Sultan Moulay Slimane University

National School of Business and

Management Beni Mellal

Course : Operations research

Professor : Hajar SABIKI

Table of contents

1 History and Introduction

2 Introduction to Linear Programming

1 / 43
History and Introduction
History
In the decades after the two world wars, the tools of operations research were more
widely applied to problems in business, industry, and society. Since that time,
operational research has expanded into a field widely used in industries ranging from
petrochemicals to airlines, finance, logistics, and government, moving to a focus on the
development of mathematical models that can be used to analyse and optimize complex
systems, and has become an area of active academic and industrial research.
In the 17th century, mathematicians Blaise Pascal and Christiaan Huygens solved
problems involving complex decisions (problem of points).
The origin of modern operations research with economic order quantity developed
by Ford W. Harris in 1913.
In 1939, the Russian mathematician Leonid Vitálievich Kantoróvich and the
Dutchman Tjalling Charles Koopmans created the mathematical theory called
"Linear Optimization", for which they received the Nobel Prize in Economics.
History and Introduction 2 / 43
History and Introduction
History
Operations research was born during the Second World War from the combined
efforts of eminent mathematicians (including von Neumann, Dantzig, Blackett) who
were asked to provide techniques for optimizing military resources. The first success
of this approach was obtained in 1940 by the Nobel Prize in Physics Patrick
Blackett who solved a problem of optimal location of surveillance radars.
From the 1950s, operational research entered companies. From the mid-1990s,
there was a strong comeback of OR, with computer tools now equal to the methods
proposed by operational research. Since then, we have witnessed an explosion in the
number of commercial software and the appearance of numerous consultancy
firms : Ilog (65 million euros turnover), Artelys (1.6 million euros turnover),
Dash-Optimization (acquired early 2008 for 32 million dollars by Fair Isaac), IBM
Optimization and many others listed by (Institute of Operations Research and
Management Science).

History and Introduction 3 / 43

Operations Research : Overview

Operation Research is a relatively new discipline. The contents and the boundaries of
the OR are not yet fixed. Therefore, to give a formal definition of the term Operations
Research is a difficult task. The OR starts when mathematical and quantitative
techniques are used to substantiate the decision being taken. The main activity of a
manager is the decision making. In our daily life we make the decisions even without
noticing them. The decisions are taken simply by common sense, judgment and expertise
without using any mathematical or any other model in simple situations. But the
decision we are concerned here with are complex and heavily responsible. Examples are
public transportation network planning in a city having its own layout of factories,
residential blocks or finding the appropriate product mix when there exists a large
number of products with different profit contributions and production requirement.

History and Introduction 4 / 43

Operations Research Definitions
Definitions
Cambridge Dictionary
Operational research, UK (US operations research). The systematic study of how
best to solve problems in business and industry.

Wikipedia
Operations research, , operational research, or simply OR, is the use of mathematical
models, statistics and algorithms to aid in decision-making.

According to the Operational Research Society of Great Britain

Operational Research, is the attack of modern science on complex problems arising in
the direction and management of large systems of men, machines, materials and money
in industry, business, government and defense.

History and Introduction 5 / 43

Operational Research : Applications

Various fields of application

Design, configuration and operation of complex technical systems :
Communication networks, information systems.
Supply chain management : transport, production, inventory. . .
Strategic investment management and also health, public education, roads,
mail collection and distribution, energy production and transmission,
telecommunications, banks, insurance. .
Production : Maximize production based on labor availability labour, market
demand, production capacity, price of comes back from the raw material. . .
Transport : Minimize total distance traveled according to the amount of materials
to be transported, capacity of transporters, points of refueling. . .

History and Introduction 6 / 43

Operations Research

Facing a practical problem of decision

Mathematical aspects
1 Constraints, Objectives, Simplifications
Modelization
1 Graphs, Linear Programming...
Model analysis and resolution
1 Complexity study : what can we expect for the
given resolution time ?
2 Development of algorithms.
3 Implementation and analysis of results :
validate against the request, iterate with the
requestor if necessary.
4 Deployment of solutions.

History and Introduction 7 / 43

Introduction to Linear
Programming

8 / 43
Introduction to Linear Programming
Although reality is often far from linear, many problems can be written in the form
linear, either directly or as a first simplification. On the other hand, a very large number
of models constitute extensions of linear programs. Understanding it is essential to
understanding more models. sophisticated.

Definition to Linear Programming (LP) :

Linear Programming is a mathematical technique for optimizing (maximizing or
minimizing) a linear objective function under constraints in the form of linear
inequalities. It aims to select among different actions the one that will most likely
achieve the desired objective.

Robert DORFMAN and Paul Samuelson, add that Linear programming is a method of
determining the best course of action to achieve given objectives in a situation where
resources are limited. It is therefore a method of solving the economic problem, either
within the framework of a global economy, or within that of the public sector, or within
a particular company.
9 / 43
Introduction to Linear Programming

Modeling :
A model, as considered in this course, is a mathematical construction used to represent
some significant aspects of real-world problems. There are many different types of
mathematical models, but we will focus first on optimization models. There are three
components principles in an optimization model :
1 Variables : these represent the components of the model that can be modified to
create configurations different.
2 Constraints : they represent the limitations on the variables.
3 Objective function : this function assigns a value to each different configuration.
The term "objective" comes from the fact that the objective is to optimize this
function.

10 / 43
Introduction to Linear Programming-Graphical Method

Objectives :
After studying this lesson, you should be able to :
Formulate Linear Programming Problem
Identify the characteristics of linear programming problem
Make a graphical analysis of the linear programming problem
Solve the problem graphically
Identify the various types of solutions

11 / 43
Solving Linear Program-Graphical Method

Formulation problem of Linear programming :

The linear programming problem formulation is illustrated through a product mix
problem. The product mix problem occurs in an industry where it is possible to
manufacture a variety of products. A product has a certain margin of profit per unit, and
uses a common pool of limited resources. In this case the linear programming technique
identifies the products combination which will maximize the profit subject to the
availability of limited resource constraints.

12 / 43
Solving Linear Program-Graphical Method

Example 1 :
A company manufactures two models of armchairs A and B. We have the following
forecast elements for year N :

Model A Model B
Possible sales in quantities 5000 units 3000 units
Assembling time per unit 1 hour 2hours
Margin on variable cost per unit 200 DH 350 DH

The factory’s production capacity is 9,000 assembly hours per year.

Objective : Seek the best production schedule allowing the highest overall margin.

13 / 43
Solving Linear Program-Graphical Method
Example 1 :
Assembly time required : (5000u × 1h) + (3000u × 2h) = 11000 hours.
Maximum production capacity of factory : 9000 hours.
Market constraints : A 6 5000 and B 6 3000.
Production constraints : A + 2B 6 9000.
Maximum margin : M.
Objective function: M = 200A + 350B.
Margin per hour : For A : 200 DH/1h=200 DH. and For B : 350 DH/2h=175 DH.
It is necessary to promote the production of A as much as possible because this product
releases the biggest margin.
Let 5000 hours ⇒ be 5000 units of A. For the production of B it will remain :
4000 hours ⇒ 2000 units of B ⇐ (9000-5000)/2h.
Maximum Margin : M,
M = (200DH × 5000u) + (175DH × 2000u) = 1700000DH.

14 / 43
Solving Linear Program-Graphical Method
Example 2 :
A company manufactures two products A and B in two specialized workshops X and Y .
The characteristics per product are given below :

Workshop X Workshop Y
Product A 3 hours 4 hours
Product B 5 hours 3 hours
Activity capacity per day 1500 hours 1200 hours
Contribution Margin unit variable For A : 1000DH For B : 500DH

For commercial reasons, the production of products A (the most profitable) cannot
exceed 200 units per day.

Objective :
1 Format the problem and the constraints.

2 Propose a mathematical solution.

15 / 43
Solving Linear Program-Graphical Method
Example 2 :
Problem and Constraint Formatting
A expresses the quantity of product A.
B expresses the quantity of product B.
Production constraints include the following :
3A + 5B 6 1500 ⇐ Constraint of workshop X .
4A + 3B 6 1200 ⇐ Constraint of workshop Y .
A 6 200, A > 0 and B > 0.
Math Solving
Search
 for the technical
 optimum (Full employment of the two workshops) :
3A + 5B = 1500 × 4 ⇒ 12A + 20B = 6000.
 
4A + 3B = 1200 × −3 ⇒ −12A − 9B = −3600.
Which gives 11B = 2400 ⇒ B = 218 units, and A = 136 units.

16 / 43
Solving Linear Program-Graphical Method

Example 3 :
Suppose an industry is manufacturing tow types of products P1 and P2. The profits per
Kg of the two products are 3$ and 2$ respectively. These two products require
processing in three types of machines. The following table shows the available machine
hours per day and the time required on each machine to produce one Kg of P1 and P2.
Formulate the problem in the form of linear programming model.

Profit/Kg P1 3$ P2 2$ DH Total Available Machine hours/day

Machine 1 1 1 80
Machine 2 2 1 100
Machine 3 1 0 40

17 / 43
Solving Linear Program-Graphical Method

Solution of example 3 :
The procedure for linear programming problem formulation is as follows : Introduce the
decision variable as follows : Let x1 = amount of P1 and x2 = amount of P2.
In order to maximize profits, we establish the objective function as 3x1 + 2x2 .
Since one Kg of P1 requires 1 hour of processing time in machine 1 while the
corresponding requirement of P2 is 1 hour. So, the first constraint can be expressed as :

x1 + x2 6 80

Similarly, corresponding to machine 2 and 3 the constraints are

2x1 + x2 6 100

x1 + 0x2 6 40.

18 / 43
Solving Linear Program-Graphical Method

Solution of example 3 :
In addition to the above there is no negative production, which may be represented
algebraically as
x1 > 0; x2 > 0.
Thus, the product mix problem in the linear programming model is as follows :
Maximize 3x1 + 2x2 .
Subject to :
x1 + x2 6 80
2x1 + x2 6 100
(1)
x1 6 40
x1 > 0, x2 > 0.

19 / 43
Solving Linear Program-Graphical Method
Solution of example 3 :
Find the feasible region. • Plot each constraint as an equation ≡ line in the plane •
Feasible points on one side of the line plug in (0, 0) to find out which

20 / 43
Solving Linear Program-Graphical Method

Solution of example 3 :

21 / 43
Solving Linear Program-Graphical Method

A corner (extreme) point X of the region R ≡ every line through X intersects R in a

segment whose one endpoint is X . Solving a linear program amounts to finding a best
corner point by the following theorem.
Theorem 1 :
If a linear program has an optimal solution, then it also has an optimal solution that is a
corner point of the feasible region.

22 / 43
Solving Linear Program-Graphical Method
Exercise 1 :
Try to find all corner points. Evaluate the objective function 3x1 +2x2 at those points.

23 / 43
Solving Linear Program-Graphical Method

Exercise 1 :
Problem : there may be too many corner points to check. There’s a better way.
Iso-value line ≡ in all points on this line the objective function has the same value.
For our objective 3x1 + 2x2 an iso-value line consists of points satisfying 3x1 + 2x2 = z
where z is some number. Graphical Method (main steps) :
1 Find the feasible region.
2 Plot an iso-value (isoprofit, isocost) line for some value.
3 Slide the line in the direction of increasing value until it only touches the region.
4 Read-off an optimal solution.

24 / 43
Solving Linear Program-Graphical Method
Exercise 1 :

25 / 43
Solving Linear Program-Graphical Method
Exercise 1 :
Optimal solution is (x1 , x2 ) = (20, 60). Observe that this point is the intersection of
two lines forming the boundary of the feasible region. Recall that lines we use to
construct the feasible region come from inequalities (the points on the line satisfy the
particular inequality with equality).
Binding constraint ≡ constraint satisfied with equality.
For solution (20, 60), the binding constraints are x1 + x2 6 80 and 2x1 + x2 6 100
because 20 + 60 = 80 and 2 × 20 + 60 = 100. The constraint x1 6 40 is not binding
because x1 = 20 < 40. The constraint is binding because changing it (a little)
necessarily changes the optimality of the solution. Any change to the binding constraints
either makes the solution not optimal or not feasible.
A constraint that is not binding can be changed (a little) without disturbing the
optimality of the solution we found. Clearly we can change x1 6 40 to x1 6 30 and the
solution (20, 60) is still optimal. We shall discuss this more in-depth when we learn
about Sensitivity Analysis. Finally, note that the above process always yields one of the
following cases
26 / 43
Solving Linear Program-Graphical Method

Theorem 2 :
Every linear program has either
a unique optimal solution, or
multiple (infinity) optimal solutions, or
is infeasible (has no feasible solution), or
is unbounded (no feasible solution is maximal)
.

27 / 43
Linear Programming-Simplex Method

Objectives :
After studying this lesson, you should be able to :
Understand the basics of simplex method
Explain the simplex calculations
Describe various solutions of Simplex Method
Understand two phase and M method.

28 / 43
Linear Programming-Simplex Method
Introduction
The Linear Programming with two variables can be solved graphically. The graphical
method of solving linear programming problem is of limited application in the business
problems as the number of variables is substantially large. If the linear programming
problem has larger number of variables, the suitable method for solving is Simplex
Method. The simplex method is an iterative process, through which it reaches ultimately
to the minimum or maximum value of the objective function.

Simplex Method
The simplex method also helps the decision maker/manager to identify the following :
• Redundant Constraints
• Multiple Solutions
• Unbounded Solution
• Infeasible Problem
29 / 43
Basics of Simplex Method

The basic of simplex method is explained with the following linear programming problem.
maximize 60x1 + 70x2 .
subject to

2x1 + x2 6 300
3x1 + 4x2 6 509
4x1 + 7x2 6 812
x1 , x2 > 0.

30 / 43
Basics of Simplex Method

Solution
First we introduce the variables s3 , s4 , s5 > 0.
So that the constraints becomes equations, thus :
2x1 + x2 + s3 = 300
3x1 + 4x2 + s4 = 509
4x1 + 7x2 + s5 = 812.

Corresponding to the three constraints, the variables s3 , s4 , s5 are called as slack

variables. Now, the system of equation has three equations and five variables. There are
two types of solutions they are basic and basic feasible, which are discussed as follows :

31 / 43
Basics of Simplex Method

Basic Solution
We may equate any two variables to zero in the above system of equations, and then the
system will have three variables. Thus, if this system of three equations with three
variables is solvable such a solution is called as basic solution.
For example suppose we take x1 = 0 and x2 = 0, the solution of the system with
remaining three variables is s3 = 300, s4 = 509 and s5 = 812, this is a basic solution and
the variables s3 , s4 , and s5 are known as basic variables where as the variables x1 , x2 are
known as non-basic variables.
The number of basic solution of a linear programming problem is depends on the
presence of the number of constraints and variables. For example if the number of
constraints is m and the number of variables including the slack variables is n then there
n
are at most Cn−m = Cm n basic solutions.

32 / 43
Basics of Simplex Method

Basic Feasible Solution

A basic solution of a linear programming problem is called as basic feasible solutions if it
is feasible it means all the variables are non-negative. The solution s3 = 300, s4 = 509
and s5 = 812 is a basic feasible solution.

Every basic feasible solution is an extreme point of the convex set of feasible solutions
and every extreme point is a basic feasible solution of the set of given constraints. It is
impossible to identify the extreme points geometrically if the problem has several
variables but the extreme points can be identified using basic feasible solutions. Since
one the basic feasible solution will maximize or minimize the objective function, the
searching of extreme points can be carry out starting from one basic feasible solution to
another.

33 / 43
Simplex Method

The Simplex Method provides a systematic search so that the objective function
increases in the cases of maximization progressively until the basic feasible solution has
been identified where the objective function is maximized.
Simplex Method Computation
This section describes the computational aspect of simplex method. Consider the
following linear programming problem : Maximize 60x1 + 70x2
Subject to :
2x1 + x2 + s3 = 300
3x1 + 4x2 + s4 = 509
4x1 + 7x2 + s5 = 812
x1 , x2 , s3 , s4 , s5 > 0.

34 / 43
Simplex Method

The profit Z = 60x1 + 70x2 i.e. Maximize 60x1 + 70x2 .

The standard form can be summarized in a compact table form as
Simplex Method Computation
In this problem the slack variables s3 , s4 , and s5 provide a basic feasible solution from
which the simplex computation starts. That is s3 = 300, s4 = 509 and s5 = 812. This
result follows because of the special structure of the columns associated with the slacks.
If Z represents profit then Z = 0 corresponding to this basic feasible solution.

35 / 43
Simplex Method

Simplex Method Computation

So that the numerical values of the basic variables are : C1 = 300, C2 = 509, C3 = 812.
The profit Z = 60x1 + 70x2 can also expressed as Z − 60x1 − 70x2 = 0. The simplex
computation starts with the first compact standard simplex table as given below :

Basic
Variable x1 x2 s3 s4 s5 Cj RC

s3 2 1 1 0 0 300 300
s4 3 4 0 1 0 509 509/4=127,25
s5 4 7 0 0 1 812 812/7=116

Z 60 70 0 0 0 0
Table 1

36 / 43
Simplex Method

Simplex Method Computation

In the objective function the coefficients of the variables are equal to 0. The topmost
row of the Table 1 denotes the coefficient of the variables x1 , x2 , s3 , s4 , s5 of the
objective function respectively. The column under x1 indicates the coefficient of x1 in the
three equations respectively. Similarly the remaining column also formed. On seeing the
equation Z = 60x1 + 70x2 we may observe that if either x1 or x2 , which is currently
non-basic is included as a basic variable so that the profit will increase. Since the
coefficient of x2 is higher we choose x2 to be included as a basic variable in the next
iteration.

37 / 43
Simplex Method

Simplex Method Computation

Thus, by bringing x2 a basic variable one of the existing basic variables becomes
non-basic. The question here is How to identify this variable ? The following statements
give the solution to this question. Consider the first equation i.e.2x1 + x2 + s3 = 300.
From this equation 2x 1 + s3 = 300 − x 2 But x1 = 0. Hence, in order that s3 > 0, we
have 300 − x2 > 0 i.e. x2 6 300.
Similarly consider the second equation i.e. 3x1 + 4x2 + s4 = 509 From this equation
3x1 + s4 = 509 − 4x2 But, x1 = 0. Hence, in order that, s4 > 0 we have 509 − 4x2 > 0
i.e. x2 6 509/4.
From this equation 4x1 + s5 = 812 − 7x2 But x1 = 0. Hence, in order that s5 > 0, we
have 812 − 7x2 > 0 i.e. x2 6 812/7 Therefore the three equation lead to x2 6 300,
x2 6 509/9, x2 6 812/7 Thus x2 = min(x2 6 300, x2 6 509/9, x2 6 812/7) it means
x2 = min(x2 6 300/1, x2 6 509/9, x2 6 812/7) = 116 Therefore x2 = 116.

38 / 43
Simplex Method
Simplex Method Computation
Therefore we obtain the elements of the RC column see Table 1.
Using the following rules the Table 2 is computed from the Table 1.
a) As x2 is the incoming basic variable we make the coefficient of x2 one
by dividing each element of row-3 by 7. Thus the numerical value of
the element corresponding to x1 is 4/7, corresponding to s5 is 1/7 in
Table 2.
b) all the variables of the pivot column become 0 except the pivot
c) if the pivot row crosses a 0, we copy the column
d) The incoming basic variable should appear only in the third row. So
we multiply the third-row of Table 2 by 1 and subtract it from the
first-row of Table 1 element by element. Thus the element
corresponding to x2 in the first-row of Table 2 is 0.

39 / 43
Simplex Method
Simplex Method Computation
Therefore the element corresponding to x1 is 2 − 1 ∗ 4/7 = 10/7 and the element
corresponding to s5 is 0 − 1 ∗ 1/7 = −1/7.
In this way we obtain the elements of the first and the second row in table 2. In table 2
the numerical values can also be calculated in a similar way.

Basic
Variable x1 x2 s3 s4 s5 Cj RC

s3 10/7 0 1 0 -1/7 184 128,8

s4 5/7 0 0 1 -4/7 45 63
x2 4/7 1 0 0 1/7 812/7=116 203

Z 20 0 0 0 -70/7 -8120
Table 2
40 / 43
Simplex Method

Simplex Method Computation

Note that zj = 20 > 0 , so that the objective function can be improved so like Table 2,
the Table 3 is computed sing the rules (a), (b), (c) and (d) as described above

Basic
Variable x1 x2 s3 s4 s5 Cj RC

s3 0 0 1 -2 1 94 94
x1 1 0 0 7/5 -4/5 63 -315/7
x2 0 1 0 -4/5 3/5 80 400/3

Z 0 0 0 -28 6 -9380
Table 3

41 / 43
Simplex Method

Simplex Method Computation

Since z5 = 6 > 0 So from the table 3, table 4 is calculated following the usual steps.

Basic
Variable x1 x2 s3 s4 s5 Cj RC

x3 0 0 1 -2 1 94
x1 1 0 4/5 7/5 -1/5 0
x2 0 1 -3/5 2/5 0 118/5

Z 0 0 -6 -16 0 -9944
Table 4

42 / 43
Linear Programming-Simplex Method

Simplex Method Computation

Note that zj 6 0 for all j, so that the objective function can’t be improved any further.
Thus, the objective function is maximized for x1 = 691/5 and x2 = 118/5 and The
maximum value of the objective function is 9944.

43 / 43