173

Chapter
Linear Programming 173

CONTENTS

6.1 Linear inequations

6.2 Terms of Linear Programming

6.3 Mathematical formulation of a linear

programming problem

6.4 Graphical solution of two variable linear G.B. Dantzing

programming problem
The linear programming originated during World
6.5 To find the vertices of simple feasible region War II, when the British and American military
management called upon a group of Scientists to
without drawing a graph study and plan the war activities, so that maximum
damages could be inflicted on the enemy camps at
6.6 Problems having infeasible solutions minimum cost and loss.

6.7 Some important points about L.P.P. Because of the success in military operations, it
quickly spread in all phases of industry and
6.8 Advantages and Limitations of L.P.P government organizations. The Russian
mathematician L.Kantorovich has, for a number of
years, been interested in the applications of
Assignment (Basic and Advance Level) mathematics to programming problems. In 1941,
F.L. Hitchcock formulated and solved the
Answer Sheet of Assignment transportation problem. In 1947 T.C. Koopmans
solved the transportation problem. The minimum
cost diet problem was studied by economist Stigler
in 1945. In 1947, G.B. Dantzig suggested an
efficient method known as the simplex method,
which is an iterative procedure to solve any linear
programming problem in a finite number of steps.
 174 Linear Programming
174

‘Linear Programming’ is a scientific tool to handle optimization problems. Here, we shall learn about some
basic concepts of linear programming problems in two variables, their applications, advantages, limitations,
formulation and graphical method of solution.

6.1 Linear Inequations .

(1) Graph of linear inequations
(i) Linear inequation in one variable : ax  b  0, ax  b  0, cy  d  0 etc. are called linear inequations in
one variable. Graph of these inequations can be drawn as follows :

Y
b
x
a
X X
O
ax+b<0 ax+b>0

Y

The graph of ax  b  0 and ax  b  0 are obtained by dividing xy-plane in two semi-planes by the line
b
x   (which is parallel to y-axis). Similarly for cy  d  0 and cy  d  0 .
a
Y
d
cy+d>0 y
c

X X
cy+d<0

Y
 Linear Programming 175

(ii) Linear Inequation in two variables : General form of these inequations are ax  by  c, ax  by  c . If any
ordered pair x 1 , y 1  satisfies some inequations, then it is said to be a solution of the inequations.

The graph of these inequations is given below (for c > 0) :

Y

ax+by=c
ax+by<c
X X
O
ax+by>c

Y

Working rule
To draw the graph of an inequation, following procedure is followed :

(i) Write the equation ax  by  c in place of ax  by  c and ax  by  c .

(ii) Make a table for the solutions of ax  by  c .

(iii) Now draw a line with the help of these points. This is the graph of the line ax  by  c .

(iv) If the inequation is > or <, then the points lying on this line is not considered and line is drawn dotted
or discontinuous.

(v) If the ineuqation is > or <, then the points lying on the line is considered and line is drawn bold or
continuous.

(vi) This line divides the plane XOY in two region.

To Find the region that satisfies the inequation, we apply the following rules:

(a) Take an arbitrary point which will be in either region.

(b) If it satisfies the given inequation, then the required region will be the region in which the arbitrary point is
located.

(c) If it does not satisfy the inequation, then the other region is the required region.

(d) Draw the lines in the required region or make it shaded.

(2) Simultaneous linear inequations in two variables : Since the solution set of a system of
simultaneous linear inequations is the set of all points in two dimensional space which satisfy all the inequations
simultaneously. Therefore to find the solution set we find the region of the plane common to all the portions
 176 Linear Programming

comprising the solution set of given inequations. In case there is no region common to all the solutions of the
given inequations, we say that the solution set is void or empty.
(3) Feasible region : The limited (bounded) region of the graph made by two inequations is called
feasible region. All the points in feasible region constitute the solution of a system of inequations. The feasible
solution of a L.P.P. belongs to only first quadrant. If feasible region is empty then there is no solution for the
problem.

Example: 1 Inequations 3 x  y  3 and 4 x  y  4

(a) Have solution for positive x and y (b) Have no solution for positive x and y

(c) Have solution for all x (d) Have solution for all y

Solution : (a) Following figure will be obtained on drawing the graphs of given inequations :
x y
From 3 x  y  3,  1 Y
1 3
x y
From 4 x  y  4 ,  1 X
1 4 O (1,0)

Clearly the common region of both the inequations is true for positive
(0,–3)
value of (x, y). It is also true for positive values of x and negative values of
(0,–4)
y.
Example: 2 The constraints [MP PET 1999]

– x1  x 2   1 ; – x1  3 x 2  9 ; x , x 2  0 define

(a) Bounded feasible space (b) Unbounded feasible space

(c) Both bounded and unbounded feasible space (d) None of these

Solution : (b)
x2
(3,4)

(0,3)

(–9,0) (0,1)
x1
O (–1,0)

It is clear from the graph, the constraints define an unbounded feasible space.

Example: 3 Shaded region is represented by [MP PET 1997]

Y
(0,20) x+y=20
C(0,16)  20 40 
B , 
 3 3 

y 2x+5y=80
Shaded
region
X
O x A(20,0) (40,0)
 Linear Programming 177

(a) 2 x  5 y  80 , x  y  20 , x  0, y  0 (b) 2 x  5 y  80 , x  y  20 , x  0, y  0

(c) 2 x  5 y  80 , x  y  20 , x  0, y  0 (d) 2 x  5 y  80 , x  y  20 , x  0, y  0

Solution: (c) In all the given equations, the origin is present in shaded area. Answer (c) satisfy this condition.

Example: 4 Shaded region is represented by

Shaded 4x–2y=–3

region A(0,3/2)
y

O x
 3 
B ,0 
 4 

(a) 4 x  2 y  3 (b) 4 x  2 y  3 (c) 4 x  2y  3 (d) 4 x  2 y  3

Solution: (b) Origin is not present in given shaded area. So 4 x  2 y  3 satisfy this condition.

6.2 Terms of Linear Programming .

The term programming means planning and refers to a process of determining a particular program.

(1) Objective function : The linear function which is to be optimized (maximized or minimized) is called
objective function of the L.P.P.

(2) Constraints or Restrictions : The conditions of the problem expressed as simultaneous equations or
inequalities are called constraints or restrictions.
 178 Linear Programming

(3) Non-negative Constraints : Variables applied in the objective function of a linear programming problem
are always non-negative. The ineqaulities which represent such constraints are called non-negative constraints.

(4) Basic variables : The m variables associated with columns of the m × n non-singular matrix which
may be different from zero, are called basic variables.

(5) Basic solution : A solution in which the vectors associated to m variables are linear and the remaining
(n  m) variables are zero, is called a basic solution. A basic solution is called a degenerate basic solution, if at
least one of the basic variables is zero and basic solution is called non-degenerate, if none of the basic variables
is zero.

(6) Feasible solution : The set of values of the variables which satisfies the set of constraints of linear
programming problem (L.P.P) is called a feasible solution of the L.P.P.

(7) Optimal solution : A feasible solution for which the objective function is minimum or maximum is
called optimal solution.

(8) Iso-profit line : The line drawn in geometrical area of feasible region of L.P.P. for which the objective
function remains constant at all the points lying on the line, is called iso-profit line.

If the objective function is to be minimized then these lines are called iso-cost lines.

(9) Convex set : In linear programming problems feasible solution is generally a polygon in first quadrant
. This polygon is convex. It means if two points of polygon are connecting by a line, then the line must be inside
to polygon. For example, A A

B B

(i) (ii) (iii) (iv)

Figure (i) and (ii) are convex set while (iii) and (iv) are not convex set

6.3 Mathemiatical formulation of a Linear programming problem .

There are mainly four steps in the mathematical formulation of a linear programming problem, as

mathematical model. We will discuss formulation of those problems which involve only two variables.

(1) Identify the decision variables and assign symbols x and y to them. These decision variables are those

quantities whose values we wish to determine.

 Linear Programming 179

(2) Identify the set of constraints and express them as linear equations/inequations in terms of the decision

variables. These constraints are the given conditions.

(3) Identify the objective function and express it as a linear function of decision variables. It might take the

form of maximizing profit or production or minimizing cost.

(4) Add the non-negativity restrictions on the decision variables, as in the physical problems, negative

values of decision variables have no valid interpretation.

6.4 Graphical solution of two variable Linear programming problem .

There are two techniques of solving an L.P.P. by graphical method. These are :
(1) Corner point method, and (2) Iso-profit or Iso-cost method.
(1) Corner point method
Working rule
(i) Formulate mathematically the L.P.P.
(ii) Draw graph for every constraint.
(iii) Find the feasible solution region.
(iv) Find the coordinates of the vertices of feasible solution region.
(v) Calculate the value of objective function at these vertices.
(vi) Optimal value (minimum or maximum) is the required solution.
Note :  If there is no possibility to determine the point at which the suitable solution found, then the
solution of problem is unbounded.
 If feasible region is empty, then there is no solution for the problem.
 Nearer to the origin, the objective function is minimum and that of further from the origin, the
objective function is maximum.
(2) Iso-profit or Iso-cost method : Various steps of the method are as follows :
(i) Find the feasible region of the L.P.P.
(ii) Assign a constant value Z1 to Z and draw the corresponding line of the objective function.
(iii) Assign another value Z2 to Z and draw the corresponding line of the objective function.
(iv) If Z1  Z 2 , (Z1  Z 2 ) , then in case of maximization (minimization) move the line P1Q1 corresponding to
Z1 to the line P2Q2 corresponding to Z2 parallel to itself as far as possible, until the farthest point within the feasible
region is touched by this line. The coordinates of the point give maximum (minimum) value of the objective function.
Note : The problem with more equations/inequations can be handled easily by this method.
 180 Linear Programming

 In case of unbounded region, it either finds an optimal solution or declares an unbounded

solution. Unbounded solutions are not considered optimal solution. In real world problems,
unlimited profit or loss is not possible.
6.5 To find the Vertices of Simple feasible region without Drawing a Graph .
(1) Bounded region: The region surrounded by the inequalities ax  by  m and cx  dy  n in first
quadrant is called bounded region. It is of the form of triangle or quadrilateral. Change these inequalities into
equation, then by putting x  0 and y  0, we get the solution also by solving the equation in which there may
be the vertices of bounded region.
The maximum value of objective function lies at one vertex in limited region.
(2) Unbounded region : The region surrounded by the inequations ax  by  m and cx  dy  n in first
quadrant, is called unbounded region.
Change the inequation in equations and solve for x  0 and y  0 . Thus we get the vertices of feasible region.
The minimum value of objective function lies at one vertex in unbounded region but there is no existence
of maximum value.

6.6 Problems having Infeasible Solutions.

In some of the linear programming problems, constraints are
inconsistent i.e. there does not exist any point which satisfies all the
constraints. Such type of linear programming problems are said to have 10
9
8
infeasible solution. 7
6
5
4
3
For Example : Maximize Z  5 x  2 y
2
1
subject to the constraints
1 2 3 4 5 6 7 8 9 10
x  y  2 , 3 x  3y  2 , x, y  0

The above problem is illustrated graphically in the fig.

There is no point satisfying the set of above constraints. Thus, the problem is having an infeasible solution.

6.7 Some important points about L.P.P..

(1) If the constraints in a linear programming problem are changed, the problem is to be re-evaluated.
(2) The optimal value of the objective function is attained at the point, given by corner points of the feasible
region.
 Linear Programming 181

(3) If a L.P.P. admits two optimal solutions, it has an infinite number of optimal solutions.
(4) If there is no possibility to determine the point at which the suitable solution can be found, then the
solution of problem is unbounded.
(5) The maximum value of objective function lies at one vertex in limited region.
Example: 5 A firm makes pants and shirt. A shirt takes 2 hour on machine and 3 hour of man labour while a pant takes 3 hour on
machine and 2 hour of man labour. In a week there are 70 hour of machine and 75 hour of man labour available. If the
firm determines to make x shirts and y pants per week, then for this the linear constraints are
(a) x  0, y  0,2 x  3 y  70 ,3 x  2 y  75 (b) x  0, y  0,2 x  3 y  70 ,3 x  2 y  75

(c) x  0, y  0,2 x  3 y  70 ,3 x  2 y   75 (d) x  0, y  0,2 x  3 y  70 ,3 x  2 y  75

Solution: (d)

Working time on Man labour

machine

Shirt (x) 2 hours 3 hours

Pant (y) 3 hours 2 hours

Availability 70 hours 75 hours

Linear constraints are 2 x  3 y  70 ,3 x  2 y  75 and x , y  0

Example: 6 The minimum value of the objective function Z  2 x  10 y for linear constraints

x  y  0, x  5 y  5 and x , y  0 is

(a) 10 (b) 15 (c) 12 (d) 8

5 5 Y
Solution: (b) Required region is unbounded whose vertex is  , 
4 4
x – y=0

x – 5y = –5

O X
5 5
Hence the minimum value of objective function is  2   10   15 .
4 4
n m
Example: 7 Minimize z   c x
j1 i1
ij ij [MP PET 1999]

n m
Subject to :  j1
x ij  ai , i  1 ........, m ; x
i 1
ij  b j , j  1 ......, n is a LPP with number of constraints

m
(a) m  n (b) m  n (c) mn (d)
n
Solution: (a) (I) Condition, i  1, x 11  x 12  x 13  .......... .  x 1n  a1
i  2, x 21  x 22  x 23  .......... .  x 2n  a 2
 182 Linear Programming

i  3, x 31  x 32  x 33  .......... ..  x 3n  a3
...................................
i  m, x m1  x m 2  x m 3  .......... .  x mn  am  m constraint s
(II) Condition j  1, x 11  x 21  x 31  ..........  x m1  b1
j  2, x 12  x 22  x 32  .......... .  x m 2  b 2
....................................
j  n, x 1n  x 2n  x 3n  .......... ....  x mn  bn  n constraints
 Total constraints  m  n .
Example: 8 Maximize z  3 x  2y, subject to x  y  1, y  5 x  0, x  y  1, x  y  6, x  3 and x , y  0
(a) x  3 (b) y3 (c) z  15 (d) All the above
Solution: (d) The shaded region represents the bounded region (3,3) satisfies, so x  3, y  3 and z  15 .

x – y = –1
y – 5x=0

(5/2, 7/2)
(3,3)

x+y=6
x=3

(0,1)

(1,0)

Example: 9 Maximum value of 4 x  5 y subject to the constraints x  y  20 , x  2 y  35 , x  3 y  12 is [Kurukshetra CEE 1996]

(a) 84 (b) 95 (c) 100 (d) 96

Solution: (b) Obviously, max. 4 x  5 y  95 . It is at (5, 15).

(0,20) x – 3y = 12

 1 (5,15) (18,2)
 0,17 
 2
(35,0) x + 2y = 35
(12,0) (20,0)
x + y = 20

Example: 10 The maximum value of Z  4 x  3 y subjected to the constraints 3 x  2 y  160 , 5 x  2 y  200 , x  2 y  80 , x , y  0 is

[MP PET 1998]

(a) 320 (b) 300 (c) 230 (d) None of these

Solution: (d) Obviously, it is unbounded. Therefore its maximum value does not exist.

3x+2y=160

x+2y=80
5x+2y=200
 Linear Programming 183

Example: 11 The objective function Z  4 x  3 y can be maximized subjected to the constraints

3 x  4 y  24 ,8 x  6 y  48 , x  5, y  6; x , y  0
(a) At only one point (b) At two points only (c) At an infinite number of points (d) None of these
Solution: (c) Obviously, the optimal solution is found on the line which is parallel to
y=6
'isoprofit line'.
Hence it has infinite number of solutions.

x=5
4x+3y =0

8x+6y=48 3x+4y =24

Isoprofit
line

Example: 12 The maximum value of µ  3 x  4 y subjected to the conditions x  y  40 , x  2 y  60 ; x , y  0 is [MP PET 2002,04 ]

(a) 130 (b) 120 (c) 40 (d) 140

Solution: (d) Obviously Max µ  3 x  4 y at (20,20 ) Y

µ  60  80  140 .
(0,40)
x+y=40
A(0,30) B(20, 20)
x+2y=60
(60,0)
X
O C (40, 0)

6.8 Advantages and Limitations of L.P.P..

(1) Advantages : Linear programming is used to minimize the cost of production for maximum output. In
short, with the help of linear programming models, a decision maker can most efficiently and effectively employ
his production factor and limited resources to get maximum profit at minimum cost.

(2) Limitations : (i) The linear programming can be applied only when the objective function and all the
constraints can be expressed in terms of linear equations/inequations.
(ii) Linear programming techniques provide solutions only when all the elements related to a problem can be
quantified.
(iii) The coefficients in the objective function and in the constraints must be known with certainty and
should remain unchanged during the period of study.
184 Linear Programming

(iv) Linear programming technique may give fractional valued answer which is not desirable in some problems.