KUNJA BIHARI (DEGREE) COLLEGE

BARANG, KHORDHA
DEPARTMENT OF MATHEMATICS
Academic session 2019-20(Semester-VI)
C.B.C.S. Final Year DSE –IV(Project Paper)

A Project on
“A SOLUTION OF LINEAR PROGRAMMING PROBLEMS”

Guide Prepared & Submitted by

Mr. Manas Kumar Patra RAJESH BEHERA
Lecturer, Dept. Of Mathematics, +3 Final Year (Science)
Department Of Mathematics, Mathematics Honours
Kunja Bihari Degree College, U.U. Roll No.: 1702010690110071
Barang Roll No.: BS-17-138
Regd No.: 1702010690110071
 KUNJA BIHARI DEGREE COLLEGE,
BARANG

Mr. Manas Kumar Patra Date :

Lecturer, Department Of Mathematics Bhubaneswar
K.B. College, Barang
CERTIFICATE
This is to certify that the project entitled “A Solution of
Linear Programming Problems” which is being submitted
by Rajesh Behera , Regn. No. -1702010690110071 carried
out by him under my supervision. The results embodied in
this project are new and have not been submitted to any
other university of institute for the award of any degree.
He has taken proper care and shown utmost sincerity in
completion of this project.
I certify that this piece of project is up to my
expectation and satisfaction. This is prepared as per the
guidelines provided by UTKAL UNIVERSITY, Vani Vihar.

Mr. Manas Kumar Patra

Lecturer, Department Of
Mathematics
K.B College, Barang
Bhubaneswar
 ACKNOWLEGDEMENT

I wish to express my deep sense of gratitude to

my supervisor Mr. Manas Kumar Patra, Lecturer,
Department of Mathematics, Kunja Bihari Degree
College, Barang for his inspiring guidance and
assistance in the preparation of this project.

I also take the opportunity to acknowledge quite

explicitly with gratitude my debt to Dr. Binaya Bhusan
Sahoo, HOD, Department of Mathematics, Kunja Bihari
Degree College, Barang who has been a constant
source of inspiration for the timely completion of my
project.

I also graceful to my friends & family members for

helping me & sparing their time to bring at this
project.

RAJESH BEHERA
Date : +3 Final Year Science Math Hons.
College Roll No.: BS-17-138
Regd. No.: 1702010690110071
 DECLARATION

I do hereby declare that this piece of research

work entitled “A SOLUTION OF LINEAR
PROGRAMMING PROBLEMS” being submitted to
Department of Mathematics, Kunja Bihari College,
Barang is of my own and preparation basing upon the
data and information gathered from various sources
during research period prepared under the supervision
of Mr. Manas Kumar Patra, Lecturer, Department of
Mathematics, Kunja Bihari College, Barang.

I further declare that it has not been submitted or

published elsewhere for any other degree in full of or
in part.

Place : Bhubaneswar
Date : (Rajesh Behera)
 CONTENT

 Chapter – 0
 Introduction
 Basic Definition
 Chapter – 1
 Graphical Method
 Chapter – 2
 Simplex Method
 Slack Variable
 Surplus Variable
 Steps of Simplex Method
 Chapter – 3
 Degeneracy Method
 Steps Of Degeneracy Method
 References
 CHAPTER – 0

INTRODUCTION

Linear Programming is a mathematical method for

determining the optimization (maximization or
minimization) of a function of variable known as objective
function, subject to set of linear equalities or inequalities
known as constraints, and followed by a set of non-negative
restrictions. In operation research or linear programming is
one of the most important and richly developed
optimization techniques. Linear programming model arises
in a variety of decision problems encountered in
government, economics, industry and engineering. These
models describe not only the practical decision problems
but also find application in sciences. For example, these
models have been in obtaining theoretical results and
computational methods in such areas of pure and applied
mathematics as graph theory, combinatorial analysis,
approximation theory and numerical analysis.
 BASIC DEFINATIONS
A linear programming problem (LPP) consists of three
components namely
a) Objective Function
b) Constraints
c) Non-negative restrictions
a) Objective function :- The linear function
z = C1X1 + C2X2 + ……………… + CnXn
which is to be minimized is called the objective function of
general LPP.
b) Constraints : - The inequalities are called the
constraints of the general LPP.
c) Non-negative Restriction : - The set of inequalities is
usually known as the set of non-nagative restriction of
the general LPP.
SOLUTION: -
An n-tuple(x1, x2,………………… xn) of real numbers which
satisfies the constraints of a general LPP. is called a solution
of general LPP.
FEASIBLE SOLUTION: -
Any solution to a general LPP which also satisfies the non-
negative restriction of the problem is called a feasible
solution of the general LPP.
OPTIMUM SOLUTION: -
Any feasible solution which optimizes the bi-jective function
of a general LPP is called an optimum solution to the
general LPP.
 Chapter -1
Solution of linear programming problems
There are many methods to find optical solution of LPP
The methods are:
(1) Graphical Method
(2) Simplex method
(3) Degeneracy method
(1) Graphical Method:
The industrial problems involving two of the variables can
be easily and effectively solved by drawing the graph for
various constrains and the objective function. It is even
simpler for two variables cases as three-dimensional
geometry is rather difficult to draw on graph paper.
The following are the steps in Graphical solution of LPP:
i) The constraints of the problems are taken as equalities and
plotted on the graph paper. The region satisfying each
constraint is determined and the region common to all the
constraint relations is located. The region is known as
feasible region for the problem as all points lying in this
region will simultaneously satisfy the given constraints. The
solution set of the system of constraints is found to be
represented by some convex polygon termed as feasible
region.
ii) Objective function is also plotted on the same graph paper
by taking some arbitrary value of Z.
iii) Trial and error methods are used to locate the optimal
solution inside the feasible region determined in Step (i)
so that the objective functions optimized. It is
experienced that the optimal solution is located near the
extreme points of the feasible region. Thus the optimum
 feasible point should be searched only at the corners of
the convex polygon.
The graph for all the constraints and the objective
function will be straight lines in LPP.
A farm is engaged in breeding pigs. The pigs are fed on various
products grown on the farm. In view of the need to ensure
certain nutrient constituents(call them X,Y and Z), it is necessary
to buy two additional products , say A and B, One unit of product
A contains 36 units of X, 3 units of Y and 20 units of Z. One unit of
product B contains 6 units of X, 12 units of Y and 10 units of Z. The
minimum requirement of X, Y and Z is 108 units, 36 units and 100
units respectively. Product A costs Rs.20 per unit and products B
Rs. 0
Formulate the above as linear programming problem to
minimize the total cost and solve the problem by using graphic
method.
Solution: Step 1 . The data of the given problem can be
summarized as follows :
Nutrient Content in product Minimum
Nutrient
amount of
constituents A B nutrient
X 36 06 108
Y 03 12 36
Z 20 10 100
Cost of Product Rs. 20 Rs. 40
Making use of the above information, the appropriate
mathematical formulation of the linear programming problem is :
Minimize z = 20x1 + 40x2 subject to the constraints:
36x1 + 6x2 > 108, 3x1 + 12x2 > 36, 20x1 + 10x2 > 100 and x1,x2 > 0.
where x1 = number of units of product A, and x2 = number of units
of product B
Step 2 . Consider now a set of Cartesian rectangular axix OX1X2 in
the plane. As each point has the coordinates of the type (x1,x2),
any point satisfying the conditions x1 > 0 and x2 > 0 lies in the first
quadrant only.
The constraints of the given problem are plotted as described
earlier by treating them as equations:
36x1 + 6x2 > 108, 3x1 + 12x2 > 36 and 20x1 + 10x2 > 100
Or x1/3 + x2/18 = 1 , x1/12 + x2/3 and x1/5 + x2/10 = 1
The area beyond these lines represents the feasible region in
respect of these constraints: any point on the straight lines or in
the region above these lines would satisfy the constraints. The
feasible region of the problem is as shown in Fig. 3.3.

Step 3 : The coordinates of the extreme points of the feasible

region are : A = (0,18), B = (2,6), C = (4,2) and D = (12,0)
Step 4 : The value of objective function at each of the extreme
points can be evaluated as follows :
Extreme Point (x1,x2) z = 20x1 + 40x2
A (0,18) 720
B (2,6) 280
C (4,2) 160 Minimum
D (12,0) 240
Hence, the optimum solution is to purchase 4 units of
product A and 2 units of product B in order to maintain a
minimum cost of Rs. 160.
(2) SIMPLEX METHOD
Simplex method id the most general and powerful technique
to solve LPP in a finite number of steps or gives an indication that
there is an unbounded solution of LPP. Simplex method is
designed to solve simultaneously a system of linear equations
where there are more/less unknowns than the number of
equations.
To solve a problem by Simplex method requires :
(i) Arrange the objective function and constraints in a
special way and
(ii) Following a systematic procedure and a set of rules in
finding the desired solution.
The first step in all LPP solved by simplex method is to
formulate the problem in the form of objective function and
constraints. Essentially the simplex method searches through
combinations of solutions until the best solution is found.
Introduction of slack, surplus and artificial variables in LPP:
(a) Slack Variables:
These variables are included in < inequalities to convert
them into equality e.g. given a1 X + a2 Y < b1
We redefine it as a1 X + a2 Y + S1 < b1
Where S1 is the Slack Variable.
i.e., slack variable = Total resource – Used resource
Slack variable are non-negative and explain the unallocated
portion of the given limited resource. These permit more
comprehensive economic interpretation of the solution.
(b) Surplus Variables:
These variable are introduced in equation to change them
into equations e.g. a3 X + a4 Y < b2
Is written as a3 X + a4 Y – S2 < b2
Where S2 is surplus variable. Here
Surplus = Product – Requirement
 The value of this variable can be interpreted as the
amount over and above the required minimum level. It is
also non- negative.
(c) Introduction of artificial variable in optimization problems:
In many cases the constraints are a combination of >, < and
= in equation and provide the initial feasible solution. To solve
the problem artificial variables are fictitious and cannot have
any physical meaning.
A very large penalty denoted by M per unit is assigned in
objective function to the artificial variables designed as -M in
the case of maximization problems and +M in the case of
minimization problems. The high value of M renders inclusion
of artificial variables in the solutions prohibitively costly e.g.
given the constraints
C1 X1 + C2 X2 < a1
C3 X1 + C4 X2 > a2
C5 X1 + C6 X2 < a3
The > and < in equations are changed into equations by
adding slack surplus variables i.e taking
C1 X1 + C2 X2 + S1 < a1 S1 is slack variable.
C3 X1 + C4 X2 – S2 > a2 S2 is surplus variable.
The need for artificial variables in > in equation arises, as the
surplus variable S2 is not satisfy the non-negative condition of
basic feasible solution.
The reason being that non of the surplus basic variables in
our problems can have a negative value. It is observed that
introduction of surplus variables with a negative sign in >
inequalities does not provide the initial feasible solution. Thus
the artificial variable A is included in the second constraint to
write it as
C3 X1 + C4 X2 – S2 + A1 > a2
 Also an artificial variable is included in the equation with
equal sign so that the equation has a basic variable.
C5 X1 + C6 X2 – A2 = a3
Note :
A variable is said to be a basic variable in an equation if it
appears with a unit coefficient in that equation and zero
coefficient in all zero coefficient in all equations.
The variables taken in the initial feasible solution are chosen
corresponding to identity vector columns i.e of the form (1,0,0),
(0,1,0),(0,0,1) initial simplex table giving priority to artificial
variables. The artificial variables are to be first removed from the
solution basis.
For this, if at any iterative stage of simplex method, the basic
feasible solution contains artificial variables but the variable to be
deleted at that stage is not an artificial variable then one can
depart from the simplex criterion for selecting the entering
variable.
In simplex method all variables must appear in all the
equations. This is done by adding the variables absent in an
equation with zero coefficients. All the variables should also
appear in objective function with zero profit or cost for surplus
and slack variables and very high penalty M for artificial variables.
M is included with a negative sign in maximization problems and
with a positive sign in minimization problems.
Simplex Procedure:
The first step in simplex procedure is to introduce slack,
surplus and artificial variables in in-equations to change them into
equation and also in the objective function.
After this the objective function and the constraints are
arranged in a special way i.e. all variables, real, surplus, slack and
artificial are arranged in the same order in all constraints and the
objective function. Thus the constraints equations are expressed
is a canonical system from which a basic feasible solution can be
easily obtained. This is known as standard form of LPP.
The next step is to be preparing the initial simplex table.
The following are the main columns of a simplex table:
Rows and columns of simplex table top row for the co-
efficient of the objective function
PROFIT PROGRAMM CONSTRAINTS REAL SLACK SURPLUS ARTIFICIAL RATIO
VARIABLES VARIABLES VARIABLES VARIABLES

The number of rows in main part of the simplex table is equal to

the number of constraints in the system. The number of column
in the variable part of the table depends on the number of slack,
surplus, artificial and real variables in the system under study. The
remaining column and rows of the simplex table are calculated
step by step.
The constraints of the initial simplex table is explained with the
help of LPP.
Key Column:
The entries of the initial simplex table can be made in the
following manner:
(i) Top Row:
This row is begins above the constant column an ends
above the column corresponding to the last variable
appearing in the problem. The column corresponding to the
last variable appearing in the problem.

The column corresponding to constant column in this

row is allocated the word C the other columns of this row
are assigned the coefficients of the corresponding variables
in the objective function. Columns corresponding to real
variables contain the co-efficient of the corresponding real
 variables in the objective function and the column above the
slack variable column constraints the co-efficient of the
corresponding real values in the objective function and the
columns above the slack variable columns contain the co-
efficient of the corresponding slack variables in the objective
function.
Thus the top row is:
Cj Constant 6 11 0 0
X1 X2 S1 S2
ROW. I. entries from constant columns to the last variable
column are respectively the constant value on right hand side and
the co-efficient of real and slack variables in equation I i.e.
Cj Constant X Y S1 S2
104 2 1 1 0
Similarly the entries of ROW if corresponds to the equation II of
the problem i.e.
Cj Constant X Y S1 S2
76 1 2 0 1

At the stage, top row and ROW I and ROW II of the initial simplex
table can be written as,

C1 6 11 0 0
PROFIT PROGRAMME CONSTANT X Y S1 S2
0 S1 104
0 S2 76

The entries in profit and programme column of initial simplex

table are made by locating basic solution variables from the initial
simplex table.
The basic solution variables can be located by following steps:
(i) Locate the unit/identity column vectors among the variable
column of the initial simplex table. These are the columns
which contain one elements as(+1) and other elements zero
e.g in the above table the columns for S1 and S2 are unit
vectors with values (1,0) and (0,1) respectively.
Now in column for S1,1 appears in 1st row is entered in
program column of 1st row. Similarly S2 entered in program
column of 2nd row. These are the variables entering in the
initial feasible solution. The variable S1 lies in programme
column of ROW I and its co-efficient in the objective function
i.e 0 is entered in profit column ROW I. similarly,
corresponding to variable S2 of programme column in ROW II,
its coefficient 0 in the objective function is entered in profit
column of ROW II.
Up to this point the initial simplex table is prepared only
with the help of the objective function and the constraint
equation with no calculations.
Now we have to determine, whether any improvement
in the solution at this stage can be made or not. Because at
this stage the feasible solution is S1 = 104, S2 =76 with
corresponding profits equal to zero. To determine this, two
more zero Zj and Zj - Cj are added in simplex table.
The value of Zj in column for X,Y,S1 and S2 were added to
the mix.
The value of Zj corresponding to S1,S2 , X and Y calculated as:
Rule 1:
Now Cj has been defined profit per unit. Thus Zj - Cj is the
net profit, which will result from introducing one unit of the
corresponding variable to the production schedule or
solution. Zj - Cj is also known as unit improvement.
Maximization Problem:
By examining Zj - Cj row, we can see the change in profit for
each variable and if this is positive for all variables then this
indicates that the solution obtained is optimum and no
improvement is possible. In case one or more values of Zj - Cj are
found to be negative, then the solution needs improvement.
Note:
Some people prefer to calculate Cj - Zj . In that case optimum
solution is reached if all Cj - Zj are negative. One or more positive
value of Cj - Zj indicate need for improvement in the solution.
Rule 2:
Simplex method is an iterative produce where each step
bring closer to the optimum solution. It is done by removing. One
variable at a time from the program column and replacing it with
a new one. The process is repeated a number of times till no
further improvement in the solution is possible i.e. all Z j - Cj are
positive.
The decision of continuing the iterative produce is taken by
examining Zj - Cj row. If no or more columns of Zj - Cj row are
found to have negative values, then there is no need to improve
the solution.
Rule 3 : Rules for choosing the program column:
Examine Zj - Cj values at each iteration and locate the
numerically largest negative value, larger the number greater the
improvement possible e.g inspection of Zj - Cj values in initial
simplex table tells us that -11 is numerically most negative value.
Hence the column containing the value -11 is designed as key
column at this iteration and the corresponding variable of this
column i.e Y is the new entering variable in the basis i.e the
program column of next table.
Rule 4 : Rules for determining the outgoing variable form the
basis:
The next step is to determine which variable is to be
replaced. For this ratio column is computed. This is done by
dividing the entry in constant column of a row by the
corresponding value in the key column e.g in our illustration.

For Row I, ratio = Entry in constant column/key column entry

= 104/1 = 104
Similarly,
For Row II, ratio = 76/2 = 38
Now we examine the ratio column of the simplex table and
select the row with smallest non-negative ratio and call it key
row. The variable lying in the program column of the key row is
the leaving variable at the iteration e.g. Row II is the key Row S 2
variable of program column is replaced by the entering variable Y.
The value lying at the intersection of key row and key column is
known as key number.
PROGRAMM CONSTANT X Y S1 S2 RATIO
S1 104 2 1 1 0 104
S2 76 1 2 0 1 38
Zj - Cj -6 -1 0 0
Entering Variable:
Rule 5 : Development of second simplex solution :
The coefficient values in the body of each iteration table are
really rates of substitution between all the variables.
In the new table the key row is replaced by the row obtained
by dividing each entry of key row by the key number. In our
example the new row II will be:
PROFIT PROGRAM CONSTANT X Y S1 S2
11 Y 76/2 1/2 2/2 0/2 1/2
 Note that variable S2 is replaced by Y and in profit column we
write the value of Cj row corresponding to key column i.e. value of
Cj = 11.
Rule 6 : To Complete The Second Tableau, New Values For
Remaining Rows are Computed by the Formula.
[Elements in the old row] – [ (Elements in the key column of that
row)x(corresponding element in the row replacing key
row)]=(new row)
e.g. in Row I of the initial simplex table the elements are
(104,2,1,1,0)
So using the above formula, Row I in new tableau will be
(104,2,1,1,0)-(1)(381/2101/2)
=(104-38,2-0.5,1-1,1-0,0-0.5)
=(66,1.501-1/2)
The new Simplex Tableau I is completed by entering the
appropriate unit costs in profit column for the mix. Variables e.g
In profit column of Tableau I, the per unit profit of Y i.e 11 is
entered. Thus the new Simplex Tableau I becomes.
Row Profit Programme Cj 0 11 0 0
Constant X Y S1 S2 Ratio
I 0
Key Row S1 66 3/2 0 1 -1/2 44
II 11 Y 38 ½ 1 0 ½ 76
Zj 11/2 11 0 11/2
-1/2
Zj - Cj 0 0 11/2
Key Column
Now Rules I,II,III are again applied to know whether optimum
solution stage has arrived or not. In Tableau I, Z j and Zj -Cj are not
positive. Thus it is not the stage of optimum solution. There is one
Zj -Cj corresponding to the column for variable X, which is
negative. So by rule 4 this column is key column. Now minimum
positive value in ratio column.
The row containing this minimum positive value in ratio
column is designated as key row at this stage. The variable
appearing in the programme column of key row is departing
variable. Thus Row I is the Key Row and X is the entering variable
and S1 is the departing variable in the basis of Tableau I.
Now Tableau is prepared from Tableau I using rules 6 and 7.
Tableau II
Cj 6 11 0 0
Row Profit Programme Constant X Y S1 S2
I 6 X 44 1 0 2/3 -1/3
II 11 Y 16 0 1 -1/3 5/6
Zj 6 11 1/3 43/6
Zj - Cj 0 0 1/3 43/6

Row II in Tableau II is given by:

(3801/21/21)-[(1/2)(442/3-1/310)]=(16-1/35/601)
Again Rules I,II and III are applied on Tableau II to find
whether the solution is optimum or not. It is observed that in
Tableau II all Zj - Cj are positive. Hence the optimum solution stage
has been attained.
The optimum solution is X = 44, Y = 16 with a maximum profit
of Rs. 440 which approximately is the same as in Example 2.
Steps in the Simplex procedure are listed below :
(i) Formulate the problem by determining the objective
function and constraints.
(ii) Introduce slack, surplus and artificial variables in the
objective function and the constraints by changing
inequalities in equalities to get the standard form.
(iii) Prepare the initial Simplex Tableau.
(iv) Apply Rules I,II and III to determine whether the solution is
optimum or not.
If the solution is not optimum, then go to step 5 to 11.
(v) Determine the entering variable (Key Column) by selecting
the column with most negative (Zj -Cj ).
(vi) Determine the departing variable (Key Row) by calculating
the Ratio column from Rule V and selecting the smallest
non-negative value.
(vii) Compute the replacing row of the improved simplex
Tableau by Rule VI.
(viii) Compute the remaining rows of the new table by Rule VII.
(ix) Go to step (iv) and continue till optimum solution is
obtained.
The application of Simplex method is illustrated with the help
of following examples.
Example 3:
Two products A and B are processed on three machines M 1, M2
and M3 .
The processing times per unit, Machine availability and profit
per unit are given below:
Machine Processing Time in Hours Availability in Hours
M1 A B 1500
M1 2 3 1500
M1 3 2 1000
Profit per 1 1
10 12
unit in Rs

Formulate the mathematical model and solve it by simplex

method. Also find the number of hours machine M 3 remains
unutilized.
Solution:
Step 1:
Let X units of A and Y units of B be produced. Then the
corresponding profit is 10X -12Y and the objective function is
Max. Z = 10X +12Y
Now the constraints for the three machines will be:
2X +3Y < 1500 for M1
3X +2Y < 1500 for M2
X +Y < 1000 for M3
X,Y > 0 Non-negative restriction.
Step 2:
Since all the constraints are having < signs, slack variables can
be introduce to change these in equalities. There are three
constraints so three slack variables S1 , S2 and S3 are introduced
in the problem.
2X +3Y + S1 +0.S2 +0.S3 = 1500
3X +2Y + 0.S1 +S2 +0.S3 = 1500
X +Y + 0.S1 +0.S2 +S3 = 1000
and Max. Z = 10X +12Y + 0.51 +0.52 +0.53
Step 3: The initial simplex tableau can now be prepared
INITIAL TABLEAU
0 Cj 10 12 0 0 0
Profit Programme Constant X Y Z S1 S2 RATIO
PROFIT S1 1500 2 3 1 0 0 500
KEY 0
ROW S2 1500 3 2 0 1 0 750
0
0 S3 1000 1 1 0 0 1 1000
Zj 0 0 0 0 0
Zj - Cj -10 -12 0 0 0

KEY COLUMN
 Program column entries are made by location unit column
vectors corresponding to the variables S1 ,S2 ,S3 ,X and Y. here
these vectors are for variables S1 (1,0,0), S2 (0,1,0) and S3 (0,0,1).
The initial feasible solution is given by the variables S1 ,S2 and S3
with total profit=0.
Step 4:
Zj and Zj - Cj are calculated by Rules I,II and III. Now, there are
two negative values in Zj - Cj row of the initial simplex tableau.
So the initial solution needs improvement.
Step 5:
From Rule IV the most negative value i.e -12 determines the
key column and the real variable Y corresponding to this
column is the entering variable.
Step 6:
Ratio column is calculated using Rule V and the least positive
Ratio i.e. 500 in the Tableau determines the key row. Thus S 1 is
departing variable at this iteration.
Step 7:
Applying Rule VI, we get the replacing row in Tableau I from
key row in initial table. The row is

12 Y 500 2/3 1 1/3 0 0

Step 8:
The remaining rows of Tableau I are obtained by applying
Rule VII on the other rows of initial Simplex Tableau and the
replaced row in Tableau I. thus the new Tableau I for the
improved solution becomes.
 TABLEAU-I
0 Cj 10 12 0 0 0
Programme Constant X Y RATIO
Profit Z S1 S2
12 Y 500 2/3 1 1/3 0 0 750
KEY S2 500 5/3 0 -2/3 1 0 300
ROW 0 S3 500 1/3 0 -1/3 0 1 1500
0
Zj 8 12 4 0 0
Zj - Cj -2 0 4 0 0
Key Column
Here Row = (1500 3,2,0 1 0) – 2(500 2/3 1 1/3 0 0) = (500 5/3
0 – 23 1 0)
Similarly Row III can be calculated.
Step 9:
It is observed from Zj - Cj row that Tableau I does not provide
an optimum solution. Hence we again go to Step $ to further
improve the solution and get Tableau II.
TABLEAU-II
0 Cj 10 12 0 0 0
Profit Programme Constant X Y Z S1 S2
12 Y 300 0 1 3/5 -2/5 0
10 X 300 1 0 -2/5 3/5 0
0 S3 400 0 0 -1/5 -1/5 1
Zj 10 12 16/5 6/5 0
Zj - Cj 0 0 16/5 6/5 0

In Tableau II all Zj - Cj are positive. Hence the optimum

solution is X = 300, Y = 300 and maximum value of Z =6600.
The time utilized for machine M3 = 300+300 =600 hours. Thus
the unutilized time for machine M3 = 1000-600 = 400 hrs.
Minimization Problem :
There may be situations where the objective may be to
minimize the cost of production, duration of production etc i.e
the objective function is to be minimized. In such cases the
problem can be visualized as maximization problem by simply
multiplying both the sides of the objective function by -1. Here
minimization of Z is equivalent to maximization of (-Z). The
constraints remain unchanged.
The method is illustrated by the following example.
Example 4:
Z = 20 X1 + 10 X2
X1 +2X2 > 40
4X1 +3X2 > 60
3X1 +X2 > 30
X1.X2 > 0
Solution :
The minimization problem is changed into maximization
problem, by taking
Max. Z* = -Z = -20 X1 - 10 X2
Introducing slack, artificial and surplus variables in the problem,
we get
2X1 + 3X2 + S1 + 0.S2 + 0.S3 + 0.A1 + 0.A2= 40
4X1 + 3X2 + 0.S2 + 0.S3 + A1 + 0.A2= 60
3X1 + X2 + 0.S1 + 0.S2 - S3 + 0.A1 + A2= 30
And Z* = -20X1 - 10X2 + 0.S1 + 0.S2 + 0.S3 - MA1 - MA2
The various Simplex Tableau are now prepared to get the
optimal solution.
 INITIAL TABLEAU
Cj -20 -10 0 0 0 M -M
Profit Programme Constant X1 X2 S1 S2 S3 A1 A2 RATIO
0 S1 40 1 2 1 0 0 0 0 40
6 Y 60 4 3 0 -1 0 1 0 15
8 Z 30 3 1 0 0 -1 0 1 10
Zj -7M -4M M 0 M -M -M
Zj - Cj -7M+20 -4M+10 0 M M 0 0

A2 Departs And X1 Enters

Tableau - I
0 S1 30 0 5/3 1 0 1/3 0 -1/3 18
-M A1 20 0 5/3 0 -1 4/3 1 -4/3 12
-20 X1 10 1 1/3 0 0 -1/3 0 1/3 30
Zj 0 M -M -M
Zj – Cj 0 M 0 0
A1 Departs and X2 Enters
Tableau –II
0 S1 10 0 0 1 0 -1 -1 1
-10 X1 12 0 1 0 -3/5 4/5 3/5 -4/5
-20 X1 6 1 0 0 1/5 -3/5 -1/5 3/5
Zj -20 -10 0 2 4 -2 -4
Zj – Cj 0 0 0 2 4 M-2 M-4
Since M is quite large so M-4 and M-2 will be positive.
Thus in Tableau II all Zj-Cj are positive. Hence the optimum
solution is X1= 6, X2=12 and minimum value of Z = 240.
Problem On Degeneracy: (Chapter-3)
In many Linear Programming Problems it is observed that at
any iteration of the Simplex Method, two or more rows in Ratio
column have identical least non-negative value. Then the
question arises that which row should be taken as key row? This
situation where at any iteration of simplex procedure two or
more variables of the basis claim themselves to be departing
variables is known as Degeneracy Problem.
The problem of degeneracy can be experienced when one of
the constraints have zero on its right hand side.
The problem of degeneracy can be resolved with the following
steps:
(i) Divide each element of the disputed rows(i.e rows with
identical non-negative minimum entry in ratio column of
the simplex Tableau) by the key column of the respective
row.
(ii) The values obtained in Step (i) are then compared step by
step from left to right. Priority is given to identity columns
for slack and artificial variables.
(iii) The search for key row terminates where the row yields
unequal ratios. Row having algebraically smaller ratio is
taken as the key row. The method is illustrated by the
following example.
Example 5:
A manufacturing firm has discontinued production of a certain
unprofitable product line. This created considerable excess
production capacity. Management is considering to devote this
excess capacity to one or more of the three products; call them
P1,P2 and P3 .
The available capacity on the machine as well as the number of
machine hours required for each unit of the respective products is
given below.
Productivity in Machine hours Available Time in Machine
Machine Type P1 P2 P3 hours per week
Milling 8 2 3 250
Lathe 4 3 - 150
Grinder 2 - 1 50
 The unit profit would be Rs. 20, Rs. 6 and Rs. 8 respectively from
P1,P2 and P3 . Find how much of each product the firm should
produce in order to maximize the profit?
Solution:
Let the firm should be produce X units of P1 ,Y units of P2 and Z
units of P3. Then the corresponding profit will be
Z = 20 X + 6 Y + 8 Z
Thus the objective function is
Max. Z = 20 X + 6 Y + 8 Z
Subject to the constraints
8 X + 2 Y + 3 Z < 250
4 X + 3 Y < 150
2 X + Z < 50
X,Y,Z < 0.
Introducing slack variables in the objective function and
constraints, the standard form becomes
Z = 20X + 6Y + 8Z + 0.S1 + 0.S2 + 0.S3
and 8X + 2Y + 3Z + S1 + 0.S2 + 0.S3 = 250
4X + 3Y + 0.Z + 0.S1 + S2 + 0.S3 = 150
2X + 0.Y + Z + 0.S1 +0.S2 + S3 = 50
The various simplex Tableau are now prepared to get the
optimal solution.
Initial Simplex Tableau
Cj 20 6 8 0 0 0
Profit Programme Constant X Y Z S1 S2 S3 Ratio
0 S1 250 8 2 3 1 0 0 31.26
0 S2 150 4 3 0 0 1 0 37.50
0 S3 50 2 0 1 0 0 1 25.00
Zj 0 0 0 0 0 0
Zj-Cj -20 -6 -3 0 0 0
 TABLEAU –I
0 S1 50 0 2 -1 1 0 -4 25
0 S2 50 0 3 -2 0 1 -2 16.67
20 X 25 1 0 ½ 0 0 ½
Zj 20 0 10 0 0 10
Zj-Cj 0 -6 2 0 0 10

TABLEAU – II
0 S1 50/3 0 0 1/3 1 -2/3 -8/3 50
6 Y 50/3 0 1 -2/3 0 1/3 -2/3 -25
20 X 25 1 0 ½ 0 0 ½ 50
Zj 20 6 6 0 2 6
Zj-Cj 0 0 -2 0 2 6
At this stage there are two non-negative minimum values in
Ratio column of Tableau II. Thus there is a problem of
degeneracy. So we divide the I and the III rows by the elements
of key column i.e row I by 1/3 and row III by ½ and observe this
ratio for S1 and S2 and S3.
S1 S2 S1
Row I 3 -2 -8
Row III 0 0 1
Comparing step by step from left to right, we observe that the
ratio corresponding to column S1 in row III is least positive.
Hence at this stage Row III is taken as the key row and X
becomes the departing variable whereas Z is entering variable.
Note :
In case both the entries corresponding to column for S 1, were
identical then we should have moved to next column
corresponding to S2.
 Tableau III
Cj 20 6 8 0 0 0
Profit Programme Constant X Y Z S1 S2 S3
0 S1 0 -2/3 0 0 1 -2/3 -3
6 Y 50 4/3 1 0 0 1/3 0
8 Z 50 2 0 1 0 0 1
Zj 24 6 8 0 2 8
Zj - Cj 4 0 6 0 2 8
Thus all Zj - Cj are positive. Hence the optimum solution is Y =6,
Z=8, X=0 and Max. Z = 700.
REFERENCES
 Linear Programming and Network Flows
Mokhtar S. Bazaraa, Jorvis and Hanif D. Sherali
 Linear Programming
G. Hadley
 Operations Research
Kanti Swarup