic feasible solution.

2. Select an entering variable using the optimality condition. Stop if there is no

entering variable; the last solution is optimal. Else, go to step 3.

3. Select a leaving variable using the feasibility condition.

4. Determine the new basic solution by using the appropriate Gauss-Jordan com-
putations. Go to step 2.

Example: The WYNDOR GLASS Co model is Maximize z = 3x1 + 5x2

subject to x1 ≤ 4 2x2 ≤ 12
3x1 + 2x2 ≤ 18 x1 , x2 ≥ 0
7.4. Some Practice Problems 44

Figure 7.4: Simplex tableau iterations for WYNDOR GLASS problem

7.4 Some Practice Problems

Question: Solve the following LP models by Simplex Method:

1. Maximize z = 2x1 − 4x2 + 5x3 − 6x4 subject to

x1 + 4x2 − 2x3 + 8x4 ≤ 2

−x1 + 2x2 + 3x3 + 4x4 ≤ 1
x1 , x2 , x3 , x4 ≥ 0

2. Minimize z = x1 + 2x2 − 3x3 − 2x4 subject to

x1 + 2x2 − 3x3 + x4 ≤ 4
x1 + 2x2 + x3 + 2x4 ≤ 4
x1 , x2 , x3 , x4 ≥ 0

Question: Solve all the problems we have considered so far in this course by simplex
method (with two variables).