Linear Programming Problem Notes

Akash Bhalekar

August 2023

1 Introduction
Linear Programming is an optimization technique which specifies the use of limited resources or capacities in a
business to attain a particular objective such as least profit (sales) maximization, cost (time) minimization, and so
on, when the resources have alternate uses.
Definition 1. LPP is the combination of a linear objective function with a set of linear constraints such that the
solution to the linear constraints optimizes (maximizes or minimizes) the objective function.
Definition 2. The variables whose values are obtained when LPP is solved are called decision variables. The
decision variables refer to the activities that are completing one another for sharing the resources available. These
variables are usually inter-related in terms of utilisation of resources and need simultaneous solution. All the
variables are considered as continuous, controllable and non-negative.
Definition 3. The linear expression of LPP which has to be optimized (maximized or minimized) is called the
objective function.

Definition 4. Constraints are the restrictions or limitations imposed on the use of resources,suah as labour,
space, raw material, money, etc. that limit the degree to which an objective can be achieved. Such constraints must
be expressed as linear inequations or equalities in terms of decision variables.
Note: Constraints are of three types:
(i) Less than or equal to (≤) (ii) Greater than or equal to (≥) (iii) Equal to (=).
Less than or equal to (≤) constraint is associated with availability of resourses or restriction on the production of a
product. Greater than or equal to (≥) constraint is generally associated with achieving minimum supply conditions
or minimum consumption conditions.
Example.

Maximize z = 3x1 + 2x2

Subject to x1 + x2 ≤ 5
3x1 − 2x2 ≤ 2
x1 , x2 ≥ 0

2 Mathematical Formulation of the LPP

1. Define the decision variables.
2. Identify the optimization criterion.
3. Define the objective function as a linear function of the decision variables.

4. Identify the constraints and define them as equations or inequations.

5. Identify the range and/or bounds on the decision variables and state it clearly.

1
Example. (1) Two foodstuffs F 1 and F 2 contain vitamins A, B, C. The minimum daily requirements of these
vitamins for a certain diet are 3 mg of A, 50 mg of B and 40 mg of C. One unit of the food-stuff F 1 contains
1 mg of A, 25 mg of B and 10 mg of C whereas one unit of the food-stuff F 2 contains 1 mg of A, 10 mg of B
and 20 mg of C. The cost of one unit of food-stuff F 1 is Re 1 and that of F 2 is Rs 2. Formulate the problem as a
linear programming problem.
Sol:

Food-stuff A B C Cost
F1 1 25 10 1
F2 1 10 20 2
Minimum 3 50 40

Let x1 be the units of F 1 and x2 be the units of F 2.

The cost of x1 units of F 1 and x2 units of F 2 is x1 + 2x2 . Since the cost is to be minimized, therefore the LPP is

Minimize z = x1 + 2x2
Subject to x1 + x2 ≥ 3
25x1 + 10x2 ≥ 50
10x1 + 20x2 ≥ 40
x1 , x2 ≥ 0.

(2) A company makes 2 products A and B; both require processing on 2 machines. Product A takes 10 and 15
minutes on the two machines per unit and product B takes 22 and 18 minutes on the two machines per unit. Both
the machines are available for 2640 minutes per week. The products are sold for Rs 200 and Rs 175 respectively per
unit. Formulate a LPP to maximize revenue.
Sol:

Processing time A B Availability

Machine 1 10 22 2640
Machine 2 15 18 2640
Selling price 200 175

Let x1 be the no. of units of product A and x2 be the no. of units of product B.
Since the revenue is to be maximized, therefore the LPP is

Maximize z = 200x1 + 175x2

Subject to 10x1 + 22x2 ≤ 2640
15x1 + 18x2 ≤ 2640
x1 , x2 ≥ 0.

(3) A manufacturer produces two types of toys for children, flutes and drums, each of which must be processed
through two machines A and B. The maximum availability of machine A and B per day are 12 and 18 hours
respectively. The manufacturing of a flutes requires 4 hours in machine A and 3 hours in machine B, whereas a
drum requires 2 hours of machine A and 6 hours of machine B, if the profit for flute is Rs 20 and per drum is
Rs 50, then formulate the problem.
Sol:

Processing time Flute Drum Availability

Machine A 4 2 12
Machine B 3 6 18
Profit 20 50

2
Let x1 be the no. of flutes and x2 be the no. of drums. Since profit is to be maximized, therefore the LPP is

Maximize z = 20x1 + 50x2

Subject to 4x1 + 2x2 ≤ 12
3x1 + 6x2 ≤ 18
x1 , x2 ≥ 0.

Definition 5. A set of values of the decision variables x1 , x2 , . . . , xn which satisfy the conditions of given linear
programming problem is called a solution of LPP.
There are two methods to find the solution of LPP: (1) Graphical Solution Method (2) Simplex Method.

3 Graphical Solution Method

Definition 6. A set of points in a plane is said to be a convex set if line segment joining any two points of the
set entirely lies within the set.

x2

x1
x1 x2

Convex set Non-convex set

Definition 7. The region of the graph where all the constraints including the non-negativity restriction are satisfied
simultaneously is called the feasible region.

Definition 8. A solution which satisfies all the constraints and the non-negativity restriction on the decision
variables is called a feasible solution.
Definition 9. Any feasible solution which optimizes (minimize or maximize) the objective function is called an
optimal solution.

Theorem 10. The set of all feasible soutions of LPP is a convex set.
Theorem 11 (Convex Polygon Theorem). The ovbjective function of LPP attains its optimum value (either max-
imum or minimum) at least at one of the vertices of convex polygon.
The major steps in the solution of a linear programming problem by graphical method are as follows:

1. Identify the problem - the decision variables, the objective and the restrictions.
2. Set up the mathematical formulation of the problem.
3. Plot a graph representing all the constraints of the problem and identify the feasible region. The feasible
region is the intersection of all the regions represented by the constraints of the problem and is restricted to
the first quadrant only.
4. The feasible region may be bounded or unbounded. Compute the coordinate of all the corner points of the
fesible region.
5. Find out the value of the objective function at each corner (solution) points.

6. Select the corner point that optimizes (minimize or maximize) the value of the objective function. It gives
the optimal fesible solution.

3
The above method is known as Serach Approach Method or Corner Point Method.

Note: Another method known as Iso-profit or Iso-cost approach, involves the folowing steps:
(a) Fisrt four steps are same as in the Corner Point Method. In the fifth step we choose a convenient profit (or
cost) and draw iso-profit (or iso-cost) line so that it falls within the feasible region.
(b) Move this iso-profit (or iso-cost) line parallel to itself farther (or closer) from (or to) the origin.
(c) Identify the optimum solution as the coordinates of that point on the feasible region touched by the highest
possible iso-profit line (or lowest possible iso-cost line).
(d) Compute the optimum feasible solution.
Example. (1) A business firm produces 2 types of products P and Q. The average profit for the product is rupees
100 per tonne and that for the product Q is rupees 70 per tonne. The plant consist of 3 production departments
A, BandC. The equipment in each department can be used for 8 hours a day. Product P requires 2 hours in
department A and 1 hour in department C per ton. Product Q requires 1 hour in department B and 1 hour in
department C per ton. Formulate this problem as a linear programming problem and solve it graphicaly to maximize
profit.
Sol:

Department Product P Product Q Availability (in hours)

A 2 0 8
B 0 1 8
C 1 1 8
Profit 100 70

Let x1 be the no. of Product P and x2 be the no. of Product Q. Since profit is to be maximized, therefore the LPP
is

Maximize z = 100x1 + 70x2

Subject to 2x1 ≤ 8
x2 ≤ 8
x1 + x2 ≤ 8
x1 , x2 ≥ 0.

Now, to draw constraint 2x1 ≤ 8. Treat it as 2x1 = 8 ⇒ x1 = 4 and x2 ≤ 8 as x2 = 8.

To draw contraint x1 + x2 ≤ 8. Treat it as x1 + x2 = 8. Then the equation x81 + x82 = 1 is in the double intercept
form. ∴ (8, 0)and (0, 8).

4
 We find intersection point of x1 = 4 and x1 + x2 = 8 by substituting the value of x1 in the equation x1 + x2 = 8.
Hence, we get x2 = 4. ∴ (4, 4). Feasible region: OABC

Vertex Coordinate Max 100x1 + 70x2

O (0, 0) ∴z =0+0=0
A (4, 0) ∴ z = 100(4) + 0 = 400
B (4, 4) ∴ z = 100(4) + 70(4) = 680
C (0, 8) ∴ z = 100(0)70(8) = 560

Hence, the optimum solution is x1 = 4 and x2 = 4 with the objective function value Rs.680.

3.1 Special Cases

1. Alternate optima: When the objective function is parallel to a binding constraint (i.e., a constraint that is
satisfied as an equation by the optimal solution), the objective function will assume the same optimum value
at more than one solution point. Hence they are called alterante optimal solutions.

Example. (1) Use a graphical method to solve the LPP

Maximize z = 2x1 + 4x2

Subject to x1 + 2x2 ≤ 5
x1 + x2 ≤ 4
x1 , x2 ≥ 0

Sol: To draw constraint x1 + 2x2 ≤ 5. Treat it as x1 + 2x2 = 5.

When x1 = 0 we get x2 = 5/2 = 2.5 and when x2 = 0 we get x1 = 5. ∴ (0, 2.5) and (5, 0).
To draw constraint x1 + x2 ≤ 4. Treat it as x1 + x2 = 4.
x1 x2
Then + = 1 is the equation in double intercept form. ∴ (0, 4) and (4, 0). We find intersection point of
4 4
x1 + 2x2 = 5 and x1 + x2 = 4 by simultaneously solving the equations. Hence, we get (3, 1).
Feasible region: OABC The maximum value of the objective function z = 10 occurs at 2 extreme points, B

Vertex Coordinate Max z = 2x1 + 4x2

O (0, 0) ∴z =0+0=0
A (4, 0) ∴ z = 2(4) + 0 = 8
B (3, 1) ∴ z = 2(3) + 4(1) = 10
C (0, 2.5) ∴ z = 2(0) + 4(2.5) = 10

and C. Any point including extreme points B and C on the same line between the B and C is an optimal
solution. Therefore, in fact an infinite number of values of x1 , x2 give the same value of objective function.
Hence, problem has multiple optimal solutions or alternate optima and maximum value of z = 10.

5
2. Unbounded solution: When the values of the decision variables may be increased indefinitely without
violating any of the constraints, the feasible region is unbounded. The value of objective function, in such
cases, may increase (for maximization) or decrease (for minimization) indefinitely. Thus both the feasible
region and the objective function values are unbounded.
Example. (1) Use a graphical method to solve the LPP

Maximize z = 6x1 + x2
Subject to 2x1 + x2 ≥ 3
x2 − x1 ≥ 0
x1 , x2 ≥ 0

Sol: To draw constraint 2x1 + x2 ≥ 3. Treat it as 2x1 + x2 = 3.

When x1 = 0 we get x2 = 3 and when x2 = 0 we get x1 = 3/2 = 1.5. ∴ (0, 3) and (1.5, 0).
To draw constraint x2 − x1 ≥ 0. Treat it as x2 − x1 = 0 ⇒ x2 = x1 . ∴ (0, 0) and (1, 1). We find intersection
point of x1 + 2x2 = 5 and x1 + x2 = 4 by simultaneously solving the equations. Hence, we get (1, 1). We

observe, that the feasible region is unbounded. The value of the objective function at the extreme points A(1, 1)
is z = 6 + 1 = 7 and at B(0, 3) is z = 0 + 3 = 3.
But there exist number of points in the feasible region for which the value of the objective function is more
than 7. For example, the point (3, 6). Thus both the variables x1 and x2 can be made arbitrarily large and the
value of z also increased. Hence, the problem has an unbounded solution.

Note: An unbounded solutions means that there exist an infinite no. of solutions to the given problem and
the optimal value of z lies at infinity.

6
3. Infeasible solution: When the constraints are not satisfied simultaneously, the linear programming problem
has no feasible solution. This situation can never occur if all the constraints are of the ‘≤’ type.
Example. (1) Use a graphical method to solve the LPP

Maximize z = x1 + x2
Subject to x1 + x2 ≤ 1
− 3x1 + x2 ≥ 3
x1 , x2 ≥ 0

Sol: There is no point (x1 , x2 ) which can lie in both the regions and satisfy all the constraints, there is no
solution to the given problem. Hence, there is infeasible solution.

7
 BIG M Method
Ex 1. Find an Initial Basic Feasible solution(IBFS) of the following LPP using Big
M method.
Maximize Z=− 2𝑥1 − 𝑥2
Subject to 3𝑥1 + 𝑥2 = 3
4𝑥1 + 3𝑥2 ≥ 6
𝑥1 + 2𝑥2 ≤ 4
𝑥1 ≥ 0, 𝑥2 ≥ 0
Solution: First we convert the above LPP into standard form as below
Maximize Z=− 2𝑥1 − 𝑥2 + 0𝑠1 + 0𝑠2
Subject to 3𝑥1 + 𝑥2 + 0𝑠1 + 0𝑠2 = 3
4𝑥1 + 3𝑥2 − 𝑠1 + 0𝑠2 = 6 👈( Note: 𝑠 is Surplus variable)
👈( Note: 𝑠 is Slack variable)
1
𝑥1 + 2𝑥2 + 0𝑠1 + 𝑠2 = 4 2
𝑥1, 𝑥2, 𝑠1, 𝑠2 ≥ 0
Here Basic Variable is only 𝑠2 and non basic variables are 𝑥1,
𝑥2 and 𝑠1. Hence 1 basic variable and 3 non basic variables.
Now, n=Number of unknowns in objective function=4
m=Number of constraints=3
So n-m=1 which is not equal to the number of non basic variables.
Therefore, IBFS is not possible in this case.
To find IBFS we need to introduce artificial variables.(Artificial variables are
added to the constraints which are of ' ≥ ' type or ‘=’ type.)
So in above standard form of LPP we add artificial variables 𝐴1 𝑎𝑛𝑑 𝐴2 in the first

and second equation respectively.

( Note: 1. Artificial variables 𝐴1 𝑎𝑛𝑑 𝐴2 are always non negative(≥ 0).

2. These variables are introduced in objective function Z with coefficient M

which is sufficiently large.
3. In case of Maximization problem artificial variables with coefficient M
are subtracted from RHS of objective function.
4. In case of Minimization problem artificial variables with coefficient M
are added to the RHS of objective function.)

So after introducing artificial variables we get the following form

Maximize Z=− 2𝑥1 − 𝑥2 + 0𝑠1 + 0𝑠2 − 𝑀𝐴1 − 𝑀𝐴2
Subject to 3𝑥1 + 𝑥2 + 0𝑠1 + 0𝑠2 + 𝐴1 + 0𝐴2 = 3
4𝑥1 + 3𝑥2 − 𝑠1 + 0𝑠2 + 0𝐴1 + 𝐴2 = 6
𝑥1 + 2𝑥2 + 0𝑠1 + 𝑠2 + 0𝐴1 + 0𝐴2 = 4
𝑥1, 𝑥2, 𝑠1, 𝑠2, 𝐴1, 𝐴2 ≥ 0
Here Basic Variable are 𝑠2, 𝐴1, 𝐴2
and non basic variables are 𝑥1, 𝑥2, 𝑠1.
Hence 3 basic variables and 3 non basic variables.
Now, n=Number of unknowns in objective function=6
m=Number of constraints=3
So n-m=3 which is equal to the number of non basic variables.
Therefore, IBFS is possible and which is as below

𝑥1 = 0, 𝑥2 = 0, 𝑠1 = 0
𝑠2 = 4, 𝐴1 = 3, 𝐴2 = 6
with Z=-3M-6M=-9M.
Ex 2. Find an Initial Basic Feasible solution(IBFS) of the following LPP using Big
M method.
Maximize Z=𝑥1 + 5𝑥2
Subject to 3𝑥1 + 4𝑥2 ≤ 6
𝑥1 + 3𝑥2 ≥ 2
𝑥1 ≥ 0, 𝑥2 ≥ 0
Ex 3. Find an Initial Basic Feasible solution(IBFS) of the following LPP using Big
M method.
Minimize Z=𝑥1 + 𝑥2 + 𝑥3
Subject to 𝑥1 − 3𝑥2 + 4𝑥3 = 5
𝑥1 − 2𝑥2 ≤ 3
2𝑥1 − 𝑥3 ≥ 4
𝑥1 ≥ 0, 𝑥2 ≥ 0, 𝑥3 ≥ 0
Ex 4. Find an Initial Basic Feasible solution(IBFS) of the following LPP using Big
M method.
Minimize Z=4𝑥1 + 2𝑥2 + 3𝑥3
Subject to 2𝑥1 + 4𝑥3 ≥ 5
2𝑥1 + 3𝑥2 + 𝑥3 ≥ 4
𝑥1 ≥ 0, 𝑥2 ≥ 0, 𝑥3 ≥ 0