B.Sc.

SEM : V CCMATH - 504C Operations Research

Unit-1 Dr. M.S.Prajapati
 Introduction to LPP: A quintessential and indispensable tool in operations
research, Linear programming has been applied to a remarkably varied number of real
problems with enormous savings in money and resources. What irony then, that this
technique, one of the most important decision-making tools used by the modern world
should owe its existence to an event responsible for horrific death and destruction – world
war II ! In 1947, George Dantzig and his associates, while working for the U.S. air Force,
had to face the problem of operational constraints due to the limited availability of
resource during wartime. It was then that Dantzig developed the theory and computational
procedure of Linear Programming for the optimization of available resource, i.e. for
determining an optional schedule of interdependent activities for the available resources.
 The Applications: Today, the technique of Linear Programming is used in many
practical industrial and business situations that require the decision maker to plan in such a
way as to make optimum use of limited resources. For example, the decision maker might
be required to maximize profit, minimize the cost of production, maximize production and
so on, all with the help of a limited pool of resources. In such situations as well as in any
sphere where constraints may be identified, LP allows the distribution of resources in the
most efficient manner and according to an objective, rational criterion.
Note that the concept of optimization and the presence of constraints are central to
such problems and hence these problems are also referred to as problems of “constrained
optimization”.
 Definition: Linear Programming refers to the technique of optimizing (maximizing /
minimizing) a linear function of variables called the “ objective function” subject to a set
of linear equations and / or inequalities called the constraints or restrictions.
 General linear programming problem: Any LPP involving two or more
variables can be expressed as follows:
Find the values of the variables x1,x2,… …,xn which optimize (maximize or minimize)
the objective function Z = c1x1+c2x2+… …+cnxn -------- (1)
Subject to the constraints,
a11x1 + a12x2 + …… + a1nxn (,=,) b1
a21x1 + a22x2 +……..+a2nxn (,=,) b2 -------- (2)
……………………………………..
am1x1 + am2x2 + …… + amnxn (,=,) bm
and the non-negative restrictions , x1,x2,… …,xn  0. -------- (3)
Each of the m constraints consists only one of the three sign , =, . The variables x1,x2,
… … ,xn are called decision variables. The real numbers c1,c2, … … ,cn are called cost-
1
coefficients. The real values b1,b2, … …,bm are different resources. Condition (3) is called
non-negative constraint.
n

 If the constraints of a LPP is, a x

j 1
ij i  bi , (i 1, 2,....k ) then, the non-negative
n

variables si which satisfy,  a x  s  b ,(i 1, 2,....k ) are called slack variables.
j 1
ij i i i

 If the constraints of a LPP is,  a x  b , (i 1, 2,....k ) then the non- negative
j 1
ij i i

variables si which satisfy,  a x  s  b ,(i 1, 2,....k ) are called surplus variables.
j 1
ij i i i

 In LPP, it is assumed that all the variables xj (j =1,2,… …,n) should be non – negative
values. In many practical situations, one or more of the variables can have either positive,
negative or zero value, are called unrestricted variables.
We have to express each of unrestricted variables as the difference of two non – negative
variables.
e.g. If variable xr be unrestricted in sign. We define two new variables say xr‟ and xr‟‟
such that xr = xr‟ - xr‟‟ ; xr‟ , xr‟‟  0.
 Canonical Form of LPP: The general canonical form of LPP is as follows:
Maximize Z = c1x1+c2x2+… …+cnxn
Subject to the constraint, ai1x1+ai2x2+…….+ainxn  bi
x1,x2,… …,xn  0, i = 1,2,……, m
It has the following characteristics:
(1) objective function is of maximization type, (2) all constraints are of () type,
(3) all variables xi are non-negative.
 Standard Form of LPP: The general standard form of LPP is as follows :
Maximize Z = c1x1+c2x2+… …+cnxn
Subject to the constraints, ai1x1+ai2x2+…….+ainxn = bi
x1,x2,… …,xn  0 , i = 1,2,……, m
It has the following characteristics:
(1) Objective function is of maximization type (2) all constraints are expressed as
equations (3) right hand side of each constraint is non-negative (4) all variables xi are non-
negative.
Note: As Minimize Z = c1x1+c2x2+… …+cnxn is equivalent to,
Maximize Z‟(= -Z) = - c1x1- c2x2 -… …- cnxn , since the objective function being the
maximization form.
2
 The inequality constraints can always be converted to equalities by adding (or
subtracting) the slack (or surplus) variables to the left hand sides of such constraints.
If the decision variables x1, x2, … … ,xn are negative , it can be expressed as the
, "
difference of two non-negative variables. e.g. xi = xi - xi , where xi, , xi"  0 .
 Some Important Definitions :
(1) Constraints: The linear inequalities or equations representing the limitations
expressed in the given problem are known as linear constraints.
(2) Objective Functions: The linear function of the variables which is to be maximized or
minimized is known as an objective function.
(3) Solution: The set of values of decision variables x1,x2,… …,xn which satisfy the given
constraints is called a solution of LPP.
(4) Feasible Solution: The set of non-negative values of the variables satisfying the given
constraints is called a feasible solution.
(5) Basic Solution: For a set of m inequalities in n unknown (n>m) variables, a solution
obtained by setting n – m variables equal to zero and the solution of the remaining m
equations in m unknown variables is called a basic solution.
(6) Basic Feasible Solution: A feasible solution which is also a basic solution is called a
basic feasible solution.
(7) Optimal Basic Feasible Solution: A basic feasible solution which optimizes the
objective function of the given L.P.P. is called optimum basic feasible solution.
(8) Non-degenerate Basic Feasible Solution: A basic solution in which all the m basic
variables are positive and remaining (n – m) variables are all zero is called a non-
degenerate basic feasible solution.
(9) Degenerate Solution: A basic feasible solution is called degenerate solution if one or
more of the basic variables have zero values.
(10) Infeasible Solution: Infeasibility arises when there is no solution to the LPP.
In graph, there is no common region which satisfies all the constraints.
(11) Unbounded Solution: A LPP in which the common feasible region is unbounded
gives unbounded solution. In such a case the optimal solution may or may not exist.
(12) Multiple Optimal Solutions: A LPP may have two or more than two optimal
solutions. In graphical method this may happens when the objective function attains
optimum value at two extreme points of a line.

3
 Graphical Method for solving LPP :
To solve a linear programming problem graphically:
Step-1. Formulate the given problem as a linear programming problem.
Step-2. Plot the given constraints as equalities on xy-coordinates plane and determine the
convex region formed by them.
Step-3. Determine the vertices of the convex region and find the value of the objective
function at each vertex. The vertex which gives the optimal (maximal or minimal)
value of the objective function gives the desired optimal solution to the problem.
Ex:1 A manufacturer produces two types of models m1 and m2 then each m1 models
requires 4 hours of grinding and 2 hours of polishing . Each m2 model requires 2 hours of
grinding and 5 hours of polishing. The manufacturer has 2 grinders and 3 polishers. Each
grinders works for 80 hour in a week and polishers works for 180 hours in a week. Profit
on m1 and m2 models is Rs.3 and Rs.4 respectively. Whatever its produced in a week is
sold in market. How should the manufacturer allocate his production capacity to the two
types of models so that he may make the maximum profit in a week.
Soln: Let x and y denote number of m1 & m2 models respectively that a manufacturer
decided to produce per week.
 Weekly profit (in Rs.) is, Z = 3x + 4y ------ (1)
Here, total number of grinding hours needed per week = 4x + 2y and total number of
polishing hours needed per week = 2x +5y
Also, 2 grinder and 3 polishers & its works in a week 80 and 180 resp.
 4x +2y  80; 2x + 5y  180  2x + y  40 ; 2x + 5y  180
Obviously x  0, y  0 .
Hence, the manufacturing problem can be put in the following form:
Maximize Z = 3x + 4y
Subject to, 2x + y  40, 2x + 5y  180 ; x  0, y  0.
The objective function and the constraints being all linear, it is a Linear Programming
Problem (LPP).
Convex region OACB is bounded by lines x = 0, y = 0, 2x+y = 40, 2x+5y = 180,
This region is called solution space or region of feasible solutions for the given
problem. Its vertices are O(0,0), A(20,0), B(0,36), C(2.5,35).
The value of the objective function (1) at these points are ,
Z(0,0) = 3(0)+4(0) = 0 , Z(20,0) = 3(20)+4(0) = 60 ,
Z(0,36) = 3(0)+4(36) = 144 , Z(2.5,35) = 3(2.5)+4(35) =147.5
Thus, the maximum value of z is 147.5 and it occurs at point C
4
Hence the optimal solution is, x = 2.5, y = 36 and Zmax=147.5
Ex:2 Solve by Graphically: Maximize Z = 2x+3y
Subject to x+y  30 , y  3, 0  y  12 , x-y  0, 0  x  20 .
Ans: ( x =18, y =12, zmax= 72 )
Note: In each of the above examples, the optimal solution was unique. But it is not
always so. The LPP may have,(1) a unique optimal solution, or (2) an infinite
number of optimal solutions, or (3) an unbounded solution, or (4) no solution.
Ex:3 Maximize Z = 100x1 +40x2,
Subject to, 10x1 + 4x2  2000 ; 3x1 + 2x2  900 ; 6x1 + 12x2  3000; x1, x2  0.
Ans: The vertices of the convex region OABC are O(0,0), A(200,0), B(125,187.5),
C(0,250). Here Z(O) = 0 , Z(A) = 20,000 , Z(B) = 20,000 , Z(C)=10,000.
Thus, the maximum value of z occurs at two vertices A and B.
 Any point on the line joining A and B will also give the same maximum value of Z .
i.e. there are infinite number of feasible solutions.

Ex:4 Maximize z = 4x1 +3x2

Subject to constraints, x1 - x2  -1, -x1 + x2  0, x1, x2  0.

Soln: There are no point (x1,x2) common to both the shaded regions, the problem cannot be
solved. Hence the solution does not exist, since the constraints are inconsistent.
5
Ex:5 Maximize Z = 2x1+3x2, subject to x1-x2  2, x1+x2  4, x1, x2  0.
Ans: Any point (x1,x2) satisfying the restrictions lies in the first quadrant only. The
solution space satisfying the constraints is the convex region shown shaded in the
following figure.
Here, the solution space is unbounded. The vertices of the feasible region are
A(3,1) and B(0,4). Value of the objective function at A, B are ZA= 9 ,ZB =12.But there are
points in this convex region for which Z will have higher values. e.g. at point (5,5) , the
value of Z =12.5
 the maximum value of Z occurs at infinity. Thus the problem has an unbounded
solution.

Exercise: Solve the following LLP by graphical method.

(1) Maximize Z = 3x + 5y [Ans: x=3, y=2, MinZ=19]
Subject to the constraints,
-3x + 4y  12; -2x + y  2; 2x + 3y  12 ; x  4 , y  2 ; x, y  0.
(2) Maximize Z = 2x1 + 4x2 [Ans: x1=0, x2=30, MaxZ=20, Degenerate soln]
Subject to the constraints,
5x1 + 4x2  200; 3x1 + 5x2  150; 5x1 + 4x2  100, 8x1 + 4x2  80, x1, x2  0.
(3) Maximize Z = 2x1 + 3x2 [Ans: x1=18, x2=12, Max Z=72]
Subject to the constraints,
x1 + x2  30; x2  3; 0  x1  20; 0  x2  12; x1 - x2  0; x1,x2 0.
(4) Maximize Z = -x1 + 2x2
Subject to the constraints, x1 - x2  -1, – 0.5x1 +x2  2 x1 , x2  0.
Comment : The maximum value of the objective function Z=4 occurs at extreme points B
and C.
every point between B and C on the line BC also gives the same value of Z.
Hence, problem has multiple optimal solutions and MaxZ = 4.

6
(5) Minimize Z = 3x1 + 2x2 [Ans: x1=1, x2=5, Min Z=13]
Subject to 5x1 + x2  10; x1 + x2  6; x1 + 4x2  12, x1, x2  0.
(6) Minimize Z = -x1 +2x2 [Ans: x1=2, x2=0, Min Z= -2]
Subject to -x1 + 3x2  10; x1 + x2  6; x1 - x2  2, x1, x2  0.
(7) Maximize Z = 7x1 + 3x2 [Ans: x1=5/2, x2=3/2, Max Z=22]
Subject to the constraints,
x1 + 2x2  3; x1 + x2  4; 0  x1  5/2 , 0  x2  3/2, x1, x2  0.
(8) Minimize Z = 20x1 +10x2 [Ans: x1=6, x2=12, Min Z= 240]
Subject to x1 + 2x2  40; 3x1 + x2  30; 4x1 +3x2  60 , x1 , x2  0.
(9) Maximize Z = 10x1 +6x2
Subject to the constraints, 5x1 + 3x2  30; x1 + 2x2  18; x1 , x2  0.
Comment: Two different extreme points B and C, the value of the objective function is
same. i.e MaxZ = 60 , two alternative solutions x1= 6/7, x2= 60/7and x1 = 6, x2 = 0 exists.
(10) Minimize Z = -x1 + 2x2 [Ans: x1=2, x2=0, Min Z=-2, Degenreate Soln]
Subject to the constraints, -x1 + 3x2  10; x1 + x2  6, x1 - x2  2, x1, x2  0.
(11) Minimize Z = 3x1 + 2x2 [Unbounded solution]
Subject to the constraints,
5x1 + x2  10; x1 + x2  6, x1 + 4x2  12 x1 , x2  0.
(12) Maximize Z = 2x1 + x2 [Ans: x1=4, x2=2, Max Z=10]
Subject to the constraints,
x1 +2x2  10; x1+x2  6, x1 - x2  2, x1 - 2x2  1, x1 , x2  0.
(13) Maximize Z = 300x1 + 400x2
Subject to the constraints,
5x1 + 4x2  200; 3x1+ 5x2  150, 5x1 + 4x2  100, 8x1 + 4x2  80, x1 , x2  0.
 Simplex Method: The graphical method is only for solving LPP involving only
two variables. Simplex method is a technique for solving LPP involving two or more
variables.
Assuming the existence of an initial basic feasible solution, an optimal solution to any
LPP by Simplex method is found as follows:
(1) Check whether the objective function is to be maximized or minimized.
If Z = c1x1+c2x2+… …+cnxn is to be minimized, then convert it into a problem of
maximization, by Maximize Z‟ = Minimize (-Z).
(2) If any value of bi is negative, then make it positive by multiplying the inequality by -1.
(3) Convert inequalities into equations using slack and surplus variables. These variables
7
 will have zero cost in objective function.
(4) Prepare a simple table with initial basic feasible solution.
(5) Compute the net evaluation j = cj-zj .
(6) If all j  0, the basic feasible solution is optimal and stop the algorithm.
(7) If one or more j are positive, locate the most positive j and the variable
corresponding to it will enter the basis. This variable is entering variable and the
corresponding column is called pivot column.
If all the elements of pivot column are zero or negative, the problem has unbounded
solution and stop the algorithm.
bj
(8) Compute replacement ratios , choose the most minimum value of replacement
aij
ratio. The variable corresponding to this minimum ratio will leave the basis. It is called
the leaving variable. For the negative or zero values of bij, the replacement ratio need
not be found out and they are not to be considered.
(9) The element which is common to elements of leaving row and entering column is
called pivot element.
(10) Construct a revised simplex table as under:
(a) Divide each element of pivotal row by the pivot element. The new row obtained in
the revised table is called replacement row.
(b) For each row other than the pivotal row, the elements are obtained by elementary
row operations..
(11) Go to step- 5 and repeat the process until an optimum solution is obtained.
(12) In the final simplex table, the values of j = Cj - Zj for basis variables are always
zeros. If the value of non-basis variable is also zero, the problem has multiple
solution.
(13) In a final simplex table if one or more basis variables have zero, then the solution is
degenerate solution.
 Solve the following problems using Simplex method:
Ex:1 Maximize Z = 7x1 +5x2
Subject to constraints, x1 + 2x2  6
4x1 + 3x2  12 ; x1 , x2  0.
Sol : By introduction slack variables s1  0, s2  0, the standard form of LPP is,
n

Maximize Z = 7x1 +5x2+0s1+0s2

Subject to x1 + 2x2+s1 = 6, 4x1 + 3x2+s2 = 12 ; x1 , x2 ,s1, s2  0.
Taking x1 = x2 = 0(non-basic variables), the initial basic feasible solution (IBFS) is,
8
s1 = 6, s2 = 12 (basic variables) and Zmax = 0.  Simplex table arrives as follows:
Cj  7 5 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 bj (bj/aij  0)
s1 0 1 2 1 0 6 6/1 = 6
s2 0 4 3 0 1 12 12/4 = 3 
(Maxi.)j = Cj-Zj 7 5 0 0 Zmax= 0
s1 0 0 5/4 1 -1/4 3
x1 7 1 3/4 0 1/4 3
j = Cj-Zj 0 -1/4 0 -7/4

Since all j  0, the current solution is optimum and unique.

the optimum solution is, x1=3, x2=0, s1=3, s2=0 and Zmax= 21
Ex:2 Maximize Z = 5x1 +7x2
Subject to x1 + x2  4; 3x1 + 8x2  24; 10x1 + 7x2  35 , x1 , x2  0.
Soln: By introduction slack variables s1 , s2, s3  0, the standard form of LPP is,
Maximize Z = 5x1 +7x2+0s1+0s2+0s3
Subject to x1 + x2+s1 = 4 , 3x1 + 8x2+s2 = 24
10x1 + 7x2+s3 = 35 ; x1 , x2 ,s1, s2 , s3  0.
Taking x1 = x2 = 0(non-basic variables), the initial basic feasible solution (IBFS) is,
s1 = 4, s2 = 24, s3 = 35 and Zmax = 0.  Simplex table arrives as follows:
Cj  5 7 0 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 s3 bj (bj/aij  0)
s1 0 1 1 1 0 0 4 4/1 = 4
s2 0 3 8 0 1 0 24 24/8 = 3
s3 0 10 7 0 0 1 35 35/7 = 5
(Maxi.)j = Cj-Zj 5 7 0 0 0 Zmax= 0
s1 0 5/8 0 1 -1/8 0 1 8/5 =1.6 
x2 7 3/8 1 0 1/8 0 3 8
s3 0 59/8 0 0 -7/8 1 14 112/59 =1.89
j = Cj-Zj 19/8 0 0 -7/8 0 Zmax= 21
x1 5 1 0 8/5 -1/5 0 8/5
x2 7 0 1 -3/5 1/5 0 12/5
s3 0 0 0 -59/5 3/5 1 11/5
j = Cj-Zj 0 0 -19/5 -2/5 0 Zmax= 124/5
9
Since all j = 0 for all basic variables and j  0 for all non-basic variables.
 the current solution is optimal and non-degenerate.
 the optimum solution is, x1 = 8/5, x2 = 12/5 , s3 = 11/5 and Zmax= 124/5 .
Ex:3 Maximize Z = 4x1 +10x2
Subject to 2x1 +x2  50; 2x1 + 5x2  100; 2x1 +3x2  90 , x1 , x2  0.
Soln: By introduction slack variables s1 , s2, s3  0, the standard form of LPP is,
Maximize Z = 4x1 +10x2+0s1+0s2+0s3
Subject to 2x1 + x2+s1 = 50
2x1 + 5x2+s2 = 100
2x1 + 3x2+s3 = 90 ; x1 , x2 ,s1, s2 , s3  0.
Taking x1 = x2 = 0(non-basic variables), the initial basic feasible solution (IBFS) is,
s1 = 50, s2 = 100, s3 = 90 and Zmax = 0.  Simplex table arrives as follows:
Cj  4 10 0 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 s3 bj (bj/aij  0)
s1 0 2 1 1 0 0 50 50/1 = 50
s2 0 2 5 0 1 0 100 100/5 = 20
s3 0 2 3 0 0 1 90 90/3 = 30
(Maxi.)j = Cj-Zj 4 10 0 0 0 Zmax= 0
s1 0 8/5 0 1 -1/5 0 30
x2 10 2/5 1 0 1/5 0 20
s3 0 4/5 0 0 -3/5 1 30
j = Cj-Zj 0 0 0 -2 0 Zmax= 200

Here j = 0 for all basic variables & a non- basic variables x1 and optimality condition is
satisfied.  the current solution is optimal but not unique. i.e. multiple optimum solution.
 the optimum solution is, x1 = 0, x2 = 20 , s1 = s3 = 30 and Zmax= 200.
Ex:4 Maximize Z = 10x1 +20x2
Subject to 5x1 + 3x2  30; 3x1 + 6x2  36; 2x1 + 5x2  20 , x1 , x2  0.
Soln: By introduction slack variables s1  0, s2  0, the standard form of LPP is,
Maximize Z = 10x1 +20x2+0s1+0s2+0s3
Subject to, 5x1 + 3x2+s1 = 30
3x1 + 6x2+s2 = 36
2x1 + 5x2+s3 = 20 ; x1 , x2 ,s1, s2 , s3  0.
Taking x1 = x2 = 0(non-basic variables), the initial basic feasible solution (IBFS) is,

10
s1 = 30, s2 = 36, s3 = 20 (basic variables) and Zmax = 0, then the LPP solve by Simplex
method as follows:

Cj  10 20 0 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 s3 bj (bj/aij  0)
s1 0 5 3 1 0 0 30 30/3 = 10
s2 0 3 6 0 1 0 36 36/6 = 6
s3 0 2 5 0 0 1 20 20/5 = 4 
(Maxi.)j = Cj-Zj 10 20 0 0 0 Zmax= 0
s1 0 19/5 0 1 0 -3/5 18 90/19
s2 0 3/5 0 0 1 -6/5 12 20
x2 20 2/5 1 0 0 1/5 4 10
j = Cj-Zj 2 0 0 0 -4 Zmax= 80
x1 10 1 0 5/19 0 -3/19 90/19
s2 7 0 0 -3/19 1 -21/19 174/19
x2 20 0 1 -2/19 0 5/19 40/19
j = Cj-Zj 0 0 -10/19 0 -70/19 Zmax= 1700/19

Since all j  0, the current solution is optimum and unique.

the optimum solution is, x1=90/19, x2=40/19, s2=174/19 and Zmax= 1700/19
Ex:5 Maximize Z = x1 +x2 +x3
Subject to 2x1 + x2-x3  2; -2x1 + x2-5x3  -6; 4x1 + x2+x3  6,x1 , x2, x3  0.
Soln: By introduction slack variables s1 , s2 , s3  0, the standard form of LPP is,
Maximize Z = x1 + x2 + x3
Subject to, 2x1 + x2 – x3 +s1 = 2
2x1 - x2+ 5x3 + s2 = 6
4x1 + x2+ x3 + s3 = 6 ; xi , si  0, i = 1,2,3.
Taking x1 = x2 =x3 =0(non-basic variables), the IBFS is, s1 =2, s2 = 6, s3 = 6 (basic -
variables) and Zmax= 0, then the LP problem solve by Simplex method as follows:

11
 Cj  1 1 1 0 0 0 Ratio(Mini.)
Basis CB x1 x2 x3 s1 s2 s3 bj (bj/aij  0)
s1 0 2 1 -1 1 0 0 2 ---
s2 0 2 -1 5 0 1 0 6 6/5 
s3 0 4 1 1 0 0 1 6 6/1 = 6
(Maxi.)j = Cj-Zj 1 1 1 0 0 0 Zmax= 0
s1 0 12/5 4/5 0 1 1/5 0 16/5 4
x3 1 2/5 -1/5 1 0 1/5 0 6/5 ---
s3 0 18/5 6/5 0 0 -1/5 1 24/5 4
j = Cj-Zj 3/5 6/5 0 0 -1/5 0 Zmax= 6/5
x2 1 3 1 0 5/4 1/4 0 4
x3 1 1 0 1 ¼ 1/4 0 2
s3 0 0 0 0 -3/2 -1/2 1 0
j = Cj-Zj -3 0 0 -3/2 -1/2 0 Zmax= 6

Since all j  0, the current solution is optimum and degenerate (since, s3=0).
 the optimum solution is, x1=0, x2=4, x3=2 and Zmax= 6
Ex:6 Minimize Z = -2x1 - x2
Subject to x1 + 2x2  10; x1 + x2  6 ; x1 - x2  2 ; x1 - 2x2  1 ,x1 , x2  0.
Soln: First we convert the given objective function to maximization by multiplying -1,
we have, Maximize Z = 2x1 + x2
Subject to x1 + 2x2  10; x1 + x2  6 ; x1 - x2  2 ; x1 - 2x2  1 ,x1 , x2  0.
By introduction slack variables si  0, (i = 1,2,3,4) the standard form of LPP is,
Maximize Z = 2x1 + x2
Subject to, x1 +2 x2 +s1 = 10
x1 + x2+ s 2 = 6
x1 - x2+ s3 = 2
x1 - 2x2 +s4 = 1 ; x1, x2 , si  0, i = 1,2,3,4.
Taking x1 = x2 =0, the IBFS is, s1 = 10, s2 = 6, s3 = 2, s4 = 1 and Zmax = 0, then the LP
problem solve by Simplex method as follows:

12
 Cj  2 1 0 0 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 s3 s4 bj (bj/aij  0)
s1 0 1 2 1 0 0 0 10 10
s2 0 1 1 0 1 0 0 6 6
s3 0 1 -1 0 0 1 0 2 2
s4 0 1 -2 0 0 0 1 1 1
(Maxi.)j = Cj-Zj 2 1 0 0 0 0 Zmax= 0
s1 0 0 4 1 0 0 -1 9 9/4
s2 0 0 3 0 1 0 -1 5 5/3
s3 0 0 1 0 0 1 -1 1 1
x1 2 1 -2 0 0 0 1 1 ---
j = Cj-Zj 0 5 0 0 0 -2 -2 Zmax= 2
s1 0 0 0 1 0 -4 3 5 5/3
s2 0 0 0 0 1 -3 2 2 1
x2 1 0 1 0 0 1 -1 1 ---
x1 2 1 0 0 0 2 -1 3 ---
j = Cj-Zj 0 0 0 0 -5 3 Zmax= 7
s1 0 0 0 1 -3/2 1/2 0 2
s4 0 0 0 0 1/2 -3/2 1 1
x2 1 0 1 0 1/2 -1/2 0 2
x1 2 1 0 0 1/2 1/2 0 4
j = Cj-Zj 0 0 0 -3/2 -1/2 0 Zmax= 10

Since all j  0, the current solution is optimum and non-degenerate.

the optimum solution is, x1=4, x2=2, s1=2, s2 = s3 =0, s4=1 and Zmini = -10
Ex:7 Maximize Z = 3x1 + 4x2
Subject to x1 - x2  1; -x1 + x2  2 ; x1 , x2  0.
Soln: By introduction slack variables s1 , s2  0, the standard form of LPP is,
Maximize Z = 3x1 +4x2
Subject to x1 - x2+s1 = 1, -x1 + x2+s2 = 2 ; x1 , x2 ,s1, s2  0.
Taking x1 = x2 = 0, the initial basic feasible solution (IBFS) is, s1 = 1, s2 = 2 and Zmax = 0.
 Simplex table arrives as follows:

13
 Cj  3 4 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 bj (bj/aij  0)
s1 0 1 -1 1 0 1 ---
s2 0 -1 1 0 1 2 2/1 = 2
(Maxi.)j = Cj-Zj 3 4 0 0 Zmax= 0
s1 0 0 0 1 1 3 ---
x2 3 -1 1 0 1 2 ---
j = Cj-Zj 7 0 0 -4 Zmax= 8

Here all elements in the pivot column are  0, the given LPP have unbounded solution.
But last one solution is, x1=0, x2=2 and Zmax= 8, which is not optimum.
Ex:8 Maximize Z = 2x1 + 3x2
Subject to -x1 + 2x2  4; x1 + x2  6; x1 + 3x2  9 and x1 , x2 unrestricted.
n
Sol : Since x1 and x2 are unrestricted, we introduce the non-negative variables x1‟, x1”, x2‟,
x2” such that x1 = x1‟- x1” and x2 = x2‟- x2”. Also introducing slack variables s1, s2, s3  0,
the standard form of LPP is,
Max Z = 2x1‟-2x1” +3x2‟-3x2”
Subject to, -x1‟+x1”+2x2‟-2x2”+s1 = 4
x1‟- x1”+ x2‟-x2”+ s2 = 6
x1‟-x1”+ 3x2‟-3x2” + s3 = 9, x1‟, x1”, x2‟,x2”, s1, s2  0
Taking x1‟= x1”= x2‟= x2” = 0, the IBFS is, s1 = 4, s2 = 6, s3 = 9 & Zmax= 0.

Cj  2 -2 3 -3 0 0 0 Ratio(Mini.)
XB CB x 1‟ x1” x2‟ x2” s1 s2 s3 bj (bj/aij  0)
s1 0 -1 1 2 -2 1 0 0 4 4/2 = 2 
s2 0 1 -1 1 -1 0 1 0 6 6/1 = 6
s3 0 1 -1 3 -3 0 0 1 9 9/3 = 3
(Maxi.)j = Cj-Zj 2 -2 3 -3 0 0 0 Zmax= 0
x2‟ 3 -1/2 1/2 1 -1 1/2 0 0 2 ---
s2 0 3/2 -3/2 0 0 -1/2 1 0 4 8/3
s3 0 5/2 -5/2 0 0 -3/2 0 1 3 6/5 
j = Cj-Zj 7/2 -7/2 0 0 -3/2 0 0 Zmax= 6
x2‟ 3 0 0 1 -1 1/5 0 1/5 13/5 65/5
s2 0 0 0 0 0 2/5 1 -3/5 11/5 11/2 
14
 x1‟ 2 1 -1 0 0 -3/5 0 2/5 9/5 ---
j = Cj-Zj 0 0 0 0 3/5 0 -7/5 Zmax= 51/5
x2‟ 3 0 0 1 -1 0 -1/2 1/2 3/2
s1 0 0 0 0 0 1 5/2 -3/2 11/2
x1‟ 2 1 -1 0 0 0 3/2 -1/2 9/2
j = Cj-Zj 0 0 0 0 0 -3/2 -1/2 Zmax= 27/2
Since all j  0, the optimum solution is x1,‟= 9/2, x2‟= 3/2 with maximum Z = 27/2.
 x1 = 9/2 – 0 = 9/2, x2 = 3/2 – 0 = 3/2 and Zmax= 18/2 + 9/2 = 27/2.
Exercise:
(1) Minimize Z = -5x1 -3x2 [Ans: x1=2, x2=0, Zmin=10]
Subject to x1 + x2  2; 5x1 + 2x2  10; 3x1 + 8x2  12 , x1 , x2  0.
(2) Maximize Z = 8x1 +9x2+5x3 [Ans: x1=1, x2=1/3,x3=0,x4=2/3, Z=11]
Subject to x1 + x2+2x3  2; 2x1 + 3x2+4x3  3; 6x1+6x2+2x3  8,x1 ,x2,x3  0.
(3) Minimize Z = x1 -3x2+2x3 [Ans: x1=14/5, x2 =22/5,x3=1, Z = -62/5]
Subject to 3x1 - x2+3x3  7; -2x1 + 4x2  12; -4x1+3x2+8x3  10,x1 ,x2,x3  0.
(4) Maximize Z = 11x1 + 9x2 [Ans: x1=2, x2=1, Zmax=31]
Subject to 3x1 + 2x2  8; 2x1 + 3x2  7 ; x1 , x2  0.
(5) Maximize Z = 4x1 + 3x2 [Ans: x1=200, x2=600, Zmax=2600]
Subject to 2x1 + x2  1000; x1 + x2  800 ; x1  400 ; x2  700 ,x1 , x2  0.
(6) Maximize Z = 10x1 +x2+2x3 [Ans: x1=5, x2=0,x3=0, Zmax =50]
Subject to x1 + x2-2x3  10; 4x1 + x2+x3  20; x1 ,x2, x3  0.
(7) Maximize Z = 3x1 -x2+3x3+4x4 [Ans: x1=0, x2=6,x3=14, Zmax =36]
Subject to x1 + 2x2 +2x3+4x4  40; 2x1-x2+x3+2x4  8;
4x1-2x2+x3-x4  10 ; x1 ,x2,x3,x4  0.
(8) Maximize Z = 2x1 +x2 [Ans: x1=3/2, x2=2, Zmax =5]
Subject to 4x1 + 3x2  12; 4x1 + x2  8; 4x1 - x2  8 , x1 , x2  0.
(9) Maximize Z = 3x1 +5x2+4x3 [Ans: x1=0, x2=8/3,x3=2, Zmax=64/3]
Subject to 2x1 + 3x2  8; 2x2 + 5x3  10; 3x1+2x2+4x3  15,x1 ,x2,x3  0.
 Tie for entering Basic variable(key column):
A situation may arise at any iteration when two or more columns may have exactly
the same Cj – Zj value. In order to break this tie, the selection for key column (entering
variable) can be made arbitrary. However, the number of iterations required to arrive at
the optimal solution can be minimized by adopting the following rules:
(1) If there is a tie between two decision variables, then the selection can be made
15
 arbitrarily.
(2) If there is a tie between a decision variable and a slack (or surplus) variable, then
select the decision variable to enter into basis first.
(3) If there is a tie between two slack (or surplus) variables, then select can be made
arbitrarily.
 Tie for leaving Basic variable (key row) – Degeneracy:
While solving an LP problem a situation may arise in which there is a tie between two
or more basic variables for leaving the basis, i.e. the minimum ratio to identify the basic
variable to leave the basis is not unique or in which value of one or more basic variables in
the „solution values‟ column (bj) become equal to zero. This causes the problem of
degeneracy. However, if the minimum ratio is zero, then the iterations of simplex method
are repeated (cycle) indefinitely without arriving at the optimal solution. Such a situation
is very rare in practical problems.
The problem of degeneracy arises due to redundant constraints, for example,
constraints such as x1  5, x2  5, and x1+x2  5 in the LP problem, make constraints
x1+x2 5 unnecessary (redundant).
Degeneracy may occur at any iteration of the Simplex method. In the most of the
cases when there is a tie in the minimum ratios, the selection is made arbitrarily. However
the number of iterations required arriving at the optimal solution can be minimized by
adopting the following rules:
(1) Divided the coefficients of slack variables in the simplex table where degeneracy is
detected by the corresponding positive numbers of the key column in the row, starting
from left to right.
(2) The row that contains the smallest ratio comparing from left to right column wise
becomes the key row.
Consider the following example:
Ex:1 Maximize Z = 5x1+3x2
Subject to x1 + x2  2, 5x1 +2x2  10, 3x1+8x2  12 , x1, x2  0.
Sol : By introduction slack variables s1 ,s2, s3  0, the standard form of LPP is,
n

Maximize Z = 5x1 +3x2+0s1+0s2+0s3

Subject to, x1 + x2+s1 = 2
5x1 + 2x2+s2 = 10
3x1 + 8x2+s3 = 12 ; x1 , x2 ,s1, s2 , s3  0.
Taking x1 = x2 = 0, the initial basic feasible solution (IBFS) is,s1 = 2, s2 = 10, s3 = 12
(basic variables) and Zmax = 0, then the LPP solve by Simplex method as follows:

16
 Cj  5 3 0 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 s3 bj (bj/aij  0)
s1 0 1 1 1 0 0 2 2
s2 0 5 2 0 1 0 10 10/5 = 2
s3 0 3 8 0 0 1 12 12/3 = 4
(Maxi.)j = Cj-Zj 5 3 0 0 0 Zmax= 0
x1 5 1 1 1 0 0 2
s2 0 0 -3 -5 1 0 0
s3 0 0 5 -3 0 1 6
j = Cj-Zj 0 -2 -5 0 0 Zmax= 10

Since all j  0, the current solution is optimum and degenerate.

the optimum solution is, x1=2, x2=0 and Zmax= 10
Ex:2 Maximize Z = 3x1+9x2
Subject to x1 + 4x2  8, x1 +2x2  4 ; x1, x2  0.
Sol : By introduction slack variables s1 ,s2  0, the standard form of LPP is,
n

Maximize Z = 3x1 +9x2+0s1+0s2

Subject to, x1 + 4x2+s1 = 8
x1 + 2x2+s2 = 4 ; x1 , x2 ,s1, s2 , s3  0.
Taking x1 = x2 = 0, the initial basic feasible solution (IBFS) is, s1 = 8, s2 = 10, s3 = 4 and
Zmax = 0, then the LPP solve by Simplex method as follows:

Cj  3 9 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 bj (bj/aij  0)
s1 0 1 4 1 0 8 8/4 = 2 
s2 0 1 2 0 1 4 4/2 = 2
(Maxi.)j = Cj-Zj 3 9 0 0 Zmax= 0
x2 9 1/4 1 1/4 0 2 8
s2 0 1/2 0 -1/2 1 0 0
j = Cj-Zj 3/4 0 -9/4 0 Zmax= 18
x2 9 0 1 1/2 -1/2 2
x1 3 1 0 -1 2 0
j = Cj-Zj 0 0 -3/2 -3/2 Zmax= 18
Since all j  0, the current solution is optimum and degenerate.  the optimum solution
is, x1=0, x2=2 and Zmax= 18
17
Theorem: Prove that the set of all feasible solutions to an L.P.P. is convex.
Proof : Let SF be the set of all feasible solutions.
Let the LPP be converts into AX = B. Let X1, X2 be two feasible solutions.
 AX1 = B, AX2 = B ------- (1) and X1, X2  0.
Let X be any convex linear combination of X1 and X2.
 X = (1- )X1 +X2 , 0    1
 X  0, since X1, X2  0. -------- (2)
Now, AX = A[(1- )X1 +X2]
= (1- ) AX1 + AX2 = (1- )B +B = B -------- (3)
From (2), (3) means that X is a feasible solution.
 The convex linear combination of every two feasible solutions is a feasible solution.
 the set of all feasible solutions SF is convex.
Theorem : Prove that a basis feasible solutions of the L.P.P. is a vertex of the convex set
of feasible solutions.
Proof : Let X = [1,2,… …m,0,0,….,0] be a basic feasible solution of eqn AX=B,
where X  0.   vectors p1,p2,… …,pm which are linearly independent such that,
1p1+2p2+… …+mpm = B ------- (1) where j  0 , j = 1,2, … … … ,m.
We have to show that X is a vector of convex set SF.
Suppose X is not a vertex of SF.
  two points X1 = (x11,x21,… …,xm1,x (m+1)1,… ….xn1) & X2 = (x12,x22,… …,xm2, x(m+1)2,
… ….xn2)SF, where X1  X , X2  X  X = X1+ (1- )X2 , 0<<1.
 j = xj1+ (1- )xj2 , j = 1,2,… ...,m
0 = xj1+ (1- )xj2 , j = m+1,… … …,n --------- (2)
Here X1, X2SF , xj1  0 , xj2  0 and 0<<1, 1- >0 , >0.
 from (2), xj1 = 0 , xj2 = 0 , j = m+1,m+2,… … …,n.
 X1 = (x11,x21,… …,xm1,0,0,… ...,0) & X2 = (x12,x22,… …,xm2,0,0,… ...0).
Here X1, X2SF  X1, X2 are solutions of AX=B.
 x11p1+x21p2+… …+xm1pm = B -------- (3)
x12p1+x22p2+… …+xm2pm = B -------- (4)
(1) – (3)  (1 – x11)p1+(2 - x21)p2+… …+(m – xm1)pm = 0.

18
Here, p1, p2,… …,pm are linearly independent.
 1 – x11 = 2 - x21 = ……..= m – xm1 = 0.
 1 = x11 , 2 = x21 , … ..., m = xm1.
 X = X1. Similarly X = X2, a contradiction with X  X1 , X  X2.
 Our supposition is wrong. Thus X is a vector of convex set SF.
Theorem : Prove that a vertex of SF is a basic feasible solution.
Proof : Let X = (1,2,… …n) be a vertex of SF.  X  SF  X  0.
We have to show that X is a basic feasible solution. i.e. we have to show that at least (n –
m) j‟s are zero. Let r of them 1, 2, … …, n be non-zero, where r  n.
Since m < n, either r  m or r > m.
If r  m then X is obviously a b.f.s. and the theorem holds.
If r > m then we have, X = (1,2,… …,r,0,0,….,0) , where j > 0 for j=1,2,…..,r.
Since X is a solution of AX = B, we have,
1p1+2p2+… …+rpr = B.------- (1)
As r > m, the vectors p1, p2,…., pr are not linearly independent.
  Scalars 1,2,3,…..,r not all zero such that 1p1+2p2+…+rpr = 0.
Multiplying by c > 0, we get c1p1+c2p2+… …+crpr = 0 --------- (2)
 eqns (1) + (2)  (1 +c1)p1+(2 +c2)p2+… …+(r +cr)pr = B
eqns (1) - (2)  (1 -c1)p1+(2 -c2)p2-… …-(r -cr)pr = B
Here j > 0 and we can choose c, so small such that , j  cpj > 0, j = 1,2,…,r.
 we have , X1 = (1+c1, 2+c2,… ..., r+cr, 0, 0,… ...,0)
X2 = (1-c1, 2-c2,… ..., r-cr, 0, 0,… ...,0)
 X1 and X2 are feasible solutions of AX = B.
Thus, we have three feasible solutions namely X , X1 ,X2 .
Now, X = (X1 +X2)/2 , then X is a convex linear combination of X1 and X2 and X1  X ,
X2  X .  X is not a vertex. (by defn of vertex) , this is contradiction, because X is a
vertex .  our supposition r > m is wrong .  r  m.
 X is a basic feasible solution. Thus vertex SF is a b.f.s.
Theorem : Prove that the objective function on a LPP attains its optimal value more than
one extreme point then for every convex combination of those points keep up the same
optimal value.
Proof : Consider the standard form of LPP, Zmax= CX ; subject to AX=B, X  0.

19
Let SF be the set of all feasible solutions of AX=B, X  0. Let the objective function
Z = CX attains same optimal value at X1 and X2 , these points are extreme points.
Zmax= CX1= CX2.
Let X3 = X2+(1- )X1 be the convex linear combination of X1 and X2, 0    1.
CX3 = C[X2+(1- )X1] = (CX2)+ (1- )(CX1) = . Zmax+ (1- ).Zmax = Zmax
i.e. X3 gives the same optimal value.  CX3 = Zmax= CX1= CX2.
Thus every convex linear combination of two extreme points gives the same value of the
objective function.

MSPMSPMSPMSPMSP

20
B.Sc. Sem : V CCMATH - 504C Operations Research
Unit-2 Dr. M.S.Prajapati
 Artificial variable techniques:
There are many problems wherein at least one of the constraints is of () or (=) type
and slack variables fail to give such a solution. In this case, we add two new variables
namely surplus variables and artificial variables. There are two similar methods for
solving such problems as below.
[1] Two-phase method:
This is another method to deal with the artificial variables wherein the LPP is solved
in two phases.
Phase - I
Step 1. Express the given problem in the standard form by introducing slack, surplus and
artificial variables.
Step 2. Formulate an artificial objective function Z* = -A1-A2- … … -Am by assigning
(-1) cost to each of the artificial variables Ai and zero cost to all other variables.
Step 3. Maximize Z* subject to the constraints of the original problem using the simplex
method. Then three cases arises:
(a) Max. Z* < 0 and at least one artificial variable appears in the optimal basis at a
positive level. In this case, the original problem doesn‟t possess any feasible solution
and the procedure comes to an end.
(b) Max. Z* = 0 and no artificial variables appears in the optimal basis.
In this case, a basis feasible solution is obtained and we proceed to phase-2 for
finding the optimal basic feasible solution to the original problem.
(c) Max. Z* = 0 and at least one artificial variable appears in the optimal basis at zero
level. Here a feasible solution to the auxiliary L.P.P. is also a feasible solution to the
original problem with all artificial variables set = 0. To obtain a basic feasible solution,
we prolong phase-1 for pushing all the artificial variables out of the basis (without
proceeding on to phase-2 ).
Phase - II The basic feasible solution found at the end of phase-1 is used as the starting
solution for the original problem in this phase. i.e.the final simplex table of phase-1
is taken as the initial simplex table of phase-2 and the artificial objective function is
replaced by the original objective function. Then we find the optimal solution.
Solve the following examples using two-phase simplex method:
Ex:1 Minimize Z = x1+x2
Subject to the constraints, 2x1 + x2  4, x1 +7x2  7, x1, x2  0.
Soln: Converting the given LPP into the maximization from and then adding surplus
variables s1 and s2 and artificial variables A1 and A2, the standard form of LPP is,
Maximize Z*=(-Z) = -x1-x2
Subject to the constraints, 2x1 + x2 – s1 + A1 = 4
x1 +7x2 - s2 + A2 = 7 ; x1, x2 , s1, s2 , A1, A2  0.
21
Phase I : The auxiliary LPP is, Maximize Z* = - A1 - A2
Subject to the constraints, 2x1 + x2 – s1 + A1 = 4
x1 +7x2 - s2 + A2 = 7 ; x1, x2 , s1, s2 , A1, A2  0.
Taking x1 = x2 = s1 = s2 = 0, the IBFS is, A1 = 4, A2 = 7 and Z*max = -11 0

Cj  0 0 0 0 -1 -1 Ratio(Mini.)
Basis CB x1 x2 s1 s2 A1 A2 bj (bj/aij  0)
A1 -1 2 1 -1 0 1 0 4 4/1=4
A2 -1 1 7 0 -1 0 1 7 7/7=1
(Maxi.)j = Cj-Zj 3 8 -1 -1 0 0 Z*max = -11
A1 -1 13/7 0 -1 1/7 1 -1/7 3 21/13
x2 0 1/7 1 0 -1/7 0 1/7 1 7
j = Cj-Zj 13/7 0 -1 1/7 0 -8/7 Z*max = -3
x1 0 1 0 -7/13 1/13 7/13 -1/13 21/13
x2 0 0 1 1/13 -2/13 -1/13 2/13 10/13
j = Cj-Zj 0 0 0 0 -1 -1 Z*max = 0

Since all j  0 corresponding to non-basic variables, the optimal solution to auxiliary LPP
is, x1 = 21/13, x2 = 10/13and Z*max= 0.
Here Max. Z* = 0 and no artificial variables appears in the optimal basis. In this case, we
proceed tophase -II for finding the optimal basic feasible solution to the original problem.
Phase II : Taking Maxi. Z = - x1 - x2 and the final simplex table of phase -I is taken as
the initial simplex table of phase –II by removing columns A1 and A2.
Cj  -1 -1 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 bj (bj/aij  0)
x1 -1 1 0 -7/13 1/13 21/13
x2 -1 0 1 1/13 -2/13 10/13
(Maxi.)j = Cj-Zj 0 0 -6/13 -1/13 Zmax = -31/13

Since all j  0 for all non-basic variables, the current solution is optimal.
 the optimal basic feasible solution to the given LPP is, x1=21/13, x2=10/13 and
Zmax = 31/13.
Ex:2 Minimize Z = x1-2x2 -3x3
Subject to the constraints, -2x1 +x2+3x3 = 2; 2x1+3x2+4x3 =1; x1, x2, x3  0.
Soln: Converting the given LPP into the maximization from and by adding artificial
variables A1 and A2 in the constraints, the standard form of LPP is,
Maximize Z*=(-Z) = -x1+2x2 +3x3
22
 Subject to the constraints, -2x1 +x2+3x3 +A1 = 2; 2x1+3x2+4x3+A2 =1;
x1, x2, x3 A1, A2  0.
*
Phase I : The auxiliary LPP is, Maximize Z = -A1-A2
Subject to the constraints, -2x1 +x2+3x3 +A1 = 2
2x1+3x2+4x3+A2 =1; x1, x2, x3 A1, A2  0.
Taking x1 = x2 = x3 = 0, the IBFS is, A1 = 2, A2 = 1 and Z*max = -3  0
Cj  0 0 0 -1 -1 Ratio(Mini.)
Basis CB x1 x2 x3 A1 A2 bj (bj/aij  0)
A1 -1 -2 1 3 1 0 2 2/3
A2 -1 2 3 4 0 1 1 1/4 
(Maxi.)j = Cj-Zj 0 4 7 0 0 Z*max = -3
A1 -1 -7/2 -5/4 0 1 -3/4 5/4 
x3 0 1/2 3/4 1 0 1/4 1/4
j = Cj-Zj -7/2 -5/4 0 0 -7/4 Z*max = -5/4

Since j  0 corresponds to non-basic variables, the optimal solution is, x1=0, x2=0,
x3=1/4, A1=5/4, A2=0 with Max Z*=-5/4. But at the same time, the value of Z* < 0 and the
artificial variable A1 appears in the basis with positive value 5/4. Hence, the given LPP
does not possess any feasible solution.
Ex:3 Maximize Z = 3x1+2x2 +2x3
Subject to, 5x1 +7x2+4x3  7
-4x1+7x2+5x3  -2
3x1+4x2-6x3  29/7 and x1, x2, x3  0.
n
Sol : Introducing slack, surplus and artificial variables, the standard form of LPP is,
Maximize Z = 3x1+2x2 +2x3
Subject to, 5x1 +7x2+4x3 +s1 = 7
4x1-7x2-5x3 +s2 = 2
3x1+4x2-6x3 -s3+A1 = 29/7 ; x1, x2, x3 , s1,s2,s3, A1 0.
Phase I : The auxiliary LPP is, Maximize Z* = - A1
Subject to, 5x1 +7x2+4x3 +s1 = 7
4x1-7x2-5x3 +s2 = 2
3x1+4x2-6x3 -s3+A1 = 29/7 ; x1, x2, x3 , s1,s2,s3, A1 0.
Taking x1 = x2 = x3 = s3 = 0, the IBFS is, s1 = 7, s2 = 2, A1= 29/7 and Z*max = -29/7  0

23
 Cj  0 0 0 0 0 0 -1 Ratio(Mini.)
Basis CB x1 x2 x3 s1 s2 s3 A1 bj (bj/aij  0)
s1 0 5 7 4 1 0 0 0 7 1
s2 0 4 -7 -5 0 1 0 0 2 ---
A1 -1 3 4 -6 0 0 -1 1 29/7 29/28
(Maxi.)j = Cj-Zj 3 4 -6 0 0 -1 0 Z*max = -29/7
x2 0 5/7 1 4/7 1/7 0 0 0 1 7/5
s2 0 9 0 -1 1 1 0 0 9 1
A1 -1 1/7 0 -58/7 -4/7 0 -1 1 1/7 1
j = Cj-Zj 1/7  0 -58/7 -4/7 0 -1 0 Z*max = -1/7
x2 0 0 1 42 3 0 5 -5 2/7
s2 0 0 0 521 37 1 63 -63 0
x1 0 1 0 -58 -4 0 -7 7 1
j = Cj-Zj 0 0 0 0 0 0 -1 Z*max = 0

Since all j  0, the optimal solution to auxiliary LPP is, x1 = 1, x2 = 2/7, s2 = 0 and
Z*max= 0. Here Max. Z* = 0 and no artificial variables appears in the optimal basis. In this
case, we proceed tophase -II for finding the optimal basic feasible solution to the original
problem.
Phase II : Taking Maximize Z = 3x1+2x2 +2x3 and the final simplex table of phase -I is
taken as the initial simplex table of phase –II by removing columns A1.
Cj  3 2 2 0 0 0 Ratio(Mini.)
Basis CB x1 x2 x3 s1 s2 s3 bj (bj/aij  0)
x2 2 0 1 42 3 0 5 2/7 1/147
s2 0 0 0 521 37 1 63 0 0
x1 3 1 0 -58 -4 0 -7 1 ---
(Maxi.)j = Cj-Zj 0 0 92 6 0 11 Zmax = 25/7
x2 2 0 1 0 9/521 -42/521 -41/521 2/7
x3 2 0 0 1 37/521 1/521 63/521 0
x1 3 1 0 0 62/521 58/521 7/521 1
(Maxi.)j = Cj-Zj 0 0 0 -278/521 -92/521 -65/521
Since all j  0 for all non-basic variables, the current solution is optimal.
 the optimal basic feasible solution to the given LPP is, x1=1, x2=2/7, x3=0 and
Zmax = 3(1)+2(2/7)+2(0) = 3 + 4/7 = 25/7 .

24
Ex:4 Maximize Z = -x1-2x2-3x3
Subject to x1 - x2 + x3  4; x1 + x2+2x3  8, x2-x3  2 , x1, x2, x3  0.
n
Sol : Introducing slack, surplus and artificial variables, the standard form of LPP is,
Maximize Z = -x1-2x2-3x3
Subject to, x1 - x2 + x3 - s1 +A1 = 4
x1 + x2+2x3 + s2 = 8
x2 - x3 –s3+A2 = 2 ; x1, x2, s1, s2, s3, A1, A2  0.
Phase I : The auxiliary LPP is, Maximize Z* = - A1 - A2
Subject to, x1 - x2 + x3 - s1 +A1 = 4
x1 + x2+2x3 + s2 = 8
x2 - x3 –s3+A2 = 2 ; x1, x2, s1, s2, s3, A1, A2  0.
Taking x1 = x2 = x3 = s1 = s3 = 0, the IBFS is, A1 = 4, s2 = 8, A2 = 2 and Z*max = -6  0.

Cj  0 0 0 0 0 0 -1 -1 Ratio(Mini.)
Basis CB x1 x2 x3 s1 s2 s3 A1 A2 bj (bj/aij  0)
A1 -1 1 -1 1 -1 0 0 1 0 4 4/1 = 4 
s2 0 1 1 2 0 1 0 0 0 8 8/1 = 8
A2 -1 0 1 -1 0 0 -1 0 1 2 ---
(Maxi.)j = Cj-Zj 1 0 0 -1 0 -1 0 0 Z*max = -6
x1 0 1 -1 1 -1 0 0 1 0 4 ---
s2 0 0 2 1 1 1 0 -1 0 4 4/2 = 2
A2 -1 0 1 -1 0 0 -1 0 1 2 2/1 = 2 
j = Cj-Zj 0 1 -1 0 -1 -1 -1 0 Z*max = -2
x1 0 1 0 0 -1 0 -1 1 1 6
s2 0 0 0 3 1 1 2 -1 -2 0
x2 0 0 1 -1 0 0 -1 0 1 2
j = Cj-Zj 0 0 0 0 0 0 -1 -1 Z*max = 0
Since all j  0. Here Max. Z* = 0 and no artificial variables appears in the optimal basis.
In this case, we proceed tophase -II for finding the optimal basic feasible solution to the
original problem.
Phase II : Taking Maximize Z = -x1-2x2-3x3 and the final simplex table of phase -I is
taken as the initial simplex table of phase –II by removing columns A1 and A2.
Cj  -1 -2 -3 0 0 0 Ratio(Mini.)
Basis CB x1 x2 x3 s1 s2 s3 bj (bj/aij  0)
x1 -1 1 0 0 -1 0 -1 6
s2 0 0 0 3 1 1 2 0
25
 x2 -2 0 1 -1 0 0 -1 2
(Maxi.)j = Cj-Zj 0 0 -5 -1 0 -3 Zmax =

Since all j  0 for all non-basic variables, the optimal basic feasible solution to the given
LPP is, x1 = 6, x2 = 2, x3 = 0 and Zmax = -1(6)-2(2)-3(0) = -10 .
Exercise:
(1) Minimize Z = -2x-y [Ans: x=4, y=0, Z=-8]
Subject to, x+y  2 ; x+y  4 , x, y 0.
(2) Maximize Z = 2x1+2x2 [Ans: Unbounded solution]
Subject to 2x1 + 4x2  1, x1 + 2x2  1, 2x1 + x2  1 , x1, x2  0.
(3) Minimize Z = -3x1+ x2 [Ans: x1=4, x2 =0, Z=-12]
Subject to x2  4; x1 +3x2  4; 2x1+x2  2 , x1, x2  0.
(4) Maximize Z=x-2y-3z [Ans: Solution does not exist]
Subject to -2x-3y-4z = -1; -2x+y+3z = 2, x, y, z  0.
(5) Maximize Z = 3x1+8x2 [Ans: x1=0, x2=5, Z=40]
Subject to 3x1 + 2x2  3; x1 +4x2  4; x1+ x2  5, x1, x2  0.
(6) Minimize Z = 4x1+x2 [Ans: x1=0, x2=5, Z=5]
Subject to 3x1 + 4x2  20; x1 +5x2  15, x1, x2  0.
(7) Maximize Z = 6x1+4x2 [Ans: x1=8, x2=0, Z=48]
Subject to 2x1 + 3x2  30; 3x1 + 2x2  24; x1+ x2  3, x1, x2  0.
(8) Minimize Z = x1 +x2 +x3 [Ans: x1=0, x2=21/5, x3=22/5, Z=43/5]
Subject to x1 -3x2+4x3 = 5; x2-2x3  3; 2x2-x3  4 , x1 , x2, x3  0.
(9) Maximize Z = -4x1 -3x2 -9x3 [Ans: x1=61/3, x2=0, x3=0, Z=-81]
Subject to , 2x1 +4x2+6x3  15; 6x1+x2+6x3  12 , x1 , x2, x3  0.
(10) Minimize Z = 3x1-x2 [Ans: x1=4/5, x2=2/5, Z=2]
Subject to 2x1 + x2  2; x1 + 3x2  2 , x2  4 ; x1, x2  0.
[2] Big - M method: (Penalty Method)
Step 1. Express the problem is standard form.
Step 2. Add non-negative variables to the left hand side of all those constraints which are
of () or (=) type. Such new variables are called artificial variables. But violation of the
corresponding constraints, we would like to get rid of this variable and not allow them to
appear in the final solution. For this, we assign a very large penalty (-M) to these artificial
variables in the objective function.
Step 3. Solve the modified L.P.P. by simplex method. On way, one of the following three
cases may arise:
(1) There remain no artificial variables in the basis and the optimality condition is
satisfied. Then the solution is an optimal basic feasible solution to the problem.
(2) There is at least one artificial variable in the basis at zero level (with zero value in

26
bj -column) and the optimality condition is satisfied. Then the solution is a degenerate
optimal basic feasible solution.
(3) There is at least one artificial variable in the basis at non-zero level (with positive
value in bj -column) and the optimality condition is satisfied. Then the problem has
no feasible solution.
 Solve the following examples using Big-M method:
Ex:1 Minimize Z = 5x1+3x2
Subject to 2x1+4x2  12, 2x1+2x2 = 10, 5x1+2x2  10, x1, x2  0.
n
Sol : Adding slack variable s1, surplus variable s2, artificial variables A1, A2 in the given
constraints, the standard form of LPP is,
Maxi. Z = -5x1-3x2 +0s1+0s2-MA1-MA2
Subject to, 2x1+4x2 +s1 = 12
2x1+2x2 +A1 = 10
5x1+2x2 –s2+A2 = 10; x1, x2 , s1, s2 , A1, A2  0.
Taking x1=x2=s2=0, the IBF solution is s1 = 12, A1 = 10, A2 = 10 & Zmax = -20M. Then the
optimum solution is shown in the following table.
Cj  -5 -3 0 0 -M -M Ratio
XB CB x1 x2 s1 s2 A1 A2 bj (bj/aij  0)
s1 0 2 4 1 0 0 0 12 12/2 = 6
A1 -M 2 2 0 0 1 0 10 10/2 = 5
A2 -M 5 2 0 -1 0 1 10 10/5 = 2 
j = Cj-Zj -5+7M -3+4M 0 -M 0 0
s1 0 0 16/5 1 2/5 0 -2/5 8 5/2 
A1 -M 0 6/5 0 2/5 1 -2/5 6 5
x1 -5 1 2/5 0 -1/5 0 1/5 2 5
j = Cj-Zj -1+ 1-
0 0 -1+(2/5)M 0
(6/5)M (7/5)M

x2 -3 0 1 5/16 1/8 0 -1/8 5/2 20

A1 -M 0 0 -3/8 1/4 1 -1/4 3 12 
x1 -5 1 0 -1/8 -1/4 0 1/4 1 ---
j = Cj-Zj 5/16 -7/8 + 13/8 +
0 0 0
-(3/8)M (1/4)M (-5/4)M
x2 -3 0 1 1/2 0 -1/2 0 1
s2 0 0 0 -3/2 1 4 -1 12
x1 -5 1 0 -1/2 0 1 0 4
j = Cj-Zj 0 0 -1 0 7/2 - M -M
Since all j  0, the optimum solution to given LPP is, x1=4, x2=1, s1=0, s2=12
and zmini = 23
27
Ex:2 Maximize Z = x1+2x2 +3x3-x4
Subject to x1+2x2 +3x3=15, 2x1+x2 +5x3=20 , x1+2x2 +x3+x4=10,
x1, x2 , x3, x4  0.
n
Sol : Since all constraints in the given problem are equations, so we add only artificial
variables A1, A2, A3 in the given constraints, the standard form of LPP is,
Maxi. Z = x1+2x2 +3x3-x4-MA1-MA2-MA3
Subject to, x1+2x2 +3x3+A1 = 15
2x1+x2 +5x3+A2 = 20
x1+2x2 +x3+x4+A3 = 10; xi , A1, A2  0, i = 1,2,3,4.
Taking x1=x2=x3=0, the IBF solution is A1 = 15, A2 = 20, A3 = 10 & Zmax = -45M. Then
the optimum solution is shown in the following table.
Cj  1 2 3 -1 -M -M -M Ratio
XB CB x1 x2 x3 x4 A1 A2 A3 bj (bj/aij )
A1 -M 1 2 3 0 1 0 0 15 5
A2 -M 2 1 5 0 0 1 0 20 4
A3 -M 1 2 1 1 0 0 1 10 10
j 3+9M 0
1+4M 2+5M -1+M 0 0

A1 -M -1/5 7/5 0 0 1 -3/5 0 3 15/7 
x3 3 2/5 1/5 1 0 0 1/5 0 4 20
A3 -M 3/5 9/5 0 1 0 -1/5 1 6 10/3
j -1/5+ 7/5+ -3/5- 0
0 -1+M 0
(2/5)M (16/5)M (9/5)M
x2 2 -1/7 1 0 0 5/7 -3/7 0 15/7 ---
x3 3 3/7 0 1 0 -1/7 2/7 0 25/7 ---
A3 -M 6/7 0 0 1 -9/7 4/7 1 15/7 15/7 
j -1- 0
(6/7)M 0 0 -1+M (-3/7)M
(16/7)M
x2 2 -1/7 1 0 0 5/7 -3/7 0 15/7 ---
x3 3 3/7 0 1 0 -1/7 2/7 0 25/7 25/3
x4 -1 6/7 0 0 1 -9/7 4/7 1 15/7 5/2 
j 6/7  0 0 0 -16/7-M 4/7-M 1-M
x2 2 0 1 0 1/6 1/2 -1/3 1/6 5/2
x3 3 0 0 1 -1/2 1/2 0 -1/2 5/2
x1 1 1 0 0 7/6 -3/2 2/3 7/6 5/2
j 0 0 0 -1 -1-M -M -M

Since all j  0, the optimum solution to given LPP is, x1=x2=x3=5/2, x4=0 & Zmax=15

28
Exercise:
(1) Maximize Z = 3x1-x2 , [Ans. x1=3, x2=0,zmax= 9]
Subject to 2x1 + x2  2; x1 +3x2  3, x2  4, x1, x2  0.
(2) Maximize Z = 6x1+4x2 [Ans: x1=12/5 x2=42/5, Max Z=48]
Subject to 2x1 + 3x2  30; 3x1 +2x2  24, x1+x2  3; x1, x2  0.
(3) Maximize Z = 2x1+x2 +3x3 [Ans: x1=2 x2=1, x3=0 Min Z=8]
Subject to x1 +x2+2x3 5; 2x1+3x2 +4x3 = 12 ; x1, x2 ,x3  0.
(4) Maximize Z = 3x1+2x2 [Ans: Infeasible solution]
Subject to, 2x1 +x2  2; 3x1+ 4x2  12, x1, x2  0.
(5) Maximize Z = 3x1+5x2 [Ans: Unbounded solution]
Subject to x1-2x2  6, x1  10, x2  1, x1, x2  0.
(6) Maximize Z = 3x1+2x2 +x3 [Ans: x1=8/3 x2=0, x3=14/3 ,Zmax=47/3]
Subject to 2x1+5x2+x3=12; 3x1+4x2 =11 ; x1 unrestricted, x2 ,x3  0.

MSPMSPMSPMSPMSP

29
B.Sc. SEM : V CCMATH - 504C Operations Research
Unit-3 Dr. M.S.Prajapati
 Duality Concept: One of the most interesting concepts in linear programming is the
duality theory. Every L.P.P has associated with it, which is another linear programming
problem involving the same data and closely related optimal solutions. Such two problems
are called the primal and dual respectively.
Formulation of dual problem: Consider the following LPP :
Maximize Z = c1x1 + c2x2 + … … + cnxn
Subject to the constraints
a11x1 + a12x2 + …… + a1nxn  b1
a21x1 + a22x2 +……..+a2nxn  b2
…..…………………………...
am1x1 + am2x2 + …… + amnxn  bm
x1,x2,… …,xn  0.
To construct the dual problem, we have the follow the steps given below :
(1) The maximization problem in the primal becomes the minimization problem in the
dual and vice versa.
(2) () type of constraints in the primal becomes () type of constraints in the dual and
vice versa.
(3) The coefficients c1,c2, … …, cn in the objective function of the primal becomes b1,b2,
… …,bn in the objective function of the dual.
(4) The constants b1,b2,… … ,bn in the constraints of the primal becomes c1,c2,……,cn in
the constraints of the dual.
(5) If the primal has n variables and m constraints, the dual will have m variables and n
constraints.
(6) The variables in both the primal and dual are non-negative.
Then the dual problem will be, Minimize W=b1y1+b2y2+ … … +bmym ,
Subject to the constraints, a11y1 + a21y2 +… …+ am1ym  c1
a12y1 + a22y2 +… …+ am2ym  c2
… … … … … … … … … … ...
a1ny1 + a2nyn +… …+ amnym  cn,
y1,y2,…. ….ym  0.
e.g. write the dual of the following LPP.
Minimize Z = 3x1-2x2+4x3
Subject to, 3x1+5x2+4x3  7
6x1+x2+3x3  4
7x1-2x2-x3  10 i.e. -7x1+2x2+x3  -10
x1-2x2+5x3  3,
4x1+7x2-2x3  2 , x1, x2, x3  0.
30
 Let y1,y2,y3,y4 and y5 be the dual variables associated with the above five constraints.
Then the dual problem is given by
Maximize W = 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, yi  0, i=1,2,..,5.
 Formation of dual problem when the primal has equality constraints:
Maximize Z = c1x1 + c2x2
Subject to a11x1 + a12x2 = b1
a21x1 + a22x2  b2 , x1,x2  0.
The equality constraint can be written as
a11x1 + a12x2  b1 and a11x1 + a12x2  b1
or a11x1 + a12x2  b1 and -a11x1 - a12x2  - b1
Then the above problem can be restated as,
Maximize Z = c1x1 + c2x2
Subject to a11x1 + a12x2  b1
-a11x1 - a12x2  - b1
a21x1 + a22x2  b2 and x1, x2  0.
The dual problem using the dual variables y1',y1'' and y2 is
Minimise W = b1(y1' - y1'') + b2y2
Subject to a11(y1' - y1'') + a21y2  c1
a12(y1' - y1'') + a22y2  c2 , y1' , y1'', y2  0.
Taking y1' - y1'' = y1 becomes unrestricted in sign being the difference of two non-negative
variables.The dual problem is,
Minimise W = b1 y1+ b2y2
Subject to a11y1+ a21y2  c1
a12y1+ a22y2  c2 , y1 is unrestricted in sign, y2  0.
In general, if the primal problem is,
Maximize Z = c1x1 + c2x2 + … … + cnxn
Subject to the constraints a11x1 + a12x2 + …… + a1nxn = b1
a21x1 + a22x2 +……..+a2nxn = b2
… … … … … … … … … ...
am1x1 + am2x2 + …… + amnxn = bm
x1,x2,… …,xn  0.
Then the dual problem is, Minimize W = b1y1 + b2y2 + … … + bmym
Subject to the constraints, a11y1 + a21y2 +… …+ am1ym  c1
a12y1 + a22y2 +… …+ am2ym  c2
… … … … … … … … … … ...
a1ny1 + a2nyn +… …+ amnym  cn
31
 y1,y2,…. ….ym are all unrestricted in sign.
Thus, the dual variables corresponding to equality constraints are unrestricted in sign.
Conversely when the primal variables are unrestricted in sign, the corresponding dual
constraints are equalities.
e.g. Construct the dual of the LPP:
Maximize Z = 4x1+9x2+2x3
Subject to 2x1+3x2+2x3  7, 3x1-2x2 +4x3 = 5, x1,x2,x3  0.
 Let y1 and y2 be the dual variables associated with the first and second constraints.
Then the dual problem is
Minimize W = 7y1+5y2
Subject to 2y1+3y2  4
3y1-2y2  9
2y1+4y2  2 , y1  0, y2 is unrestricted in sign.
Exercise: Write the dual of the following problems:
(1) Maximize Z = 3x1+16x2+7x3
Subject to x1-x2+3x3  3
-3x1+2x2  1
2x1+x2 -x3 = 4 , x1, x2, x3  0.
(2) Maximize Z = 3x1+x2+2x3
Subject to x1+x2+x3  6
3x1-2x2+3x3 = 3
-4x1+3x2 -6x3 = 4 , x1, x2, x3  0.
(3) Minimize Z = 2x1+3x2+4x3
Subject to 2x1+3x2+5x3  2
3x1+x2+7x3 = 3 , x1+4x2 +6x3  5 , x1, x2  0, x3 is unrestricted.
 The fundamental Theorem of Duality Principle : If the primal and the dual problems
have feasible solutions then both have optimal solutions and the optimal value of the
primal objective function is equal to the optimal value of the dual objective function. i.e.
Max.Z = Min.W
 Rules for obtaining an optimal solution to the primal (dual) problem from that of
the dual(primal).
Suppose we have already found an optimal solution to the dual (primal) by simplex
method.
Rule (1). If the primal variable corresponds to slack starting variables in the dual problem,
then its optimal value is directly given by the coefficient of the slack variables with
changed sign, in the Cj row of the optimal dual simplex table and vice-versa.
Rule (2). If the primal variable corresponds to artificial starting variables in the dual
problem, then its optimal value is directly given by the coefficient of the artificial

32
variables, with changed sign, in the Cj row of the optimal dual simplex table, after deleting
the constant M and vice-versa.
If the primal has an unbounded solution, then the dual problem will have no feasible
solution and vice-versa.
 Construct the dual of the following problem & Solve the primal and
dual:
Ex:1 Primal Problem : Maximize Z = 2x1+x2
Subject to -x1+2x2  2
x1+x2  4
x1  3 ; x1, x2  0.
n
Sol : Using Simples method or graphical method, optimal solution to given LPP is x1 = 3,
x2 = 1, Zmax= 7.
Now the Dual of the given LPP is, Mini. Z*= 2y1+4y2+3y3
Subject to, -y1+y2+y3  2
2y1+y2  1 ; y1, y2, y3  0
By introducing surplus and artificial variables, the standard form of LPP is,
Maxi. Z*= -2y1-4y2-3y3+0s1+0s2- MA1-MA2
Subject to, -y1+y2+y3-s1+A1 = 2
2y1+y2 –s2 +A2 = 1 ; y1, y2, y3 ,s1, s2, A1, A2  0
Taking y1= y2= y3 = s1= s2 = 0, the IBF solution is, A1 = 2, A2 = 1 and Mini. Z*= -3M  0.
Solve this LPP by Big-M method.

Cj  -2 -4 -3 0 0 -M -M Ratio(mini.)
Basis CB y1 y2 y3 s1 s2 A1 A2 bj (bj/aij  0)
A1 -M -1 1 1 -1 0 1 0 2 2/1 = 2
A2 -M 2 1 0 0 -1 0 1 1 1/1 = 1
(maxi)j = Cj-Zj -2+M -4+2M -3+M -M -M 0 0 Z*max = -3M
A1 -M -3 0 1 -1 1 1 -1 1 1/1 = 1
y2 -4 2 1 0 0 -1 0 1 1 ---
j = Cj-Zj 6-3M 0 -3+M -M -4+M 0 4-2M Z*max = -4-M
y3 -3 -3 0 1 -1 1 1 -1 1
y2 -4 2 1 0 0 -1 0 1 1
j = Cj-Zj -3 0 0 -3 -1 3-M 1-M Z*max = -7
Since all j  0, the current optimal solution to the dual is, y1=0, y2=1, y3=1, Min. Z*=7.
To derive the optimal basic feasible solution to the primal problem, we note that the
primal variables x1, x2 corresponds to the artificial starting dual variables A1, A2
respectively. In the final simplex table of the dual problem, j corresponding to A1and A2
33
are 3 and 1 respectively after ignoring M. Thus by rule II, we get opti. soln to the primal is
x1 = 3, x2 =1 and Zmax= 7.
Ex:2 Primal Problem : Minimize Z = 0.7x1+0.5x2
Subject to, x1  4
x2  6
x1+2x2  20,
2x1+x2  18 ; x1, x2  0.
[Hint: Write dual of the given LP problem, and solve by simplex method. The optimal
basic feasible solution is, y1 = 0, y2 = 0, y3 = 20, y4 =18 and Wmax= 7.4
We note that the primal variables x1, x2 corresponds to the slack starting dual variables
s1, s2 respectively. In final simplex table of the dual problem, j corresponding to s1and s2
are -16/3 and -22/3 respectively. Thus by rule I, an optimal basic feasible solution to the
given primal is, x1 = 16/3, x2 = 22/3 and Zmin= 7.4]
 Working procedure for dual simplex method :
Step 1.(1) Convert the problem to maximization form, if it is not so.
(2) Convert () type constraints, if any to () type by multiplying such
constraints by -1.
(3) Express the problem in standard form by introducing slack variables.
Step 2. Find the initial basic solution and express this information in the form of dual
simplex table.
Step 3. Test the nature of j = Cj – Zj:
(a) If all j  0 and all bi  0, then optimal basic feasible solution has been attained.
(b) If all j  0 and at least one bi < 0, then go to step 4.
(c) If any j  0, the method fails.
(d) If all j  0 and all bi  0, then apply Simplex method for further procedure.
Step 4. Make the outgoing variable. Select the row that contains the most negative bi.
This will be the key row and the corresponding basic variable is the outgoing
variable.
Step 5. Test the nature of key row elements:
(a) If all these elements are  0, the problem does not have a feasible solution.
(b) If at least one element < 0, find the ratios of the corresponding elements of j-
row to these elements. Choose the smallest of these ratios. The corresponding
column is the key column and the associated variable is the incoming variable.
Step 6. Iterate towards optimal feasible solution. Make the key element unity.
Perform row operations as in the regular simplex method and repeat iterations until
either an optimal feasible solution is attained or there is an indication of non-existence
of a feasible solution.

34
 Using dual simplex method, solve the following LPP.
Ex:1 Minimize Z = 3x1+x2
Subject to x1+x2  1, 2x1+3x2  2 ; x1, x2  0
n
Sol : Here Maximize Z‟ = -3x1-x2
Subject to, -x1-x2  -1, -2x1-3x2  -2 ; x1, x2  0
Introducing slack variables s1, s2  0, the standard form of LPP is,
Maximize Z‟ = -3x1-x2+0s1+0s2
Subject to, -x1-x2 + s1 = -1
-2x1-3x2 + s2 = -2 ; x1, x2 , s1, s2  0
Taking x1, x2 = 0, the initial basic solution s1 = -1, s2 = -2 and maxi Z‟ = 0. The problem
can be solve by Dual Simplex method as under.
Cj  -3 -1 0 0
Basis CB x1 x2 s1 s2 bj
s1 0 -1 -1 1 0 -1
s2 0 -2 -3 0 1 -2
j = Cj-Zj -3 -1 0 0
Ratio(j / aij0) 3/2 1/3 -- --
s1 0 -1/3 0 1 -1/3 -1/3
x2 -1 2/3 1 0 -1/3 2/3
j = Cj-Zj -7/3 0 0 -1/3
Ratio(j / aij0) 7 -- -- 1
s2 0 1 0 -3 1 1
x2 -1 1 1 -1 0 1
j = Cj-Zj -2 0 -1 0

Here all j  0 and all bi  0, the optimum basic feasible solution is, x1= 0, x2=1& Zmin= 1.
Ex:2 Maximize Z = -3x1-2x2
Subject to x1+x2  1, x1+x2  7 , x1+2x2  10, x2  3; x1, x2  0
n
Sol : The canonical form of LPP is,
Maximize Z = -3x1-2x2
Subject to -x1-x2  -1, x1+x2  7 , -x1-2x2  -10, x2  3; x1, x2  0
Introducing slack variables s1, s2, s3, s4  0, the standard form of LPP is,
Maximize Z = -3x1-2x2
Subject to the constraints, -x1-x2 +s1 = -1
x1+x2 +s2 = 7
-x1-2x2 +s3 = -10
x2 +s4 = 3 and x1, x2 , s1, s2, s3, s4  0
35
Taking x1, x2 = 0, the initial basic solution s1 = -1, s2 = 7, s3 = -10, s4 = 3 and maxi Z = 0.
The problem can be solve by Dual Simplex method.
Cj  -3 -2 0 0 0 0
Basis CB x1 x2 s1 s2 s3 s4 (Mini)bj
s1 0 -1 -1 1 0 0 0 -1
s2 0 1 1 0 1 0 0 7
s3 0 -1 -2 0 0 1 0 -10
s4 0 0 1 0 0 0 1 3
j = Cj-Zj -3 -2 0 0 0 0
Ratio(j / aij0) 3 1 -- -- -- --
s1 0 -1/2 0 1 0 -1/2 0 4
s2 0 1/2 0 0 1 1/2 0 2
x2 -2 1/2 1 0 0 -1/2 0 5
s4 0 -1/2 0 0 0 1/2 1 -2
j = Cj-Zj -2 0 0 0 1 0
Ratio(j / aij0) 4 -- -- -- -- --
s1 0 0 0 1 0 -1 -1 6
s2 0 0 0 0 1 1 1 0
x2 -2 0 1 0 0 0 1 3
x1 -3 1 0 0 0 -1 -2 4
j = Cj-Zj 0 0 0 0 -3 -4

Since all j  0 and all bi  0, the optimum basic feasible solution is, x1= 4, x2= 3 &
Zmax= -18.
Ex:3 Minimize Z = 2x1+2x2+4x3
Subject to 2x1+3x2+5x3  2
3x1+x2 +7x3  3 , x1+4x2+6x3  5 ; x1, x2, x3  0
n
Sol : The canonical form of LPP is,
Maximize Z = -2x1-2x2-4x3
Subject to -2x1-3x2-5x3  -2
3x1+x2 +7x3  3
x1+4x2+6x3  5 and x1, x2, x3  0
Introducing slack variables s1, s2, s3  0, the standard form of LPP is,
Maximize Z = -2x1-2x2-4x3
Subject to -2x1-3x2-5x3 +s1 = -2
3x1+x2 +7x3 + s2 = 3
36
 x1+4x2+6x3 +s3 = 5 and xi, si  0, i = 1,2,3
Taking x1, x2, x3 = 0, the initial basic solution s1 = -2, s2 = 3, s3 = 5, and maxi Z = 0. Then
the problem can be solve by Dual Simplex method as follows.
Cj  -2 -2 -4 0 0 0
Basis CB x1 x2 x3 s1 s2 s3 (Mini)bj
s1 0 -2 -3 -5 1 0 0 -2
s2 0 3 1 7 0 1 0 3
s3 0 1 4 6 0 0 1 5
j = Cj-Zj -2 -2 -4 0 0 0
Ratio(j / aij0) 1 2/3 4/5 -- -- --
x2 -2 2/3 1 5/3 -1/3 0 0 2/3
s2 0 7/3 0 16/3 1/3 1 0 7/3
s3 0 -5/3 0 -2/3 4/3 0 1 7/3
j = Cj-Zj -2/3 0 -2/3 -2/3 0 0
Since all j  0 and all bi  0, the optimum basic feasible solution is, x1=0, x2=2/3 , x3=0
and Zmin= 4/3
Exercise: Solve the following LPP by Dual simplex method.
Ex-1 Minimize Z = 2x1+x2 [Ans: x1=3/5, x2=6/5 and Zmin=12/5]
Subject to 3x1+x2  3, x1+2x2  3 , 4x1+3x2 6, x2 3 ; x1, x2  0
Ex-2 Minimize Z = x1+2x2+3x3 [Ans: x1=6, x2=2,x3=0 and Zmin=10]
Subject to 2x1-x2+x3  4
x1+x2 +2x3  8, x2-x3  2 ; x1, x2, x3  0.
Ex-3 Max. Z = 2x+y [Use step-3(d)] [Ans: x=3, y=0 and Zmax =6]
Subject to, x+5y  3 , 2x-3y  2 ; x, y  0.
Ex-4 Maximize Z = -2x1 - x3 [Ans: x1= 0, x2=14,x3= 9 and Zmax= -9]
Subject to x1+x2-x3  5
x1-2x2 +4x3  8 ; x1, x2, x3  0.
Ex-5 Minimize Z = 6x1+x2 [Ans: x1= 1, x2=1and Zmin= 7]
Subject to 2x1+x2  3, x1- x2  0, x1, x2  0.
Theorem : Prove that the dual of the dual is the primal.
Proof: Consider the primal problem,
Minimize ZX = CX , subject to, AX  B, X  0 --------- (1)
where A is an mn matrix BTEm and C, XTEn.
Applying the transformation rules, the dual of this problem is,
Maximize ZY = BTY , Subject to, ATY  CT, Y  0 --------- (2)
37
This dual problem can also be written as,
Minimize = (-BT)Y , Subject to, (-AT)Y  (-CT), Y  0 --------- (3) where = -ZY.
If we consider LP problem (3) as primal, then its dual can be constructed by considering X
as the dual variable. Thus, we have
Maximize = (-CT)TX=(-C)X ,
Subject to, (-AT)TX  (-BT)T or (-A)X  (-B) and X  0 --------- (4)
But LP problem (4) is identical to the given primal LP problem (1). This completes the
proof of the theorem.
Theorem: Let X* be any feasible solution to the primal LPP, Maximize ZX = CX , subject
to, AX  B, X  0 and Y* be any feasible solution of the dual LP problem, Minimize ZY =
BTY , subject to, ATY  CT, Y  0, of the above primal problem, then prove that CX* 
BTY*. i.e. Zx  Zy .
Proof: Since X* and Y* are the feasible solutions to the primal and dual LP problems
respectively, then from the constraints in primal and dual, we have
AX*  B , X*  0 and ATY*  CT , Y*  0.
CT  ATY*  C  (Y*)TA.
Multiplying both sides by X*, we get
CX*  (Y*TA) X* = Y*T( AX*)  Y*TB = BTY*.
 CX*  BTY*. This completes the proof of theorem.
Theorem: The value of the objective function f(x) for any feasible solution of the primal
is not less than the value of the objective function (y) for any feasible solution of the
dual.
Proof: Consider the primal LP problem,
Mini f(x) = c1x1+c2x2+… …+cnxn
Subject to, AX  B, X  0.
This can be written after introducing surplus variables,
a11x1+a12x2+… …+a1nxn-xn+1= b1
a21x1+a22x2+… …+a2nxn-xn+2= b2 --------- (1)
……………………………
am1x1+am2x2+… …+amnxn-xn+m= bm,
where x1,x2,… …,xn,xn+1,… …,xn+m  0.
The dual of the above primal is,

38
 Max (y) = b1y1+b2y2+… …+bmym
Subject to, a11y1+a21y2+… …+am1ym+ym+1= c1
a12y1+a22y2+… …+am2ym+ym+2= c2 --------- (2)
……………………………
a1ny1+a2ny2+… …+amnym+ym+n= cn,
where y1,y2,… …,ym,ym+1,… …,ym+n  0.
Let x1,x2,… …,xn,xn+1,… …,xn+m and y1,y2,… …,ym,ym+1,… …,ym+n be any feasible
solution of the primal and its dual respectively.
Multiply (1) by y1,y2,… …,ym respectively and add them. Similarly, multiply (2) by
x1,x2,… …,xn respectively and then subtract (1) from (2), we shall get,
f(x) - (y) = x1ym+1+x2ym+2+… …+xnym+n+y1xn+1+y2xn+2+… …+ymxn+m.
As all the variables on R.H.S are non-negative, we have f(x) - (y)  0.
 f(x)  (y). Hence the theorem.
Theorem: Prove that the primal problem has a finite optimal solution if and only if the
dual problem does too in which case the optimal value of their objective functions are the
same.
Proof: H.W
 Integer Programming Problem: In many practical problems the decision
variables makes sense only if they have integer value. For example if the variables are the
number of buses on different routes in a town, the number of bank branches in different
regions of a country ,…etc, then the fractional answer have no meaning. Mathematical
programming, subject to the constraint that the variables are integers is called integer
programming.
If some of the variables are restricted to be integers while others are real numbers, the
problem is said to be mixed integer programming. The integer programming problem is a
special case of LPP, where all or some variables are constraints to assume non negative
integer values.
A vector XEn is called an integer vector if its component xi, for all i = 1,2,3,….,n
are integers. It will be called a mixed integer vector if xi is integer for i{1,2,…n}.
The standard form of IPP is, Maximize Z = f(x) = CX
Subject to, AX = B , X  0
X is an integer or a mixed integer vector.

39
 Gomory’s cutting plane method for all ILP problem:
Consider the given ILP, Maximize Z = f(x) = CX …….(1)
Subject to AX = B ………(2)
X  0 an integer………(3)
The related LP problem is, Maximize Z = f(x) = CX
Subject to AX = B
X0
Let the basic optimal solution of the related LPP is,
B.V. XB x1 x2 x3 ……… xm xm+1 …….. xn
x1 b1' 1 0 0 ………………… 0 a1,' m+1 ……………. a1,n
'

x2 b2' 1 0 0 …………………0 a2,' m+1 …………… a2,n

'

. . . . . . . .
. . . . . . . .
. . . . . . . .
xm bm' 0 0 0 1 am' ,m+1 ………. am' ,n
That is,
x1  a1,' m 1.xm 1  ........  a1,' n .xn  b1'
x2  a2,' m 1.xm 1  ........  a2,' n .xn  b2'
..... ..... ..... ..... ..... ...... ...... ………. (5)
xm  am' ,m 1.xm 1  ........  am' ,n .xn  bm'
Since the solution is necessarily feasible.  xi = bi‟  0 , i = 1,2,….,m.
If all the bi‟ are integers, then we have solution of the ILP problem.
If not, let a particular bi‟ is non-integer. The corresponding equation in (5) is,
xi  ai',m1.xm1  ........  ai',n .xn  bi' ………. (6)
Let bi‟= [bi‟] + i ……… (7) and a‟ij = [a‟ij] + ij ……... (8) , for j = m+1,….,n.
where [b‟i]  0 and 0  i  1, 0  ij  1.
n

So, equation (6) can be written as, xi  a

j  m 1
'
ij x j  bi'
n

 {[ai' j ]  i j }x j  [bi' ]  i  xi 
n n
 xi 
j  m 1
 [a
j  m 1
'
ij ]x j  
j  m 1
ij x j  [bi' ]  i
n n
 xi   [a
j  m 1
'
ij ]x j  [b ]  i 
i
'

j  m 1
ij xj .

40
This equation being one of the constraints must be satisfied by every feasible solution of
the ILP as well as related LP problem. For an integer feasible solution, the left side should
be an integer and so the right side too should be an integer.
n

Also 0 ij 1 , xij  0 . Hence i  

j  m 1
x is an integer and it is less than or equal to
ij j

i . But 0 j  1, so for integer feasible solution,

n
   i j x j   i ………. (9)
n
i  
j  m 1
ij xj  0 or
j  m 1
But for the optimal solution of the related problem with which xj = 0, j = m+1,….n.
n
i  
j  m 1
ij x j  i  0 ………. (10)

So, we have linear constraints (9) , which is satisfied by all integer solutions of the
problem but cuts out the optimal solution of the LP problem provided it is non-integer.
n

So, we add constraint (9) with equality sign. i.e.

   i j x j  s   i , where s is the
j  m 1

slack variable and solve the modified problem by dual simplex method.
If the optimal solution is integer, we stop. Otherwise we again obtain a cutting plane to cut
out this optimum solution, but not any of the integer feasible solutions. We go doing this,
till we get an integer optimal solution.
 Solve the following integer programming problem using the Gomory’s
cutting plane method:
Ex:1 Maximize Z = 4x1 + 3x2
Subject to the constraints, x1 + 2x2  4
2x1 + x2  6 ; x1, x2  0 and all are integers
n
Sol : Using Simplex method, the optimum solution of the related LPP without integer
requirements is given below.
Introducing slack variables s1, s2  0, the standard form of LPP is,
Maximize Z = 4x1 + 3x2
Subject to the constraints, x1 + 2x2 + s1 = 4
2x1 + x2 + s2 = 6 ; x1, x2 , s1, s2  0
Taking x1= x2= 0, the IBFS is, s1 = 4, s2 = 6 and Zmax = 0.

41
 Cj  4 3 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 bj (bj/aij  0)
s1 0 1 2 1 0 4 4
s2 0 2 1 0 1 6 3
j = Cj-Zj 4 3 0 0 Zmax= 0
s1 0 0 3/2 1 -1/2 1 2/3 
x1 7 1 1/2 0 1/2 3 6
j = Cj-Zj 0 1 0 -2 Zmax=12
x2 3 0 1 2/3 -1/3 2/3
x1 4 1 0 -1/3 2/3 8/3
j = Cj-Zj 0 0 -2/3 -5/3 Zmax= 38/3

Since all j  0, the current solution is optimal but not integer solution. The
fractional part for both x1 and x2 is 2/3. So anyone can be selected. Consider the first row
for constructing a cut. We have
x2 + (2/3)s1+(-1/3)s2 = 2/3  (0+1)x2 + (0+2/3)s1+(-1+2/3)s2 = (0+2/3)
 x2 - s2 = 2/3 - (2/3)s1 - (2/3)s2  0
 - (2/3)s1 - (2/3)s2  -2/3, which is Gomory‟s constraint. Using gomory‟s slack
variable sg1  0 such that - (2/3)s1 - (2/3)s2 + sg1 = -2/3. Insert this constraint at the bottom
of the previous optimal table and then solve the problem by dual simplex method.

Cj  4 3 0 0 0
Basis CB x1 x2 s1 s2 sg1 bj
x2 3 0 1 2/3 -1/3 0 2/3
x1 4 1 0 -1/3 2/3 0 8/3
sg1 0 0 0 -2/3 -2/3 1 -2/3 
j = Cj-Zj 0 0 -2/3 -5/3 0
j /aij  0 -- -- 1 5/2
x2 3 0 1 0 -1 1 0
x1 4 1 0 0 1 -1/2 3
s1 0 0 0 1 1 -3/2 1
j = Cj-Zj 0 0 0 -1 -1
Since all j  0 and all bj  0 and all integers, the solution is optimal.
 an optimum integer solution is, x1= 3, x2= 0 and Zmax= 12.
42
Ex:2 Maximize Z = 2x1 + 2x2
Subject to the constraints 5x1 + 3x2  8
x1 + 2x2  4 ; x1, x2  0 are integers.
n
Sol : Using the simplex method, the optimum solution of the LPP without integer
requirements is as follows.

Cj  2 2 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 bj (bj/aij  0)
x1 2 1 0 5/7 -3/14 4/7
x2 2 0 1 -1/7 5/14 12/ 7 =1+5/7
j = Cj-Zj 0 0 -8/7 -2/7 Zmax=

Since all j  0, the current non-inter optimum solution is, x1=4/7 , x2=12/7 & Zmax= 32/7.
The fraction part for x1 and x2 are 4/7 & 5/7. We select the largest fractional part which
corresponds to the second row for constructing a Gomorian cut. Hence, we have
x2 – (1/7)s1+(5/14)s2 = 12/7
 x2 +(-1+6/7)s1+(5/14)s2 = 1+(5/7)
 5/7 - (6/7)s1 - (5/14)s2 = x2 – s1 + s2  0
 -(6/7)s1 - (5/14)s2  -5/7, which is Gomory‟s cut. We add Gomory‟s slack
variable sg1  0 such that - (6/7)s1 +(7/12)s2 + sg1 = -5/7.
Adding this constraint at the bottom of the previous optimal table, we get
Cj  2 2 0 0 0
Basis CB x1 x2 s1 s2 sg1 bj
x2 2 1 0 5/7 -3/14 0 4/7
x1 2 0 1 -1/7 5/14 0 12/7
sg1 0 0 0 -6/7 -5/14 1 -5/7
j = Cj-Zj 0 0 -8/7 -2/7 0
j /aij  0 -- -- 4/3 4/5 --
x2 2 1 0 61/35 0 -3/5 1
x1 2 0 1 -1 0 1 1
s2 0 0 0 12/5 1 -14/5 2
j = Cj-Zj 0 0 -16/35 0 -4/5

Since all j  0 and all bj  0 and all integers, the current solution is optimal.
 an optimum integer solution is, x1= 1, x2= 1 and Zmax= 4.

43
Ex:3 Maximize Z = x1 + x2
Subject to the constraints, 2x1 + 3x2  16
6x1 + 5x2  30 ; x1, x2  0 and are integers
n
Sol : Using the simplex method, the optimum solution is,
Cj  1 1 0 0 Ratio(Mini.)
Basis CB x1 x2 s1 s2 bj (bj/aij  0)
x2 1 0 1 3/10 -1/10 9/5=1+4/5
x1 1 1 0 -1/4 1/4 7/2=3+1/2
j = Cj-Zj 0 0 -8/7 -2/7 Zmax= 53/10

Since x1, x2 is not the integer-valued. A fractional cut is formatted with the first row as,
x2 + (3/10)s1 + (-1/10)s2 = 9/5
 x2 + (0+3/10)s1 + (-1+9/10)s2 = 1+4/5
 4/5 – (3/10)s1 – (9/10)s2 = x2 – s2 -1 0
 – (3/10)s1 – (9/10)s2  -4/5, which is Gomory‟s constraint.
Using Gomorian slack variable sg1  0, we have – (3/10)s1 – (9/10)s2 + sg1 = -4/5.
Add this constraint at the bottom of the previous optimal table and solve by dual simplex
method.
Cj  1 1 0 0 0
Basis CB x1 x2 s1 s2 sg1 bj
x2 1 0 1 3/10 -1/10 0 9/5
x1 1 1 0 -1/4 1/4 0 7/2
sg1 0 0 0 -3/10 -9/10 1 -4/5
j = Cj-Zj 0 0 -1/20 -3/20 0
j /aij  0 -- -- 1/3 1/6
x2 1 0 1 1/3 0 -1/9 17/9
x1 1 1 0 -1/3 0 5/18 59/8
s2 0 0 0 1/3 1 -10/9 8/9
j = Cj-Zj 0 0 0 0 -1/6 Zmax= 31/6
Steel the solution is not integer-valued. Insert the fractional cut from the third row as,
(1/3)s1 + s2 + (-2+8/9)sg1 = 8/9
 8/9 – (1/3)s1- (8/9)sg1 = s2-2sg1 0
 – (1/3)s1- (8/9)sg1  -8/9  – (1/3)s1- (8/9)sg1 + sg2 = -8/9, where sg2  0.
Add this constraints in the previous table, we get

44
 Cj  1 1 0 0 0 0
Basis CB x1 x2 s1 s2 sg1 sg2 bj
x2 1 0 1 1/3 0 -1/9 0 17/9
x1 1 1 0 -1/3 0 5/18 0 59/8
s2 0 0 0 1/3 1 -10/9 0 8/9
sg2 0 0 0 -1/3 0 -8/9 1 -8/9
j = Cj-Zj 0 0 0 0 -1/6 0
j /aij  0 -- -- 0 -- 3/16 --
x2 1 0 1 0 0 -1 1 1
x1 1 1 0 0 0 7/6 -1 25/6
s2 0 0 0 0 1 -2 1 0
s1 0 0 0 1 0 8/3 -3 8/3
j = Cj-Zj 0 0 0 0 -1/6 0 Zmax= 31/6

Since x1 is not integer, we need one more fractional cut from the fourth row,
s1+ (8/3) sg1-3sg2 = 8/3  s1+ (2+2/3) sg1-3sg2 = 2+2/3
2/3 – 2/3 sg1 = s1+ 2sg1-3sg2  0  – 2/3 sg1  -2/3
 – 2/3 sg1 + sg3 = -2/3, where sg3  0.
Add this constraint at the bottom of the previous table and solve by dual simplex method.
Cj  1 1 0 0 0 0 0
Basis CB x1 x2 s1 s2 sg1 sg2 sg3 bj
x2 1 0 1 0 0 -1 1 0 1
x1 1 1 0 0 0 7/6 -1 0 25/6
s2 0 0 0 0 1 -2 1 0 0
s1 0 0 0 1 0 8/3 -3 0 8/3
sg3 0 0 0 0 0 -2/3 0 1 -2/3
j = Cj-Zj 0 0 0 0 -1/6 0 0 Zmax= 31/6
j /aij  0 -- -- -- -- 1/4 -- --
x2 1 0 1 0 0 0 1 -3/2 2
x1 1 1 0 0 0 0 -1 7/4 3
s2 0 0 0 0 1 0 1 -3 2
s1 0 0 0 1 0 0 -3 4 0
sg1 0 0 0 0 0 1 0 0 1

45
 j = Cj-Zj 0 0 0 0 0 0 -1/4 Zmax= 5

Since all j  0 and all bj  0 and all integers, the current solution is optimal.
 an optimum integer solution is, x1= 3, x2= 2 and Zmax= 5.
Exercise:
(1) Maximize Z = 2x1 + 20x2 – 10x3 [Ans : x1=2, x2=0, x3=2 & MaxZ = -16]
Subject to the constraints 2x1 + 20x2 + 4x3  15
6x1 + 20x2 +4x3 = 20,
x1, x2, x3  0 and all are integers.
(2) Maximize Z = x1 + 2x2 [Ans : x1=4,x2=3, MaxZ=10]
Subject to the constraints, 2x2  7
x1 + x2  7
2x1  11; x1, x2  0 and all are integers.
(3) Maximize Z = x1 + x2 [Ans : x1=0, x2=2, MaxZ=2]
Subject to the constraints 3x1 + 2x2  5
x2  2 ; x1, x2  0 and all are integers.
(4) Maximize Z = 3x1 + 4x2 [Ans : x1=0, x2=4, MaxZ=16]
Subject to the constraints 3x1 + 2x2  8
x1 + 4x2  10 ; x1, x2  0 and all are integers.
(5) Maximize Z = 3x1 + 12x2
Subject to the constraints 2x1 + 4x2  7
5x1 + 3x2  15 ; x1, x2  0 and all are integers.
(6) Maximize Z = x + 4y
Subject to the constraints 2x+ 4y  7
5x+3y  15 ; x1, x2  0 and all are integers.
(7) Maximize Z = 5x1 + 5x2
Subject to the constraints 2x1 + 5x2  16
6x1 + 5x2 = 30 ; x1, x2  0 and all are integers.

MSPMSPMSPMSPMSP

46