0% found this document useful (0 votes)
80 views190 pages

Operations Research PDF

The document provides an overview of operations research (OR) including: 1) OR aims to optimize real-world systems represented as mathematical models. It was first used in World War II and later applied to business. 2) An OR model includes decision variables, constraints, and an objective function. It seeks values for variables that optimize the objective within constraints. 3) A 7-step OR process includes defining the problem, observing the system, modeling the problem mathematically, verifying the model, selecting alternatives, presenting results, and implementing recommendations.

Uploaded by

ola nasser
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
80 views190 pages

Operations Research PDF

The document provides an overview of operations research (OR) including: 1) OR aims to optimize real-world systems represented as mathematical models. It was first used in World War II and later applied to business. 2) An OR model includes decision variables, constraints, and an objective function. It seeks values for variables that optimize the objective within constraints. 3) A 7-step OR process includes defining the problem, observing the system, modeling the problem mathematically, verifying the model, selecting alternatives, presenting results, and implementing recommendations.

Uploaded by

ola nasser
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 190

Aswan University

Faculty of Commerce

Department of Statistics

Operations Research
LECTURE NOTES

Prepared by

Dr.Mohamed Abd el Sattar

0
List of Contents

Chapter one : Introduction 2


Chapter Two: Linear Programming(Simplex method) 21
Chapter Three: Transportation Problem 94
Chapter Four: Network Analysis 140
References 192

1
Chapter One
What Is Operations Research?

(1-1): Introduction:

The first formal activities of Operations Research (OR) were


initiated in England during World War II , when a team of
British scientists set out to make scientifically based decisions
regarding the best utilization of war materiel, After the war, the
ideas advanced in military operations were adapted to improve
efficiency and productivity in the civilian sector.

(1-2): Operations Research Models:

Consider forming a maximum - area rectangle out of a


piece of wire of length L inches, What should be the best width
and height of the rectangle?

We can look at the situation as a decision – making


problem whose solution requires answering three questions:

1. What are the decision variables?

2. Under what restrictions is the decision made?

3. What is an appropriate objective criterion for


evaluating the variables?

2
To formalize this observation, the decision variables;
namely, the width and height of the rectangle are identified by
defining the width and height algebraically as:

W= width of the rectangle in inches

H= height of the rectangle in inches

Based on these definitions, the restrictions of the


situation can be expressed verbally as:

 Width of rectangle +Height of rectangle = Half the


length of the wire

 Width and height cannot be negative

These restrictions are translated algebraically as:

1. ( )
2.

The only remaining component now is the objective of the


problem; namely, maximization of the area of the rectangle. Let
z be the area of the rectangle; then the complete model
becomes:

( )

3
The general OR model can be organized in the following
general format:

Subject to:

Constraints

A solution of the model is feasible if it satisfies all the


constraints. It is optimal if, in addition to being feasible, it
yields the best (maximum or minimum) value of the objective
function. When OR is used to solve a problem of an
organization, the following seven step procedure should be
followed:

Step 1. Formulate the Problem

OR analyst first defines the organization's problem. Defining


the problem includes specifying the organization's objectives
and the parts of the organization (or system) that must be
studied before the problem can be solved.

Step 2. Observe the System

Next, the analyst collects data to estimate the values of


parameters that affect the organization's problem. These
estimates are used to develop (in Step 3) and evaluate (in Step
4) a mathematical model of the organization's problem.

4
Step 3. Formulate a Mathematical Model of the Problem

The analyst, then, develops a mathematical model (in other


words an idealized representation) of the problem. In this class,
we describe many mathematical techniques that can be used to
model systems.

Step 4. Verify the Model and Use the Model for Prediction

The analyst now tries to determine if the mathematical model


developed in Step 3 is an accurate representation of reality. To
determine how well the model fits reality, one determines how
valid the model is for the current situation.

Step 5. Select a Suitable Alternative

Given a model and a set of alternatives, the analyst chooses the


alternative (if there is one) that best meets the organization's
objectives.

Sometimes the set of alternatives is subject to certain


restrictions and constraints. In many situations, the best
alternative may be impossible or too costly to determine.

Step 6. Present the Results and Conclusions of the Study

In this step, the analyst presents the model and the


recommendations from Step 5 to the decision making
individual or group. In some situations, one might present
several alternatives and let the organization choose the decision
maker(s) choose the one that best meets her/his/their needs.
5
After presenting the results of the OR study to the decision
maker(s), the analyst may find that s/he does not (or they do
not) approve of the recommendations. This may result from
incorrect definition of the problem on hand or from failure to
involve decision maker(s) from the start of the project. In this
case, the analyst should return to Step 1, 2, or 3.

Step 7. Implement and Evaluate Recommendation

If the decision maker(s) has accepted the study, the analyst aids
in implementing the recommendations. The system must be
constantly monitored (and updated dynamically as the
environment changes) to ensure that the recommendations are
enabling decision maker(s) to meet her/his/their objectives.

1.3 BASIC OR CONCEPTS


"OR is the representation of real-world systems by
mathematical models together with the use of quantitative
methods (algorithms) for solving such models, with a view to
optimizing."

We can also define a mathematical model as consisting of:

 Decision variables, which are the unknowns to be


determined by the solution to the model.

 Constraints to represent the physical limitations of the


system

 An objective function
6
 An optimal solution to the model is the identification of a
set of variable values which are feasible (satisfy all the
constraints) and which lead to the optimal value of the
objective function.

An optimization model seeks to find values of the decision


variables that optimize (maximize or minimize) an objective
function among the set of all values for the decision variables
that satisfy the given constraints.

The Reddy Mikks Company

Reddy Mikks produces both interior and exterior paints from


two raw materials M1 and M2.The following table provides the
basic data of the problem

Exterior paint Interior paint Maximum daily

Availability (tons)

M1 6 4 24

M2 1 2 6

Profit per 5 4
ton($1000)

A market survey indicates that the daily demand for interior


paint cannot exceed that for exterior paint by more than 1 tons.
Also, the maximum daily demand for interior paint is 2 tons.
Reddy Mikks wants to determine the optimum (best) product
7
mix of interior and exterior paints that maximizes the total daily
profit

All OR models, LP included, consist of three basic components

1. Decision variables that we seek to determine.

2. Objective function that we need to optimize (maximizes


or minimizes).

3. Constraints that the solution must satisfy.

The proper definition of the decision variables is an essential


first step in the development of the model. Once done, the task
of constructing the objective function and the constraints
becomes more straightforward.

For the Reddy Mikks problem, we need to determine the


daily amounts to be produced of exterior and interior paints.
Thus the variables of the model are defined as

The goal of Reddy Mikks is to maximize (i.e., increase as much


as possible) the total daily profit of both paints. The two
components of the total daily profit are expressed in terms of
the variables as:

Profit from exterior paint ( )

Profit from interior paint ( )

8
Let Z represent the total daily profit (in thousands of dollars),
the objective (or goal) of Reddy Mikks is expressed as

Next, we constraints that restrict raw material usage and


product demand. The raw material restrictions are expressed
verbally as:

( ) ( )

The daily usage of raw material is 6 tons per ton of


exterior paint and 4 tons per ton of interior paint. Thus

In a similar manner,

The maximum daily availabilities of raw materials

Thus, the raw material constraints are

( )

( )

The first restriction on product demand stipulates that the daily


production of interior paint cannot exceed that of exterior paint
by more than 1 ton, which translates to

( )
9
The second restriction limits the daily demand of interior paint
to 2 tons – that is,

( )

Nonnegative constraints: This restriction is expressed as

The complete Reddy Mikks model is

Two Mines Company

Two Mines Company own two different mines that produce an


ore which, after being crushed, is graded into three classes:
high, medium and low-grade. The company has contracted to
provide a smelting plant with 12 tons of high-grade, 8 tons of
medium-grade and 24 tons of low-grade ore per week. The two
mines have different operating characteristics as detailed
below.

10
Mine Cost per day £ Production (tons/day)

High Medium Low


X 180 6 3 4
Y 160 1 1 6
Consider that mines cannot be operated in the weekend. How
many days per week should each mine be operated to fulfill the
smelting plant contract?

Solution

This process is often called formulating the problem (or more


strictly formulating a mathematical representation of the
problem).

Variables

These represent the "decisions that have to be made" or the


"unknowns". We have two decision variables in this problem:

x = number of days per week mine X is operated

y = number of days per week mine Y is operated

Note here that x ≥ 0 and y ≥ 0.

Constraints

High 6x + 1y ≥ 12

Medium 3x + 1y ≥ 8

Low 4x + 6y ≥ 24

11
days per week constraint - we cannot work more than a
certain maximum number of days a week e.g. for a 5 day
week we have

x≤5

y≤5

Objective

Again in words our objective is (presumably) to minimize


cost which is given by 180x + 160y

Hence we have the complete mathematical representation of


the problem:

Minimize 180x + 160y

Subject to

6x + y ≥ 12

3x + y ≥ 8

4x + 6y ≥ 24

X ≤5

Y ≤5

x, y ≥ 0

12
LINEAR PROGRAMMING

It can be recalled from the Two Mines example that the


conditions for a mathematical model to be a linear program (LP)
were:

 all variables continuous (i.e. can take fractional values)

 a single objective (minimize or maximize)

 The objective and constraints are linear i.e. any term is


either a constant or a constant multiplied by a unknown.

LP's are important - this is because:

 many practical problems can be formulated as LP's

 there exists an algorithm (called the simplex algorithm)


which enables us to solve LP's numerically relatively
easily

We will return later to the simplex algorithm for solving LP's


but for the moment we will concentrate upon formulating LP's.

Some of the major application areas to which LP can be applied


are:

 Work scheduling

 Production planning & Production process

 Capital budgeting

 Financial planning

 Blending (e.g. Oil drilling and production )


13
 Farm planning

 Distribution

 Multi-period decision problems

o Inventory model

o Financial models

o Work scheduling

Note that the key to formulating LP's is practice. However a


useful hint is that common objectives for LP's are maximized
profit/minimized cost.

1.4 FORMULATING LP

1.4.1 Giapetto Example

Giapetto's wooden soldiers and trains. Each soldier sells for


$27, uses $10 of raw materials and takes $14 of labor &
overhead costs. Each train sells for $21, uses $9 of raw
materials, and takes $10 of overhead costs. Each soldier needs
2 hours finishing and 1 hour carpentry; each train needs 1 hour
finishing and 1 hour carpentry. Raw materials are unlimited,
but only 100 hours of finishing and 80 hours of carpentry are
available each week. Demand for trains is unlimited; but at
most 40 soldiers can be sold each week. How many of each toy
should be made each week to maximize profits?

Answer

14
Decision variables completely describe the decisions to be
made (in this case, by Giapetto). Giapetto must decide how
many soldiers and trains should be manufactured each week.
With this in mind, we define:

x1= the number of soldiers produced per week

x2= the number of trains produced per week

Objective function is the function of the decision variables that


the decision maker wants to maximize (revenue or profit) or
minimize (costs). Giapetto can concentrate on maximizing the
total weekly profit (z).

Here profit equals to (weekly revenues) – (raw material


purchase cost) – (other variable costs). Hence Giapetto’s
objective function is:

Maximize z = 3x1 + 2x2

Constraints show the restrictions on the values of the decision


variables. Without constraints Giapetto could make a large
profit by choosing decision variables to be very large. Here
there are three constraints:

Finishing time per week

Carpentry time per week

Weekly demand for soldiers

Sign restrictions are added if the decision variables can only

15
assume nonnegative values (Giapetto can not manufacture
negative number of soldiers or trains!)

All these characteristics explored above give the following


Linear Programming

(LP) model

Max Z = 3x1+2x2 (The Objective function)

s.t. 2x1 + x2 ≤100 (Finishing constraint)

x1+ x2 ≤80 (Carpentry constraint)

x1 ≤ 40 (Constraint on demand for soldiers)

x1, x2 >0 (Sign restrictions)

A value of (x1, x2) is in the feasible region if it satisfies all the


constraints and sign restrictions.

Graphically and computationally we see the solution is (x1, x2)


= (20, 60) at which z=180. (Optimal solution)

1.4.2 Advertisement Example

Dorian makes luxury cars and jeeps for high-income men and
women. It wishes to advertise with 1 minute spots in comedy
shows and football games. Each comedy spot costs $50K and is
seen by 7M high-income women and 2M high-income men.
Each football spot costs $100K and is seen by 2M high-income
women and 12M high-income men. How can Dorian reach
28M high-income women and 24M high- income men at the
16
least cost?

Answer

The decision variables are

x1= the number of comedy spots

x2= the number of football spots

The model of the problem:

Min z = 50 x1 + 100 x2

st 7x1 + 2x2 ≥ 28

2x1 +12x2 ≥ 24

x1, x2 ≥ 0

The graphical solution is z = 320 when (x1, x2) = (3.6, 1.4).


From the graph, in this problem rounding up to (x1, x2) = (4, 2)
gives the best integer solution.

1.4.3 (Production planning )

Suppose an industry is manufacturing two types of


products P1 and P2. The profits per Kg of the two products are
Rs.30 and Rs.40 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.

17
Profit/Kg P1 P2 Total available Machine

Rs.30 Rs.40 hours/day

Machine 1 3 2 600

Machine 2 3 5 800

Machine 3 5 6 1100

Solution:

The procedure for linear programming problem formulation is


as follows:

Introduce the decision variable as follows:

Let x1 = amount of P1

x2 = amount of P2

In order to maximize profits, we establish the objective function


as

30x1 + 40x2

Since one Kg of P1 requires 3 hours of processing time in


machine 1 while the corresponding requirement of P2 is 2 hours.
So, the first constraint can be expressed as

3x1 + 2x2 ≤ 600

Similarly, corresponding to machine 2 and 3 the constraints are

3x1 + 5x2 ≤ 800


18
5x1 + 6x2 ≤ 1100

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

30x1 + 40x2

Subject to:

3x1 + 2x2 ≤ 600

3x1 + 5x2≤ 800

5x1 + 6x2 ≤ 1100

x1≥ 0, x2 ≥ 0

1.4.4 (Product mix problem)

A company owns two flour mills viz. A and B, which have


different production capacities for high, medium and low quality
flour. The company has entered a contract to supply flour to a
firm every month with at least 8, 12 and 24 quintals of high,
medium and low quality respectively. It costs the
companyRs.2000 and Rs.1500 per day to run mill A and B
respectively. On a day, Mill A produces 6, 2 and 4quintals of
high, medium and low quality flour, Mill B produces 2, 4 and 12
19
quintals of high, medium and low quality flour respectively.
How many days per month should each mill be operated in order
to meet the contract order most economically.

Solution:

Let us define x1 and x2 are the mills A and B. Here the objective
is to minimize the cost of the machine runs and to satisfy the
contract order. The linear programming problem is given by

Minimize

2000x1 + 1500x2

Subject to:

6x1 + 2x2≥ 8

2x1 + 4x2≥ 12

4x1 + 12x2≥ 24

x1≥ 0, x2≥ 0

20
Chapter two
General Linear Programming Problem

General Linear Programming Problem (GLPP)


Maximize / Minimize Z = c1x1 + c2x2 + c3x3 +………..+ cnxn
Subject to constraints
a11x1 + a12x2 + …..........+a1nxn (≤ or ≥) b1
a21x1 + a22x2 + ………..+a2nxn (≤ or ≥) b2
.
.
.
am1x1 + am2x2 + ……….+amnxn (≤ or ≥) bm
and
x1 ≥ 0, x2 ≥ 0,…, xn≥ 0

Where cj, bi, and aij (i = 1, 2 , .... m; j = 1.2 .... , n) are constants
determined from the technology or the problem, and xj are the
decision variables. Only one sign (≤, =, or ≥) holds for each
constraint. Although all variables are declared nonnegative, the
preceding section shows that every unrestricted variable can be
converted equivalently to nonnegative variables. The non-
negative restriction is essential for the development of the
solution method for linear programming.

After formulating a linear programming model, the


analyst's next step is to solve the model. Because linear
21
programming models are presented in a variety of forms
(maximization or minimization for the objective function and (≤
, = , and / or ≥ ) for the constraints) , it is necessary to modify
these forms to fit the solution procedure that will be presented
in the next chapter. Two forms are introduced for this purpose:
the canonical forms and the standard form. The standard form
is used directly for solving the model; the canonical form is
particularly useful in presenting duality theory. The details of
the two forms are now presented.

1) The canonicaL form:

the general linear programming problem defined above


can always be put in the following form, which will be referred
to as the canonical form:

Subject to :

∑∑

The characteristics of this form are :

1- All decision variables are nonnegative.

22
2- All constraints are of the (≤) type.

3- The objective function is of the maximization type.

A linear programming problem can be put in the canonical form


by the use of five elementary transformations.

1- The minimization of a function, ƒ (x), is mathematically


equivalent to the maximization of the negative expression of
this function, - ƒ (x) , for example, the linear objective function.

Minimize Z = c1 x1 + c2 x2 + ..... + cn xn

Is équivalent to

Maximize g0 = - Z = - c1 x1 - c2x2- ..... - cn xn

with Z = -g0 Consequently, in any linear programming problem


the objective function can be put in the maximization form.

2- An inequality in one direction (≤ or ≥) may be changed to an


inequality in the opposite direction (≥or ≤) by multiplying both
sides of the inequality by (-1). For example, the linear
constraint. 4x1 + 6x2 ≥ 7

Is equivalent to : -4x1 -6x2 ≤ - 7

Also, 2x1 + 3x2 ≤ -5

Is equivalent to : -2x1 - 3x2 ≥ 5

3- An equation maybe replaced by two (weak) inequalities in


opposite directions. For example,
23
a 1 x 1 + a2 x 2 = b

is equivalent to the two simultaneous constraints.

a 1 x 1 + a2 x 2 ≤ b and a1 x 1 + a2 x 2 ≥ b

a 1 x 1 + a2 x 2 ≤ b and -a1x1 - a2x2 ≤ -b

4- An inequality constraint with its left-hand side in the


absolute form can be changed into two regular
inequalities .thus, for b ≥ 0.

│a1x1 + a2 x2 │ ≤ b

is equivalent to

a 1 x 1 + a2 x 2 ≥ - b and a1x1 + a2 x2 ≤ b

Similarly, for q ≥ 0.

│p1x1 + p2 x2 │ ≥ q

is equivalent to either

p1x1 + p2 x2 ≥ q or p 1 x 1 + p2 x 2 ≤ - q

5- A variable which is unconstrained in sign (that is, positive,


negative, or zero) is equivalent to the difference between
two nonnegative variables. Thus, if x is unconstrained in
sign, it can be replaced by ( ̿ ̅ ) where ̅ and
̿ .

24
* Example:

Consider the linear programming problem.

Minimize z = 3x1 – 3x2 + 7x3

Subject to

x1 + x2 + 3x2 ≤ 40

x1 + 9 x2 - 7x3 ≥ 50

5x1 + 3x2 = 20

│5x2 + 8x3 │ ≤ 100

x1 ≥ 0 , x2 ≥ 0

x3 is unconstrained in sign

The problem can be put in the canonical form as follows.

By the second transformation,

x1 + 9 x2 - 7x3 ≥ 50

is equivalent to

-x1 - 9 x2 + 7x3 -50

By the third transformation,

5x1 + 3x2 = 20

is equivalent to

5x1 + 3x2 20

25
And -5x1 - 3x2 -20

By the fourth transformation,

│5x2 + 8x3 │ ≤ 100

is equivalent to

5x2 + 8x3 ≤ 100 and 5x2 + 8x3 ≥ - 100

by the fifth transformation,

̿̿̿ ̅̅̅

Where ̿̿̿ ≥ 0 and ̅̅̅ ≥ 0. Finally, if the objective function is


transformed to maximization, the canonical form becomes.

Maximize g0 = (-x0) = -3x1 + 3x2 – 7 (̿̿̿ ̅̅̅)

Subject to :

x1 + x2 + 3 (̿̿̿ ̅̅̅) ≤ 40

-x1 + 9 x2 + 7(̿̿̿ ̅̅̅) ≤-50

5x1 + 3x2 ≤ 20

-5x1 - 3x2 ≤ -20

5x2 + 8 (̿̿̿ ̅̅̅ ) ≤ 100

-5x2 - 8 (̿̿̿ ̅̅̅) ≤ 100

x1≥ 0 , x2 ≥ 0 , ̿̿̿ ≥ 0 and ̅̅̅ ≥ 0

The only difference between the original and the


canonical forms in the above example occurs in the objective
26
function where x0 in the original problem becomes (-g0) in the
canonical form. The values of the variables are the same in
both cases, however, since the constraints are mathematically
equivalent.

II) the standard form :

The characteristics of the standard form are:

1- All constraints are equations except for the nonnegativity


constraints which remain inequalities ( ≥ 0)

2- The right- hand side element of each constraint equation


is nonnegative.

3- All variables are nonnegative.

4- The objective function is of the maximization or the


minimization type.

Inequality constraints can be changed to equations by


augmenting (adding or subtracting) the left-hand side of each
such constraint by a nonnegative variable. These new variable
are called slack variables and are added if the constraint is (≤)
or subtracted if the constraint is (≥). The right- hand side can be
made always positive by multiplying both sides of the resulting
equation by (-1) whenever necessary. The remaining
characteristics can be realized by using the elements
transformations introduced with the canonical form.
27
To illustrate the concept of the slack variables, the
constraint.

A1x1 + a2x2 ≥ b , b ≥ 0

Is changed in the standard form to

A1x1 + a2x2 – S1= b

Where S1 ≥ 0. Also, the constraint.

p1x1 + p2x2 ≤ q , q ≥ 0

is changed to

P1x1 + p2x2 + S2 = q

Where S2 ≥ 0

The standard form plays an important role in the solution


of the linear programming problem. For example, consider the
following linear program with all (≤) constraints:

Subject to :

∑∑

This is expressed in the standard form as :


28

Subject to :

∑∑

For example, consider the following linear program with all ( )


constraints:

Subject to :

∑∑

This is expressed in the standard form as :

Subject to :

∑∑

29
The standard form basically reduces the linear program to
a set of (m) equations in (m + n) unknowns. A solution to these
equations is of interest only if it is feasible, that is , if it satisfies
the nonnegativity constraints xj ≥ 0 and Si ≥ 0. for all i and j . the
optimum solution is then given by the feasible solution that
maximizes x0.

The idea given above is simple. But the difficulty is that a


set of (m) equations in (m + n) unknowns usually yields infinity
of solutions. Consequently, since it is computationally
impossible to determine every feasible point, a procedure is
needed which locates the optimum after checking a finite
number of solution points. Chapter 2 presents the simplex
method, which is an interactive algorithm proven to converge
to the optimum (when it exists) in a finite number of iteration.

In summary, the following table represents the different


characteristics between each of the canonical and the standard
form as follows:

30
Canonical form Standard form

1- The objective function f(x) 1- The objective function may


must be in the maximization be either in Maximization or in
form. minimization form.

2- each constraint must be in 2-The right hand side for each


the form (≤) constraint (constant) must be
positive

3-each decision variable must 3-Each constraint must be in


be nonnegative. the equation form (=).

4- Each decision variable must


be nonnegative.

Note that :

If there is a constraint in the absolute value form, i.e., we


have as the following form:

│ax1 ± bx2 │< C

Then we have the two forms:

1- If │ax1 ± bx2 │≤ C

i.e., -C ≤ ax1 ± bx2 ≤ C

put y = ax1 ± bx2

31
then : - C ≤ y ≤ C

The intersection between the two arrows which are


represent the area by which satisfied the two sub constraints y
= ax1 ± bx2 ≥ - c and y = ax1 ± bx2 ≤ c means that we must take
the two constraints resulted form the constraint of the
absolute value in the solution.

II – if │ax1 ± bx2 │ ≥ c meaning that

c ≤ ax1 ± bx2 ≤ -c

i.e., c ≤ y ≤ -c, where y = ax1 ± bx2

Since there is no intersection between the two arrows, (i.e.,


there is no feasible solution between the two arrows) , then ,
we have to take only one of the two sub constraints

y = ax1 ± bx2 ≥ c , or y = ax1 ± bx2 ≤ -c in the solution.

Steps to convert GLPP to SLPP (Standard LPP)


Step 1 The objective function can be in the maximization form.
Or minimization form.
Step 2 Convert all inequalities as equations.
o If an equality of ‘≤’ appears then by adding a
variable called Slack variable. We can convert it to
an equation. For example x1 +2x2 ≤ 12, we can write
as
x1 +2x2 + s1 = 12.

32
o If the constraint is of ‘≥’ type, we subtract a variable
called Surplus variable and convert it to an
equation. For example
2x1 +x2 ≥ 15
2x1 +x2 – s2 = 15
Step 3 – The right side element of each constraint should be
made non-negative
2x1 +x2 – s2 = -15
-2x1 - x2 + s2 = 15 (That is multiplying throughout by -1)
Step 4 – All variables must have non-negative values.
For example: x1 +x2 ≤ 3
x1> 0, x2 is unrestricted in sign
Then x2 is written as x2 = ̿̿̿ - ̅̅̅ ‫׳‬where ̿̿̿, ̅̅̅ ≥ 0
Therefore the inequality takes the form of equation as
x1 + (̿̿̿ - ̅̅̅) + s1 = 3
Using the above steps, we can write the GLPP in the form of
SLPP.
Write the Standard LPP (SLPP) of the following
Example
Maximize Z = 3x1 + x2
Subject to
2 x 1 + x2 ≤ 2
3 x1 + 4 x2 ≥ 12
and x1 ≥ 0, x2 ≥ 0
SLPP
33
Maximize Z = 3x1 + x2
Subject to
2 x 1 + x 2 + s1 = 2
3 x1 + 4 x2 – s2 = 12
x1≥ 0, x2 ≥ 0, s1 ≥ 0, s2 ≥ 0
Example
Minimize Z = 4x1 + 2 x2
Subject to
3x1 + x2 ≥ 2
x1 + x2 ≥ 21
x1 + 2x2 ≥ 30
and x1 ≥ 0, x2 ≥ 0
SLPP
Minimize Z = 4x1 + 2 x2
Subject to
3x1 + x2 – s1 = 2
x1 + x2 – s2 = 21
x1 + 2x2 – s3 = 30
x1≥ 0, x2 ≥ 0, s1 ≥ 0, s2 ≥ 0, s3 ≥ 0
Example
Minimize Z = x1 + 2 x2 + 3x3
Subject to
2x1 + 3x2 + 3x3 ≥ – 4
3x1 + 5x2 + 2x3 ≤ 7
and x1 ≥ 0, x2 ≥ 0, x3 is unrestricted in sign
34
SLPP put : ̿̿̿ ̅̅̅

Minimize Z = x1 + 2 x2 + 3(̿̿̿ ̅̅̅)


Subject to
–2x1 – 3x2 – 3(̿̿̿ ̅̅̅) + s1= 4
3x1 + 5x2 + 2(̿̿̿ ̅̅̅) + s2 = 7
‫׳‬
x1≥ 0, x2 ≥ 0, ̿̿̿ , ̅̅̅ , s1 ≥ 0, s2 ≥ 0

Example :

If you have the following linear programming model:

ƒ (x) = 2x1 – 3x2 +5x3 (minimization)

Subject to :

x1 + 2x2 + x3 ≤ 35

2x1 + 7x2 – 5x3 ≥ 70

3x1 – 2x2 = 5

│5x1 + 12x2 – 7x3 │ ≤ 90

│7x1 - 8x2 –x3 │ ≥ 12

x1 , x3 ≥ 0

x2 is unrestricted in sign.

Required :

Put the linear programming model in the canonical form


and the standard form.

35
Solution :

1- The canonical form :

Since is unrestricted in sign,then, we have to substitute that:

̿̿̿ ̅̅̅

Where ̿̿̿ ≥ 0 and̅̅̅ ≥ 0

Then : Z = -2x1 + 3(̿̿̿ ̅̅̅) -5x3 (maximization)

Subject to :

x1 + 2(̿̿̿ ̅̅̅) + x3

-2x1 - 7(̿̿̿ ̅̅̅) + 5x3

3x1 – 2(̿̿̿ ̅̅̅)

-3x1 + 2(̿̿̿ ̅̅̅)

( (̿̿̿ ̅̅̅) )

The intersection between the two arrows which are


represent the area by which satisfied the two sub constraints
Then, we have the following two constraints in the solution:

(̿̿̿ ̅̅̅)

(̿̿̿ ̅̅̅)

( (̿̿̿ ̅̅̅) )

36
Since there is no intersection between the two arrows, (i.e.,
there is no feasible solution between the two arrows) , then ,
we have to take only one of the two sub constraints

(̿̿̿ ̅̅̅)

Or :

- (̿̿̿ ̅̅̅)

Where ̿̿̿ ≥ 0 and ̅̅̅

II) The standard form :

Put ̿̿̿ ̅̅̅, then the standard form can be summarized


as: F(x) Z= 2x1 - 3(̿̿̿ ̅̅̅) +5x3 (minimization)

Subject to:

x1 + 2(̿̿̿ ̅̅̅) + x3+

-2x1 -7(̿̿̿ ̅̅̅) + 5x3 -

3x1 – 2(̿̿̿ ̅̅̅)

( (̿̿̿ ̅̅̅) )

Then, we have the following two constraints in the


solution:

(̿̿̿ ̅̅̅)

(̿̿̿ ̅̅̅)

( (̿̿̿ ̅̅̅) )

37
We have to take only one of the two sub constraints

(̿̿̿ ̅̅̅)

Or :

(̿̿̿ ̅̅̅)

Where: ̿̿̿ ,̅̅̅

Some Basic Definitions


Solution of LPP
Any set of variable (x1, x2… xn) which satisfies the given
constraint is called solution of LPP.
Basic solution
It is a solution obtained by setting any ‘n’ variable equal to zero
and solving remaining ‘m’ variables. Such ‘m’ variables are
called Basic variables and ‘n’ variables are called Non-basic
variables.
Basic feasible solution
A basic solution that is feasible (all basic variables are non-
negative) is called basic feasible solution. There are two types of
basic feasible solution.
1. Degenerate basic feasible solution
If any of the basic variable of a basic feasible solution is zero
than it is said to be degenerate basic feasible solution.
2. Non-degenerate basic feasible solution

38
It is a basic feasible solution which has exactly ‘m’ positive
xi, where i=1, 2 … m. In other words all ‘m’ basic variables
are positive and remaining ‘n’ variables are zero.
Optimum basic feasible solution
A basic feasible solution is said to be optimum if it optimizes
(max / min) the objective function.
Introduction to Simplex Method
It was developed by G. Danztig in 1947. The simplex method
provides an algorithm (a rule of procedure usually involving
repetitive application of a prescribed operation) which is based
on the fundamental theorem of linear programming.

In order to determine the optimal solution for a linear


programming model, firstly, put the right hand side for all the
constraint are positive values. Then, we have the following
results:

(1) If all the constraints of the LPM are in the relation form
less than or equal to, i.e., (≤) , then we must use the
ordeinary simplex method.

(2) If there is at least one constraint in an equation form (=)


or in the form (≥) , then, we must use one of the following
two methods:

a. The M – technique.

B. The Two – phase technique.


39
All algorithms for the simplex methods started from an
initial solution (or table) which is called a basic solution. If the
solution yields all nonnegative basic variables, it is called a basic
feasible solution; otherwise, it is infeasible. A feasible extreme
point is thus defined by a basic feasible solution. This
fundamental property shows how the extr3eme point of the
solution space is translated algebraically as the basic solutions
of the equations representing in the linear program.

Optimality and feasibility conditions of the simplex methods:

The simplex methods are considered as an extension for


the algebra solution which is search about the optimum
solution to a general linear program in (m) equations and (n)
unknowns may be obtained by solving = n! / I m! I sets of
simultaneous equations. This procedure is inefficient. First, the
number of possible basic solution may be too long. Second,
many of these solutions may be infeasible or nonexistent.
Third, the object function plays a passive role in the calculation,
since it is used only after all the basic feasible solutions have
been determined.

The simplex method is designed specifically to avoid these


inefficiencies. The overall approach is to start from a basic
feasible solution and then move successively through a
sequence of basic feasible solutions such that each new
40
solution has the potential to improve the value of the objective
function. The basis of the simplex method, which guarantees
generating such a sequence of basic solutions, is two
fundamental conditions:

(1) The Feasibility condition:

the feasibility condition guarantees, that starting with a


basic feasible solution, only basic feasible solutions are
encountered during computation. This condition implies the
values for the basic variables in any solution must verify the
nonnegativity constraints.

(2) The Optimality condition :

The optimality condition ensures that no inferior solutions


(relative to the current solution point) are even encountered.
This condition implies that row of the test of optimality must be
zero or negative coefficients in case of maximization the value
of the objective function, and vice versa, these coefficients
must be zero or positive values in case of minimization the
value of the objective function.

* The Ordinary simplex method:

If the general linear model have positive values in the


R.H.S of the constraints, and all these constraints are
inequalities in the form (≤) except for the nonnegativity
41
constraint, then we have to use the ordinary simplex method
passes through the following steps:

1- Put the LPM in the standard form , then we have a set of


slack variables equal to the number of constraints :

2- The set of the slack variables are considered as the basic


variable for the set of constraints. Then construct the
initial solution tableau as the following table :

Coef.Of Z Cj C1 C2 Cn …. 0 0 ... 0 Ratio.

Basic variable CB XB ... .... …


(B.V)

0 The Coefficient of all the decision

0 Feasibility and slack variables in

: : Condition the set of constraints

: :

Zj= CBXj Max (Min)


Z=CBXB

Δj= Cj - Zj Optimality condition

42
Note that, the special arrangement of the above tableau
provides useful information. The columns associated with the
starting basic variable always appear immediately to the left of
the solution column, and their constraints coefficients will
constitute an identity matrix.

The next step is to determine a new basic feasible


solution with an improved value of the objective function. The
simplex method does this by selecting a current non-basic
variable to be increased above zero providing its coefficient in
the objective function has potential to improve the value of the
objective function Z or ƒ(x).

3- Solution improvement :

If the optimality condition is not satisfied, then one can


improve to solution according the following steps:

a) Determin the pivot column (Entering variable):

the non-basic variable which have the most positive


coefficient in the row of the optimality condition (Δj) in case of
maximization the value of the objective function ƒ(x) or Z is
called the entering variable and its column is called the pivot
column. Vice versa, the non-basic variable which have the most
negative coefficient in the row of the optimality condition (Δj)
in case of minimization f(x) orZ is called the entering variable.

43
b) Determine the pivot row (Leaving variable):

The leaving variable may be secured directly from the


preceding tableau by taking the ratios of the value of solution
column to the positive constraint coefficients under the
entering variable. The basic variable associated with the
minimum ratio (smallest intercept) is the leaving variable and
its row is the pivot row.

C) Determine the Pivot Elements :

The intersection between the pivot column and the pivot


row determine the pivot element in any interaction for the
solution.

 The improvement of the simplex tableau according


to the Gauss- Jordan alimentation passes through
the following steps:

 The first step is to divide the pivot row by the pivot


element and replace the basic leaving variable by
the entering variable.

 The detailed row operations are given as follows :

new row =current row - (Its pivot column coefficient) * new


pivot row .
44
The result of these calculations gives the second tableau, which
is further improved since all the two conditions (Feasibility and
optimality) are satisfied.

Note that, the only difference between maximization and


minimization occurs in the optimality condition: In
minimization. The entering variable is the one with the most
negative coefficient in the optimality condition (Δj) and in
maximization. The entering variable is the one with the most
positive coefficient in the optimality condition (Δj). The
feasibility condition remains the same since it depends on the
constraints and not the objective function.

To conclude the preceding section, the following is a


summary of the optimality and feasibility conditions.

Optimality condition. Given the coefficient of (Δj)


expressed in terms of the non-basic variables only, one selects
the entering variable in maximization (minimization) as the
non-basic variable having the most positive (most negative)
coefficient in the (Δj) row. A tie between two non-basic
variables may be broken arbitrarily, when all the left hand side
coefficients of the (Δj) are negative (positive), the optimum is
reached.

45
Feasibility condition. The leaving variable is the basic
variable corresponding to the smallest ratio of the current
value of the basic variables to the positive constraint coefficient
of the entering variable.

Finally, a summary of the computational procedure of the


simplex method is shown as in the following steps:

Step (1) : express the standard from of the linear program in a


tableau as we show in the preceding section.

Step (2) : Select a starting basic feasible solution. This step


involves two cases:

1- If all the constraints in the original problem are (≤), the


slack variables are used for a starting solution.

2- If the constraints in the original problem include (≥) or (=)


, then, the two techniques (M- technique or two phase
technique) are used the artificial technique to give a
starting basis.

Step (3) : generate new basic feasible solution using the


optimality and feasibility conditions until the optimal solution is
attained. This step assumes that the optimal solution exists and
is bounded. The cases of nonexisting and unbounded solutions
are discussed in the later section.

46
The following examples are now introduced to sum up the
basic features of the ordinary simplex method.
Computational Procedure of Simplex Method
Consider an example
Maximize Z = 3x1 + 2x2
Subject to
x 1 + x2 ≤ 4
x1 – x2 ≤ 2
and x1 ≥ 0, x2 ≥ 0
Solution
Step 1 Write the given GLPP in the form of SLPP
Maximize Z = 3x1 + 2x2 + 0s1 + 0s2
Subject to
x1 + x2+ s1= 4
x1 – x2 + s2= 2
x1≥ 0, x2 ≥ 0, s1 ≥ 0, s2 ≥ 0
Step 2 Present the constraints in the matrix form
x 1 + x 2+ s 1 = 4
x 1 – x 2 + s2 = 2

( )[ ] * +

47
Step 3 – Construct the starting simplex table using the notations
(TABLEAU 1)

Coef.Z Cj 3 2 0 0 Min
Ratio.
(B.V) CB XB X1 X2 S1 S2

0 4 1 1 1 0 4

0 2 1 -1 0 1 2

Zj=CBXj Z=0 0 0 0 0

Δj= Cj - Zj 3 2 0 0
Max Z= CBXB =0

Step 4 – Calculation of Z and Δj and test the basic feasible


solution for optimality by the rules given.
Z= CB XB
= 0 *4 + 0 * 2 = 0
Δj= Cj - Zj
= Cj -CBXj
Δ1 = C1 - CB X1 = 3-0 * 1 + 0 * 1 = 3
Δ2 = C2 - CB X2 = 2-0 * 1 + 0 * -1 = 2
Δ3 = C3 - CB X3 =0- 0 * 1 + 0 * 0 = 0
Δ4 = C4 - CB X4 =0- 0 * 0 + 0 * 1 = 0
Procedure to test the basic feasible solution for optimality by the
rules given
48
Rule 1 – If at least one Δj , the solution is not optimal and
then proceeds to improve the solution in the next step.
Rule 2 – If all Δj , the solution under the test will be
optimal.
Rule 3 – If corresponding to any positive Δj, all elements of the
column Xj are negative or zero, then the solution under
test will be unbounded.
In this problem it is observed that Δ1 and Δ2 are positive. Hence
proceed to improve this solution
Step 5 – To improve the basic feasible solution, the vector
entering the basis matrix and the vector to be removed from the
basis matrix are determined.
 Incoming vector
The incoming vector X1 is always selected corresponding
to the most positive value of Δj. It is indicated by (↑). It is
called pivot column.
 Outgoing vector
The outgoing vector is selected corresponding to the
least positive value of minimum ratio. It is indicated by
(→). It is called pivot row.
Step 6 – Mark the key element or pivot element by ‘1’‘.The
element at the intersection of outgoing vector ( )and incoming
vector (X1) is the pivot element. (Pivot element is the element at
the intersection of pivot column and pivot row)

49
Any new table will form:
The row that replaces the outgoing row (V)
= pivot row / pivot element
The remaining rows:
New row = old row – corresponding element at pivotcolumn * V
(TABLEAU 2)

Coef.Z Cj 3 2 0 0 Min
Ratio.
(B.V) CB XB X1 X2 S1 S2

0 2 0 2 1 -1 1

3 2 1 -1 0 1 -2

Zj=CBXj Z=6 3 -3 0 3

Δj= Cj - Zj 0 5 0 -3
Max Z= CBXB =6

Step 7 – Now repeat step 4 through step 6 until an optimal


solution is obtained.

50
(TABLEAU 3)

Coef.Z Cj 3 2 0 0 Min
Ratio.
(B.V) CB XB X1 X2 S1 S2

2 1 0 1 ⁄ ⁄

3 3 1 0 ⁄ ⁄

Zj=CBXj Z=11 3 1 ⁄ ⁄

Δj= Cj - Zj 0 0 ⁄ ⁄
Max Z= CBXB =11

Since all Δj , optimal basic feasible solution is obtained


Therefore the solution is Max Z = 11, x 1 = 3 and x2 = 1
Example
Maximize Z = 5x1 + 7x2
Subject to
x 1 + x2 ≤ 4
3x1 – 8x2 ≤ 24
10x1 + 7x2 ≤ 35
and x1 ≥ 0, x2 ≥ 0
Solution
SLPP
Maximize Z = 5x1 + 7x2 + 0s1 + 0s2 + 0s3
51
Subject to
x1 + x2 + s1= 4
3x1 – 8x2 + s2= 24
10x1 + 7x2 + s3= 35
x1≥ 0, x2 ≥ 0, s1 ≥ 0, s2 ≥ 0, s3 ≥ 0

( ) [ ]

[ ]
(TABLEAU 1)

Coef.Z Cj 5 7 0 0 0 Min
Ratio.
(B.V) CB XB X1 X2 S1 S2 S3

0 4 1 1 1 0 0 4

0 24 3 -8 0 1 0 -3

0 35 10 7 0 0 1 5

Zj=CBXj Z=0 0 0 0 0 0

Δj= Cj - Zj 5 7 0 0 0
Max Z= CBXB =0

52
(TABLEAU 2)

Coef.Z Cj 5 7 0 0 0 Min
Ratio.
(B.V) CB XB X1 X2 S1 S2 S3

7 4 1 1 1 0 0

0 56 11 0 8 1 0

0 7 3 0 -7 0 1

Zj=CBXj Z=0 7 7 7 0 0

Δj= Cj - Zj -2 0 -7 0 0
Max Z= CBXB =0

Since all Δj , optimal basic feasible solution is obtained


Therefore the solution is Max Z = 28, x 1 = 0 and x2 = 4
Example
Maximize Z = 2x – 3y + z
Subject to
3x + 6y + z ≤ 6
4x + 2y+ z ≤ 4
x– y +z ≤3
and x≥ 0, y≥ 0, z ≥ 0
Solution : SLPP
Maximize Z = 2x – 3y + z + 0s1 + 0s2 + 0s3

53
Subject to
3x + 6y+ z + s1 = 6
4x + 2y + z + s2 = 4
x – y+ z + s3= 3

( ) [ ]

[ ]
x≥ 0, y≥ 0, z ≥ 0 s1 ≥ 0, s2 ≥ 0, s3 ≥ 0
(TABLEAU 1)

Coef.Z Cj 2 -3 1 0 0 0 Min
Ratio.
(B.V) CB XB X Y Z S1 S2 S3

0 6 3 6 1 1 0 0 2

0 4 4 2 1 0 1 0 1

0 3 1 -1 1 0 0 1 3

Zj=CBXj Z=0 0 0 0 0 0 0

Δj= Cj - Zj 2 -3 1 0 0 0
Max Z= CBXB =0

(TABLEAU 2)
54
Coef.Z Cj 2 -3 1 0 0 0 Min
Ratio.
(B.V) CB XB X Y Z S1 S2 S3

0 3 0 1 0
12

2 1 1 0 0
4

0 2 0 0 1

Zj=CBXj Z=2 2 1 0 0

Δj= Cj - Zj 0 -4 0 0
Max Z= CBXB =0

(TABLEAU 3)

Coef.Z Cj 2 -3 1 0 0 0 Min
Ratio.
(B.V) CB XB X Y Z S1 S2 S3

0 0 5 0 1

2 1 1 0 0

1 0 -2 1 0

55
Zj=CBXj Z= 2 0 1 0

Δj= Cj - Zj 0 -3 0
Max Z= CBXB =0

Since all Δj , optimal basic feasible solution is obtained


Therefore the solution is Max Z = 10/3, x = 1/3, y = 0 and z =
8/3
Example
Minimize Z = x1 – 3x2 + 2x3
Subject to
3x1 – x2 + 3x3 ≤ 7
-2x1 + 4x2 ≤ 12
-4x1 + 3x2 + 8x3 ≤ 10
and x1 ≥ 0, x2 ≥ 0, x3 ≥ 0
Solution: SLPP
Min Z = x1 -3x2 +2x3 + 0s1 + 0s2 + 0s3
Subject to
3x1 – x2 + 3x3 + s1 = 7
-2x1 + 4x2 + s2 = 12
-4x1 + 3x2 + 8x3 + s3 = 10
x1≥ 0, x2 ≥ 0, x3 ≥ 0 s1 ≥ 0, s2 ≥ 0, s3 ≥ 0

56
( ) [ ]

[ ]
(TABLEAU 1)

Coef.Z Cj 1 -3 2 0 0 0 Min
Ratio.
(B.V) CB XB S1 S2 S3

0 7 3 -1 3 1 0 0 _

0 12 -2 44 0 0 1 0 3

0 10 -4 3 8 0 0 1

Zj=CBXj Z=0 0 0 0 0 0 0

Δj= Cj - Zj 1 -3 2 0 0 0
Max Z= CBXB =0

(TABLEAU 2)

Coef.Z Cj 1 -3 2 0 0 0 Min
Ratio.
(B.V) CB XB S1 S2 S3

0 10 3 1 0
4

57
-3 3 0 0 0
-

0 1 8 0 1

Zj=CBXj Z -3 0 0 0
=-9

Δj= Cj - Zj 0 0 0
Max Z= CBXB =-9

(TABLEAU 3)

Coef.Z Cj 1 -3 2 0 0 0 Min
Ratio.
(B.V) CB XB S1 S2 S3

1 4 1 0 0

-3 0 1
5

0 0 0 11 1
11

Zj=CBXj Z= 1 -3

Δj= Cj - Zj 0 0 0
Max Z= CBXB =-11
Since all Δj , optimal basic feasible solution is obtained
Therefore the solution is MinZ = -11, x1 = 4, x2 = 5, x3 = 0
Exercise
58
Solve by simplex method
1. Maximize Z = 5x1 + 3x2
Subject to
3x1 + 5x2 ≤ 15
5x1 + 2x2 ≤ 10
and x1 ≥ 0, x2 ≥ 0
[Ans. Max Z = 235/19, x1= 20/19, x2= 45/19]
2. Maximize Z = 5x1 + 7x2
Subject to
x1 + x 2 ≤ 4
3x1 - 8x2 ≤ 24
10x1 + 7x2 ≤ 35
and x1 ≥ 0, x2 ≥ 0
[Ans. Max Z = 28, x1= 0, x2= 4]
3. Maximize Z = 2x1 + 4x2 + x3+ x4
Subject to
x1 + 3x2 + x4 ≤ 4
2x1 + x2 ≤ 3
x2+ 4x3 + x4 ≤ 3
and x1 ≥ 0, x2 ≥ 0, x3 ≥ 0, x4 ≥ 0
[Ans. Max Z = 13/2, x1= 1, x2= 1, x3= 1/2, x4= 0]
4. Maximize Z = 7x1 + 5x2
Subject to
-x1 - 2x2 ≥-6
4x1 + 3x2 ≤ 12
and x1 ≥ 0, x2 ≥ 0
[Ans. Max Z = 21, x1= 3, x2= 0]

59
Computational Procedure of Big – M Method (Charne’s
Penalty Method)
Step 1 – Express the problem in the standard form.
Step 2 – Add non-negative artificial variable to the left side of
each of the equations corresponding to the constraints of the
type ‘≥’ or ‘=’.
When artificial variables are added, it causes violation of the
corresponding constraints. An artificial variable is added to form
a starting solution similar to the all slack basic solution
.However ,the artificial variables are not part of the original
problem ,and a modeling (trick) is needed to force them to zero
by the time ,the optimum iteration is reached . The desired goal
is achieved by penalizing these variables in the objective
function using the following rule :
Penalty rule for artificial variables
Given M , a sufficiently large positive value (mathematically
,M ) ,The objective coefficient of an artificial represents an
appropriate penalty if :

60
Step 3 – In the last, use the artificial variables for the starting
solution and proceed with the usual simplex routine until the
optimal solution is obtained.

Example : Determine the value of x1 and x2 by which :

Max Z = 3x1 + 2x2

Subject to:

2 x 1 + x2 2

3x1 + 4x2 ≥ 12

x1 ≥ 0, x2 ≥ 0

Solution :

Since all the constant (R. H. S) For all constraints are


positive numbers, and there are at least one constraint is in
the type ( ≥ ), therefore , either the M-technique or the
two-phase technique are suitable for the solution of this
LPM.

Now, we will use the M- technique:

Step 1: Express the model in its standard form.

Max Z = 3x1+2x2 +0 -0

61
Subject to:

2x1 + x2 + s1= 2

3x1 + 4x2 – s2= 12

x1 , x2 , s1, s2 ≥ 0

Step 2 : Add an artificial variable corresponding to the


constraint of the type ( ≥) and express these artificial
variable with (-M) coefficient in the equation of the
objective function in case of maximization the value of f (x),
where M > 0 : then , we have the following model :

Max Z = 3x1+2x2 +0 + 0 -M

2 x1 + x2 + s1 =2

3x1 + 4x2 – s2+ a1= 12

x1 , x2 , s1, s2, a1,≥ 0

( ) * +

[ ]
Step 3 : construct the initial solution tableau by considering
that the slack variable corresponding to the constraint in the
form (≤) is the basic variables, and the artificial variable
corresponding to the constraint are in either the form (= or ≥) is
62
the basic variable for these constraints. Then, we have the
following tableau:

Starting (or initial ) solution tableau:

Coef.Z Cj 3 2 0 0 -M Min
Ratio.
(B.V) CB XB S1 S2 a1

0 2 2 1 1 0 0 2

-M 3 4 0 4
12 -1 1

Zj=CBXj -12M -3M -4M 0 M -M

Δj= Cj - Zj 3+3M 2+4M 0 -M 0


Max Z=-12M

1st iprovement :

2nd tableau

Coef.Z Cj 3 2 0 0 -M Min
Ratio.
(B.V) CB XB S1 S2 a1

2 2 2 1 1 0 0

-M -5 0 -4
4 -1 1

Zj=CBXj 4-4M 4+5M 2 2+4M M -M

63
Δj= Cj - Zj -1-5M 0 -2-4M -M 0
Max Z=4-4M

Now, from the preceding tableau note that :

 Since , all the values for the basic variables (solution


column) are negative values, so that this tableau is
considered a basic feasible solution (feasibility
condition).

 Since all the coefficients corresponding to the basic


variables and a1 are zeros only, and negative
coefficient corresponding to all the nonbasic
variables (x1 , S1 and S2) , so that the 2nd tableau is
considered a unique optimal solution (optimality
condition).

Henceforth, the optimal solution for this problem :

X1=0, X2 = 2, s1 =0 , s2 =0,a1 =4,

Z = 3*0 + 2*2 + 0*2 + 0*0 - M*4 =4-4M

Example :

Determine the value of x1 and x2 by which :

Z = 4x1 + x2 minimize

Subject to :
64
3x1 + x2 = 3

4x1 + 3x2 ≥ 6

x1 + 2x2 ≤ 4

x 1 , x2 ≥ 0

Solution :

Since all the constant (R. H . S) for all constraints are


positive numbers, and there are at least one constraint is in the
type ( ≥ or = ), therefore , either the M-technique or the two-
phase technique are suitable for the solution of this LPM.

Now, we will use the M- technique:

Step 1: Express the model in its standard form.

Z = 4x1 + x2 (Min)

Subject to :

3x1 + x2 = 3

4x1 + 3x2 - S1 = 6

x1 + 2x2 + S2 = 4

x 1 , x 2 , S 1 , S2 ≥ 0

Step 2 : Add an artificial variable corresponding to the


constraint of the type (= or ≥) and express these artificial

65
variable with (M) coefficient in case of minimization value
of f (x), where M > 0 : then , we have the following model :

Z = 4x1 + x2 + 0 (S1 + S2) + M ( + ) (Min)

Subject to :

3x1 + x2 + = 3

4x1 + 3x2 – S1 + = 6

x1 + 2x2 + S2 = 4

x1 ≥ 0 for i = 1 : 6

( ) [ ]

[ ]
Step 3 : construct the initial solution tableau by considering
that the slack variable corresponding to the constraint in the
form (≤) is the basic variables, and the artificial variable
corresponding to the constraint are in either the form (= or ≥) is
the basic variable for these constraints.

Starting (or initial ) solution tableau:

Coef.Z Cj 4 1 0 0 M M Min
Ratio.
(B.V) CB XB S1 S2

66
M 3 3 1 0 0 1 0 1

M 4 3 -1 0 ⁄
6 0 1

0 1 2 0 1 0 4
4 0

Zj=CBXj Z= 7M 4M -M 0 M M

Δj= Cj - Zj M 0 M
Min Z= 9M

2nd tableau

Coef.Z Cj 4 1 0 0 M M Min
Ratio.
(B.V) CB XB S1 S2

4 1 1 ⁄ 0 0 ⁄ 0
3

M 0 ⁄ -1 0 ⁄
2 ⁄ 1

0 0 ⁄ 0 1 0 ⁄
3 ⁄

Zj=CBXj Z= 4 ⁄ -M 0 ⁄ M

67
Δj= Cj - Zj 0 M 0 0
⁄ ⁄
Min Z= 4+2M

3nd tableau

Coef.Z Cj 4 1 0 0 M M Min
Ratio.
(B.V) CB XB S1 S2

4 ⁄ 1 0 ⁄ 0 ⁄ ⁄
3

1 0 1 ⁄ 0
⁄ ⁄ ⁄ -

0 0 0 1 1 -1 1
1 1

Zj=CBXj ⁄ 4 1 ⁄ 0 ⁄ ⁄

Δj= Cj - Zj 0 0 0
⁄ ⁄ ⁄
Min Z= ⁄

68
4th tableau

Coef.Z Cj 4 1 0 0 M M Min
Ratio.
(B.V) CB XB S1 S2

4 ⁄ 1 0 0 ⁄ ⁄ 0

1 0 1 0 ⁄
⁄ ⁄ 0

0 0 0 1 1 -1
1 1

Zj=CBXj ⁄ 4 1 0 ⁄ ⁄ 0

Δj= Cj - Zj 0 0 0 M
⁄ ⁄
Min Z= ⁄

Now, from the preceding tableau note that :

 Since , all the values for the basic variables (solution


column) are positive values, so that this tableau is
considered a basic feasible solution (feasibility
condition).

 Since all the coefficients corresponding to the basic


variables x1 , x2 and are zeros only, and positive
coefficient corresponding to all the nonbasic
variables( , , so that the 4th tableau is
69
considered a unique optimal solution (optimality
condition).

Henceforth, the optimal solution for this problem :

⁄ ⁄ ⁄

Example
Max Z =3x1 + 2x2 + x3
Subject to:
2x1 + x2 + x3 = 12
\3x1 + 4x2 = 11
x1 is unrestricted, x2 ≥ 0, x3 ≥ 0

Solution :

Now, we will use the M- technique:

Step 1: Express the model in its standard form.

Let (̿̿̿ ̅̅̅)

Max Z =(̿̿̿ ̅̅̅)+ 2 +


Subject to:
2(̿̿̿ ̅̅̅) + + = 12
3(̿̿̿ ̅̅̅) + 4 = 11
̿̿̿ ̅̅̅, , ≥0

Step 2 : Add an artificial variables :


70
Max Z =(̿̿̿ ̅̅̅)+ 2 + -M( )
Subject to:
2(̿̿̿ ̅̅̅) + + + = 12
3(̿̿̿ ̅̅̅) + 4 = 11
̿̿̿ ̅̅̅, , ≥0

̿̿̿
̅̅̅
( ) * +

[ ]
Step 3 : construct the initial solution tableau.

Starting (or initial ) solution tableau:

Coef.Z Cj 3 -3 2 1 -M -M Min
Ratio.
(B.V) CB XB ̿̿̿ ̅̅̅ X3

-M 12 2 -2 1 1 1 0 6

-M 3 -3 4 0 ⁄
11 0 1

Zj=CBXj -23M -5M 5M -5M -M -M -M

Δj= Cj - Zj 3+5M -3-5M 2+5M 1+M 0 0


Max Z= -23M

71
2nd tableau

Coef.Z Cj 3 -3 2 1 -M - Min
M Ratio.

(B.V) CB XB ̿̿̿ ̅̅̅ X3

-M ⁄ 0 0 ⁄ 1 1 x ⁄

̿̿̿ 3 1 -1 ⁄ 0 -
⁄ 0 x

Zj=CBXj 3 -3 -M -M x

Δj= Cj - Zj 0 0 1+M 0 x
Max Z= -23M

3nd tableau

Coef.Z Cj 3 -3 2 1 -M - Min
M Ratio.

(B.V) CB XB ̿̿̿ ̅̅̅ X3

1 ⁄ 0 0 ⁄ 1 x x ⁄

72
̿̿̿ 3 1 -1 ⁄ 0 -
⁄ x x

Zj=CBXj ⁄ 3 -3 ⁄ 1 x x

Δj= Cj - Zj 0 0 0 x x
MaxZ= ⁄

Now, from the preceding tableau note that :

̿̿̿̿ ̅̅̅̅

Then ( )

Therefore the solution is Max Z = 47/3, x1 = 11/3, x2 = 0, x3 =


14/3

Example
Max Z = 8x2
Subject to
x1 - x2 ≥ 0
2x1 + 3x2 ≤ - 6
x1 , x2 unrestricted

Solution :

Now, we will use the M- technique:

Step 1: Express the model in its standard form.

Let (̿̿̿ ̅̅̅) (̿̿̿ ̅̅̅)


73
Max Z = 8(̿̿̿ ̅̅̅)
Subject to
(̿̿̿ ̅̅̅) - (̿̿̿ ̅̅̅) - =0
-2(̿̿̿ ̅̅̅) - 3(̿̿̿ ̅̅̅) - = 6
̿̿̿ ̅̅̅, ̿̿̿ , ̅̅̅ , ≥0

Step 2 : Add an artificial variables :

Max Z = 8 (̿̿̿ ̅̅̅) +0( ) - M( )


Subject to:
Subject to
(̿̿̿ ̅̅̅) - (̿̿̿ ̅̅̅) - + =0
-2(̿̿̿ ̅̅̅) - 3(̿̿̿ ̅̅̅) - + = 6
̿̿̿ ̅̅̅, ̿̿̿ , ̅̅̅ , ≥0

̿̿̿
̅̅̅
̿̿̿
( ) ̅̅̅ * +

[ ]
Step 3 : construct the initial solution tableau.

74
Starting (or initial ) solution tableau:

Coef.Z Cj 0 0 8 -8 0 0 -M -M Min
Ratio.
(B.V) CB XB ̿̿̿ ̅̅̅ ̿̿̿ ̅̅̅

-M 0 1 -1 -1 1 -1 0 1 0 0

-M -2 2 -3 3 2
6 0 -1 0 1

Zj=CBXj -6M M -M 4M -4M M M -M -M

Δj= Cj - Zj -M M 8-4M -8+4M -M -M 0 0


Max Z= -6M

2nd tableau

Coef.Z Cj 0 0 8 -8 0 0 -M -M Min
Ratio.
(B.V) CB XB ̿̿̿ ̅̅̅ ̿̿̿ ̅̅̅

̅̅̅ -8 0 1 -1 -1 1 -1 0 x 0 -

-M -5 5 0 0 6/5
6 3 -1 x 1

Zj=CBXj -6M 5M-8 8-5M 8 -8 8-3M M x -M

Δj= Cj - Zj 8-5M 5M-8 0 0 3M-8 M x 0


Max Z= -6M

75
3nd tableau

Coef.Z Cj 0 0 8 -8 0 0 -M -M Min
Ratio.
(B.V) CB XB ̿̿̿ ̅̅̅ ̿̿̿ ̅̅̅

̅̅̅ -8 ⁄ 0 0 -1 1 ⁄ ⁄ X X

̅̅̅ 0 ⁄ -1 1 0 0 ⁄ ⁄ X X

Zj=CBXj ⁄ 0 0 8 -8 ⁄ ⁄ X X

Δj= Cj - Zj 0 0 0 0 x x
⁄ ⁄
Max Z=

Now, from the preceding tableau note that :

̿̿̿̿ ̅̅̅̅ ̿̿̿ ̅̅̅

Then ( ) , ( )

Therefore the solution is Max Z = -48/5, x1 = -6/5, x2 = -6/5

76
Exercise
Solve by Big-M method
1. Min Z = 4x1 + 2x2
Subject to: 3x1 + x2 ≥ 27
x1 + x2 ≥ 21
and x1 ≥ 0, x2 ≥ 0
[Ans. Min Z = 48, x1 = 3, x2 =18]
2. Min Z = x1 + x2 + 3x3
Subject to: 3x1 + 2x2 + x3 ≤ 3
2x1 + x2+ 2x3 ≥ 3
and x1 ≥ 0, x2 ≥ 0, x3 ≥ 0
[Ans. Min Z = 3, x1 = 3/4, x2 =0, x3 = 3/4]
3. Max Z = 3x1 - x2
Subject to
2x1 + x2 ≥ 2
x1 + 3x2 ≤ 2
x2≤ 4
and x1 ≥ 0, x2 ≥ 0
[Ans. Max Z = 6, x1 = 2, x2 =0]
4. Max Z = 5x1 - 2x2 +3x3
Subject to: 2x1 + 2x2 - x3 ≥ 2
3x1 - 4x2 ≤ 3
x2 + 3x3 ≤ 5
and x1 ≥ 0, x2 ≥ 0, x3 ≥ 0
[Ans. Max Z = 85/3, x1 = 23/3, x2 =5, x3 =0]

77
Duality in LPP
Every LPP called the primal is associated with another LPP
called dual. Either of the problems is primal with the other one
as dual. The optimal solution of either problem reveals the
information about the optimal solution of the other.
Let the primal problem be
Max Zx = c1x1 + c2x2 + … +cnxn
Subject to restrictions
a11x1 + a12x2 + … + a1nxn ≤ b1
a21x1 + a22x2 + … + a2nxn ≤ b2
.
.
am1x1 + am2x2 + … + amnxn≤ bn
and
x1 ≥ 0, x2 ≥ 0,…, xn≥ 0
The corresponding dual is defined as
Min Zw = b1w1 + b2w2 + … + bmwm
Subject to restrictions
a11w1 + a21w2 + … + am1wm ≥ c1
a12w1 + a22w2 + … + am2wm ≥ c2
.
.
a1nw1 + a2nw2 + ……….+amnwm ≥ cn
and
w1, w2, …,wm≥ 0
78
Matrix Notation
Primal Dual
Max Zx = CX Min Zy = bTY
Subject to Subject to
AX ≤ b , X ≥ 0 AT Y ≥ CT , Y ≥ 0

Min Zx = CX Max Zy = bTY


Subject to Subject to
AX ≥ b , X ≥ 0 AT Y ≤ CT , Y ≥ 0

Important characteristics of Duality


1. Dual of dual is primal
2. If either the primal or dual problem has a solution then the
other also has a solution and their optimum values are
equal.
Advantages and Applications of Duality
1. Sometimes dual problem solution may be easier than
primal solution, particularly when the number of decision
variables is considerably less than slack / surplus variables.
2. In the areas like economics, it is highly helpful in
obtaining future decision in the activities being
programmed.
3. In physics, it is used in parallel circuit and series circuit
theory.
79
4. When a problem does not yield any solution in primal, it
can be verified with dual.
5. Economic interpretations can be made and shadow prices
can be determined enabling the managers to take further
decisions.
Steps for a Primal Form
Step 1 – If the objective function is Maximization form, then all
constraints must have an inequality sign ‘≤’. And if the objective
function is Minimization form, then all constraints must have an
inequality sign ‘≥’.
Step 2 – If the constraint has an ‘=’ sign then replace it by two
constraints involving the inequalities going in opposite
directions. For example: in case of maximization objective
function, and x1+ 2x2 = 4 is written as
x1+2x2 ≤ 4
x1+2x2 ≥ 4 → - x1-2x2 ≤ - 4
but in case of minimization objective function, and
x1+ 2x2 = 4 is written as
-x1-2x2 ≥ -4
x1+2x2 ≥ 4
Step 3 – Every unrestricted variable is replaced by the
difference of two non-negative variables.
Step4 – We get the standard primal form of the given LPP in
which.

80
o All constraints have ‘≤’ sign, where the objective
function is of maximization form.
o All constraints have ‘≥’ sign, where the objective
function is of minimization from.
Rules for Converting any Primal into its Dual
1. Minimize the objective function instead of maximizing it,
and Maximize the objective function instead of
minimizing it.
2. Transpose the rows and columns of the constraint co-
efficient.
3. Transpose the co-efficient (c1,c2,…cn) of the objective
function and the right side constants (b1,b2,…bn)
4. Change the inequalities from ‘≤’ to ‘≥’ sign.
Example Problems
Write the dual of the given problems
Example :
Min Zx= 2x2 + 5x3
Subject to:
x1+x2 ≥ 2
2x1+x2+6x3 ≤ 6
x1- x2 +3x3 = 4
x1, x2 , x3 ≥ 0
Solution
Primal
Min Zx= 2x2 + 5x3
81
Subject to:
x1+ x2 ≥2 y1
-2x1- x2- 6x3 ≥ - 6 y2
-x1-+ x2 - 3x3 ≥ - 4 y3
x1- x2 + 3x3 ≥ 4 y4

x1, x2 , x3 ≥ 0
Dual
Max Zw = 2 y1 - 6 y2 - 4 y3 + 4 y4
Subject to: y1 – 2y2 - y3 +y4 ≤ 0
y1 - y2 + y3 -y4 ≤ 2
-6y2 – 3y3 +3y4 ≤ 5
y1, y2, y3, y4 ≥ 0
Example :
Min Zx= 3x1 - 2x2 + 4x3
Subject to: 3x1+5x2 + 4x3 ≥ 7
6x1+x2+3x3 ≥ 4
7x1 - 2x2 -x3 ≥ 10
x1- 2x2 + 5x3 ≥ 3
4x1 + 7x2 - 2x3 ≥ 2
x1, x2 , x3 ≥ 0
Solution
Primal
Min Zx= 3x1 - 2x2 + 4x3
Subject to:
82
3x1+ 5x2 + 4x3 ≥ 7 y1
6x1+x2+ 3x3 ≥ 4 y2
7x1 - 2x2 - x3 ≥ 10 y3
x1- 2x2 + 5x3 ≥ 3 y4
4x1 + 7x2 - 2x3 ≥ 2 y5
x1, x2 , x3 ≥ 0
Dual
Max Zw = 7y1 +4y2 + 10y3 + 3y4 +2y5
Subject to: 3y1 + 6y2 + 7y3 +y4 + 4y5 ≤ 3
5y1 +y2 – 2y3 – 2y4 –+7y5 ≤ - 2
4y1 + 3y2 - y3 +5y4 – 2y5 ≤ - 4
y1, y2, y3, y4, y5 ≥ 0
Example :
Max Z= 2x1+ 3x2 + x3
Subject to:
4x1+ 3x2 + x3 = 6
x1+ 2x2 + 5x3 = 4
x1, x2 ≥ 0
Solution
Primal
Max Zx= 2x1+ 3x2 + x3
Subject to :

83
4x1+ 3x2 + x3 ≤ 6 y1
-4x1 - 3x2 - x3 ≤ -6 y2
x1+ 2x2 + 5x3 ≤ 4 y3
-x1- 2x2 - 5x3 ≤ -4 y4
x1, x2 ,x3≥ 0
Dual
Min Zw = 6y1 – 6y2 + 4y3 –4y4
Subject to
4y1 – 4y2 + y3 –y 4 ≥ 2
3y1 – 3y2 + 2y3 – 2y4 ≥ 3
y1 - y2 + 5y3 – 5y4 ≥ 1
y1, y2, y3, y4≥ 0
Example ;
Min Zx= x1+ x2 + x3
Subject to: x1 - 3x2 + 4x3 = 5
x1- 2x2 ≤ 3
2x2 - x3 ≥ 4
x1, x2 ≥ 0 ,x3 is unrestricted in sign
Solution
Primal
Let (̿̿̿ ̅̅̅)

Min Zx= X1+ X2 + (̿̿̿ ̅̅̅)


Subject to;

84
X1- 3X2 + 4(̿̿̿ ̅̅̅) ≥ 5 Y1
-X1+ 3X2 - 4(̿̿̿ ̅̅̅) ≥ -5 Y2
-X1+ 2X2 ≥ -3 Y3
2X2 - (̿̿̿ ̅̅̅) ≥ 4 Y4
X1, X2 , ̿̿̿ ̅̅̅
Dual
Max Zw = 5Y1 – 5Y2 - 3Y3 + 4Y4
Subject to
Y1 - Y2 - Y3 ≤1
-3Y1 + 3Y2 + 2Y3 – 2Y4 ≤ 1
4Y1 – 4Y2 -Y4 ≤ 1
-4Y1 + 4Y2 +Y4 ≤ -1
Y1, Y2, Y3, Y4, ≥ 0
Primal –Dual Relationship
Weak duality property
If x is any feasible solution to the primal problem and w is any
feasible solution to the dual problem then CX ≤ bT Y.
i.e. ZX ≤ Zy
Strong duality property
If x* is an optimal solution for the primal problem and y* is the
optimal solution for the dual problem then CX* = bT Y*
i.e. ZX = Zy

85
Complementary optimal solutions property
At the final iteration, the simplex method simultaneously
identifies an optimal solution x* for primal problem and a
complementary optimal solution w* for the dual problem where
ZX = Zy.
Symmetry property
For any primal problem and its dual problem, all relationships
between them must be symmetric because dual of dual is primal.
Complementary basic solutions property
Each basic solution in the primal problem has a complementary
basic solution in the dual problem where ZX = Zy.
Complementary slackness property
The variables in the primal basic solution and the
complementary dual basic solution satisfy the complementary
slackness relationship as shown in the table.
Primal variable Associated dual variable
Decision variable (xj) Cj– Zj (surplus variable) j = 1, 2, ..n
Slack variable (Si) Yi (decision variable) i = 1, 2, .. n
Duality and Simplex Method
Example ; If you have the following LPM:

F (x) = 2x1 + x2 (minimize)

Subject to :

3X1 + x2 ≥ 3

4x1 + 3x2 ≥ 6
86
x1+ 2x2 ≤ 3

x 1 , x2 ≥ 0

Required : 1- Determine the Dual problem.

2- Solve the dual problem with the suitable simplex method.

3- From your preceding result in (2), how can you constraint


to optimal solution tableau for the primal problem.

4- Determine the shadow prices for the different resources


by using two different methods.

1- Verify between the two optimal solutions of the primal


and dual problem.

Solution :

1- In order to determine the dual problem, we must put all


the constraints for the primal LPM in the same type either
(≤) or (≥), then, we have the following model:

f (x) = 2x1 + x2 minimize

Subject to :

3X1 + x2 ≥ 3 (y1)

4x1 + 3x2 ≥ 6 (y2)

-x1 - 2x2 ≥ -3 (y3)

x 1 , x2 ≥ 0

87
Then, the dual problem becomes: find y1 , y2 and y3 by which :

f (y) = 3y1 + 6y2 – 3y3 (maximize)

Subject to :

3y1 + 4y2- y3 ≤ 2

y1 + 3y2 – 2y3 ≤ 1

y1 , y2, y3 ≥ 0

2- In order to determine the optimal solution tableau for


the dual problem. Put the dual problem in its standard
form, and then we have.

f (y) = 3y1 + 6y2 – 3y3 + 0 (S1 + S2 ) (maximize)

Subject to :

3y1 + 4y2- y3 + S1 = 2

y1 + 3y2 –y3 + S2 = 1

y1 ≥ 0 for all i = 1 : 3and S1,S2≥ 0

( ) * +

[ ]

88
1st tableau

Coef.Z Cj 3 6 -3 0 0 Min
Ratio.
(B.V) CB XB

0 2 3 4 -1 1 0 ⁄

0 1 -2 ⁄
1 3 0 1

Zj=CBXj 0 0 0 0 0 0

Δj= Cj - Zj 3 6 -3 0 0
Max Z= 0

Coef.Z Cj 3 6 -3 0 0 Min
Ratio.
(B.V) CB XB

0 ⁄ 0 ⁄ 1 ⁄ ⁄

6 ⁄ ⁄ 1 ⁄ ⁄
0 1

Zj=CBXj 2 1 0 1 0 -2

Δj= Cj - Zj 0 0 -2
1 1
Max Z= 2

89
Coef.Z Cj 3 6 -3 0 0 Min
Ratio.
(B.V) CB XB

3 ⁄ 1 0 1 ⁄ ⁄

6 ⁄ 0 1 -1 ⁄ ⁄

Zj=CBXj ⁄ 3 6 -3 ⁄ ⁄

Δj= Cj - Zj 0 0 0
⁄ ⁄
Max Z=

 then, the last tableau is one of the multiple optimal


solution, and hence :

Y*1 = 2/5 , y*2 = 1/5 , y*3 = S*1 = S*2 = 0 , f (y*) = 12/5

3-From the optimal solution tableau for the dual problem, we can get
the optimal solution tableau for the primal problem as follows :

Since the optimal values for the decision variable to the primal problem
can be derived from the optimality condition row (Δj) to the dual
problem, specifically under the slack variables for the dual problem, i.e.,

X*1 = the coefficient of the 1st slack variable in the dual problem

= the coefficient of the variable y4 in the Δj row= 3/5

X*2 = the coefficient of the 2nd slack variable in the Δj row = 6/5

90
And , the optimal values for the slack variables in the primal problem are
the same coefficient for the decision variables in the dual problem, i.e.,

= the coefficient of the 1st decision variable in the Δj for the dual
problem.

= the coefficient of y1 in the Δj row = 0

= the coefficient of y2 in the Δj row = 0,

= the coefficient of y3 in the Δj row = 0 ,

ƒ(x*) = ƒ(y*) = 12/5

Finally, the optimal solution tableau for the primal problem :

Coef.Z Cj 2 1 0 0 0 Min
Ratio.
(B.V) CB XB

2 ⁄ 1 0
0

1 ⁄ 0 1
0

0 0 0 0 1

Zj=CBXj 2 1 ⁄ ⁄
0

Δj= Cj - Zj 0 0
⁄ ⁄
Min Z= 0

91
4-Determining the shadow prices for the different resources
(constraints) to the primal problem by using two different
methods:

Firstly : From the optimal solution for the primal problem.


Hence the ith shadow price for the ith resource is equal to the
coefficient for the ith slack variable in the optimality row
condition Δj. Therefore,

The 1st shadow price = the coefficient of the 1st slack variable in
the optimality row condition Δj

= the coefficient of x3 in the Δj row = 2/5 (neglect the sign)

the 2nd shadow price = the coefficient of the 2nd slack variable

= the coefficient of x4 in the Δj row = 1/5

The 3rd shadow price = the coefficient of x5 in the Δj row = 0

Secondly : One can determine the shadow prices for the set of
resources (constraints) to the primal problem form the optimal
solution for the dual problem from the optimal value for the
decision variable (solution) for it. Therefore;

The 1st shadow price for the 1st resource = the optimum value
for the 1st decision variable. i.e.,

The 1st shadow price = y*1 = 2/5 , and

The 2nd shadow price = y*2 = 1/5

92
The 3rd shadow price = y*3 = 0

.Exercise
1. Max f(x) = 5x1+ 4x2

Subject to:

6 x1 + 4x2 24

x1 + 2x2 6

- x 1 + x2 1

x2 2

x1 ≥ 0, x2 ≥ 0

Required : 1- Determine the Dual problem.

2- Solve the dual problem with the suitable simplex method.

3- From your preceding result in (2), how can you constraint


to optimal solution tableau for the primal problem.

4- Determine the shadow prices for the different resources by


using two different methods.

93
Chapter Three
Transportation Problem

Introduction
The transportation problem is to transport various
amounts of a single homogeneous commodity that are
initially stored at various origins, to different
destinations in such a way that the total transportation
cost is a minimum.
It can also be defined as to ship goods from various
origins to various destinations in such a manner that the
transportation cost is a minimum.
The availability as well as the requirements is finite. It
is assumed that the cost of shipping is linear.

Mathematical Formulation
Let there be m origins, ith origin possessing ai units of a
certain product
Let there be n destinations, with destination j requiring
bj units of a certain product

97
Let cij be the cost of shipping one unit from ith source to
jth destination
Let xij be the amount to be shipped from ith source to jth
destination
It is assumed that the total availabilities Σai satisfy the
total requirements Σbji.e.
Σai = Σbj(i = 1, 2, 3 … m and j = 1, 2, 3 …n)
The problem now, is to determine non-negative of xij
satisfying both the availability constraints

as well as requirement constraints

and the minimizing cost of transportation (shipping)

This special type of LPP is called as Transportation


Problem.

Tabular Representation
Let ‘m’ denote number of factories (F1, F2 … Fm)
Let ‘n’ denote number of warehouse (W1, W2 … Wn)
98
W→
F Capacities
W1 W2 … Wn
↓ (Availability)
F1 c11 c12 … c1n a1
F2 c21 c22 … c2n a2
. . . . . .
. . . . . .
Fm cm1 cm2 … cmn am
Required b1 b2 … bn Σai = Σbj
W→
F Capacities
W1 W2 … Wn
↓ (Availability)
F1 x11 x12 … x1n a1
F2 x21 x22 … x2n a2
. . . . . .
. . . . . .
Fm xm1 xm2 … xmn am
Required b1 b2 … bn Σai = Σbj

In general these two tables are combined by inserting


each unit cost cij with the corresponding amount xij in
the cell (i, j). The product cijxij gives the net cost of
shipping units from the factory Fi to warehouse Wj.

Some Basic Definitions


99
 Feasible Solution
A set of non-negative individual allocations (xij ≥
0) which simultaneously removes deficiencies is
called as feasible solution.
 Basic Feasible Solution
A feasible solution to ‘m’ origin, ‘n’ destination
problem is said to be basic if the number of
positive allocations are m+n-1. If the number of
allocations is less than m+n-1 then it is called as
Degenerate Basic Feasible Solution. Otherwise
it is called as Non- Degenerate Basic Feasible
Solution.
 Optimum Solution
A feasible solution is said to be optimal if it
minimizes the total transportation cost.

Methods for Initial Basic Feasible Solution


Some simple methods to obtain the initial basic feasible
solution are
1. North-West Corner Rule
2. Row Minima Method
3. Column Minima Method
100
4. Lowest Cost Entry Method (Matrix Minima
Method)
5. Vogel’s Approximation Method (Unit Cost
Penalty Method)
North-West Corner Rule
Step 1
 The first assignment is made in the cell
occupying the upper left-hand (north-west)
corner of the table.
 The maximum possible amount is allocated here
i.e. x11 = min (a1, b1). This value of x11 is then
entered in the cell (1,1) of the transportation
table.
Step 2
i. If b1 > a1, move vertically downwards to the
second row and make the second allocation of
amount x21 = min (a2, b1 - x11) in the cell (2, 1).
ii. If b1 < a1, move horizontally right side to the
second column and make the second allocation of
amount x12 = min (a1 - x11, b2) in the cell (1, 2).
iii. If b1 = a1, there is tie for the second allocation.
One can make a second allocation of magnitude
101
x12 = min (a1 - a1, b2) in the cell (1, 2) or x21 = min
(a2, b1 - b1) in the cell (2, 1)
Step 3: Start from the new north-west corner of the
transportation table and repeat steps 1 and 2 until all the
requirements are satisfied.
Example: Find the initial basic feasible solution by
using North-West Corner Rule for the following
problems
Problem (1):
W→
F Factory
W1 W2 W3 W4
↓ Capacity
F1 19 30 50 10 7
F2 70 30 40 60 9
F3 40 8 70 20 18
Warehouse
5 8 7 14 34
Requirement

102
Solution
W1 W2 W3 W5 Availability

5 2
F1 7 2 0
(19) (30)
6 3
F2 9 3 0
(30) (40)
4 14
F3 18 14 0
(70) (20)

5 8 7 14
Requirement 0 6 4 0
0 0

Initial Basic Feasible Solution


x11 = 5, x12 = 2, x22 = 6, x23 = 3, x33 = 4, x34 = 14
The transportation cost is 5 (19) + 2 (30) + 6 (30) + 3
(40) + 4 (70) + 14 (20) = 1015

Problem (2):

103
D1 D2 D3 D4 Supply
O1 1 5 3 3 34
O2 3 3 1 2 15
O3 0 2 2 3 12
O4 2 7 2 4 19
Demand 21 25 17 17 80

Solution

D1 D2 D3 D4 Supply
21 13
O1 (1) (5) 34 13 0
12 3
O2 (3) (1) 15 3 0
12
O3 (2) 12 0
2 17
O4 (2) (4) 19 17
Demand 21 25 17 17
0 12 14 0
0 2
0
Initial Basic Feasible Solution

104
x11 = 21, x12 = 13, x22 = 12, x23 = 3, x33 = 12, x43 = 2, x44 =
17
The transportation cost is 21 (1) + 13 (5) + 12 (3) + 3 (1) +
12 (2) + 2 (2) + 17 (4) = 221
Problem (3):
From To Supply
2 11 10 3 7 4
1 4 7 2 1 8
3 1 4 8 12 9
Demand 3 3 4 5 6
Solution
From To Supply
3 1
(2) (11) 4 1 0
2 4 2
8 6 2
(4) (7) (2)
0
3 6
(8) (12) 9 6 0
3 3 4 5 6
Demand 0 2 0 3 0
0 0

Initial Basic Feasible Solution

105
x11 = 3, x12 = 1, x22 = 2, x23 = 4, x24 = 2, x34 = 3, x35 = 6
The transportation cost is 3 (2) + 1 (11) + 2 (4) + 4 (7) + 2
(2) + 3 (8) + 6 (12) = 153

Row Minima Method


Step 1
 The smallest cost in the first row of the
transportation table is determined.
 Allocate as much as possible amount xij = min (a1, bj)
in the cell (1, j) so that the capacity of the origin or
the destination is satisfied.
Step 2
 If x1j = a1, so that the availability at origin O1 is
completely exhausted, cross out the first row of the
table and move to second row.
 If x1j = bj, so that the requirement at destination Dj is
satisfied, cross out the jth column and reconsider the
first row with the remaining availability of origin O1.
 If x1j = a1 = bj, the origin capacity a1 is completely
exhausted as well as the requirement at destination
Dj is satisfied. An arbitrary tie-breaking choice is
made. Cross out the jth column and make the second
allocation x1k = 0 in the cell (1, k) with c 1k being the

106
new minimum cost in the first row. Cross out the
first row and move to second row.
Step 3
Repeat steps 1 and 2 for the reduced transportation table
until all the requirements are satisfied
Example:
Determine the initial basic feasible solution using Row
Minima Method

W1 W2 W3 W4 Availability
F1 19 30 50 10 7
F2 70 30 40 60 9
F3 40 80 70 20 18
Requirement 5 8 7 14
Solution
W1 W2 W3 W4
7
X
F1 (19) (30) (50) (10)

F2 9
(70) (30) (40) (60)

F3 18
(40) (80) (70) (20)
5 8 7 7

107
W1 W2 W3 W4
7
X
F1 (19) (30) (50) (10)
8
F2 (70) (30) (40) (60) 1

F3 18
(40) (80) (70) (20)
5 X 7 7

W1 W2 W3 W4
7
X
F1 (19) (30) (50) (10)
8 1
X
F2 (70) (30) (40) (60)

F3 (40) (80) (70) (20) 18


5 X 6 7
W1 W2 W3 W4
7
X
F1 (19) (30) (50) (10)
8 1
X
F2 (70) (30) (40) (60)
5 6 7
X
F3 (40) (80) (70) (20)
X X X X
Initial Basic Feasible Solution
108
x14 = 7, x22 = 8, x23 = 1, x31 = 5, x33 = 6, x34 = 7
The transportation cost is 7 (10) + 8 (30) + 1 (40) + 5 (40)
+ 6 (70) + 7 (20) = 1110

Lowest Cost Entry Method (Matrix Minima Method)


Step 1: Determine the smallest cost in the cost matrix of
the transportation table. Allocate xij = min (ai, bj) in the cell
(i, j)
Step 2
 If xij = ai, cross out the ith row of the table and
decrease bj by ai. Go to step 3.
 If xij = bj, cross out the jth column of the table and
decrease ai by bj. Go to step 3.
 If xij = ai= bj, cross out the ith row or jth column but
not both.
Step 3: Repeat steps 1 and 2 for the resulting reduced
transportation table until all the requirements are satisfied.
Whenever the minimum cost is not unique, make an
arbitrary choice among the minima.
Example:Find the initial basic feasible solution using
Matrix Minima method

109
W1 W2 W3 W4 Availability
F1 19 30 50 10 7
F2 70 30 40 60 9
F3 40 8 70 20 18
Requirement 5 8 7 14
Solution

W1 W2 W3 W4

F1 7
(19) (30) (50) (10)

F2 9
(70) (30) (40) (60)

F3 8 10
(40) (8) (70) (20)
5 X 7 14

W1 W2 W3 W4
7
F1 X
(19) (30) (50) (10)

F2 9
(70) (30) (40) (60)
8
F3 10
(40) (8) (70) (20)
5 X 7 7

W1 W2 W3 W4

110
7
F1 X
(19) (30) (50) (10)

F2 9
(70) (30) (40) (60)
8 7
F3 3
(40) (8) (70) (20)
5 X 7 X

W1 W2 W3 W4
7
F1 X
(19) (30) (50) (10)

F2 9
(70) (30) (40) (60)
3 8 7
F3 X
(40) (8) (70) (20)
2 X 7 X

W1 W2 W3 W4
7
F1 X
(19) (30) (50) (10)
2 7
F2 X
(70) (30) (40) (60)
3 8 7
F3 X
(40) (8) (70) (20)
X X X X

111
Initial Basic Feasible Solution

x14 = 7, x21 = 2, x23 = 7, x31 = 3, x32 = 8, x34 = 7


The transportation cost is 7 (10) + 2 (70) + 7 (40) + 3 (40)
+ 8 (8) + 7 (20) = 814
Example:
To Availability
2 11 10 3 7 4
From 1 4 7 2 1 8
3 9 4 8 12 9
Requirement 3 3 4 5 6
Solution
To
4
4 0
(3)
3 5
8 5 0
(1) (1)
From 3 4 1 1
9 5 4 1 0
(9) (4) (8) (12)
3 3 4 5 6
0 0 0 1 1
0 0
Initial Basic Feasible Solution
x14 = 4, x21 = 3, x25 = 5, x32 = 3, x33 = 4, x34 = 1, x35 = 1
112
The transportation cost is 4 (3) + 3 (1) + 5(1) + 3 (9) + 4 (4)
+ 1 (8) + 1 (12) = 78

Vogel’s Approximation Method (Unit Cost


Penalty Method)

Step1: For each row of the table, identify the smallest and
the next to smallest cost. Determine the difference
between them for each row. These are called penalties. Put
them aside by enclosing them in the parenthesis against the
respective rows. Similarly compute penalties for each
column.

Step 2:Identify the row or column with the largest penalty.


If a tie occurs then use an arbitrary choice. Let the largest
penalty corresponding to the ith row have the cost cij.
Allocate the largest possible amount xij = min (ai, bj) in the
cell (i, j) and cross out either ith row or jth column in the
usual manner.

Step 3: Again compute the row and column penalties for


the reduced table and then go to step 2. Repeat the
procedure until all the requirements are satisfied.

113
Example:

Find the initial basic feasible solution using vogel’s


approximation method

Problem (1):

W1 W2 W3 W4 Availability
F1 19 30 50 10 7
F2 70 30 40 60 9
F3 40 8 70 20 18
Requirement 5 8 7 14

Solution

W1 W2 W3 W4 Av. Penalty
F1 19 30 50 10 7 19-10=9
F2 70 30 40 60 9 40-30=10
F3 40 8 70 20 18 20-8=12
Req. 5 8 7 14
Penalty 40-19=21 30-8=22 50-40=10 20-10=10

W1 W2 W3 W4 Availability Penalty
F1 (19) (30) (50) (10) 7 9
F2 (70) (30) (40) (60) 9 10

114
F3 (40) 8(8) (70) (20) 18/10 12
Requirement 5 8/0 7 14
Penalty 21 22 10 10

W1 W2 W3 W4 Availability Penalty
F1 5(19) (30) (50) (10) 7/2 9
F2 (70) (30) (40) (60) 9 20
F3 (40) 8(8) (70) (20) 18/10 20
Requirement 5/0 X 7 14
Penalty 21 X 10 10

W1 W2 W3 W4 Availability Penalty
F1 5(19) (30) (50) (10) 7/2 40
F2 (70) (30) (40) (60) 9 20
F3 (40) 8(8) (70) 10(20) 18/10/0 50
Requirement X X 7 14/4
Penalty X X 10 10

W1 W2 W3 W4 Availability Penalty
F1 5(19) (30) (50) 2(10) 7/2/0 40
F2 (70) (30) (40) (60) 9 20
F3 (40) 8(8) (70) 10(20) X X

115
Requirement X X 7 14/4/2
Penalty X X 10 50

W1 W2 W3 W4 Availability Penalty
F1 5(19) (30) (50) 2(10) X X
F2 (70) (30) 7(40) 2(60) X X
F3 (40) 8(8) (70) 10(20) X X
Requirement X X X X
Penalty X X X X

Initial Basic Feasible Solution


x11 = 5, x14 = 2, x23 = 7, x24 = 2, x32 = 8, x34 = 10
The transportation cost is 5 (19) + 2 (10) + 7 (40) + 2 (60)
+ 8 (8) + 10 (20) = 779
Problem (2):
Stores Availability
I II III IV
A 21 16 15 13 11
Warehouse B 17 18 14 23 13
C 32 27 18 41 19
Requirement 6 10 12 15

Solution

116
Stores Availability Penalty
I II III IV
A (21) (16) (15) (13) 11 2
Warehouse B (17) (18) (14) (23) 13 3
C (32) (27) (18) (41) 19 9

Requirement 6 10 12 15
Penalty 4 2 1 10

Stores Availability Penalty


I II III IV
A (21) (16) (15) 11(13) 11/0 2
Warehouse B (17) (18) (14) (23) 13 3
C (32) (27) (18) (41) 19 9
Requirement 6 10 12 15/4
Penalty 4 2 1 10

Stores Availability Penalty

I II III IV

A (21) (16) (15) 11(13) X X


Warehouse B (17) (18) (14) 4(23) 13/9 3
C (32) (27) (18) (41) 19 9
Requirement 6 10 12 15/4/0

Penalty 15 9 4 18

Stores Av. Pen.

I II III IV

Warehouse A (21) (16) (15) 11(13) X X

117
B 6(17) (18) (14) 4(23) 13/9/3 3
C (32) (27) (18) (41) 19 9
Req. 6/0 10 12 X
Penalty 15 9 4 X

Stores Av. Pen.


I II III IV
A (21) (16) (15) 11(13) X X
Warehouse B 6(17) 3(18) (14) 4(23) 13/9/3/0 4
C (32) (27) (18) (41) 19 9
Req. X 10/7 12 X
Penalty X 9 4 X

Stores Av. Pen.


I II III IV
A (21) (16) (15) 11(13) X X
Warehouse B 6(17) 3(18) (14) 4(23) X X
C (32) 7(27) 12(18) (41) X X
Req. X X X X
Penalty X X X X
Initial Basic Feasible Solution
x14 = 11, x21 = 6, x22 = 3, x24 = 4, x32 = 7, x33 = 12

118
The transportation cost is 11 (13) + 6 (17) + 3 (18) + 4 (23)
+ 7 (27) + 12 (18) = 796
Exercise
1. Determine an initial basic feasible solution to the
following transportation problem using north-west corner
rule.
I II III IV Supply
A 13 11 15 20 2000
From B 17 14 12 13 6000
C 18 18 15 12 7000
Demand 3000 3000 4000 5000

[Ans. x11 = 2, x21 = 1, x22 = 3, x23 = 2, x34 = 5]

2. Obtain an initial basic feasible solution to the following


transportation problem using matrix minima method.
D1 D2 D3 D4 Capacity
O1 1 2 3 4 6
From O2 4 3 2 0 8
O3 0 2 2 1 10
Demand 4 6 8 6
[Ans. x12 = 6, x23 = 2, x24 = 6, x31 = 4, x32 = 0, x33 = 6]
3. Determine the minimum cost to the following
transportation problem using Vogel’s method.

119
D1 D2 D3 D4 D5 Capacity
O1 2 11 10 3 7 4
From O2 1 4 7 2 1 8
O3 3 9 4 8 12 9
Demand 3 3 4 5 6 21
[Ans. Min cost = 68]
4. Determine the minimum cost to the following
transportation problem using matrix minima method and
vogel’s method
D1 D2 D3 D4 Capacity
O1 1 2 1 4 30
From O2 3 3 2 1 50
O3 4 2 5 9 20
Demand 20 40 30 10
[Ans. Min cost = 180]
Examining the Initial Basic Feasible Solution for
Non-Degeneracy
Examine the initial basic feasible solution for non-
degeneracy. If it is said to be non-degenerate then it has
the following two properties
 Initial basic feasible solution must contain
exactly m + n – 1 number of individual
allocations.

120
 These allocations must be in independent
positions
Independent Positions
  
  
 

 
 
 

Non-Independent Positions

121
Transportation Algorithm for Minimization
Problem (MODI Method)
Step1: Construct the transportation table entering the origin
capacities ai, the destination requirement bj and the cost cij
Step 2: Find an initial basic feasible solution by vogel’s
method or by any of the given method.
Step 3: For all the basic variables xij, solve the system of
equations ui + vj = cij, for all i, j for which cell (i, j) is in the
basis, starting initially with some ui = 0, calculate the
values of ui and vj on the transportation table
Step 4: Compute the cost differences dij = cij – ( ui + vj ) for
all the non-basic cells
Step 5: Apply optimality test by examining the sign of each
dij
 If all dij ≥ 0, the current basic feasible solution is
optimal
 If at least one dij< 0, select the variable x rs (most
negative) to enter the basis.
 Solution under test is not optimal if any dij is
negative and further improvement is required by
repeating the above process.

122
Step 6: Let the variable xrs enter the basis. Allocate an
unknown quantity Ө to the cell (r, s). Then construct a loop
that starts and ends at the cell (r, s) and connects some of
the basic cells. The amount Ө is added to and subtracted
from the transition cells of the loop in such a manner that
the availabilities and requirements remain satisfied.
Step 7: Assign the largest possible value to the Ө in such a
way that the value of at least one basic variable becomes
zero and the other basic variables remain non-negative. The
basic cell whose allocation has been made zero will leave
the basis.
Step 8: Now, return to step 3 and repeat the process until
an optimal solution is obtained.
Worked Examples
Example 1
Find an optimal solution

W1 W2 W3 W4 Availability
F1 19 30 50 10 7
F2 70 30 40 60 9
F3 40 8 70 20 18
Requirement 5 8 7 14
Solution

123
1. Applying vogel’s approximation method for finding the
initial basic feasible solution

W1 W2 W3 W4 Availability Penalty
F1 5(19) (30) (50) 2(10) X X
F2 (70) (30) 7(40) 2(60) X X
F3 (40) 8(8) (70) 10(20) X X
Requirement X X X X
Penalty X X X X

Minimum transportation cost is 5 (19) + 2 (10) + 7 (40) + 2


(60) + 8 (8) + 10 (20) = 779
2. Check for Non-degeneracy
The initial basic feasible solution has m + n – 1 i.e. 3 + 4 –
1 = 6 allocations in independent positions. Hence
optimality test is satisfied.
3. Calculation of ui and vj :ui + vj = cij

ui
 (19)  (10) u1= -10
 (40)  (60) u2 = 40
 (8)  (20) u3 = 0
vj v1= 29 v2 = 8 v3 = 0 v4 = 20
Assign a ‘u’ value to zero. (Convenient rule is to select the
ui, which has the largest number of allocations in its row)

124
Let u3 = 0, then
u3 + v4= 20 which implies 0 + v 4 = 20, so v4 = 20
u2 + v4= 60 which implies u2 + 20 = 60, so u2 = 40
u1 + v4= 10 which implies u1 + 20 = 10, so u1 = -10
u2 + v3= 40 which implies 40 + v3 = 40, so v3 = 0
u3 + v2= 8 which implies 0 + v2 = 8, so v2 = 8
u1 + v1= 19 which implies -10 + v1= 19, so v1 = 29

4.Calculation of cost differences for non basic cells


dij = cij – ( ui + vj )
cij ui + vj
 (30) (50)   -2 -10 
(70) (30)   69 48  
(40)  (70)  29  0 

dij = cij – ( ui + vj )
 32 60 
1 -18  
11  70 
5.Optimality test
dij< 0 i.e. d22 = -18
so x22 is entering the basis

125
6.Construction of loop and allocation of unknown
quantity Ө

We allocate Ө to the cell (2, 2). Reallocation is done by


transferring the maximum possible amount Ө in the marked
cell. The value of Ө is obtained by equating to zero to the
corners of the closed loop. i.e. min(8-Ө, 2-Ө) = 0 which
gives Ө = 2. Therefore x 24 is outgoing as it becomes zero.
5 (19) 2 (10)
2 (30) 7 (40)
6 (8) 12 (20)

Minimum transportation cost is 5 (19) + 2 (10) + 2 (30) + 7


(40) + 6 (8) + 12 (20) = Rs. 743
7.Improved Solution

ui
 (19)  (10) u1= -10
 (30)  (40) u2 = 22
 (8)  (20) u3 = 0
vj v1= 29 v2 = 8 v3 = 18 v4 = 20
126
cij ui + vj
 (30) (50)   -2 8 
(70)   (60) 51   42
(40)  (70)  29  18 

dij = cij – ( ui + vj )
 32 42 
19   18
11  52 
Since dij> 0, an optimal solution is obtained with minimal
cost 743

Example 2
Solve by lowest cost entry method and obtain an optimal
solution for the following problem

Available
50 30 220 1
From 90 45 170 3
250 200 50 4
Required 4 2 2

Solution
By lowest cost entry method
127
Available
1(30) 1/0
From 2(90) 1(45) 3/2/0
2(250) 2(50) 4/2/0
Required 4/2/2 2/1/0 2/0

Minimum transportation cost is 1 (30) + 2 (90) + 1 (45) + 2


(250) + 2 (50) = Rs. 855
Check for Non-degeneracy
The initial basic feasible solution has m + n – 1 i.e. 3 + 3 –
1 = 5 allocations in independent positions. Hence
optimality test is satisfied.

Calculation of ui and vjby equationui + vj =cij

ui
 (30) u1= -15
 (90)  (45) u2 = 0
 (250)  (50) u3 = 160
vj v1= 90 v2 = 45 v3 = -110

128
Calculation of cost differences for non-basic cellsdij = cij – ( ui + vj
)

cij u i + vj
50  220 75  -125
  170   -110
 200   205 

dij = cij – ( ui + vj )
-25  345
  280
 -5 

Optimality test
dij< 0 i.e. d11 = -25 is most negative
So x11 is entering the basis
Construction of loop and allocation of unknown
quantity Ө

min(2-Ө, 1-Ө) = 0 which gives Ө = 1. Therefore x 12 is


outgoing as it becomes zero.
129
1(50)

1(90) 2(45)

2(250) 2(50)

Minimum transportation cost is 1 (50) + 1 (90) + 2 (45) + 2


(250) + 2 (50) = 830
II Iteration
Calculation of ui and vjby equationui + vj = cij

ui
 (50) u1= -40
 (90)  (45) u2 = 0
 (250)  (50) u3 = 160
vj v1= 90 v2 = 45 v3 = -110

Calculation of dij = cij – ( ui + vj )

cij u i + vj
 30 220  5 -150
  170   -110
 200   205 

dij = cij – ( ui + vj )
 25 370

130
  280
 -5 

Optimality test
dij< 0 i.e. d32 = -5
So x32 is entering the basis
Construction of loop and allocation of unknown
quantity Ө

2 – Ө = 0 which gives Ө = 2. Therefore x 22 and x31 is


outgoing as it becomes zero.

1(50)

3(90) 0(45)

2(200) 2(50)

Minimum transportation cost is 1 (50) + 3 (90) + 2 (200) +


2 (50) = Rs. 820

131
III Iteration
Calculation of ui and vjby equationui + vj = cij

ui
 (50) u1= -40
 (90)  (45) u2 = 0
 (200)  (50) u3 = 155
vj v1= 90 v2 = 45 v3 = -105

Calculation of dij = cij – ( ui + vj )

u i + vj
cij
 30 220  5 -145
  170   -105
250   245  

dij = cij – ( ui + vj )
 25 365
  275
5  

Since dij> 0, an optimal solution is obtained with minimal


cost Rs.820

Example 3

132
Is x13 = 50, x14 = 20, x21 = 55, x31 = 30, x32 = 35, x34 = 25 an
optimal solution to the transportation problem.
Available
6 1 9 3 70
From 11 5 2 8 55
10 12 4 7 90
Required 85 35 50 45
Solution

Available
50(9) 20(3) X
From 55(11) X
30(10) 35(12) 25(7) X
Required X X X X

Minimum transportation cost is 50 (9) + 20 (3) + 55 (11) +


30 (10) + 35 (12) + 25 (7) = 2010
Check for Non-degeneracy
The initial basic feasible solution has m + n – 1 i.e. 3 + 4 –
1 = 6 allocations in independent positions. Hence
optimality test is satisfied.
Calculation of ui and vjby equationui + vj = cij

133
ui
 (9)  (3) u1= -4
 (11) u2 = 1
 (10)  (12)  (7) u3 = 0
vj v1= 10 v2 = 12 v3 = 13 v4 = 7
Calculation of cost differences for non-basic cells
dij = cij – ( ui + vj )
cij ui + vj
6 1   6 8  
 5 2 8  13 14 8
  4    13 

dij = cij – ( ui + vj )
0 -7  
 -8 -12 0
  -9 
Optimality test
dij< 0 i.e. d23 = -12 is most negative
So x23 is entering the basis
Construction of loop and allocation of unknown
quantity Ө

134
min(50-Ө, 55-Ө, 25-Ө) = 25 which gives Ө = 25. Therefore
x34 is outgoing as it becomes zero.
25(9) 45(3)
30(11) 25(2)
55(10) 35(12)
Minimum transportation cost is 25 (9) + 45 (3) + 30 (11) +
25 (2) + 55 (10) + 35 (12) = Rs. 1710

II iteration
Calculation of ui and vjby equationui + vj = cij

ui
 (9)  (3) u1= 8
 (11)  (2) u2 = 1
 (10)  (12) u3 = 0
vj v1= 10 v2 = 12 v3 = 1 v4 = -5

Calculation of cost differences for non-basic cells


dij = cij – ( ui + vj )

135
cij ui + vj

6 1   20  
18
-
 5  8  13 
4
-
  4 7   1
5

dij = cij – ( ui + vj )
-12 -19  
 -8  12
  3 12
Optimality test
dij< 0 i.e. d12 = -19 is most negative
So x12 is entering the basis
Construction of loop and allocation of unknown
quantity Ө

min(25-Ө, 30-Ө, 35-Ө) = 25 which gives Ө = 25. Therefore


x13 is outgoing as it becomes zero.

136
25(1) 45(3)
5(11) 50(2)
80(10) 10(12)
Minimum transportation cost is 25 (1) + 45 (3) + 5 (11) +
50 (2) + 80 (10) + 10 (12) = 1235
III Iteration
Calculation of ui and vjby equationui + vj = cij

ui
 (1)  (3) u1= -11
 (11)  (2) u2 = 1
 (10)  (12) u3 = 0
vj v1= 10 v2 = 12 v3 = 1 v4 = 14

Calculation of cost differences for non-basic cells


dij = cij – ( ui + vj )
cij ui + vj
-
6  9   -10 
1

 5  8  
13 15

  4 7   1
14

137
dij = cij – ( ui + vj )
7  19 
 -8  -7
  3 -7

Optimality test
dij< 0 i.e. d22 = -8 is most negative
So x22 is entering the basis
Construction of loop and allocation of unknown
quantity Ө

min(5-Ө, 10-Ө) = 5 which gives Ө = 5. Therefore x 21 is


outgoing as it becomes zero.

25(1) 45(3)
5(5) 50(2)
85(10) 5(12)
Minimum transportation cost is 25 (1) + 45 (3) + 5 (5) + 50
(2) + 85 (10) + 5 (12) = Rs. 1195
IV Iteration
Calculation of ui and vj : - ui + vj = cij

138
ui
 (1)  (3) u1= -11
 (5)  (2) u2 = -7
 (10)  (12) u3 = 0
vj v1= 10 v2 = 12 v3 = 9 v4 = 14
Calculation of cost differences for non-basic cells
dij = cij – ( ui + vj )
cij ui + vj
-
6  9   -2 
1

  8 3   7
11

  4 7   9
14

dij = cij – ( ui + vj )
7  11 
8   1
  -5 -7
Optimality test
dij< 0 i.e. d34 = -7 is most negative
So x34 is entering the basis
Construction of loop and allocation of unknown
quantity Ө

139
min(5-Ө, 45-Ө) = 5 which gives Ө = 5. Therefore x 32 is
outgoing as it becomes zero.

30(1) 40(3)
5(5) 50(2)
85(10) 5(7)
Minimum transportation cost is 30 (1) + 40 (3) + 5 (5) + 50
(2) + 85 (10) + 5 (7) = Rs. 1160

V Iteration
Calculation of ui and vjby equationui + vj = cij

ui
 (1)  (3) u1= -4
 (5)  (2) u2 = 0
 (10)  (7) u3 = 0
vj v1= 10 v2 = 5 v3 = 2 v4 = 7

Calculation of cost differences for non-basic cells


dij = cij – ( ui + vj )

140
cij ui + vj
6  9  6  -2 

   
11 8 10 7
 12 4   5 2 

dij = cij – ( ui + vj )
0  11 
1   1
 7 2 
Since dij> 0, an optimal solution is obtained with minimal
cost Rs.1160. Furthermore d11 = 0 which indicates that
alternative optimal solution also exists.
Exercises

1. Determine the optimal solution of the given


transportation problem

To Supply
2 3 11 7 6
1 0 6 1 1
From
5 8 15 10 10
Demand 7 5 3 2 17

141
[Ans. x12 = 5, x13 = 1, x24 = 1, x31 = 7, x33 = 2, x34 = 1 Min
cost = Rs 102]
2. Using North-West Corner rule for initial basic
feasible solution, obtain an optimum basic feasible
solution to the following problem

To Available
7 3 4 2
From 2 1 3 3
3 4 6 5
Demand 4 1 5 10

[Ans. x13 = 2, x22 = 1, x23 = 2, x31 = 4, x33 = 1 Min cost =33]


3. Determine the optimal solution of the given
transportation problem

To Supply
10 7 3 6 3
1 6 7 3 5
From
7 4 5 3 7
Demand 3 2 6 4
[Ans. x13 = 3, x21 = 3, x24 = 2, x32 = 2, x33 = 3, x34 = 2, Min
cost = 47]

142
Chapter Four
Network Analysis
Shortest Route Problem
The criterion of this method is to find the shortest
distance between two nodes with minimal cost.
Example 1
Find the shortest path

Solution

143
Solved
nodes
Closest

Last connection
nth nearest node

Min. distance
directly Total
connected
n connected distance
unsolved
to involved
node
unsolved
nodes
1 a c 7 c 7 a-c
a b 13 b 13 a-b
2
c e 7+6 =13 e 13 c-e
b d 13+5 =18 d 18 b-d
3 c f 7+11 =18 f 18 c-f
e h 13+8 =21 - - -
e h 13+8 =21 h 21 e-h
4 d g 18+9 =27 - - -
f h 18+5 =23 - - -
e g 13+10 =23 g 23 e-g
5 h i 21+10 =31 - - -
d g 18+9 =27 - - -
g i 23+6 =29 i 29 g-i
6
h i 21+10 =31 - - -

144
The shortest path from a to i is a → c →e →g → i
Distance = 7 + 6 + 10 + 6 = 29 units
Example 2

145
Solution:
Solved
nodes
Closest
directly Total nth
connected Min. Last
n connected distance nearest
unsolved distance connection
to involved node
node
unsolved
nodes
1 1 3 1 3 1 1-3
1 2 5 - - -
2
3 2 1+2 =3 2 3 3-2
2 5 3+1 =4 5 4 2-5
3
3 4 1+6 =7 - - -
2 6 3+6 =9 - - -
4 3 4 1+6 =7 4 7 3-4
5 4 4+3 =7 4 7 5-4
2 6 3+6 =9 6 9 2-6
5 4 6 7+4 =11 - - -
5 6 4+5 =9 6 9 5-6
4 7 7+6 =13 - - -
6 5 7 4+9 =13 - - -
6 7 9+2 =11 7 11 6-7

The shortest path from 1 to 7 can be

146
1 →3 → 2 → 6 →7
Total distance is 11 units

1 → 3 → 2 →5 → 6 →7
Total distance = 11 units
Minimal Spanning Tree Problem
A tree is defined to be an undirected, acyclic and
connected graph. A spanning tree is a sub graph of G
(undirected, connected graph), is a tree and contains all
147
the vertices of G. A minimum spanning tree is a
spanning tree but has weights or lengths associated
with edges and the total weight is at the minimum.
Prim’s Algorithm
 It starts at any vertex (say A) in a graph and finds
the least cost vertex (say B) connected to the start
vertex.
 Now either from A or B, it will find the next least
costly vertex connection, without creating cycle
(say C)
 Now either from A, B or C find the next least
costly vertex connection, without creating a cycle
and so on.
 Eventually all the vertices will be connected
without any cycles and a minimum spanning tree
will be the result.

Example 1
Suppose it is desired to establish a cable
communication network that links major cities, which
is shown in the figure. Determine how the cities are

148
connected such that the total cable mileage is
minimized.

Solution
C = {LA} C' = {SE, DE, DA, EH, NY, DC}
C = {LA, SE} C' = {DE, DA, EH, NY, DC}
C = {LA, SE, DE} C' = {DA, EH, NY, DC}
C = {LA, SE, DE, DA} C' = {EH, NY, DC}
C = {LA, SE, DE, DA, EH} C' = {NY, DC}
C = {LA, SE, DE, DA, EH, NY} C' = {DC}
C = {LA, SE, DE, DA, EH, NY, DC} C' = { }

The resultant network is

149
Thus the total cable mileage is 1100 + 1300 + 780 +
900 + 800 + 200 = 5080
Example 2
For the following graph obtain the minimum spanning
tree. The numbers on the branches represent the cost.

Solution
C = {A} C' = {B, C, D, E, F, G}
C = {A, D} C' = {B, C, E, F, G}
C = {A, D, B} C' = {C, E, F, G}
150
C = {A, D, B, C} C' = {E, F, G}
C = {A, D, B, C, G} C' = {E, F}
C = {A, D, B, C, G, F} C' = {E}
C = {A, D, B, C, G, F, E} C' = { }
The resultant network is

Cost = 2 + 1 + 4 + 3 + 3 + 5 = 18 units
Example 3
Solve the minimum spanning problem for the given
network. The numbers on the branches represent in
terms of miles.

151
Solution
C = {1} C' = {2, 3, 4, 5, 6}
C = {1, 2} C' = {3, 4, 5, 6}
C = {1, 2, 5} C' = {3, 4, 6}
C = {1, 2, 5, 4} C' = {3, 6}
C = {1, 2, 5, 4, 6} C' = {3}
C = {1, 2, 5, 4, 6, 3} C' = {}
The resultant network is

1 + 4 + 5+ 3 + 3 = 16 miles

152
Maximal Flow Problem
Algorithm
Step1: Find a path from source to sink that can
accommodate a positive flow of material. If no path
exists go to step 5
Step2: Determine the maximum flow that can be
shipped from this path and denote by ‘k’ units.
Step3:Decrease the direct capacity (the capacity in the
direction of flow of k units) of each branch of this path
‘k’ and increase the reverse capacity k1. Add ‘k’ units
to the amount delivered to sink.
Step4: Go to step1
Step5: The maximal flow is the amount of material
delivered to the sink. The optimal shipping schedule is
determined by comparing the original network with the
final network. Any reduction in capacity signifies
shipment.
Example 1
Consider the following network and determine the
amount of flow between the networks.

153
Solution
Iteration 1: 1 – 3 – 5

Iteration 2: 1 – 2 – 3 – 4 – 5

154
Iteration 3: 1 – 4 – 5

Iteration 4: 1 – 2 – 5

155
Iteration 5: 1 – 3 – 2 – 5

Maximum flow is 60 units. Therefore the network can


be written as

156
Example 2
Solve the maximal flow problem

Solution
Iteration 1: O – A – D – T

157
Iteration 2: O – B – E – T

Iteration 3: O – A – B – D – T

158
Iteration 4: O – C – E – D – T

Iteration 5: O – C – E – T

159
Iteration 6: O – B – D – T

Therefore there are no more augmenting paths. So the


current flow pattern is optimal. The maximum flow is
13 units.

160
Exercise
1. Find the shortest path

2. Solve the maximal flow problem

3. Solve the minimal spanning tree

161
Introduction to CPM / PERT Techniques
CPM/PERT or Network Analysis as the technique is
sometimes called, developed along two parallel
streams, one industrial and the other military.
CPM (Critical Path Method)was the discovery of
M.R.Walker of E.I.Du Pont de Nemours & Co. and
J.E.Kelly of Remington Rand, circa 1957. The
computation was designed for the UNIVAC-I
computer. The first test was made in 1958, when CPM
was applied to the construction of a new chemical
plant. In March 1959, the method was applied to
maintenance shut-down at the Du Pont works in
Louisville, Kentucky. Unproductive time was reduced
from 125 to 93 hours.
PERT (Project Evaluation and Review
Technique)was devised in 1958 for the POLARIS
missile program by the Program Evaluation Branch of
the Special Projects office of the U.S.Navy, helped by
the Lockheed Missile Systems division and the
Consultant firm of Booz-Allen & Hamilton. The
calculations were so arranged so that they could be
carried out on the IBM Naval Ordinance Research
Computer (NORC) at Dahlgren, Virginia.
162
The methods are essentially network-oriented
techniques using the same principle. PERT and CPM
are basically time-oriented methods in the sense that
they both lead to determination of a time schedule for
the project. The significant difference between two
approaches is that the time estimates for the different
activities in CPM were assumed to be deterministic
while in PERT these are described probabilistically.
These techniques are referred as project scheduling
techniques.
In CPM activities are shown as a network of
precedence relationships using activity-on-node
network construction
– Single estimate of activity time
– Deterministic activity times
USED IN: Production management - for the jobs of
repetitive in nature where the activity time estimates
can be predicted with considerable certainty due to the
existence of past experience.
In PERT activities are shown as a network of
precedence relationships using activity-on-arrow
network construction
– Multiple time estimates
– Probabilistic activity times
USED IN: Project management - for non-repetitive
jobs (research and development work), where the time
and cost estimates tend to be quite uncertain. This
technique uses probabilistic time estimates.

Benefits of PERT/CPM
 Useful at many stages of project management
163
 Mathematically simple
 Give critical path and slack time
 Provide project documentation
 Useful in monitoring costs
Limitations of PERT/CPM
 Clearly defined, independent and stable activities
 Specified precedence relationships
 Over emphasis on critical paths

Applications of CPM / PERT


These methods have been applied to a wide variety of
problems in industries and have found acceptance even
in government organizations. These include
 Construction of a dam or a canal system in a
region
 Construction of a building or highway
 Maintenance or overhaul of airplanes or oil
refinery
 Space flight
 Cost control of a project using PERT / COST
 Designing a prototype of a machine
 Development of supersonic planes

Basic Steps in PERT / CPM


Project scheduling by PERT / CPM consists of four
main steps
1. Planning
 The planning phase is started by splitting the total
project in to small projects. These smaller
projects in turn are divided into activities and are
analyzed by the department or section.
164
 The relationship of each activity with respect to
other activities are defined and established and
the corresponding responsibilities and the
authority are also stated.
 Thus the possibility of overlooking any task
necessary for the completion of the project is
reduced substantially.
2. Scheduling
 The ultimate objective of the scheduling phase is
to prepare a time chart showing the start and
finish times for each activity as well as its
relationship to other activities of the project.
 Moreover the schedule must pinpoint the critical
path activities which require special attention if
the project is to be completed in time.
 For non-critical activities, the schedule must
show the amount of slack or float times which
can be used advantageously when such activities
are delayed or when limited resources are to be
utilized effectively.
3. Allocation of resources
 Allocation of resources is performed to achieve
the desired objective. A resource is a physical
variable such as labour, finance, equipment and
space which will impose a limitation on time for
the project.
 When resources are limited and conflicting,
demands are made for the same type of resources
a systematic method for allocation of resources
become essential.

165
 Resource allocation usually incurs a compromise
and the choice of this compromise depends on the
judgment of managers.
4. Controlling
 The final phase in project management is
controlling. Critical path methods facilitate the
application of the principle of management by
expectation to identify areas that are critical to
the completion of the project.
 By having progress reports from time to time and
updating the network continuously, a better
financial as well as technical control over the
project is exercised.
 Arrow diagrams and time charts are used for
making periodic progress reports. If required, a
new course of action is determined for the
remaining portion of the project.
The Framework for PERT and CPM
Essentially, there are six steps which are common to
both the techniques. The procedure is listed below:
I. Define the Project and all of its significant
activities or tasks. The Project (made up of
several tasks) should have only a single start
activity and a single finish activity.
II. Develop the relationships among the activities.
Decide which activities must precede and which
must follow others.
III. Draw the "Network" connecting all the activities.
Each Activity should have unique event numbers.
Dummy arrows are used where required to avoid
giving the same numbering to two activities.
166
IV. Assign time and/or cost estimates to each activity
V. Compute the longest time path through the
network. This is called the critical path.
VI. Use the Network to help plan, schedule, and
monitor and control the project.
The Key Concept used by CPM/PERT is that a small
set of activities, which make up the longest path
through the activity network control the entire project.
If these "critical" activities could be identified and
assigned to responsible persons, management resources
could be optimally used by concentrating on the few
activities which determine the fate of the entire project.
Non-critical activities can be replanned, rescheduled
and resources for them can be reallocated flexibly,
without affecting the whole project.
Five useful questions to ask when preparing an activity
network are:
 Is this a Start Activity?
 Is this a Finish Activity?
 What Activity Precedes this?
 What Activity Follows this?
 What Activity is Concurrent with this?
Network Diagram Representation
In a network representation of a project certain
definitions are used
1. Activity
Any individual operation which utilizes resources and
has an end and a beginning is called activity. An arrow
is commonly used to represent an activity with its head
indicating the direction of progress in the project. These
are classified into four categories
167
1. Predecessor activity – Activities that must be
completed immediately prior to the start of
another activity are called predecessor activities.
2. Successor activity – Activities that cannot be
started until one or more of other activities are
completed but immediately succeed them are
called successor activities.
3. Concurrent activity – Activities which can be
accomplished concurrently are known as
concurrent activities. It may be noted that an
activity can be a predecessor or a successor to an
event or it may be concurrent with one or more of
other activities.
4. Dummy activity – An activity which does not
consume any kind of resource but merely depicts
the technological dependence is called a dummy
activity.

The dummy activity is inserted in the network to clarify


the activity pattern in the following two situations
 To make activities with common starting and
finishing points distinguishable
 To identify and maintain the proper precedence
relationship between activities that is not
connected by events.
For example, consider a situation where A and B are
concurrent activities. C is dependent on A and D is
dependent on A and B both. Such a situation can be
handled by using a dummy activity as shown in the
figure.

168
2. Event
An event represents a point in time signifying the
completion of some activities and the beginning of new
ones. This is usually represented by a circle in a
network which is also called a node or connector.
The events are classified in to three categories
1. Merge event – When more than one activity
comes and joins an event such an event is known
as merge event.
2. Burst event – When more than one activity
leaves an event such an event is known as burst
event.
3. Merge and Burst event – An activity may be
merge and burst event at the same time as with
respect to some activities it can be a merge event
and with respect to some other activities it may
be a burst event.

3. Sequencing
The first prerequisite in the development of network is
to maintain the precedence relationships. In order to

169
make a network, the following points should be taken
into considerations
 What job or jobs precede it?
 What job or jobs could run concurrently?
 What job or jobs follow it?
 What controls the start and finish of a job?
Since all further calculations are based on the network,
it is necessary that a network be drawn with full care.
Rules for Drawing Network Diagram
Rule 1
Each activity is represented by one and only one arrow
in the network

Rule 2
No two activities can be identified by the same end
events

Rule 3
In order to ensure the correct precedence relationship in
the arrow diagram, following questions must be
checked whenever any activity is added to the network
 What activity must be completed immediately
before this activity can start?
170
 What activities must follow this activity?
 What activities must occur simultaneously with
this activity?
In case of large network, it is essential that certain good
habits be practiced to draw an easy to follow network
 Try to avoid arrows which cross each other
 Use straight arrows
 Do not attempt to represent duration of activity
by its arrow length
 Use arrows from left to right. Avoid mixing two
directions, vertical and standing arrows may be
used if necessary.
 Use dummies freely in rough draft but final
network should not have any redundant dummies.
 The network has only one entry point called start
event and one point of emergence called the end
event.
Common Errors in Drawing Networks
The three types of errors are most commonly observed
in drawing network diagrams
1. Dangling
To disconnect an activity before the completion of all
activities in a network diagram is known as dangling.
As shown in the figure activities (5 – 10) and (6 – 7)
are not the last activities in the network. So the diagram
is wrong and indicates the error of dangling

171
2. Looping or Cycling
Looping error is also known as cycling error in a
network diagram. Drawing an endless loop in a
network is known as error of looping as shown in the
following figure.

3. Redundancy
Unnecessarily inserting the dummy activity in network
logic is known as the error of redundancy as shown in
the following diagram

172
Advantages and Disadvantages
PERT/CPM has the following advantages:
 A PERT/CPM chart explicitly defines and makes
visible dependencies (precedence relationships)
between the elements,
 PERT/CPM facilitates identification of the
critical path and makes this visible,
 PERT/CPM facilitates identification of early
start, late start, and slack for each activity,
 PERT/CPM provides for potentially reduced
project duration due to better understanding of
dependencies leading to improved overlapping of
activities and tasks where feasible.

PERT/CPM has the following disadvantages:


 There can be potentially hundreds or thousands
of activities and individual dependency
relationships,
 The network charts tend to be large and unwieldy
requiring several pages to print and requiring
special size paper,
 The lack of a timeframe on most PERT/CPM
charts makes it harder to show status although
colours can help (e.g., specific colour for
completed nodes),
 When the PERT/CPM charts become unwieldy,
they are no longer used to manage the project.
173
Critical Path in Network Analysis
Basic Scheduling Computations
The notations used are
(i, j) = Activity with tail event i and head event j
Ei = Earliest occurrence time of event i
Lj = Latest allowable occurrence time of event j
Dij = Estimated completion time of activity (i, j)
(Es)ij = Earliest starting time of activity (i, j)
(Ef)ij = Earliest finishing time of activity (i, j)
(Ls)ij = Latest starting time of activity (i, j)
(Lf)ij = Latest finishing time of activity (i, j)
The procedure is as follows:
1. Determination of Earliest time (Ej): Forward
Pass computation
 Step 1: The computation begins from the start
node and move towards the end node. For
easiness, the forward pass computation starts by
assuming the earliest occurrence time of zero for
the initial project event.
 Step 2

174
i. Earliest starting time of activity (i, j) is the
earliest event time of the tail end event i.e. (Es)ij
= Ei
ii. Earliest finish time of activity (i, j) is the earliest
starting time + the activity time i.e. (Ef)ij =
(Es)ij + Dijor (Ef)ij = Ei + Dij
iii. Earliest event time for event j is the maximum of
the earliest finish times of all activities ending in
to that event i.e. Ej = max [(Ef)ij for all
immediate predecessor of (i, j)] or Ej =max [Ei +
Dij]
2. Backward Pass computation (for latest allowable
time)
 Step 1: For ending event assume E = L.
Remember that all E’s have been computed by
forward pass computations.
 Step 2: Latest finish time for activity (i, j) is
equal to the latest event time of event j i.e. (Lf)ij =
Lj
 Step 3: Latest starting time of activity (i, j) = the
latest completion time of (i, j) – the activity time
or (Ls)ij =(Lf)ij - Dij or (Ls)ij = Lj - Dij
175
 Step 4: Latest event time for event ‘i’ is the
minimum of the latest start time of all activities
originating from that event i.e. Li = min [(Ls)ij for
all immediate successor of (i, j)] = min [(Lf) ij -
Dij] = min [Lj - Dij]
3. Determination of floats and slack times
There are three kinds of floats
 Total float – The amount of time by which the
completion of an activity could be delayed
beyond the earliest expected completion time
without affecting the overall project duration
time.
Mathematically
(Tf)ij = (Latest start – Earliest start) for activity (i–j)
(Tf)ij = (Ls)ij - (Es)ij or (Tf)ij = (Lj - Dij) - Ei
 Free float – The time by which the completion of
an activity can be delayed beyond the earliest
finish time without affecting the earliest start of a
subsequent activity.
Mathematically
(Ff)ij = (Earliest time for event j – Earliest time for
event i) – Activity time for ( i, j)
176
(Ff)ij = (Ej - Ei) - Dij
 Independent float – The amount of time by
which the start of an activity can be delayed
without effecting the earliest start time of any
immediately following activities, assuming that
the preceding activity has finished at its latest
finish time.
Mathematically
(If)ij = (Ej - Li) - Dij
The negative independent float is always taken as
zero.
 Event slack - It is defined as the difference
between the latest event and earliest event times.
Mathematically
Head event slack = Lj – Ej,
Tail event slack = Li - Ei
4. Determination of critical path
 Critical event – The events with zero slack times
are called critical events. In other words the event
i is said to be critical if Ei = Li
 Critical activity – The activities with zero total
float are known as critical activities. In other
177
words an activity is said to be critical if a delay in
its start will cause a further delay in the
completion date of the entire project.
 Critical path – The sequence of critical activities
in a network is called critical path. The critical
path is the longest path in the network from the
starting event to ending event and defines the
minimum time required to complete the project.

Exercise
1. What is PERT and CPM?

2. What are the advantages of using PERT/CPM?

3. Mention the applications of PERT/CPM?

4. Explain the CPM in network analysis.


5. What are the rules for drawing network diagram?
6. What is the difference between PERT and CPM?
7. What are the uses of PERT and CPM?
8. Explain the basic steps in PERT/CPM techniques.

178
Worked Examples on CPM
Example 1: Determine the early start and late start in
respect of all node points and identify critical path for
the following network.

Solution: Calculation of E and L for each node is


shown in the network

179
Network Analysis Table
Earliest Time Latest Time Float
Normal
Activity(i, Finish Start Time
Time Start Finish
j) (Ei + (Li - (Li - Dij )
(Dij) (Ei) (Li)
Dij ) Dij ) - Ei
(1, 2) 10 0 10 0 10 0
(1, 3) 8 0 8 1 9 1
(1, 4) 9 0 9 1 10 1
(2, 5) 8 10 18 10 18 0
(4, 6) 7 9 16 10 17 1
(3, 7) 16 8 24 9 25 1
(5, 7) 7 18 25 18 25 0
(6, 7) 7 16 23 18 25 2
(5, 8) 6 18 24 18 24 0
(6, 9) 5 16 21 17 22 1
(7, 10) 12 25 37 25 37 0
(8, 10) 13 24 37 24 37 0
(9, 10) 15 21 36 22 37 1

From the table, the critical nodes are (1, 2), (2, 5), (5,
7), (5, 8), (7, 10) and (8, 10)
From the table, there are two possible critical paths
i. 1 → 2 → 5 → 8 → 10
ii. 1 → 2 → 5 → 7 → 10

180
Example 2: Find the critical path and calculate the
slack time for the following network

Solution: The earliest time and the latest time are


obtained below
Earliest Time Latest Time Float
Normal
Activity(i, Finish Start Time
Time Start Finish
j) (Ei + (Li - (Li - Dij
(Dij) (Ei) (Li)
Dij ) Dij ) ) - Ei
(1, 2) 2 0 2 5 7 5
(1, 3) 2 0 2 0 2 0
(1, 4) 1 0 1 6 7 6
(2, 6) 4 2 6 7 11 5
(3, 7) 5 2 7 3 8 1
(3, 5) 8 2 10 2 10 0
(4, 5) 3 1 4 7 10 6
(5, 9) 5 10 15 10 15 0
(6, 8) 1 6 7 11 12 5
(7, 8) 4 7 11 8 12 1
(8, 9) 3 11 14 12 15 1

181
From the above table, the critical nodes are the
activities (1, 3), (3, 5) and (5, 9)

The critical path is 1 → 3 → 5 → 9


Example 3
A project has the following times schedule
Times in Times in
Activity Activity
weeks weeks
(1 – 2) 4
(5 – 7) 8
(1 – 3) 1
(6 – 8) 1
(2 – 4) 1
(7 – 8) 2
(3 – 4) 1
(8 – 9) 1
(3 – 5) 6
(8 – 10) 8
(4 – 9) 5
(9 – 10) 7
(5 – 6) 4

182
Construct the network and compute
1. TE and TL for each event
2. Float for each activity
3. Critical path and its duration
Solution
The network is

Event
1 2 3 4 5 6 7 8 9 10
No.:
TE: 0 4 1 5 7 11 15 17 18 25
TL: 0 12 1 13 7 16 15 17 18 25

Float = TL (Head event) – TE (Tail event) –


Duration

183
TL (Head
Activity Duration TE (Tail event) Float
event)
(1 – 2) 4 0 12 8
(1 – 3) 1 0 1 0
(2 – 4) 1 4 13 8
(3 – 4) 1 1 13 11
(3 – 5) 6 1 7 0
(4 – 9) 5 5 18 8
(5 – 6) 4 7 16 5
(5 – 7) 8 7 15 0
(6 – 8) 1 11 17 5
(7 – 8) 2 15 17 0
(8 – 9) 1 17 18 0
(8 – 10) 8 17 25 0
(9 – 10) 7 18 25 0
The resultant network shows the critical path

The two critical paths are


184
i. 1 → 3 → 5 →7 → 8 → 9 →10
ii. 1 → 3 → 5 → 7 → 8 →10
Project Evaluation and Review Technique (PERT)
The main objective in the analysis through PERT is to
find out the completion for a particular event within
specified date. The PERT approach takes into account
the uncertainties. The three time values are associated
with each activity
1. Optimistic time – It is the shortest possible time
in which the activity can be finished. It assumes
that every thing goes very well. This is denoted
by t0.
2. Most likely time – It is the estimate of the
normal time the activity would take. This
assumes normal delays. If a graph is plotted in
the time of completion and the frequency of
completion in that time period, then most likely
time will represent the highest frequency of
occurrence. This is denoted by tm.
3. Pessimistic time – It represents the longest time
the activity could take if everything goes wrong.
As in optimistic estimate, this value may be such
185
that only one in hundred or one in twenty will
take time longer than this value. This is denoted
by tp.
In PERT calculation, all values are used to obtain the
percent expected value.
1. Expected time – It is the average time an activity
will take if it were to be repeated on large number
of times and is based on the assumption that the
activity time follows Beta distribution, this is
given by
te = ( t0 + 4 tm + tp ) / 6
2. The variance for the activity is given by
σ2 = [(tp – to) / 6] 2
Worked Examples
Example 1:For the project

186
Task: A B C D E F G H I J K

Least time: 4 5 8 2 4 6 8 5 3 5 6

Greatest
8 10 12 7 10 15 16 9 7 11 13
time:

Most
likely 5 7 11 3 7 9 12 6 5 8 9
time:

Find the earliest and latest expected time to each event


and also critical path in the network.
Solution
Expected
Greatest Most
Least time
Task time likely time
time(t0) (to + tp +
(tp) (tm)
4tm)/6
A 4 8 5 5.33
B 5 10 7 7.17
C 8 12 11 10.67
D 2 7 3 3.5
E 4 10 7 7
F 6 15 9 9.5
G 8 16 12 12
H 5 9 6 6.33
I 3 7 5 5
J 5 11 8 8
K 6 13 9 9.17

187
Expected Start Finish Total
Task
time (te) Earliest Latest Earliest Latest float
A 5.33 0 0 5.33 5.33 0
B 7.17 0 8.83 7.17 16 8.83
C 10.67 5.33 5.33 16 16 0
D 3.5 0 10 3.5 13.5 10
E 7 16 16 23 23 0
F 9.5 3.5 13.5 13 23 10
G 12 3.5 18.5 15.5 30.5 15
H 6.33 23 23 29.33 29.33 0
I 5 23 25.5 28 30.5 2.5
J 8 28 30.5 36 38.5 2.5
K 9.17 29.33 29.33 31.5 38.5 0

The network is

The critical path is A →C →E → H → K


188
General Applications
1) The following table represents the unit cost
transported to a specific commodity from the set of sources
to the set of destinations and the number of supply and
demand units from this commodity:

Destinations B1 B2 B3 Supply
Sources units
A1 3 4 8 500
A2 2 6 10 600
A3 5 7 9 800
Demand 400 650 550
units
Required:

a- Formulate the transportation problem as a LPM


b- Compare between the total cost resulted from the
three different techniques for determining the initial
solution tableaus.
c- From your preceding result in (b) determine the
optimum solution.
2) The following table represents the unit cost
transported to a specific commodity from the set of sources
to the set of destinations and the number of supply and
demand units from this commodity:
Destinations B1 B2 B3 Supply
Sources units
A1 5 1 0 40
A2 3 2 4 20
A3 7 5 2 30
Demand 10 20 30
units
189
Required:

a) Formulate the transportation problem as a LPM


b) Determine the optimum solution for this
transportation problem.
3) If you have the following LPM:

Max Z = 3x1-5x2
Subject to:
7x1- 2x2 ≥ -14
x1 ≤4
x2≤ 6
3x1+ 2x2 ≤ 18
and x1 ≥ 0, x2 ≥ 0
Required:
a) Put the LPM in its canonical form.
b) using the suitable simplex method, solve the
LPM.
c) Determine the dual problem and find the
optimum value for its objective function.
4) If you have the following LPM:
Max Z = 2x1 + 4x2
Subject to:
2x1+ x2 ≤ 400
2x1+ 2x2 ≤ 500
x1 ≤ 300
190
x2≤ 400
and x1 ≥ 0, x2 ≥ 0
Required:
a- By using the suitable simplex method, solve
the LPM.
b- Find the dual problem and determinates
optimum solution tableau .
c- By using two different methods, determine the
shadow prices.
5) If you have the following LPM:
Min Z = -x2
Subject to:
x 1 + x2 ≤ 2
| x1 - x2 | ≤ 1
and x1 ≥ 0, x2 ≥ 0
Required:
a) Put the LPM in its canonical form.
b) By using the suitable simplex method, solve
the LPM.
c) Determine the dual problem and find the
optimum value for its objective function.

191
References
Alrobh, M. Abd-elhamid.Operations Research: An
introduction. 2015, Faculty of commerce, South
Valley University
Hillier, Frederick S. and Lieberman, Gerald J.
Introduction to Operations Research.
SeventhEdition, 2000
Paul, A. Jensen and Jonathan, F. Bard.Operations
Research Models and Methods.John Wiley and
Sons 2003.
Reid, Dan R. and Sanders, Nada R. Operations
Management. 4th Edition, Wiley 2010
Taha,Hamdy A.Operations Research: An
Introduction. Eighth Edition, Pearson Prentice Hall
2007

192

You might also like