CHAPTER – 2

Linear Programming Models

(Resource Allocation Models)
2.1. INTRODUCTION
A model, which is used for optimum allocation of scarce or limited resources to competing
products or activities under such assumptions as certainty, linearity, fixed technology, and
constant profit per unit, is linear programming.
Linear Programming is one of the most versatile, powerful and useful techniques for making
managerial decisions. Linear programming technique may be used for solving broad range of problems
arising in business, government, industry, hospitals, libraries, etc. Whenever we want to allocate the
available limited resources for various competing activities for achieving our desired objective, the
technique that helps us is LINEAR PROGRAMMING. As a decision making tool, it has demonstrated
its value in various fields such as production, finance, marketing, research and development and
personnel management. Determination of optimal product mix (a combination of products, which
gives maximum profit), transportation schedules, Assignment problem and many more. In this chapter,
let us discuss about various types of linear programming models.

2.2. PROPERTIES OF LINEAR PROGRAMMING MODEL

Any linear programming model (problem) must have the following properties:
(a) The relationship between variables and constraints must be linear.
(b) The model must have an objective function.
(c) The model must have structural constraints.
(d) The model must have non-negativity constraint.
Let us consider a product mix problem and see the applicability of the above properties.
Example 2.1. A company manufactures two products X and Y, which require, the following
resources. The resources are the capacities machine M1, M2, and M3. The available capacities
are 50,25,and 15 hours respectively in the planning period. Product X requires 1 hour of
machine M2 and 1 hour of machine M3. Product Y requires 2 hours of machine M1, 2 hours of
machine M2 and 1 hour of machine M3. The profit contribution of products X and Y are Rs.5/-
and Rs.4/- respectively.
Linear Programming Models (Resource Allocation Models) 23

The contents of the statement of the problem can be summarized as follows:

Machines Products Availability in hours
X Y
M1 0 2 50
M2 1 2 25
M3 1 1 15
Profit in Rs. Per unit 5 4

In the above problem, Products X and Y are competing candidates or variables.

Machine capacities are available resources. Profit contribution of products X and Y are given.
Now let us formulate the model.
Let the company manufactures x units of X and y units of Y. As the profit contributions of X and
Y are Rs.5/- and Rs. 4/- respectively. The objective of the problem is to maximize the profit Z, hence
objective function is:
Maximize Z = 5x + 4y OBJECTIVE FUNCTION.
This should be done so that the utilization of machine hours by products x and y should not
exceed the available capacity. This can be shown as follows:
For Machine M1 0x + 2y ≤ 50
For Machine M2 1x + 2y ≤ 25 and LINEAR STRUCTURAL CONSTRAINTS.
For machine M3 1x + 1y ≤ 15

But the company can stop production of x and y or can manufacture any amount of x and y. It
cannot manufacture negative quantities of x and y. Hence we have write,
Both x and y are ≥ 0 . NON -NEGATIVITY CONSTRAINT.
As the problem has got objective function, structural constraints, and non-negativity constraints
and there exist a linear relationship between the variables and the constraints in the form of inequalities,
the problem satisfies the properties of the Linear Programming Problem.

2.2.1. Basic Assumptions

The following are some important assumptions made in formulating a linear programming model:
1. It is assumed that the decision maker here is completely certain (i.e., deterministic conditions)
regarding all aspects of the situation, i.e., availability of resources, profit contribution of the
products, technology, courses of action and their consequences etc.
2. It is assumed that the relationship between variables in the problem and the resources available
i.e., constraints of the problem exhibits linearity. Here the term linearity implies proportionality
and additivity. This assumption is very useful as it simplifies modeling of the problem.
3. We assume here fixed technology. Fixed technology refers to the fact that the production
requirements are fixed during the planning period and will not change in the period.
4. It is assumed that the profit contribution of a product remains constant, irrespective of level
of production and sales.
24 Operations Research

5. It is assumed that the decision variables are continuous. It means that the companies
manufacture products in fractional units. For example, company manufacture 2.5 vehicles,
3.2 barrels of oil etc. This is referred too as the assumption of divisibility.
6. It is assumed that only one decision is required for the planning period. This condition
shows that the linear programming model is a static model, which implies that the linear
programming problem is a single stage decision problem. (Note: Dynamic Programming
problem is a multistage decision problem).
7. All variables are restricted to nonnegative values (i.e., their numerical value will be ≥ 0).

2.2.2. Terms Used in Linear Programming Problem

Linear programming is a method of obtaining an optimal solution or programme (say, product mix in a
production problem), when we have limited resources and a good number of competing candidates to
consume the limited resources in certain proportion. The term linear implies the condition of
proportionality and additivity. The programme is referred as a course of action covering a specified
period of time, say planning period. The manager has to find out the best course of action in the interest
of the organization. This best course of action is termed as optimal course of action or optimal solution
to the problem. A programme is optimal, when it maximizes or minimizes some measure or criterion of
effectiveness, such as profit, sales or costs.
The term programming refers to a systematic procedure by which a particular program or plan of
action is designed. Programming consists of a series of instructions and computational rules for solving
a problem that can be worked out manually or can fed into the computer. In solving linear programming
problem, we use a systematic method known as simplex method developed by American mathematician
George B. Dantzig in the year 1947.
The candidates or activity here refers to number of products or any such items, which need the
utilization of available resources in a certain required proportion. The available resources may be of any
nature, such as money, area of land, machine hours, and man-hours or materials. But they are limited
in availability and which are desired by the activities / products for consumption.

2.2.3. General Linear Programming Problem

A general mathematical way of representing a Linear Programming Problem (L.P.P.) is as given below:

Z = c1x1 + c2 x2 + … cn xn subjects to the conditions, OBJECTIVE FUNCTION

≥, =, ≤) b1
a11x1 + a12 x2 + a13 x3 + …+ a1j xj + …+..a1n xn ((≥
a21 x1 + a22 x2 + a23 x3 + …………. + a2j xj + ……+ a2nxn ( ≥, =, ≤) b2 Structural
……………………………………………………………………………………… Constraints
..........................................................................................................
≥, =, ≤ ) bm
am1x1 + am2x2 + am3x3 +…+ amj xj… + amn xn ((≥
and all xj are = 0 NON NEGETIVITY CONSTRINT
CONSTRINT..
Where j = 1, 2, 3, … n

Where all cj s, bi s and aij s are constants and xj s are decision variables.
To show the relationship between left hand side and right hand side the symbols ≤ , = , ≥ are
used. Any one of the signs may appear in real problems. Generally ≤ sign is used for maximization
Linear Programming Models (Resource Allocation Models) 25

problems and ≥ sign is used for minimization problems and in some problems, which are known as
mixed problems we may have all the three signs. The word optimize in the above model indicates either
maximise or minimize. The linear function, which is to be optimized, is the objective function. The
inequality conditions shown are constraints of the problem. Finally all xi s should be positive, hence the
non-negativity function.
The steps for formulating the linear programming are:
1. Identify the unknown decision variables to be determined and assign symbols to them.
2. Identify all the restrictions or constraints in the problem and express them as linear
equations or inequalities of decision variables.
3. Identify the objective or aim and represent it also as a linear function of decision variables.
Construct linear programming model for the following problems:

2.3. MAXIMIZATION MODELS

Example 2.2. A retail store stocks two types of shirts A and B. These are packed in attractive
cardboard boxes. During a week the store can sell a maximum of 400 shirts of type A and a
maximum of 300 shirts of type B. The storage capacity, however, is limited to a maximum of
600 of both types combined. Type A shirt fetches a profit of Rs. 2/- per unit and type B a profit
of Rs. 5/- per unit. How many of each type the store should stock per week to maximize the
total profit? Formulate a mathematical model of the problem.
Solution: Here shirts A and B are problem variables. Let the store stock ‘a’ units of A and ‘b’ units of
B. As the profit contribution of A and B are Rs.2/- and Rs.5/- respectively, objective function is:
Maximize Z = 2a + 5b subjected to condition (s.t.)
Structural constraints are, stores can sell 400 units of shirt A and 300 units of shirt B and the
storage capacity of both put together is 600 units. Hence the structural constraints are:
1a + 0b ≥ 400 and 0a + 1b ≤ 300 for sales capacity and 1a + 1b ≤ 600 for storage capacity.
And non-negativity constraint is both a and b are ≥ 0. Hence the model is:
Maximize Z = 2a + 5b s.t.
1a + 0b ≤ 400
0a + 1b ≤ 300
1a + 1b ≤ 600 and
Both a and b are ≥ 0.
Problem 2.3. A ship has three cargo holds, forward, aft and center. The capacity limits are:
Forward 2000 tons, 100,000 cubic meters
Center 3000 tons, 135,000 cubic meters
Aft 1500 tons, 30,000 cubic meters.
The following cargoes are offered, the ship owners may accept all or any part of each commodity:

Commodity Amount in tons. Volume/ton in cubic meters Profit per ton in Rs.
A 6000 60 60
B 4000 50 80
C 2000 25 50
26 Operations Research

In order to preserve the trim of the ship the weight in each hold must be proportional to the
capacity in tons. How should the cargo be distributed so as to maximize profit? Formulate this as linear
programming problem.
Solution: Problem variables are commodities, A, B, and C. Let the shipping company ship ‘a’ units of
A and ‘b’ units of B and ‘c’ units of C. Then Objective function is:
Maximize Z = 60a + 80b + 50c s.t.
Constraints are:
Weight constraint: 6000a + 4000b +2000c ≤ 6,500 ( = 2000+3000+1500)
The tonnage of commodity is 6000 and each ton occupies 60 cubic meters, hence there are 100
cubic meters capacity is available.
Similarly, availability of commodities B and C, which are having 80 cubic meter capacities each.
Hence capacity inequality will be:
100a +80b + 80c ≤ 2,65,000 (= 100,000+135,000+30,000). Hence the l.p.p. Model is:
Maximise Z = 60a+80b+50c s.t. 100a = 6000/60 = 100
6000a + 4000b + 2000c ≤ 6,500 80b = 4000/50 = 80
100a+80b+80c ≤ 2,65,000 and 80c = 2000/25 = 80 etc.
a,b,c all ≥ 0

2.4. MINIMIZATION MODELS

Problem 2.4. A patient consult a doctor to check up his ill health. Doctor examines him and advises
him that he is having deficiency of two vitamins, vitamin A and vitamin D. Doctor advises him to
consume vitamin A and D regularly for a period of time so that he can regain his health. Doctor
prescribes tonic X and tonic Y, which are having vitamin A, and D in certain proportion. Also advises
the patient to consume at least 40 units of vitamin A and 50 units of vitamin Daily. The cost of tonics
X and Y and the proportion of vitamin A and D that present in X and Y are given in the table below.
Formulate l.p.p. to minimize the cost of tonics.
Vitamins Tonics Daily requirement in units.
X Y
A 2 4 40
D 3 2 50
Cost in Rs. per unit. 5 3

Solution: Let patient purchase x units of X and y units of Y.

Objective function: Minimize Z = 5x + 3y
Inequality for vitamin A is 2x + 4y ≥ 40 (Here at least word indicates that the patient can
consume more than 40 units but not less than 40 units of vitamin A daily).
Similarly the inequality for vitamin D is 3x + 2y ≥ 50.
For non–negativity constraint the patient cannot consume negative units. Hence both x and y
must be ≥ 0.
Now the l.p.p. model for the problem is:
Linear Programming Models (Resource Allocation Models) 27

Minimize Z = 5x + 3y s.t.
2x + 4y ≥ 40
3x + 2y ≥ 50 and
Both x and y are ≥ 0.
Problem 2.5. A machine tool company conducts a job-training programme at a ratio of one for every
ten trainees. The training programme lasts for one month. From past experience it has been found that
out of 10 trainees hired, only seven complete the programme successfully. (The unsuccessful trainees
are released). Trained machinists are also needed for machining. The company's requirement for the
next three months is as follows:
January: 100 machinists, February: 150 machinists and March: 200 machinists.
In addition, the company requires 250 trained machinists by April. There are 130 trained machinists
available at the beginning of the year. Pay roll cost per month is:
Each trainee Rs. 400/- per month.
Each trained machinist (machining or teaching): Rs. 700/- p.m.
Each trained machinist who is idle: Rs.500/- p.m.
(Labour union forbids ousting trained machinists). Build a l.p.p. for produce the minimum cost
hiring and training schedule and meet the company’s requirement.
Solution: There are three options for trained machinists as per the data given. (i) A trained machinist
can work on machine, (ii) he can teach or (iii) he can remain idle. It is given that the number of trained
machinists available for machining is fixed. Hence the unknown decision variables are the number of
machinists goes for teaching and those who remain idle for each month. Let,
‘a’ be the trained machinists teaching in the month of January.
‘b’ be the trained machinists idle in the month of January.
‘c’ be the trained machinists for teaching in the month of February.
‘d’ be the trained machinists remain idle in February.
‘e’ be the trained machinists for teaching in March.
‘f ’ be the trained machinists remain idle in the month of March.
The constraints can be formulated by the rule that the number of machinists used for (machining
+ teaching + idle) = Number of trained machinists available at the beginning of the month.
For January 100 + 1a + 1b ≥ 130
For February, 150 + 1c + 1d = 130 + 7a (Here 7a indicates that the number of machinist trained
is 10 × a = 10a. But only 7 of them are successfully completed the training i.e. 7a).
For the month of March, 200 + 1e + 1f ≥ 130 + 7a +7c
The requirement of trained machinists in the month of April is 250, the constraints for this will be
130 + 7a + 7c + 7e ≥ 250 and the objective function is
Minimize Z = 400 (10a + 10c + 10e) + 700 (1a +1c + 1e) + 400 (1b + 1d +1f) and the non-
negativity constraint is a, b, c, d, e, f all ≥ 0. The required model is:
Minimize Z = 400 (10a + 10c + 10e) + 700 (1a +1c + 1e) + 400 (1b + 1d + 1f) s.t.
100 + 1a + 1b ≥ 130
150 + 1c + 1d ≥ 130 + 7a
28 Operations Research

200 + 1e + 1f ≥ 130 + 7a + 7c
130 + 7a + 7c + 7e ≥ 250 and
a, b, c, d, e, f all ≥ 0.

2.5. METHODS FOR THE SOLUTION OF A LINEAR PROGRAMMING PROBLEM

Linear Programming, is a method of solving the type of problem in which two or more candidates or
activities are competing to utilize the available limited resources, with a view to optimize the objective
function of the problem. The objective may be to maximize the returns or to minimize the costs. The
various methods available to solve the problem are:
1. The Graphical Method when we have two decision variables in the problem. (To deal with
more decision variables by graphical method will become complicated, because we have to
deal with planes instead of straight lines. Hence in graphical method let us limit ourselves to
two variable problems.
2. The Systematic Trial and Error method, where we go on giving various values to variables
until we get optimal solution. This method takes too much of time and laborious, hence this
method is not discussed here.
3. The Vector method. In this method each decision variable is considered as a vector and
principles of vector algebra is used to get the optimal solution. This method is also time
consuming, hence it is not discussed here.
4. The Simplex method. When the problem is having more than two decision variables, simplex
method is the most powerful method to solve the problem. It has a systematic programme,
which can be used to solve the problem.
One problem with two variables is solved by using both graphical and simplex method, so as to
enable the reader to understand the relationship between the two.

2.5.1. Graphical Method

In graphical method, the inequalities (structural constraints) are considered to be equations. This is
because; one cannot draw a graph for inequality. Only two variable problems are considered, because
we can draw straight lines in two-dimensional plane (X- axis and Y-axis). More over as we have non-
negativity constraint in the problem that is all the decision variables must have positive values always
the solution to the problem lies in first quadrant of the graph. Some times the value of variables may fall
in quadrants other than the first quadrant. In such cases, the line joining the values of the variables must
be extended in to the first quadrant. The procedure of the method will be explained in detail while
solving a numerical problem. The characteristics of Graphical method are:
(i) Generally the method is used to solve the problem, when it involves two decision variables.
(ii) For three or more decision variables, the graph deals with planes and requires high imagination
to identify the solution area.
(iii) Always, the solution to the problem lies in first quadrant.
(iv) This method provides a basis for understanding the other methods of solution.
Problem 2.6. A company manufactures two products, X and Y by using three machines A, B, and C.
Machine A has 4 hours of capacity available during the coming week. Similarly, the available capacity
of machines B and C during the coming week is 24 hours and 35 hours respectively. One unit of
Linear Programming Models (Resource Allocation Models) 29

product X requires one hour of Machine A, 3 hours of machine B and 10 hours of machine C. Similarly
one unit of product Y requires 1 hour, 8 hour and 7 hours of machine A, B and C respectively. When
one unit of X is sold in the market, it yields a profit of Rs. 5/- per product and that of Y is Rs. 7/- per
unit. Solve the problem by using graphical method to find the optimal product mix.
Solution: The details given in the problem is given in the table below:

Machines Products Available capacity in hours.

(Time required in hours).
X Y
A 1 1 4
B 3 8 24
C 10 7 35
Profit per unit in Rs. 5 7

Let the company manufactures x units of X and y units of Y, and then the L.P. model is:
Maximise Z = 5x + 7y s.t.
1x + 1y ≤ 4
3x + 8y ≤ 24
10x + 7y ≤ 35 and
Both x and y are ≥ 0.
As we cannot draw graph for inequalities, let us consider them as equations.
Maximise Z = 5x + 7y s.t.
1x + 1y = 4
3x + 8y = 24
10x + 7y = 35 and both x and y are ≥ 0
Let us take machine A. and find the boundary conditions. If x = 0, machine A can manufacture
4/1 = 4 units of y.

Figure 2.1 Graph for machine A

30 Operations Research

Similarly, if y = 0, machine A can manufacture 4/1 = 4 units of x. For other machines:

Machine B When x = 0 , y = 24/8 = 3 and when y = 0 x = 24/3 = 8
Machine C When x = 0, y = 35/10 = 3.5 and when y = 0, x = 35 / 7 = 5.
These values we can plot on a graph, taking product X on x-axis and product Y on y- axis.
First let us draw the graph for machine A. In figure 2. 1 we get line 1 which represents 1x + 1y
= 4. The point P on Y axis shows that the company can manufacture 4 units of Y only when does not
want to manufacture X. Similarly the point Q on X axis shows that the company can manufacture 4
units of X, when does not want to manufacture Y. In fact triangle POQ is the capacity of machine A
and the line PQ is the boundary line for capacity of machine A.
Similarly figure 2.2 show the Capacity line RS for machine B. and the triangle ROS shows the
capacity of machine B i.e., the machine B can manufacture 3 units of product Y alone or 8 units of
product X alone.

Figure 2.2. Graph for machine B

The graph 2.3 shows that the machine C has a capacity to manufacture 5 units of Y alone or 3.5
units of X alone. Line TU is the boundary line and the triangle TOU is the capacity of machine C.
The graph is the combined graph for machine A and machine B. Lines PQ and RS intersect at M.
The area covered by both the lines indicates the products (X and Y) that can be manufactured by using
both machines. This area is the feasible area, which satisfies the conditions of inequalities of machine
A and machine B. As X and Y are processed on A and B the number of units that can be manufactured
will vary and the there will be some idle capacities on both machines. The idle capacities of machine A
and machine B are shown in the figure 2.4.
Linear Programming Models (Resource Allocation Models) 31

Figure 2.3. Graph for machine C

Figure 2.4. Graph of Machines A and B

32 Operations Research

Figure 2.5 shows the feasible area for all the three machines combined. This is the fact because
a products X and Y are complete when they are processed on machine A, B, and C. The area covered
by all the three lines PQ. RS, and TU form a closed polygon ROUVW. This polygon is the feasible area
for the three machines. This means that all the points on the lines of polygon and any point within the
polygon satisfies the inequality conditions of all the three machines. To find the optimal solution, we
have two methods.

Figure 2.5. Graph for machine A, B and C combined

Method 1. Here we find the co-ordinates of corners of the closed polygon ROUVW and substitute
the values in the objective function. In maximisaton problem, we select the co-ordinates giving maximum
value. And in minimisaton problem, we select the co-ordinates, which gives minimum value.
In the problem the co-ordinates of the corners are:
R = (0, 3.5), O = (0,0), U = (3.5,0), V = (2.5, 1.5) and W = (1.6,2.4). Substituting these
values in objective function:
Z( 0,3.5) = 5 × 0 + 7 × 3.5 = Rs. 24.50, at point R
Z (0,0) = 5 × 0 + 7 × 0 = Rs. 00.00, at point O
Z(3.5,0) = 5 × 3.5 + 7 × 0 = Rs. 17.5 at point U
Z (2.5, 1.5) = 5 × 2.5 + 7 × 1.5 = Rs. 23.00 at point V
Z (1.6, 2.4) = 5 × 1.6 + 7 × 2.4 = Rs. 24.80 at point W
Hence the optimal solution for the problem is company has to manufacture 1.6 units of product
X and 2.4 units of product Y, so that it can earn a maximum profit of Rs. 24.80 in the planning period.
Method 2. Isoprofit Line Method: Isoprofit line, a line on the graph drawn as per the objective
function, assuming certain profit. On this line any point showing the values of x and y will yield same
profit. For example in the given problem, the objective function is Maximise Z = 5x + 7y. If we assume
a profit of Rs. 35/-, to get Rs. 35, the company has to manufacture either 7 units of X or 5 units of Y.
Linear Programming Models (Resource Allocation Models) 33

Hence, we draw line ZZ (preferably dotted line) for 5x + 7y = 35. Then draw parallel line to this line ZZ
at origin. The line at origin indicates zero rupees profit. No company will be willing to earn zero rupees
profit. Hence slowly move this line away from origin. Each movement shows a certain profit, which is
greater than Rs.0.00. While moving it touches corners of the polygon showing certain higher profit.
Finally, it touches the farthermost corner covering all the area of the closed polygon. This point where
the line passes (farthermost point) is the OPTIMAL SOLUTION of the problem. In the figure 2.6. the
line ZZ passing through point W covers the entire area of the polygon, hence it is the point that yields
highest profit. Now point W has co-ordinates (1.6, 2.4). Now Optimal profit Z = 5 × 1.6 + 7 × 2.4 =
Rs. 24.80.

Points to be Noted:
(i) In case lsoprofit line passes through more than one point, then it means that the
problem has more than one optimal solution, i.e., alternate solutions all giving
the same profit. This helps the manager to take a particular solution depending
on the demand position in the market. He has options.
(ii) If the lsoprofit line passes through single point, it means to say that the problem
has unique solution.
(iii) If the Isoprofit line coincides any one line of the polygon, then all the points on
the line are solutions, yielding the same profit. Hence the problem has
innumerable solutions.
(iv) If the line do not pass through any point (in case of open polygons), then the
problem do not have solution, and we say that the problem is UNBOUND.

Figure 2.6. ISO profit line method.

34 Operations Research

Now let us consider some problems, which are of mathematical interest. Such problems may not
exist in real world situation, but they are of mathematical interest and the student can understand the
mechanism of graphical solution.
Problem 2.7. Solve graphically the given linear programming problem. (Minimization Problem).
Minimize Z = 3a + 5b S.T
–3a + 4b ≤ 12
2a – 1b ≥ – 2
2a + 3b ≥ 12
1a + 0b ≥ 4
0a + 1b ≥ 2
And both a and b are ≥ 0.
Points to be Noted:
(i) In inequality –3a + 4b ≤ 12, product/the candidate/activity requires –3 units of
the resource. It does not give any meaning (or by manufacturing the product A
the manufacturer can save 3 units of resource No.1 or one has to consume –3
units of A. (All these do not give any meaning as far as the practical problems or
real world problems are concerned).
(ii) In the second inequality, on the right hand side we have –2. This means that –2
units of resource is available. It is absolutely wrong. Hence in solving a l.p.p.
problem, one must see that the right hand side we must have always a positive
integer. Hence the inequality is to be multiplied by –1 so that the inequality sign
also changes. In the present case it becomes: –2a + 1b ≤ 2.
Solution: Now the problem can be written as:
Minimize Z = 3a + 5b S.T.
When converted into equations they can be written as Min. Z = 3a + 5b S.T.
–3a + 4b ≤ 12 –3a + 4b = 12
–2a + 1b ≤ 2 –2a + 1b = 2
2a – 3b ≥ 12 2a – 3b = 12
1a + 0b ≤ 4 1a + 0b = 4
0a + 1b ≥ 2 and both a and b are ≥ = 0.0a + 1b ≥ 2 and both a and b are ≥ 0.
The lines for inequalities –3a + 4b ≤ 12 and –2a + 1b ≤ 2 starts from quadrant 2 and they are to
be extended into quadrant 1. Figure 2.7 shows the graph, with Isocost line.
Isocost line is a line, the points on the line gives the same cost in Rupees. We write
Isocost line at a far off place, away from the origin by assuming very high cost in objective
function. Then we move line parallel towards the origin (in search of least cost) until it passes
through a single corner of the closed polygon, which is nearer to the origin, (Unique Solution),
or passes through more than one point, which are nearer to the origin (more than one solution)
or coincides with a line nearer to the origin and the side of the polygon (innumerable solution).
The solution for the problem is the point P (3,2,) and the Minimum cost is Rs. 3 × 3 + 2 × 5 =
Rs. 19/-
Linear Programming Models (Resource Allocation Models) 35

Problem 2.8. The cost of materials A and B is Re.1/- per unit respectively. We have to manufacture an
alloy by mixing these to materials. The process of preparing the alloy is carried out on three facilities X,
Y and Z. Facilities X and Z are machines, whose capacities are limited. Y is a furnace, where heat
treatment takes place and the material must use a minimum given time (even if it uses more than the
required, there is no harm). Material A requires 5 hours of machine X and it does not require processing
on machine Z. Material B requires 10 hours of machine X and 1 hour of machine Z. Both A and B are
to be heat treated at last one hour in furnace Y. The available capacities of X, Y and Z are 50 hours, 1
hour and 4 hours respectively. Find how much of A and B are mixed so as to minimize the cost.

Figure 2.7. Graph for the problem 2.7

Solution: The l.p.p. model is:

Figure 2.8. Graph for the problem 2.8

36 Operations Research

Minimize Z = 1a + 1b S.T. Equations are: Minimise Z = 1a + 1b S.T

5a + 10b ≤ 50, 5a + 10b = 50
1a + 1b ≥ 1 1a + 1b = 1
0a + 1b ≤ 4 and both a and b are ≥ 0. 0a + 1b = 4 and both a and b are ≥ 0.
Figure 2.8 shows the graph. Here Isocost line coincides with side of the polygon, i.e., the line
MN. Hence the problem has innumerable solutions. Any value on line (1,1) will give same cost. Optimal
cost is Re.1/-
Problem 2.9. Maximise Z = 0.75 a + 1b S.T.
1a + 1b ≥ 0
–0.5 a + 1b ≤ 1 and both a and b are ≥ 0.
Solution: Writing the inequalities as equations,
1a + 1b = 0 i.e., a = b = 1 which is a line passing through origin at 45°
0.5 a + 1 b = 1 and both a and b are ≥ 0. Referring to figure 2.9.
The polygon is not closed one i.e., the feasible area is unbound. When Isoprofit line is drawn, it
passes through open side of the polygon and it does not coincide with any corner or any line. Hence the
line can be moved indefinitely, still containing a part of the feasible area. Thus there is no finite maximum
value of Z. That the value of Z can be increased indefinitely. When the value of Z can be increased
indefinitely, the problem is said to have an UNBOUND solution.

Figure 2.9. Graph for the problem 2.9

Problem 2.10. A company manufactures two products X and Y on two facilities A and B. The data
collected by the analyst is presented in the form of inequalities. Find the optimal product mix for
maximising the profit.
Maximise Z = 6x – 2y S.T. Writing in the equation form: Maximise Z = 6x – 2y S.T.
2x – 1y ≤ 2 2x – 1y = 2
1x + 0y ≤ 3 and both x and y are ≥ 0 1x + 0y =3 and both x and y are ≥ 0
Linear Programming Models (Resource Allocation Models) 37

Solution: The straight line for 2x – 1y = 2 starts in 4th quadrant and is to be extended into first
quadrant. The polygon is not a closed one and the feasible area is unbound. But when an Isoprofit line
is drawn it passes through a corner of the feasible area that is the corner M of the open polygon. The
(figure 2.10) coordinates of M are (3, 4) and the maximum Z = Rs. 10/-

Figure 2.10. Graph for the problem 2.10

Problem 2.11. A company manufactures two products X and Y. The profit contribution of X and Y are
Rs.3/- and Rs. 4/- respectively. The products X and Y require the services of four facilities. The
capacities of the four facilities A, B, C, and D are limited and the available capacities in hours are 200
Hrs, 150 Hrs, and 100 Hrs. and 80 hours respectively. Product X requires 5, 3, 5 and 8 hours of
facilities A, B, C and D respectively. Similarly the requirement of product Y is 4, 5, 5, and 4 hours
respectively on A, B, C and D. Find the optimal product mix to maximise the profit.
Solution: Enter the given data in the table below:

products
Machines X Y Availability in hours.
(Time in hours)
A 5 4 200
B 3 5 150
C 5 4 100
D 8 4 80
Profit in Rs. Per unit: 3 4

The inequalities and equations for the above data will be as follows. Let the company manufactures
x units of X and y units of Y. (Refer figure 2.11)
38 Operations Research

Maximise Z 3x + 4y S.T. Maximise Z = 3x + 4y S.T.

5x + 4y ≤ 200 5x + 4y = 200
3x + 5y ≤ 150 3x + 5y = 150
5x + 4y ≤ 100 5x + 4y = 100
8x + 4y ≤ 80 8x + 4y = 80
And both x and y are ≥ 0 And both x and y are ≥ 0

In the graph the line representing the equation 8x + 4y is out side the feasible area and hence it is
a redundant equation. It does not affect the solution. The Isoprofit line passes through corner T of the
polygon and is the point of maximum profit. Therefore ZT = Z(32,10) = 3 × 32 + 4 × 10 = Rs. 136/.
Problem 2.12. This problem is of mathematical interest.
Maximise Z = 3a + 4b S.T. Converting and writing in the form of equations,
1a – 1b ≤ –1. Maximise Z = 3a + 4b S.T
– 1a + 1b ≤ 0 1a – 1b = 0
And both a and b are ≥ 0 –1a + 1b = 0
And both a and b are ≥ 0

Referring to figure 2.11, the straight line for equation 1 starts in second quadrant and extended in
to first quadrant. The line for equation 2 passes through the origin. We see that there is no point, which
satisfies both the constraints simultaneously. Hence there is no feasible solution. Given l.p.p. has no
feasible solution.

Figure 2.11. Graph for the problem 2.11

Linear Programming Models (Resource Allocation Models) 39

Figure 2.12. Graph for the problem 2.12

Problem 2.13. Solve the l.p.p. by graphical method.

Maximise Z = 3a + 2b S.T.
1a + 1b ≤ 4
1a – 1b ≤ 2 and both a and b are ≥ 0.
Solution: The figure 2.13 shows the graph of the equations.
Equations are: Maximise Z = 3a + 2b S.T.
1a + 1b = 4
1a – 1b = 2 and both a and b are ≥ 0.
In the figure the Isoprofit line passes through the point N (3,1). Hence optimal Profit Z = 3 × 3
+ 2 × 1 = Rs.11 /-

Figure 2.13. Graph for the problem 2.13

40 Operations Research

Problem 2.14: Formulate the l.p.p. and solve the below given problem graphically.
Old hens can be bought for Rs.2.00 each but young ones costs Rs. 5.00 each. The old hens lay 3 eggs
per week and the young ones lay 5 eggs per week. Each egg costs Rs. 0.30. A hen costs Rs.1.00 per
week to feed. If the financial constraint is to spend Rs.80.00 per week for hens and the capacity
constraint is that total number of hens cannot exceed 20 hens and the objective is to earn a profit more
than Rs.6.00 per week, find the optimal combination of hens.
Solution: Let x be the number of old hens and y be the number of young hens to be bought. Now the
old hens lay 3 eggs and the young one lays 5 eggs per week. Hence total number of eggs one get is
3x + 5y.
Total revenues from the sale of eggs per week is Rs. 0.30 (3x + 5y) i.e., 0.90 x + 1.5 y
Now the total expenses per week for feeding hens is Re.1 (1x + 1y) i.e., 1x + 1y.
Hence the net income = Revenue – Cost = (0.90 x + 1.5 y) – (1 x + 1 y) = – 0.1 x + 0.5 y or 0.5y
– 0.1 x. Hence the desired l.p.p. is
Maximise Z = 0.5 y – 0.1 × S.T.
2 x + 5 y ≤ 80
1x + 1y ≤ 20 and both x and y are ≥ 0
The equations are:
Maximise Z = 0.5 y – 0.1 × S.T.
2 x + 5 y = 80
1x + 1y = 20 and both x and y are ≥ 0
In the figure 2.13, which shows the graph for the problem, the isoprofit line passes through the
point C. Hence Zc = Z (0,16) = Rs.8.00. Hence, one has to buy 16 young hens and his weekly profit
will be Rs.8.00

Figure 2.14. Graph for the problem 2.14

Linear Programming Models (Resource Allocation Models) 41

Point to Note: In case in a graphical solution, after getting the optimal solution,
one more constraint is added, we may come across following situations.
(i) The feasible area may reduce or increase and the optimal solution point may be
shifted depending the shape of the polygon leading to decrease or increase in
optimal value of the objective function.
(ii) Some times the new line for the new constraint may remain as redundant and
imposes no extra restrictions on the feasible area and hence the optimal value
will not change.
(iii) Depending on the position of line for the new constraint, there may not be any
point in the feasible area and hence there may not be a solution. OR the isoprofit
line may coincide with a line and the problem may have innumerable number of
solutions.

SUMMARY
1. The graphical method for solution is used when the problem deals with 2 variables.
2. The inequalities are assumed to be equations. As the problem deals with 2 variables, it is easy
to write straight lines as the relationship between the variables and constraints are linear. In
case the problem deals with three variables, then instead of lines we have to draw planes and
it will become very difficult to visualize the feasible area.
3. If at all there is a feasible solution (feasible area of polygon) exists then, the feasible area
region has an important property known as convexity Property in geometry. (Convexity
means: Convex polygon consists of a set points having the property that the segment joining
any two points in the set is entirely in the convex set. There is a mathematical theorem,
which states, “The points which are simulations solutions of a system of inequalities of the
≤ type form a polygonal convex set”.
The regions will not have any holes in them, i.e., they are solids and the boundary will not
have any breaks. This can be clearly stated that joining any two points in the region also
lies in the region.
4. The boundaries of the regions are lines or planes.
5. There will be corners or extreme points on the boundary and there will be edges joining the
various corners. The closed figure is known as polygon.
6. The objective function of a maximisation is represented by assuming suitable outcome
(revenue) and is known as Isoprofit line. In case of minimisation problem, assuming suitable
cost, a line, known as Isocost line, represents the objective function.
7. If isoprofit or isocost line coincides with one corner, then the problem has unique solution.
In case it coincides with more than one point, then the problem has alternate solutions. If
42 Operations Research

the isoprofit or isocost line coincides with a line, then the problem will have innumerable
number of solutions.
8. The different situation was found when the objective function could be made arbitrarily
large. Of course, no corner was optimal in that case.

QUESTIONS

1. An aviation fuel manufacturer sells two types of fuel A and B. Type A fuel is 25 % grade 1
gasoline, 25 % of grade 2 gasoline and 50 % of grade 3 gasoline. Type B fuel is 50 % of
grade 2 gasoline and 50 % of grade 3 gasoline. Available for production are 500 liters per
hour grade 1 and 200 liters per hour of grade 2 and grade 3 each. Costs are 60 paise per liter
for grade 1, 120 paise for grade 2 and100 paise for grade 3. Type A can be sold at Rs. 7.50
per liter and B can be sold at Rs. 9.00 per liter. How much of each fuel should be made and
sold to maximise the profit.
2. A company manufactures two products X1 and X2 on three machines A, B, and C. X1 require
1 hour on machine A and 1 hour on machine B and yields a revenue of Rs.3/-. Product X2
requires 2 hours on machine A and 1 hour on machine B and 1 hour on machine C and yields
revenue of Rs. 5/-. In the coming planning period the available time of three machines A, B,
and C are 2000 hours, 1500 hours and 600 hours respectively. Find the optimal product mix.
3. Maximize Z = 1 x + 1 y S.T.
1 x + 2 y ≤ 2000
1 x + 1 y ≤ 1500
0 x + 1 y ≤ 600 and both x and y are ≥ 0.
4. Maximise Z = 8000a + 7000b S.T.
3 a + 1 b ≤ 66
1 a + 1 b ≤ 45
1 a + 0 b ≤ 20
0 a + 1 b ≤ 40 and both a and b are ≥ 0.
5. Minimise Z = 1.5 x + 2.5 y S.T.
1x+3y ≥ 3
1 x + 6 y ≥ 2 and both x and y ≥ 0
6. Maximise Z = 3 a + 2 b S.T.
1a – 1 b ≤ 1
1 a + 1b ≥ 3 and both x and y are ≥ 0
7. Maximise Z = –3 x + 2 y S.T.
1x+0y ≤ 3
1 x – 1 y ≤ 0 and both x and y are ≥ 0
Linear Programming Models (Resource Allocation Models) 43

8. Maximize Z = – 1 a + 2b S.T.
–1a+1b ≤ 1
– 1 a + 2 b ≤ 4 and both a and b are ≥ 0.
9. Maximise Z = 3 x – 2 y S.T.
1x + 1 y ≤ 1
2 x + 2 y ≥ 4 and both x and y are ≥ 0
10. Maximize Z = 1x + 1 y S.T.
1 x – 1y ≥ 0
–3x + 1 y ≥ 3 and both x and y ≥ 0
 CHAPTER – 3

Linear Programming Models:

(Solution by Simplex Method)
Resource Allocation Model – Maximisation Case

3.1. INTRODUCTION
As discussed earlier, there are many methods to solve the Linear Programming Problem, such as
Graphical Method, Trial and Error method, Vector method and Simplex Method. Though we use
graphical method for solution when we have two problem variables, the other method can be used
when there are more than two decision variables in the problem. Among all the methods, SIMPLEX
METHOD is most powerful method. It deals with iterative process, which consists of first designing
a Basic Feasible Solution or a Programme and proceed towards the OPTIMAL SOLUTION and
testing each feasible solution for Optimality to know whether the solution on hand is optimal or not.
If not an optimal solution, redesign the programme, and test for optimality until the test confirms
OPTIMALITY. Hence we can say that the Simplex Method depends on two concepts known as
Feasibility and optimality.
The simplex method is based on the property that the optimal solution to a linear programming
problem, if it exists, can always be found in one of the basic feasible solution. The simplex method
is quite simple and mechanical in nature. The iterative steps of the simplex method are repeated until a
finite optimal solution, if exists, is found. If no optimal solution, the method indicates that no finite
solution exists.
3.2. COMPARISION BETWEEN GRAPHICAL AND SIMPLEX METHODS
1. The graphical method is used when we have two decision variables in the problem. Whereas
in Simplex method, the problem may have any number of decision variables.
2. In graphical method, the inequalities are assumed to be equations, so as to enable to draw
straight lines. But in Simplex method, the inequalities are converted into equations by:
(i) Adding a SLACK VARIABLE in maximisation problem and subtracting a SURPLUS
VARIABLE in case of minimisation problem.
3. In graphical solution the Isoprofit line moves away from the origin to towards the far off
point in maximisation problem and in minimisation problem, the Isocost line moves from far
off distance towards origin to reach the nearest point to origin.
4. In graphical method, the areas outside the feasible area (area covered by all the lines of
constraints in the problem) indicates idle capacity of resource where as in Simplex method,
the presence of slack variable indicates the idle capacity of the resources.
Linear Programming Models : Solution by Simplex Method 45

5. In graphical solution, if the isoprofit line coincides with more than one point of the feasible
polygon, then the problem has second alternate solution. In case of Simplex method the net-
evaluation row has zero for non-basis variable the problem has alternate solution. (If two
alternative optimum solutions can be obtained, the infinite number of optimum, solutions can
be obtained).
However, as discussed in the forth coming discussion, the beauty of the simplex method lies in
the fact that the relative exchange profitabilities of all the non -basis variables (vectors) can be determined
simultaneously and easily; the replacement process is such that the new basis does not violate the
feasibility of the solution.

3.3. MAXIMISATION CASE

Problem 3.1: A factory manufactures two products A and B on three machines X, Y, and Z. Product
A requires 10 hours of machine X and 5 hours of machine Y a one our of machine Z. The requirement
of product B is 6 hours, 10 hours and 2 hours of machine X, Y and Z respectively. The profit contribution
of products A and B are Rs. 23/– per unit and Rs. 32 /– per unit respectively. In the coming planning
period the available capacity of machines X, Y and Z are 2500 hours, 2000 hours and 500 hours
respectively. Find the optimal product mix for maximizing the profit.
Solution:
The given data is:
Products
Machines A B Capacity in hours
Hrs. Hrs.
X 10 6 2500
Y 5 10 2000
Z 1 2 500
Profit/unit Rs. 23 32 —

Let the company manufactures a units of A and b units of B. Then the inequalities of the
constraints (machine capacities) are:
Maximise Z = 23 a + 32 b S.T. OBJECTIVE FUNCTION
10 a + 6 b ≤ 2500
5 a + 10 b ≤ 2000 STRUCTURAL CONSTRAINTS.
1 a + 2 b ≤ 500
And both a and b are ≥ 0. NON-NEGATIVITY CONSTRAINT.

Now the above inequalities are to be converted into equations.

Take machine X: One unit of product A requires 10 hours of machine X and one unit of product
B require 6 units. But company is manufacturing a units of A and b units of B, hence both put together
must be less than or equal to 2,500 hours. Suppose a = 10 and b = 10 then the total consumption is
10 × 10 + 6 × 10 = 160 hours. That is out of 2,500 hours, 160 hours are consumed, and 2,340 hours
46 Operations Research

are still remaining idle. So if we want to convert it into an equation then 100 + 60 + 2,340 = 2,500. As
we do not know the exact values of decision variables a and b how much to add to convert the
inequality into an equation. For this we represent the idle capacity by means of a SLACK VARIABLE
represented by S. Slack variable for first inequality is S1, that of second one is S2 and that of ‘n’th
inequality is Sn.
Regarding the objective function, if we sell one unit of A it will fetch the company Rs. 23/– per
unit and that of B is Rs. 32/– per unit. If company does not manufacture A or B, all resources remain
idle. Hence the profit will be Zero rupees. This clearly shows that the profit contribution of each hour
of idle resource is zero. In Linear Programming language, we can say that the company has capacity of
manufacturing 2,500 units of S1, i.e., S1 is an imaginary product, which require one hour of machine
X alone. Similarly, S2 is an imaginary product requires one hour of machine Y alone and S3 is an
imaginary product, which requires one hour of machine Z alone. In simplex language S1, S2 and S3 are
idle resources. The profit earned by keeping all the machines idle is Rs.0/–. Hence the profit contributions
of S1, S2 and S3 are Rs.0/– per unit. By using this concept, the inequalities are converted into equations
as shown below:
Maximise Z = 23 a + 32 b + 0S1 + 0S2 + 0S3 S.T.
10 a + 6 b + 1S1 = 2500
5 a + 10 b + 1S2 = 2000
1 a + 2 b + 1S3 = 500 and a, b, S1, S2 and S3 all ≥ 0.
In Simplex version, all variables must be available in all equations. Hence the Simplex format of
the model is:
Maximise Z = 23 a + 32 b + 0S1 + 0S2 + 0S3 S.T.
10 a + 6 b + 1S1 + 0S2 + 0S3 = 2500
5 a + 6 b + 0S1 + 1S2 + 0S3 = 2000
1 a + 2 b + 0S1 + 0S2 + 1S3 = 500 and a, b, S1, S2 and S3 all ≥ 0.
The above data is entered in a table known as simplex table (or tableau). There are many
versions of table but in this book only one type is used.
In Graphical method, while finding the profit by Isoprofit line, we use to draw Isoprofit line at
origin and use to move that line to reach the far off point from the origin. This is because starting from
zero rupees profit; we want to move towards the maximum profit. Here also, first we start with zero
rupees profit, i.e., considering the slack variables as the basis variables (problem variables) in the initial
programme and then improve the programme step by step until we get the optimal profit. Let us start
the first programme or initial programme by rewriting the entries as shown in the above simplex table.
Linear Programming Models : Solution by Simplex Method 47

This column shows Basic or Problem variables.

This column shows objective co-efficients corresponding to basic variables in the
programme This column shows the values of the basic variables, the value of such non
basic variable is = 0
This row shows Cj above each variable, the respective
objective coefficient.
Variable row lists all the variable in the problem.
The numbers under non-basic variables
represent substitution Ratios.
Every simplex tableau contains an
identity Matrix under the basic variables.

Programme Profit per Quantity or 23 32 0 0 0

Variable or unit in Rs. Capacity a b S1 S2 S3
Basic variable
S1 0 2500 10 6 1 0 0
S2 0 2000 5 10 0 1 1
S3 0 500 1 2 0 0 1
Zj 0 0 0 0 0
Net Evaluation 23 32 0 0 0
Cj – Zj

The number in Zj row under each column variable gives the total gross
amount of outgoing profit when we consider the exchange between one
unit of the column variable and the basic Variable.
The numbers in the net-evaluation row, under each column represent the opportunity cost of not
having one unit of the respective column variables in the solution. In other words, the number
represent the potential improvement in the objective function that will result by introducing into
the programme one unit of the respective column variable.

Table: 1. Initial Programme

Solution: a = 0, b = 0, S1= 2500, S2 = 2000 and S3 = 500 and Z = Rs. 0.00.

Programme Profit per unit Quantity in C 23 32 0 0 Replacement
j
(Basic variables) In Rs. Cb Units. a b S1 S2 S3 Ratio.

S1 0 2500 10 6 1 0 0 2500/6 = 416.7

S2 0 2000 5 10 0 1 0 2000/10 = 200
S3 0 500 1 2 0 0 1 500/2 = 250
Zj 0 0 0 0 0
Cj– Zj = Opportu- 23 32 0 0 0
nity cost in Rs.
Net evaluation
row.
48 Operations Research

The interpretation of the elements in the first table

1. In the first column, programme column, are the problem variables or basis variables that are
included in the solution or the company is producing at the initial stage. These are S1, S2 and
S3, which are known as basic variables.
2. The second column, labeled as Profit per unit in Rupees shows the profit co-efficient of the
basic variables i.e., Cb. In this column we enter the profit co-efficient of the variables in the
program. In table 1, we have S1, S2 and S3 as the basic variables having Rs.0.00 as the profit
and the same is entered in the programme.
3. In the quantity column, that is 3rd column, the values of the basic variables in the programme
or solution i.e., quantities of the units currently being produced are entered. In this table, S1,
S2 and S3 are being produced and the units being produced (available idle time) is entered i.e.,
2500, 2000 and 500 respectively for S1, S2 and S3. The variables that are not present in this
column are known as non-basic variables. The values of non-basis variables are zero; this
is shown at the top of the table (solution row).
4. In any programme, the profit contribution, resulting from manufacturing the quantities of
basic variables in the quantity column is the sum of product of quantity column element
and the profit column element.
In the present table the total profit is Z = 2500 × 0 + 2000 × 0 + 500 × 0 = Rs. 0.00.
5. The elements under column of non-basic variables, i.e., a and b (or the main body of the
matrix) are interpreted to mean physical ratio of distribution if the programme consists of
only slack variables as the basic variables. Physical ratio of distribution means, at this stage,
if company manufactures one unit of ‘a’ then 10 units of S1, 5 units of S2 and 1 unit of S3
will be reduced or will go out or to be scarified. By sacrificing the basic variables, the
company will lose the profit to an extent the sum of product of quantity column element and
the profit column element. At the same time it earns a profit to an extent of product of profit
co-efficient of incoming variable and the number in the quantity column against the
just entered (in coming) variable.
6. Coming to the entries in the identity matrix, the elements under the variables, S1, S2 and S3
are unit vectors, hence we apply the principle of physical ratio of distribution, one unit of
S1 replaces one unit of S1 and so on. Ultimately the profit is zero only. In fact while doing
successive modifications in the programme towards getting optimal; solution, finally the unit
matrix transfers to the main body. This method is very much similar with G.J. method
(Gauss Jordan) method in matrices, where we solve simultaneous equations by writing
in the form of matrix. The only difference is that in G.J method, the values of variables
may be negative, positive or zero. But in Simplex method as there is non-negativity
constraint, the negative values for variables are not accepted.
7. Cj at the top of the columns of all the variables represent the coefficients of the respective
variables I the objective function.
8. The number in the Zj row under each variable gives the total gross amount of outgoing profit
when we consider the exchange between one unit of column, variable and the basic variables.
9. The number in the net evaluation row, Cj – Zj row gives the net effect of exchange
between one unit of each variable and basic variables. This they are zeros under columns
of S1, S2 and S3. A point of interest to note here is the net evaluation element of any
basis variable (or problem variable) is ZERO only. Suppose variable ‘a’ becomes basis
Linear Programming Models : Solution by Simplex Method 49

variable, the entry in net evaluation row under ‘a’ is zero and so on. Generally the
entry in net evaluation row is known as OPPORTUNITY COST. Opportunity cost
means for not including a particular profitable variable in the programme, the
manufacturer has to lose the amount equivalent to the profit contribution of the
variable. In the present problem the net evaluation under the variable ‘a’ is Rs. 23 per unit
and that of ‘b’ is Rs, 32 per unit. Hence the if the company does not manufacture ‘a’ at this
stage it has to face a penalty of Rs. 23/– for every unit of ‘a’ for not manufacturing and the
same of product variable ‘b’ is Rs. 32/–. Hence the opportunity cost of product ‘b’ is higher
than that of ‘a’, hence ‘b’ will be the incoming variable. In general, select the variable,
which is having higher opportunity cost as the incoming variable (or select the variable,
which is having highest positive number in the net evaluation row.
In this problem, variable ‘b’ is having higher opportunity cost; hence it is the incoming
variable. This should be marked by an arrow ( ↑ ) at the bottom of the column and enclose the
elements of the column in a rectangle this column is known as KEY COLUMN. The elements
of the key column show the substitution ratios, i.e., how many units of slack variable goes
out when the variable enters the programme.
Divide the capacity column elements by key column numbers to get REPLACEMENT
RATIO COLUMN ELEMENTS, which show that how much of variable ‘b’ can be
manufactured in each department, without violating the given constraints. Select the
lowest replacement ratio and mark a tick (√) at the end of the row, which indicates OUT
GOING VARIABLE. Enclose the elements of this column in a rectangle, which indicates
KEY ROW, indicating out going variable. We have to select the lowest element because this
is the limiting ratio, so that, that much of quantity of product can be manufactured on all
machines or in all departments as the case may be. In the problem 200 units is the limiting
ratio, which falls against S2, i.e., S2 is the outgoing variable. This means that the entire
capacity of machine Y is utilized. By manufacturing 200 units of ‘b’, 6 × 200 = 1200 hours
of machine X is consumed and 2 × 200 = 400 hours of machine Z is consumed. Still 2500 –
1200 = 1300 hours of machine X and 500 – 400 = 100 units of machine Z remains idle. This
is what exactly we see in graphical solution when two lines of the machines are superimposed.
The element at the intersection of key column and key row is known as KEY NUMBER.
This is known as key number because with this number we have to get the next table.
For getting the revised programme, we have to transfer the rows of table 1 to table 2. To
do this the following procedure is used.
Step 1: To Write the incoming variable ‘b’ in place of out going variable S2. Enter the profit of ‘b’
in profit column. Do not alter S1 and S3. While doing so DO NOT ALTER THE POSITION
OF THE ROWS.
Step 2: DIVIDING THE ELEMENTS OF OLD COLUMN BY KEY COLUMN ELEMENTS obtains
capacity column elements.
Step 3: Transfer of key row: DIVIDE ALL ELEMENTS OF KEY ROW BY RESPECTIVE KEY
COLUMN NUMBER.
Step 4: Transfer of Non-Key rows: NEW ROW NUMBER = (old row number – corresponding key
row number) × fixed ratio.
Fixed ratio = Key column number of the row/key number.
50 Operations Research

Step 5: Elements of Net evaluation row are obtained by:

Objective row element at the top of the row – Σ key column element × profit column
element.
Step 6: Select the highest positive element in net evaluation row or highest opportunity cost and
mark the column by an arrow to indicate key column (incoming variable).
Step 7: Find the replacement ratios by dividing the capacity column element in the row by key
column element of the same row and write the ratios in replacement ratio column. Select the
limiting (lowest) ratio and mark with a tick mark to indicate key row (out going variable).
The element at the intersection of key column and key row is known as key number.
Continue these steps until we get:
(i) For maximisation problem all elements of net evaluation row must be either zeros or
negative elements.
(ii) For Minimisation problem, the elements of net evaluation row must be either zeros or
positive elements.
Table: 2.

Solution: S1 = 1,300, S2 = 0, S3 = 100, a = 0, b = 200, Z = 32 × 200 = Rs. 6400.

Problem Profit in Capacity C j 23 32 0 0 0 Replacement

variable. Rs. a b S1 S2 S3 Ratio (R.R)
S1 0 1,300 7 0 1 -0.6 0 1300/7 = 185.7
b 32 200 0.5 1 0 0.10 0 400
S3 0 100 0 0 0 -0.5 1 --
Zj 16 32 0 3.2 0
C j- Z j = net evaluation 7 0 0 - 3.2 0

1. Transfer of Key row: 2000/10 , 5 /10, 10 /10, 0 /10, 1 / 10, 0 / 10

2. Transfer of Non key rows:
Rule: (Old row Number – corresponding key row number) – key column number / key number = new
row no.
1st row. 2500 – 2000 × 6/10 = 1300 2nd row: 500 – 2000 × 2/10 = 100
10 – 10 × 6/10 = 0
6 – 10 × 6/10 = 0 1 – 5 × 2/10 = 0
1 – 0 × 6/10 = 1 2 – 10 × 2/10 = 0
0 – 1 × 6/10 = – 0.6 0 – 0 × 2/10 = 0
0 – 0 × 6/10 = 0 0 – 1 × 2/10 = – 0.2
1 – 0 × 2 / 10 = 1
Replacement ratios: 1300/7 = 185.7, 200/0.5 = 400, 100/0 = Infinity.
Linear Programming Models : Solution by Simplex Method 51

Net evaluation row elements =

Column under 'a' = 23 – (7 × 0 + 0.5 × 32 + 0 × 0) = 23 – 16 = 7
'b' = 32 – (0 × 0 + 1 × 32 + 0 × 0) = 32 – 32 = 0
S1 = 0 – (1 × 0 + 0 × 32 + 0 × 0) = 0
S2 = 0 – (– 0.6 × 0 + 0.1 × 32 + –0.2 × 0) = – 3.2
S3 = 0 – (0 × 0 + 0 × 32 + 1 × 0) = 0
In the above table, the net evaluation under S2 is – 3.2. This resource is completely utilized to
manufacture product B. The profit earned by manufacturing B is Rs. 6400/–. As per the law of
economics, the worth of resources used must be equal to the profit earned. Hence the element 3.2
(ignore negative sign) is known as economic worth or artificial accounting price (technically it can
be taken as MACHINE HOUR RATE) of the resources or shadow price of the resource. (In fact
all the elements of reevaluation row under slack variables are shadow prices of respective
resources). This concept is used to check whether the problem is done correctly or not. To do
this MULTIPLY THE ELEMENTS IN NET EVALUATION ROW UNDER SLACK VARIABLES
WITH THE ORIGINAL CAPACITY CONSTRAINTS GIVEN IN THE PROBLEM AND FIND
THE SUM OF THE SAME. THIS SUM MUST BE EQUAL TO THE PROFIT EARNED BY
MANUFACTRUING THE PRODUCT.
Σ Shadow prices of resources used must be equal to the profit earned.
Table: 3.

Problem Profit in Capacity Cj 23 32 0 0 0 Replacement

variable Rs. a b S1 S2 S3 ratio
a 23 185.7 1 0 0.143 – 0.086 0
b 32 107.14 0 1 – 0.07 0.143 0
S3 0 100 0 0 0 – 0.02 1
Zj 23 32 1 2.6 0
Cj – Z j Net evaluation. 0 0 –1.0 – 2.6 0

Transfer of key row: 1300/7 = 185.7, 7/7 = 1, 0/7 = 0, 1/7 = 0.143, –3 /5 = – 0.086 0/7 = 0
Row No. 2 Row No.3
200 – 1300 × 1/14 = 107.14 As the fixed ratio will be zero for this row
0.5 – 7 × 1/14 = 0 the row elements will not change.
1 – 0 × 1/14 = 1
0 – 1 × 1/14 = – 0.07
0.1 – (– 0.6) × 1/14 = 0.143
0 – 0 × 1/14 = 0
Net evaluation row elements:
For ‘a’ = 23 – 1 × 23 + 0 × 32 + 0 × 0 = 0
For ‘b’ = 32 – 0 × 23 + 1 × 32 +0 × 0 = 0
For S1 = 0 – 0.143 × 23 + ( – 0.07 × 32) + 0 × 0 = –1
For S2 = 0 – (–0.086 × 23) + 0.143 × 32 + (– 0.02 × 0) = – 2.6
52 Operations Research

For S3 = 0 – 0 × 23 + 0 × 32 + 1 × 0 = 0
Profit Z = 185.7 × 23 + 107.14 × 32 = Rs. 7,700
Shadow price = 1 × 2500 + 2.6 × 2000 = Rs. 2500 + 5200 = Rs. 7700/–
As all the elements of net evaluation row are either negative elements or zeros, the solution is
optimal.
Also the profit earned is equal to the shadow price.
The answer is the company has to manufacture:
185.7 units of A and 107.14 units of B and the optimal return is Z = Rs. 7,700/–

3.4. MINIMISATION CASE

Above we have discussed how to solve maximisation problem and the mechanism or simplex method
and interpretation of various elements of rows and columns. Now let us see how to solve a minimization
problem and see the mechanism of the simplex method in solving and then let us deal with some typical
examples so as to make the reader confident to be confident enough to solve problem individually.

Comparison between maximisaton case and minimisation case

[Link]. Maximisation case Minimisation case

Similarities:
1. It has an objective function. This too has an objective function.
2. It has structural constraints. This too has structural constraints.
3. The relationship between variables and Here too the relationship between and variables
constraintsis linear. constraints is linear.
4. It has non-negativity constraint. This too has non-negativity constraints.
5. The coefficients of variables may be positive The coefficient of variables may be positive,
or negative or zero. Negative or zero.
6. For selecting out going variable (key row) For selecting out going variable (key row) lowest
lowest replacement ratio is selected. replacement ratio is selected.
Differences:
1. The objective function is of maximisation The objective function is of minimisation type.
type.
2. The inequalities are of ≤ type. The inequalities are of ≥ type.
3. To convert inequalities into equations, slack To convert inequalities into equations, surplus
variables are added. Variables are subtracted and artificial surplus
variables are added.
4. While selecting incoming variable, highest While slecting, incoming variable, lowest element
positive Opportunity cost is selected from in the net evaluation row is selected (highest number
net evaluation Row. with negative sign).
5. When the elements of net evaluation row are When the element of net evaluation row are either
either Negative or zeros, the solution is positive or zeros the solution is optimal.
optimal
Linear Programming Models : Solution by Simplex Method 53

It is most advantageous to introduce minimisation problem by dealing with a well-known problem,

known as diet problem.
Problem 3.2: In this problem, a patient visits the doctor to get treatment for ill health. The doctor
examines the patient and advises him to consume at least 40 units of vitamin A and 50 units of vitamin
B daily for a specified time period. He also advises the patient that to get vitamin A and vitamin B he has
to drink tonic X and tonic Y that have both vitamin A and vitamin B in a proportion. One unit of tonic
X consists 2 units of vitamin A and 3 units of vitamin B and one unit of tonic Y consists of 4 units of
vitamin A and 2 units of vitamin B. These tonics are available in medical shops at a cost of Rs.3.00 and
Rs.2.50 per unit of X and Y respectively. Now the problem of patient is how much of X and how much
of Y is to be purchased from the shop to minimise the total cost and at the same time he can get
required amounts of vitamins A and B.
First we shall enter all the data in the form of a table.
Vitamin Tonic Requirement
X Y
A 2 4 40
B 3 2 50
Cost in Rs. 3 2.50

Let the patient purchase ‘x’ units of X and ‘y’ units of Y then the inequalities are (Note: the
condition given in the problem is AT LEAST hence the inequalities are of ≥ type)
Inequalities:
For vitamin A: Minimize Z = 3 x + 2.5 y S.T
2 x + 4 y ≥ 40
3 x + 2 y ≥ 50
And both x and y are ≥ 0.
In the above inequalities, say 2 x + 4 y ≥ 40, if we give values to x and y such that the sum is
greater than or equal to 40, for example, x = 10 and y = 10 then 2 x + 4 y = 60 which is > 40. To make
it equal to 40 we have to subtract 20, so that 20 + 40 – 20 = 40. When we know the values, we can do
this. But as we do not know the values of x and y we have to subtract a SURPLUS VARIABLE,
generally represented by ‘p’, ‘q’, ‘r’……. etc. If we do this then the inequality 2 x + 4 y ≥ 40 will be
2 x + 4 y – 1p = 40.
Now if we allocate value zero to x and y then 0 + 0 – 1p = 40 or p = – 40. Which is against to the
rules of l.p.p. as every [Link] the values of variables must be ≥ 0. Hence in minimization problem,
we introduce one more Surplus variable, known as ARTIFICIAL SURPLUS VARIABLE generally
represented by A1, A2, A3 … etc. Now by introducing artificial surplus variable, we can write 2x + 4y
= 40 as 2 x + 4 y –1p + 1A1 = 40.
If values of x, y, and p are equal to zero, then 1A1 = 40. The artificial surplus variable has the value
40, a positive integer. Hence we start our initial programme with the artificial variables, A1, A2, A3 etc.
and go on replacing them by x, y, z etc. that is decision variables.
Coming to the cost coefficients of surplus and artificial surplus variables, for example, p is very
similar to vitamin A and one unit of p consists of only one unit of vitamin A. It will come as give
away product when we purchase vitamin A. That is the cost coefficient of ‘p’ is zero (it is very much
54 Operations Research

similar to slack variable in maximization problem). But the artificial surplus variable has to be purchased
by paying a very high price for it. In character it is very much similar to surplus variable ‘p’ because
one unit of A1 consists of one unit of vitamin A. The cost coefficient of A1 is represented by a very high
value represented by M (which means one unit of A1 cost Millions or Rupees). As we are introducing
CAPITAL ‘M’, THIS METHOD IS KNOWN AS BIG ‘M’ METHOD.
By using the above concept, let us write the equations of the inequalities of the problem.
Minimise Z = 3 x + 2.5 y + 0p + 0q + M A1 + M A2 S.T. Objective Function.
2 x + 4 y – 1 p + 1A1 = 40
3 x + 2 y – 1 q + 1 A2 = 50 Structural Constraints.
And x, y, p, q, A1, A2 all ≥ 0 Non negativity Constraint.
Simplex format of the above is:
Minimise Z = 3x + 2.5y + 0p + 0q + MA1 + MA2 S.T.
2x + 4y – 1p + 0q + 1A1 + 0A2 = 40
3x + 2y + 0p – 1q + 0A1 + 1A2 = 50
And x, y, p, q, A1, A2 all = 0.
Let us enter the data in the Initial table of Simplex method.
Table: 1.
x = 0, y = 0, p = 0, q = 0 A1 = 40, Z = Rs. 40M + 50M = 90M

Programme Cost per Cost Cj 3 2.5 0 0 M M Replacement

variable unit in Rs. requirement x y p q A1 A2 ratio

A1 M 40 2 4 –1 0 1 0 40/4 = 10
A2 M 50 3 2 0 –1 0 1 50/2 = 25
Zj 5M 6M –M –M M M
Net Cj – Zj 3– 5 M 2.5 – 6 M M M 0 0
evaluation

Note: As the variables A1 and A2 are basis variables, their Net evaluation is zero.
Now take 6M and 5M, 6 M is greater and if we subtract 2.5 from that it is negligible. Hence –6m
will be the lowest element. The physical interpretation is if patient purchases Y now, his cost will be
reduced by an amount 6M. In other words, if the patient does not purchase the Y at this point, his
penalty is 6M, i.e., the opportunity cost is 6M. As the non-basis variable Y has highest opportunity
cost (highest element with negative sign), Y is the incoming variable. Hence, the column under Y is
key column. To find the out going variable, divide requirement column element by key column
element and find the replacement ratio. Select the lowest ratio, i.e., here it is 10, falls in first row, hence
A1 is the out going variable.
To transfer key row, divide all the elements of key row by key number (= 4).
40/4 = 10, 2/4 = 0.5, –1/4 = – 0.25, 0/25 = 0, 1/25 = 0.25, 0/4 = 0.
Linear Programming Models : Solution by Simplex Method 55

To transfer non-key row elements:

New row element = old row element – corresponding Key row element × (Key column
number/key number).
50 – 40 × 2/4 = 30
3 – 2 × 0.5 = 2
2 – 4 × 0.5 = 0
0 – (–1) × 0.5 = 0.5 .
–1 – 0 × 0.5 = – 1
0 – 1 × 0.5 = – 0.5
1 – 0 × 0.5 = 1
Note:
(i) The elements under A1 and A2 i.e., artificial variable column are negative versions of
elements under artificial variable column.
(ii) The net evaluation row elements of basis variables are always zero. While writing
the second table do not change the positions of the rows).
Let us now enter the new elements of changed rows in the second simplex table.
Table: 2.

x = 0, y = 10, p = 0, q = 0, A1 = 0, A2 = 30 and Z = Rs. 10 × 2.5 = Rs. 25.00

Programme Cost per Cost Cj 3 2.5 0 0 M M Replacement

variable unit in Rs. requirement x y p q A1 A2 ratio
y 2.5 10 0.5 1 – 0.25 0 0.25 0 10/0.5 = 20
A2 M 30 2 0 0.5 –1 –0.5 1 30/2 = 15
Zj 1.25 + 2.5 0.5 M – – M 0.625 – 0
2M 0.625 0.5 M
Cj – Z j 1.75 – 0 0.625 M 1.5 M – 0
2M – .5 M 0.625

Changing the key row: 30/2 = 15, 2/2 = 1, 0/2 = 0, 0.5/2 = 0.25, –1/2 = –0.5, –0.5/2 = –0.25,
1/2 = 0.5.
Changing the non key row:
10 – 30 × 0.5/2 = 2.5
0.5 – 2 × 0.25 = 0
1 – 0 × 0.25 = 1
–0.25 – 0.5 × 0.25 = –0.375
0 – (–1) × 0.25 = 0.25
0.25 – (–0.5) × 0.25 = 0.375
0 – 1 × 0.25 = – 0.25
Entering the above in the simplex table 3.
56 Operations Research

Table: 3.

x = 15, y = 2.5, p = 0, q = 0 , A1 = 0, A2 = 0 and Z = Rs. 15 × 3 + Rs. 2.5 × 2.5 = 45 + 6.25 =

Rs. 51.25
Programme Cost per Cost Cj 3 2.5 0 0 M M Replacement
Variable unit in Rs. Requirement x y p q A1 A2 ratio

y 2.5 2.5 0 1 – 0.25 0.375 – 0.25 —

0.375
Ax 3 15 1 0 0.25 –0.5 –0.25 0.5 —
Zj 3 2.5 – 0.875 0.188 0.875
0.188
Cj – Z j 0 0 0.188 0.875 M– M–
0.188 0.875

Optimal Cost = Z* = 3 × 15 + 2.5 × 2.5 = 45 + 6.25 = Rs. 51.25

Imputed value = 0.1875 × 40 + 0.875 × 50 = 7.5 + 43.75 = Rs. 51.25.
As all the elements of net evaluation row are either zeros or positive elements the solution
is optimal.
The patient has to purchase 15 units of X and 2.5 units of Y to meet the requirement and
the cost is Rs. 51.25/–
While solving maximisation problem, we have seen that the elements in net evaluation row, i.e.,
(Cj – Zj) row associated with slack variables represent the marginal worth or shadow price of the
resources. In minimisation problem, the elements associated with surplus variables in the optimal table,
represent the marginal worth or imputed value of one unit of the required item. In minimisation
problem, the imputed value of surplus variables in optimal solution must be equal to the optimal
cost.

Point to Note:
1. In the mechanics of simplex method of minimization problem, once an artificial
surplus variable leaves the basis, its exit is final because of its high cost coefficient
(M), which will never permit the variable to reenter the basis. In order to save
time or to reduce calculations, we can cross out the column containing the artificial
surplus variable, which reduces the number of columns.
2. A better and easier method is to allocate a value for M in big M method; this value
must be higher than the cost coefficients of the decision variables. Say for example
the cost coefficients of the decision variable in the above problem are for X it is
Rs.3/– and for Y it is Rs. 2.5. We can allocate a cost coefficient to M as Rs.10,
which is greater than Rs.3/– and Rs. 2.5. Depending the value of decision variables,
this value may be fixed at a higher level (Preferably the value must be multiples
of 10 so that the calculation part will be easier.

By applying the above note, let us see how easy to work the same problem:
Linear Programming Models : Solution by Simplex Method 57

Table: 1.

x = 0, y = 0, p = 0, q = 0 = A1 = 40, A2 = 50 and Z = 10 × 40 + 10 × 50 = Rs.900/–

Problem Cost Cj 3 2.5 0 0 10 10 Replacement

variable requirement x y p q A1 A2 ratio

A1 10 40 2 4 –1 0 1 0 10
A2 10 50 3 2 0 –1 0 1 25
NER – 47 – 57.5 10 10 0 0

Table: 2.

x = 0, y = 25, p = 0, q = 0, A1 = 0, A2 = 30 and Z = 25 × 10 + 30 × 10 = 250 + 300 = Rs. 550/–

Problem Cost per C j 3 2.5 0 0 10 Replacement

variable requirement x y p q A2 ratio

y 2.5 10 0.5 1 – 0.5 0 0 20

A2 10 30 2 0 0.5 –1 1 15
– 18.75 0 12.5 10 0

Table: 3.

x = 15, y = 2.5, p =0, q = 0, A1 = 0, A2 = 0 and Z = 15 × 3 + 2.5 × 2.5 = Rs. 51.75

Problem Cost per C j 3 2.5 0 0 Replacement

variable requirement x y p q ratio

y 2.5 2.5 0 1 – 0.375 0.25 —

x 3 15 1 0 – 0.25 – 0.5 —
NER 0 0 0.1875 0.875

Optimal Cost = 15 × 3 + 2.5 × 2.5 = Rs. 51.25 /–

Imputed value = 0.1875 × 40 + 0.875 × 50 = Rs. 51.25/–
58 Operations Research

CERTAIN IMPORTANT POINTS TO BE REMEMBERED WHILE SOLVING LINEAR

PROGRAMMING PROBLEMS BY SIMPLEX METHOD:
1. In the given inequalities, there should not be any negative element on right hand
side (bi ≥ 0). If any bi is negative, multiply the inequality by –1 and change the
inequality sign.
2. Sometimes, the objective function may be maximisation type and the inequalities
may be ≥ type. In such cases, multiply the objective function by –1 and convert
it into minimisation type and vice versa.
3. While selecting, the incoming variable, i.e., key column, in maximisation case,
we have to select the highest positive opportunities cost and in minimisation
case, select the highest element with negative sign (smallest element). While
doing so, sometimes you may find the highest positive element in maximisation
case or lowest element in minimisation case falls under the slack variable in
maximisation case or under surplus variable in minimisation case. Do not worry.
As per rule, select that element and take the column containing that element as
key column.
4. Some times the columns of non-basis variables (decision variables) may have
their net evaluation elements same. That is the net evaluation elements are
equal. This is known as a TIE in Linear Programming Problem. To break the
time, first select any one column of your choice as the key column. In the next
table, everything will go right.
5. While selecting the out going variable i.e., key row, we have to select limiting
ratio (lowest ratio) in net evaluation row. In case any element of key column is
negative, the replacement ratio will be negative. In case it is negative, do not
consider it for operation. Neglect that and consider other rows to select out going
variable.
6. Sometimes all the replacement ratios for all the rows or some of the rows may be
equal and that element may be limiting ratio. This situation in Linear Programming
Problem is known as DEGENERACY. We say that the problem is degenerating.
When the problem degenerate, the following precautions are taken to get rid of
degeneracy.
(a) Take any one ratio of your choice to select key row or out going variable. If
you do this, there is a possibility that the problem may cycle. Cycling means,
after doing many iterations, you will get the first table once again. But it
may not be the case all times.
(b) Select the variable, whose subscript is small. Say S1 is smaller than S2 and S2
is smaller than S3 or X1 is smaller than X2 an so on or x is smaller than y or
‘a’ is smaller than ‘b’ and so on.
(c) If we do above two courses of action, we may encounter with one problem.
That one of the remaining variable in the next table (the one corresponding
to the tied variable that was not considered) will be reduced to a magnitude
of zero. This causes trouble in selecting key column in the next table.
Linear Programming Models : Solution by Simplex Method 59

(d) Identify the tied variable or rows. For each of the columns in the identity
(starting with the extreme left hand column of the identity and proceeding
one at a time to the right), compute a ratio by dividing the entry in each tied
row by the key column number in that row.
Compare these ratios, column by column, proceeding to the right. The first
time the ratios are unequal, the tie is broken. Of the tied rows, the one in
which the smaller algebraic ratio falls is the key row.
(e) If the ratios in the identity do not break the tie, form similar ratios in the
columns of the main body and select the key row as described in (d) above.
The application of the above we shall see when we deal with degeneracy
problems.
7. While solving the linear programming problems, we may come across a situation
that the opportunity cost of more than one non- basic variables are zero, then we
can say that the problem has got ALTERNATE SOLUTIONS.
8. If in a simplex table only one unfavourable Cj – Zj identifying the only incoming
variable and if all the elements of that column are either negative elements or
zeros, showing that no change in the basis can be made and no current basic
variable can be reduced to zero. Actually, as the incoming arable is introduced,
we continue to increase, without bounds, those basic variables whose ratios of
substitutions are negative. This is the indication of UNBOUND SOLUTION.
9. In a problem where, the set of constraints is inconsistent i.e., mutually exclusive,
is the case of NO FEASIBLE SOLUTION. In simplex algorithm, this case will
occur if the solution is optimal (i.e., the test of optimality is satisfied) but some
artificial variable remains in the optimal solution with a non zero value.

3.5. WORKED OUT PROBLEMS

Example 3.3. A company manufactures two products X and Y whose profit contributions are Rs.
10 and Rs. 20 respectively. Product X requires 5 hours on machine I, 3 hours on machine II and
2 hours on machine III. The requirement of product Y is 3 hours on machine I, 6 hours on
machine II and 5 hours on machine III. The available capacities for the planning period for
machine I, II and III are 30, 36 and 20 hours respectively. Find the optimal product mix.
Solution: The given data:

Products
Machine (Time required in hours) Availability in hours
X Y

I 5 3 30
II 3 6 36
III 2 5 20

Profit per unit in Rs. 10 20

60 Operations Research

Inequalities: Simplex format:

Maximize Z = 10x + 20y s.t. Maximize Z = 10x + 20y + 0S1 + 0S2 + 0S3 s.t.
5x + 3 y ≤ 30 5x + 3y + 1S1 + 0S2 + 0S3 = 30
3x + 6y ≤ 36 3x + 6y + 0S1 + 1S2 + 0S3 = 36
2x + 5y ≤ 20 and 2x + 5y + 0S1 + 0S2 + 1 S3 = 20 and
Both x and y are ≥ 0. x, y, S1, S2 and S3 all ≥ 0.
Table: I. x = 0, y = 0 S1 = 30, S2 = 36 and S3 = 20, Z = Rs. 0

Problem Profit Capacity C j =10 20 0 0 0 Replacement

variable in Rs. x y S1 S2 S3 ratio

S1 0 30 5 3 1 0 0 30/3 = 10
S2 0 36 3 6 0 1 0 36/6 = 6
S3 0 20 2 5 0 0 1 20/5 = 4
Opportunity 10 20 0 0 0
cost.

Step 1: For the first table the net evaluation row elements are same as profit row element. For
successive rows, the key column element is given by Profit – (Sum of key column element
x respective object row element). For maximization problem, select the highest element
among the key column element and mark an arrow as shown to indicate incoming variable.
For minimization problems select the lowest element or highest element with negative sign
and write an arrow to indicate the incoming variable. Enclose key column elements in a
rectangle. Here they are shown in red colour. It is known as key row because it gives the
clue about incoming variable.
Step 2: Divide the capacity column numbers with respective key column number to get the replacement
ratio. Select the lowest ratio as the indicator of out going variable. The lowest ratio is also
known as limiting ratio. In the above table the limiting ratio elements are 10, 6, 4. We select
4 as the indicator of outgoing variable. It is because, if we select any other number the third
machine cannot compete more than 4 units of product. Though the machine has got capacity
to manufacture 10 units and second machine has got capacity of 6 units, only 4 units can be
manufactured completely. Rest of the capacity of machine 1 and 2 will become idle resource.
Enclose the elements of key row in a rectangle. It is known as key row because it gives the
clue about out going variable. Mark this row with a tick mark. (Here the elements are
marked in thick.
Step 3: The element at the intersection of key row and key column is known as Key number, in the
table it is marked in bold thick number. It is known as the key number, because, the next
table is written with the help of this key element.
Linear Programming Models : Solution by Simplex Method 61

Table: II. x = 0, y = 4, S1 = 18, S2 = 12, S3 = 0. Z = Rs. 4 × 20 = Rs. 80.

Problem Profit Capacity Cj = 10 20 0 0 0 Replacement

variable in Rs. x y S1 S2 S3 ratio

S1 0 18 3.8 0 1 0 – 0.6 18/3.8 = 4.8

S2 0 12 0.6 0 0 1 – 1.2 12/0.6 = 2.0
y 20 4 0.4 1 0 0 0.2 4/0.4 = 10
Opportunity 2 0 0 0 –4
cost.

Step 4: To improve the progeramme or to get the new table the following procedure is adopted:
(i) Transfer of key row of old tableau: Divide all the elements of key row of old tableau by
key number, which gives the key row elements of the new tableau.
(ii) To transfer non key rows: New row number = old row number – (corresponding key
row number × fixed ratio.)
Here, fixed ratio = (key column number of the row/key number).
While transferring remembers you should not alter the positions of the rows. Only incoming
variable replaces the outgoing slack variable or any other outgoing basis variable as the case may be.
The net evaluation row element of the variable entered into the programme will be zero. When all
the variables are transferred, the identity matrix will come in the position of main matrix.
To check whether, the problem is done in a correct manner, check that whether profit earned at
present stage is equal to shadow price at that stage. Multiplying the net evaluation row element under
non-basis can get shadow price variable (identity matrix) by original capacities of resources given in
the problem.
Above explained procedure of transferring key row and non-key rows is worked out below:
Transfer of Key row: 20 / 5 = 4, 2 / 5 = 0.4, 5 / 5 = 1, 0 /5 = 0, 0 / 5 = 0, and 1 / 5 = 0.2.
Transfer of non-key rows:
30 – 20 × 3/5 = 18 36 – 20 × 6 / 5 = 12
5 – 2 × 3/5 = 3.8 3 – 2 × 1.2 = 0.6
3 – 5 = 3/5 = 0 6 – 5 × 1,2 = 0
1 – 0 × 3/5 = 1 0 – 0 × 1.2 = 0
0 – 0 × 3/5 = 0 1 – 0 × 1.2 = 1
0 – 1 × 3/5 = – 0.6 0 – 1 × 1.2 = – 1.2
Step 5: Once the elements of net evaluation row are either negatives or zeros (for maximization
problem), the solution is said to be optimal. For minimization problem, the net evaluation
row elements are either positive elements or zeros.
As all the elements of net evaluation row are either zeros or negative elements, the solution is
optimal. The company will produce 4.8 units of X and 3.6 units of Y and earn a profit of Rs. 120/–.
Shadow price is 2.6 × 30 + 2 × 20 = Rs. 128/–. The difference of Rs. 8/– is due to decimal values,
which are rounded off to nearest whole number.
62 Operations Research

Table: III. x = 4.8, y = 3.6, S2 = 9, S1 = 0, S3 = 0 and Z = 4.8 × 10 + 3.6 × 20 = Rs. 120/–

Problem Profit Capacity Cj = 10 20 0 0 0 Replacement

variable in Rs. x y S1 S2 S3 ratio

x 10 4.8 1 0 0.26 0 – 0.16

S2 0 9 0 0 – 0.16 1 – 1.1
y 20 3.6 0 1 0 0 0.18
Opportunity 0 0 – 2.6 0 –2.0
cost.

Problem. 3.4: A company manufactures three products namely X, Y and Z. Each of the product
require processing on three machines, Turning, Milling and Grinding. Product X requires 10 hours of
turning, 5 hours of milling and 1 hour of grinding. Product Y requires 5 hours of turning, 10 hours of
milling and 1 hour of grinding, and Product Z requires 2 hours of turning, 4 hours of milling and 2
hours of grinding. In the coming planning period, 2700 hours of turning, 2200 hours of milling and 500
hours of grinding are available. The profit contribution of X, Y and Z are Rs. 10, Rs.15 and Rs. 20 per
unit respectively. Find the optimal product mix to maximize the profit.
Solution: The given data can be written in a table.

Product
Machine Time required in hours per unit Available hours
X Y X
Turning. 10 5 2 2,700
Milling 5 10 4 2,200
Grinding. 1 1 2 500
Profit contribution in 10 15 20
Rs. per unit.

Let the company manufacture x units of X, y units of Y and z units of Z

Inequalities: Equations:
Maximise Z = 10x + 15 y + 20 z S.T. Maximise Z = 10x +15y +20x S.T
10 x+ 5y + 2z ≤ 2,700 10x + 5y + 2z +1S1 = 2700
5x + 10y + 4z ≤ 2,200 5x + 10y + 4z + 1S2 = 2200
1x + 1y + 2z ≤ 500 and 1x + 1y + 2z + 1S3 = 500 and
All x, y and z are ≥ 0 x, y and z all ≥ 0
Simplex format:
Maximise Z = 10x + 15y + 20z + 0S1 + 0S2 + 0S3 S.t.
10x +5y + 2z + 1S1 + 0S2 + 0S3 = 2700
Linear Programming Models : Solution by Simplex Method 63

5x + 10y + 4z + 0S1 + 1S2 + 0S3 = 2200

1x + 1y + 2z + 0S1 + 0S2 + 1S3 = 500
And all x, y, z, S1, S2, S3 are ≥ 0

Table I. x = 0, y = 0, z = 0, S1 = 2700, S2 = 2200, S3 = 500. Profit Z = Rs. 0

Programme Profit Capacity C j =10 15 20 0 0 0 Replacement Check

x y z S1 S2 S3 ratio column.
S1 0 2700 10 5 2 1 0 0 2700/2 = 13500 2718
S2 0 2200 5 10 4 0 1 0 2200/4 = 550 2220
S3 0 500 1 1 2 0 0 1 500/2 = 250 505
Net 10 15 20 0 0 0
evaluation

{Note: The check column is used to check the correctness of arithmetic calculations. The check
column elements are obtained by adding the elements of the corresponding row starting from capacity
column to the last column (avoid the elements of replacement ration column). As far as treatment of
check column is concerned it is treated on par with elements in other columns. In the first table add the
elements of the row as said above and write the elements of check column. In the second table
onwards, the elements are got by usual calculations. Once you get elements, add the elements of
respective row starting from capacity column to the last column of identity, then that sum must be
equal to the check column element.}
Table: II. x = 0, y = 0, z = 250 units, S1 = 2200, S2 = 1200, S3 = 0 and Z = Rs. 20 × 250
= Rs. 5,000.

Programme Profit Capacity C j =10 15 20 0 0 0 Check Replacement

x y z S1 S2 S3 column. ratio

S1 0 2210 9 4 0 1 0 –1 2213 552.5

S2 0 1200 3 8 0 0 1 –2 1210 150
Z 20 250 0.5 0.5 1 0 0 0.5 500 500
Net 0 5 0 0 0 – 10
Evaluation.

Profit at this stage = Rs. 20 × 250 = Rs. 5,000 and Shadow price = 10 × 500 = Rs. 5000.
64 Operations Research

Table: III. x = 0, y = 150, z = 174.4, S1 = 1600, S2 = 0, S3 = 0 and Z = Rs. 5738/–

Programme Profit Capacity C j =10 15 20 0 0 0 Check Replacement

x y z S1 S2 S3 column. Ratio
S1 0 1600 7.5 0 0 1 –0.5 0 1608
Y 15 150 0.375 1 0 0 0.125 –0.25 151.25
Z 20 174.4 0.311 0 1 0 –0.063 0.626 423.7

Net Evn. –1.85 0 0 0 –0.615 – 8.77

As all the elements of Net evaluation row are either zeros or negative elements, the solution is
optimal. The firm has to produce 150 units of Y and 174.4 units of Z. The optimal profit = 15 × 150 +
20 × 174.4 = Rs.5738 /–
To check the shadow price = 0.615 × 2200 + – 8.77 × 500 = 1353 + 4385 = Rs. 5738 /–.
Problem 3.5: A company deals with three products A, B and C. They are to be processed in
three departments X, Y and Z. Products A require 2 hours of department X, 3 hours of department Y
and product B requires 3 hours,2 hours and 4 hours of department X, Y and Z respectively. Product C
requires 2 hours in department Y and 5 hours in department Z respectively. The profit contribution of
A, B and C are Rs. 3/–, Rs.5/– and Rs. 4/– respectively. Find the optimal product mix for maximising
the profit. In the coming planning period, 8 hours of department X, 15 hours of department Y and 10
hours of department Z are available for production.

The Data Product

Departments Hours required per unit Available capacity in hours
A B C
X 2 3 0 8
Y 3 2 4 15
Z 0 2 5 10

Profit per unit in Rs.. 3 5 4

Inequalities: Equations.
Maximise Z = 3a +5b + 4c s.t. Maximise Z = 3a + 5b + 4c + 0S1 + 0S2 + 0S3 s.t.
2a + 3b + 0c ≤ 8 2a + 3b + 0c + 1S1 + 0S2 + 0S3 = 8
3a + 2b + 4c ≤ 15 3a + 2b + 4c + 0S1 + 1S2 + 0S3 = 15
0a + 2b + 5c ≤ 10 and 0a + 2b + 5c + 0S1 + 0S2 + 1S3 = 10 and
a, b, and c all ≥ 0 a, b, c, S1, S2, S3 all ≥ 0.
Linear Programming Models : Solution by Simplex Method 65

Table: I. a = 0, b = 0, c = 0 , S1 = 8, S2 = 15 and S3 = 10 and Z = Rs.0

Programme Profit Capacity Cj = 3 5 4 0 0 0 Check Replacement

a b c S1 S2 S3 column. Ratio
S1 0 8 2 3 0 1 0 0 14 2.6
S2 0 15 3 2 4 0 1 0 25 7.5
S3 0 10 0 2 5 0 0 1 18 5

[Link]. 3 5 4 0 0 0

Table: II. a = 0, b = 2.6, c = 0, S1 = 0, S2 = 9.72, S3 = 4.72 and Z = Rs.5 × 2.6 = Rs. 13/–

Programme Profit Capacity Cj = 3 5 4 0 0 0 Check Replacement

a b c S1 S2 S3 column. Ratio
B 5 2.6 0.66 1 0 0.33 0 0 4.6 —
S2 0 9.72 1.68 0 4 0.66 1 0 15.76 2.43
S3 0 4.72 – 1.32 0 5 – 0.66 0 1 8.76 0.944

[Link] 3 0 4 – 1.65 0 0

Note: as the key column element is equal to 0 for column under ‘c’ replacement ratio is not
calculated, as it is equals to zero.
Profit at this stage is Rs. 2 × 2.6 = Rs. 13/ –. Shadow price = 1.65 × 8 = Rs. 13 / –.
Table: III. a = 0, b = 2.6, c = 0.944, S2 = 5.94, S1 = 0, S3 = 0, Profit Z = Rs. 5 × 2.6 + 4 × 0.944
= Rs. 16.776 /–

Programme Profit Capacity Cj = 3 5 4 0 0 0 Check Replacement

a b c S1 S2 S3 column. ratio
B 5 2.6 0.66 1 0 0.33 0 0 4.6
S2 0 5.94 2.74 0 0 1.19 1 – 0.8 8.72
C 4 0.944 – 0.264 0 1 – 0.132 0 0.2 1.792
Net Evn. – 8.64 0 0 – 1.122 0 – 0.8 —

As the elements of net evaluation row are either zeros or negative elements, the solution is optimal.
Now the optimal profit = Z = Rs. 5 × 2.6 + Rs. 4 × 0.944 = Rs. 16.776. The company manufactures
2.6 units of B and one unit of C.
The shadow price = 1.122 × 8 + 0.8 × 10 = Rs. 17.976. The small difference is due to decimal
calculations. Both of them are approximately equal hence the solution is correct.
66 Operations Research

Problem 3.6: A company manufactures two types of products X and Y on facilities, A, B, C, D,

E, and F having production capacities as under.
Facilities. Production capacity to produce
A 100 of X OR 150 of Y
B 80 of X OR 80 of Y
C 100 of X OR 200 of Y
D 120 of X OR 90 of Y
E 60 of X only (Testing facility for product X)
F 60 of Y only.(Testing facility for product Y)
If the profit contribution of product X is Rs.40/– per unit and that of Y is Rs. 30 per unit, find the
optimal product mix for maximising the profit.
Solution:
Let the company manufactures x units of X and y units of Y.
Each unit of product ‘x’ uses1/100 capacity of A therefore capacity A used by product ‘x’ is (A
/ 100) x.
Similarly capacity of A used by ‘y’ is (A / 150) y. Therefore,
(A/100) x + (A/150) y ≤ A i.e., 150 x + 100 y ≤ 15000 OR 3 x + 2 y ≤ 300 is the equation for A.
Similarly equations for other facilities can be written. Objective function is to Maximise Z = 40x + 30y
Inequalities: Simplex version:
Maximise Z = 40x + 30y s.t. Maximise Z = 40x + 30y + 0S1 + 0S2 + 0S3 + 0S4 + 0S5 + 0S6 s.t.
3x + 2y ≤ 300 3x + 2y + 1S1 + 0S2 + 0S3 + 0S4 + 0S5 + 0S6 = 300
1x + 1y ≤ 80 1x + 1y + 0S1 + 1S2 + 0S3 + 0S4 + 0S5 + 0S6 = 80
2x + 1y ≤ 200 2x + 1y + 0S1 + 0S2 + 1S3 + 0S4 + 0S5 + 0S6 = 200
3x + 4y ≤ 360 3x + 4y + 0S1 + 0S2 + 0 S3 + 1S4 +0S5 + 0S6 = 360
1x + 0y ≤ 60 1x + 0y + 0S1 + 0S2 + OS3 + 0S4 + 1S5 + 0S6 = 60
0x + 1y ≤ 60 and 0x + 1y + 0S1 + 0S2 + 0S3 + 0S4 + 0S5 + 1S6 = 60 and
Both x and y are ≥ 0. All x, y, Si (i = 1,2,3,4,5,6) are ≥ 0

Table: I. x = 0, y = 0, S1 = 300, S2 = 80, S3 = 200, S4 = 360, S5 = 60, S6 = 60 and Profit Z = Rs.0

Prog. Profit Capacity Cj = 40 30 0 0 0 0 0 0 Check Replacement

x y S1 S2 S3 S4 S5 S6 Col. ratio.
S1 0 300 3 2 1 0 0 0 0 0 306 102
S2 0 80 1 1 0 1 0 0 0 0 83 80
S3 0 200 2 1 0 0 1 0 0 0 203 100
S4 0 360 3 4 0 0 0 1 0 0 367 120
S5 0 60 1 0 0 0 0 0 1 0 61 60
S6 0 60 0 1 0 0 0 0 0 1 61 Infinity.
Net Evaluation 40 30 0 0 0 0 0 0 —
Linear Programming Models : Solution by Simplex Method 67

Table: III. x = 60, y = 20, S1 = 80, S2 = 0, S3 = 60, S4 = 100, S5 = 0, S6 = 40, Profit: Rs. 40 × 60
+ 30 × 20 = Rs. 3000 /–

Prog. Profit Capacity Cj = 40 30 0 0 0 0 0 0 Check Replacement

x y S1 S2 S3 S4 S5 S6 col. ratio.
S1 0 80 0 0 1 –2 0 0 –1 0 78
y 30 20 0 1 0 1 0 0 –1 0 21
S3 0 60 0 0 0 –1 1 0 –1 0 60
S4 0 100 0 0 0 –4 1 1 1 0 99
X 40 60 1 0 0 0 0 0 1 0 59
S6 0 40 0 0 0 –1 0 0 1 1 39
Net Evaluation 0 0 0 –30 0 0 – 10 0 —

As all the elements of net evaluation row are either negative elements or zeros the solution is
optimal. The company will produce 60 units of X and 20 units of Y and the optimal Profit = Z Rs. 40
× 60 + Rs. 30 × 20 = Rs. 3000/–
Shadow price = 30 × 80 + 10 × 6 = 2400 + 600 = Rs. 3000 /–. Shadow price and profit are equal.
Problem 3.7: A company produces three products A, B and C by using two raw materials X and
Y. 4000 units of X and 6000 units of Z are available for production. The requirement of raw materials
by each product is given below:

Requirement per unit of product

Raw material
A B C
X 2 3 5

Y 4 2 7

The labour time for each unit of product A is twice that of product B and three times that of
product C. The entire labour force of the company can produce the equivalent of 2500 units of product
A. A market survey indicates the minimum demand of the three products are 500, 500 and 375 respectively
for A, B and C. However, their ratio of number of units produced must be equal to 3: 2: 5. Assume that
the profit per units of product A, B and C are Rupees 60/–, 40/– and 100 respectively. Formulate the
L.P.P. for maximizing the profit.
Solution:
Let the company manufactures a units of A, b units of B and c units of C. The constraints for
raw materials are
2a + 3b + 5c ≤ 4000 …(1)
4a +2b + 7c ≤ 6000 ...(2)
Now let 't' be the labour time required for one unit of product A, then the time required for per
unit of product B is t/2 and that for product C is t/3. As 2500 units of A are produced, the total time
available is 2500 t. Hence the constraints for time are:
ta + t/2 b + t/3 c ≤ 2500t
68 Operations Research

i.e., a + 1/2 t b + 1/3 t c ≤ 2500 ...(3)

Now the market demand constraints are: a ≥ 500, b ≥ 500 and c ≥ 375 ...(4)
As the ratio of production must be 3 : 2 : 5,
A = 3k. b = 2k and c = 5k which gives the equations:
1/3 a = 1/2/ b and 1/2 b = 1/5 c …(5)
The objective function is Maximise Z = 60 a + 40 b + 100 c
Hence the Linear programme in the form inequalities for the above problem is:
Maximise Z = 60 a + 40b + 100c s.t.
2a + 3b + 5c ≤ 4000
4a + 2b + 7c ≤ 6000
1a + 1/2 b + 1/3 c ≤ 2500
1/3 a = 1/3 b
½ b = 1/5 c
a ≥ 500, b ≥ 500 and c ≥ 375
As the last constraint shows that the values of all the variables are ≥ 0, the same constraint will
become non-negativity constraint.
Problem 3.8: A product consists of two components A and B. These components require two
different raw materials X and Y. 100 units of X and 200 units of Y are available for production. The
materials are processed in three departments. The requirement of production time in hours and materials
in units are given in the table below.
Departments Raw material input Output of components
per run in unit per run (units)
X Y A B
1 7 5 6 4
2 4 8 5 8
3 2 7 7 3

Formulate a progrmme to determine the number of production runs for each department, which
will maximise the total number of components A and B for the product.
Solution: Formulation of L.P.P
Let a, b and c is the production runs for departments 1, 2 and 3 respectively. Therefore the total
production is 6a + 5b + 7c of components of A and 4a + 8b + 3c of components B.
The raw material restrictions are :
7a + 4b + 2c ≤ 100
5a + 8b + 7c ≤ 200
Now the final product requires 4 units of A and 3 units of B for assembly. Hence the total production
of final product will be the smaller of the quantities: 1/4 (6a + 5b + 7c) and 1/3 (4a + 8b + 3c).
Our objective is to maximize the production of final product. Hence the objective function would be:
Maximise Z = Minimum of {¼ (6a + 5b + 7c), 1/3 (4a + 8b + 3c)}
Let Minimum {1/4 (6a + 5b + 7c), 1/3 (4a + 8b + 3c)} = v i.e.
Linear Programming Models : Solution by Simplex Method 69

¼ (6a + 5b + 7c) ≥ v and 1/3 (4a + 8b + 3c) = v

Then the required l.p.p. is : Find a, b, c and v which maximise Z = v , subject to the constraints
7a + 8b + 2c ≤ 100
5a + 8b + 7c ≤ 200
6a + 5b + 7c – 4v ≥ 0
4a + 8b + 3c – 3v ≥ 0, and
a, b, c and v all ≥ 0.
Problem 3.9: A firm manufactures three types of coils each made of a different alloy. The flow
process chart is given in the figure below. The problem is to determine the amount of each alloy to
produce, within the limitations of sales and machine capacities, so as to maximise the profits.

The further data given is:

Table: I.

Machine Number of machines 8–hour shift per week Down time %

A 3 18 5
B 2 26 10
C 1 22 0

Table: II.

Alloy Operation Machine rate Sales potential Profit per ton in Rs.

A 30 hours/30 ton
C (1) 40 feet per minute 1500 tons per month. 80
1 B 25 feet per minute
C (2) 30 feet per minute
A 25 hours per 10 tons
2 B 25 feet per minute 800 tons per month. 400
C 30 feet per minute
3 B 15 feet per minute 1000 tons per 250
C 20 feet per minute
70 Operations Research

Coils for each alloy are 400 feet long and weigh 5 tons. Set up objective function and restrictions
to set up matrix.
Solution: Let the company produce ‘a’ units of alloy 1, b units of alloy 2 and c units of alloy 3.
Then the objective functions is Maximise Z = 80a + 400b + 250c subject to the limitations imposed by
the available machine capacity and sales potential.
Constraint of Machine Capacity per month:
Table: III.

Process Number shifts % of useful Capacity in

of machines per week time hours per month
A 3 18 95 3 × 18 × 8 × 4 ½ × 0.95 = 1778.3
B 2 26 90 2 × 26 × 8 × 4 ½ × 0.90 = 1662.4

C 1 22 100 1 × 22 × 8 × 4 ½ × 1 = 726.7

To convert machine rates into tons per hour:

Table: IV.

Alloy Process Machine rates

A 30 hour per 10 tons = 0.333 tons per hour.
1 C (1) 40 feet per minute = (40 × 60 × 5)/400 = 30 tons per hour.
B 25 feet per minute = (25 × 60 × 5)/400 = 18.75 tons per hour.
C (2) 30 feet per minute = (30 × 60 × 5)/400 = 22.5 tons per hour.
A 25 hours per 10 tons. = 0.4 tons per hour.
2 B 25 feet per minute = (25 × 60 × 5)/400 = 18.75 tons per hour.
C 30 feet per minute = (30 × 60 × 5)/400 = 22.5 tons per hour.
3 B 15 feet per minute = (15 × 60 × 5)/400 = 11.25 tons per hour.
C 20 feet per minute = (20 × 60 × 5)/400 = 15 tons per hour.

Now let us calculate the times required for a, b and c tons of alloys and use these machine times
for a formal statement of capacity constraints.
For process A : (a/0.333) + (b/0.4) ≤ 1778.3
For process B : (a/18.75) + (b/18.75) + (c/11.25) ≤ 1662.4
For process C : (a/30) + (a/22.5) + (b/22.5) + (c/15) ≤ 762.7 OR
(7a/90) + (b/22.5) + 1c ≤ 762.7
Limitations imposed by sales potential:
a ≤ 1500, b ≤ 800 and c ≤ 1000.
Hence the l.p.p is:
Maximise Z = 80a + 400 b + 250c s.t.
(a/0.333) + (b/0.4) ≤ 1778.3
(a / 18.75) + (b/18.75) + (c / 11.25) ≤ 1662.4
Linear Programming Models : Solution by Simplex Method 71

(7a/90) + (b/22.5) + 1c ≤ 762.7

a ≤ 1500, b ≤ 800 and c ≤ 1000. And
a,b and c all ≥ 0

3.7. MINIMISATION PROBLEMS

Problem 3.10: A small city of 15,000 people requires an average of 3 lakhs of gallons of water
daily. The city is supplied with water purified at a central water works, where water is purified by
filtration, chlorination and addition of two chemicals softening chemical X and health chemical Y.
Water works plans to purchase two popular brands of products, product A and product B, which
contain these two elements. One unit of product A gives 8 Kg of X and 3 Kg of Y. One unit of product
B gives 4 Kg of X and 9 Kg of Y. To maintain the water at a minimum level of softness and meet a
minimum in health protection, it is decided that 150 Kg and 100 Kg of two chemicals that make up each
product must be added daily. At a cost of Rs. 8/– and Rs. 10/– per unit respectively for A and B, what
is the optimum quantity of each product that should be used to meet consumer standard?
Before discussing solution, let us have an idea of what is known as Big M–n Method, which is
generally used to solve minimization problems.
While solving the linear programming problems by graphical method, we have seen an isoprofit
line is drawn and at the origin and then it is moved away from the origin to find the optima point.
Similarly an isocost line is drawn away from the origin in minimization problem and moved towards the
origin to find the optimal point.
But in simplex method of solving the minimization problem, a highest cost is allocated to artificial
surplus variable to remove it form the matrix. This high cost is Big – M. M stands for millions of
rupees. If we use big M some times we feel it difficult while solving the problem. Hence, we can
substitute a big numerical number to M, which is bigger than all the cost coefficients given in the
problem. This may help us in numerical calculations.
Solution: Let the water works purchase x units of X and y units of Y, then:
Inequalities: Simplex Format:
Minimise Z = 8x + 10y s.t Minimise Z = 8x + 10y + 0p + 0q + M A1 + MA2 s.t.
3x + 9y ≥ 100 3x + 9y – 1p + 0q + 1A1 + 0A2 = 100
8x + 4y ≥ 150 and 8x + 4y + 0p – 1q + 0A1 + 1A2 = 150 and
Both x and y ≥ 0 x, y, p, q, A1, A2 all ≥ 0
Table: I. x = 0, y = 0, p = 0, a = 0, A1 = 100, A2 = 150 and Z = 100M + 150M = Rs. 250 M.

Programe Cost in Cj = 8 10 0 0 M M Replacement

Rs. requirement x y p q A1 A2 ratio

A1 M 100 3 9 –1 0 1 0 100/9 = 11.11

A2 M 150 8 4 0 –1 0 1 150 / 4 = 37.5

Net Evaluation 8 – 11M 10 – 13M M M 0 0 —

72 Operations Research

Table II: x = 0, y = 11.11, p = 0, q = 0, A1 = 1.32, A2 = 0 , Z = Rs. 11.11 x 10 + 1.32M =

111.1 +1.32 M

Program Cost in Cj = 8 10 0 0 M M Replacement

Rs. requirement x y p q A1 A2 ratio

y 10 11.1 0.33 1 –0.11 0 0.11 0 33.6

A2 M 106 6.88 0 0.44 –1 –0.44 1 15.4

Net Evaluation 4.3–6.88M 0 –1.1+0.44M M –1.1+5.4M 0

Table III. x = 0.5, y = 15.4, p = 0, q = 0 , A1 = 0, A2 = 0, Z = Rs. 10 x 0.50 + 8 x 15.4 =

Rs. 128.20

Program Cost in Cj = 8 10 0 0 M M Replacement

Rs. requirement x y p q A1 A2 ratio
y 10 0.5 0 1 –0.154 0.1 0.154 –0.1 —
x 8 15.4 1 0 0.06 –0.14 –0.06 0.14 —

Net Evaluation — 0 0 1.062 0.12 M–1.06 M– 0.12

Water works purchases 0.5 Kg of Y and 15.4 Kg of X at a cost of Rs. 128.20. The shadow price
will be Rs. 107/–. The difference is due to decimal numbers. (Note: We can avoid the artificial variables
as and when they go out to reduce the calculations. We can use a numerical value for M, which is
higher than the cost of variables given in the problem so that we can save time.).
Problem 3.11: 10 grams of Alloy A contains 2 grams of copper, 1 gram of zinc and 1 gram of
lead. 10 grams of Alloy B contains 1 gram of copper, 1 gram of zinc and 1 gram of lead. It is required
to produce a mixture of these alloys, which contains at least 10 grams of copper, 8 grams of zinc, and
12 grams of lead. Alloy B costs 1.5 times as much per Kg as alloy A. Find the amounts of alloys A and
B, which must be mixed in order to satisfy these conditions in the cheapest way.
Solution: The given data is: (Assume the cost of Alloy A as Re.1/– then the cost of Alloy B will
be Rs. 1.50 per Kg.

Alloys
(In grams per 10 grams) Requirement in Grams
Metals
A B
Copper 2 1 10
Zinc 1 1 8
Lead 1 1 12
Cost in Rs. per Kg. 1 1.5

Let the company purchase x units of Alloy A and y units of Alloy B. (Assume a value of 10 for M)
Linear Programming Models : Solution by Simplex Method 73

Inequalities: Simplex Format:

Minimise Z = 1x + 1.5 y s.t. Minimise Z = 1x + 1.5y + 0p + 0q + 0r + 10A1 + 10 A2 +10A3 s.t.
2x + 1y ≥ 10 2x + 1y – 1p + 0q + 0r + 1A1 + 0A2 + 0A3 = 10
1x + 1y ≥ 8 1x + 1y + 0p – 1q + 0r + 0A1 + 1A2 + 0A3 = 8
1x + 1y ≥ 12 and 1x + 1y + 0p + 0q – 1r + 0A1 + 0A2 + 1A3 = 12 and
x, y both ≥ 0 x, y, p, q, r, A1, A2, A3 all ≥ 0

Table: I. x = 0, y = 0 , p = 0, q = 0, r = 0, A1 = 10, A2 8 and A3 = 12 and Profit Z = Rs. 10 × 10 +

10 × 8 + 10 × 12 = Rs. 300

Cost in Cj = 1 1.5 0 0 0 10 10 10 Replace–

Program Rs. Require– x y p q r A1 A2 A3 ment ratio
ment

A1 10 10 2 1 –1 0 0 1 0 0 5
A2 10 8 1 1 0 –1 0 0 1 0 8
A3 10 12 1 1 0 0 –1 0 0 1 12
Net Evaluation –39 –28.5 10 10 10 0 0 0

Table: II. x = 5, y = 0, p = 0, q = 0, r = 0, A1 = 0, A2 = 3, A3 = 7 and Z = Rs. 1 × 5 + 10 x 3 + 7 ×

10 = Rs. 105/–

Cost in Cj = 1 1.5 0 0 0 10 10 10 Replace–

Program Rs. Require– x y p q r A1 A2 A3 ment ratio
ment
x 1 5 1 0.5 – 0.5 0 0 0.5 0 0 – 10 (neglect)
A2 10 3 0 0.5 0.5 –1 0 – 0.5 1 0 6
A3 10 7 0 0.5 0.5 0 –1 – 0.5 0 1 14
Net Evaluation 0 –9 –9.5 10 10 19.5 0 0

Table: III. x = 8, y = 0, p = 6, q = 0, r = 0, A1 = 0, A2 = 0, A3 = 4, Z = Rs. 8 + 40 = Rs. 48 /–

Cost in Cj = 1 1.5 0 0 0 10 10 10 Replace–

Program Rs. Require– x y p q r A1 A2 A3 ment ratio
ment

x 1 8 1 1 0 –1 0 0 1 0 – 8 (neglect)
p 0 6 0 1 1 –2 0 –1 2 0 – 3 (neglect)
A3 10 4 0 0 0 1 –1 0 –1 1 4
Net Evaluation 0 0.5 0 –9 10 10 19 0
74 Operations Research

Table: IV. x = 12, y = 0, p = 14, q = 4, r = 0, Z = Rs. 1 × 12 = Rs. 12/ –

Cost in Cj = 1 1.5 0 0 0 10 10 10 Replace–

Program Rs. Require– x y p q r A1 A2 A3 ment ratio
ment

x 1 12 1 1 0 0 –1 0 0 1
p 0 14 0 1 1 0 –2 –1 1 2
q 0 4 0 0 0 1 –1 0 –1 1
Net Evaluation 0 0.5 0 0 1 0 0 9

As all the net evaluation elements are either zeros or positive element, the solution is optimal. The
company can purchase 12 units of X at a cost of Rs. 12/–
Problem 3.12: Minimise Z = 4a + 2b s.t.
3a + 1b ≥ 27
–1a – 1b ≤ –21
1a + 2b ≥ 30 and both a and b are ≥ 0.
Inequalities: Equations:
Minimise Z = 4a + 2b ≥ 27 Minimise Z = 4a + 2b + 0p +0q + 0r + MA1 + MA2 + MA3 s.t.
3a + 1b ≥ 27 3a + 1b – 1p + 0q + 0r + 1A1 + 0A2 + 0A3 = 27
1a + 1b ≥ 21 1a + 1b + 0p – 1q + 0r + 0A1 + 1A2 + 0A3 = 21
1a + 2b ≥ 30 1a + 2b + 0p +– + 0q – 1r + 0A1 + 0A2 + 1A3 = 30
And a, b both ≥ 0 a, b, p, q, r, A1, A2 and A3 all ≥ 0
(Note: converting the objective function conveniently we can solve the minimization or
maximization problems. For example, if the objective function given is minimization type, we
can convert it into maximization type by multiplying the objective function by –1. For example,
in the problem 3.12, the objective function may be written as Maximise Z = –4a – 2b s.t. But the
inequalities are in the form of ≥ type. In such cases when artificial surplus variable (AI) is
introduced then the cost co–efficient of the artificial surplus variable will be –M instead of +M.
Rest of the procedure of solving the problem is same. Similarly, any maximization problem can
be converted into minimization problem by multiplying the objective function by –1. If the
inequalities are in ≥ form, subtracting the surplus variable and adding the artificial surplus
variable is done to inequalities to convert them into equations. In case the inequalities are of ≤
type, slack variable is added to convert them into equations. Let us see this in next example).

Solution: Let M be represented by a numerical value Rs.10/– that is higher than the cost coefficients
given in the problem. (i.e. 4 and 2).
Linear Programming Models : Solution by Simplex Method 75

Table: I. a = 0, b = 0, p = 0, q = 0, r = 0, A1 = 27, A2 = 21 and A3 = 30 and the cost Z = Rs. 780/–

Cost in Cj 4 2 0 0 0 10 10 10 Replace–
Program Rs. Require– a b p q r A1 A2 A3 ment ratio
ment

A1 10 27 3 1 –1 0 0 1 0 0 9
A2 10 21 1 1 0 –1 0 0 1 0 21
A3 10 30 1 2 0 0 –1 0 0 1 30
Net Evaluation – 46 – 38 10 10 10 0 0 0 —

Table: II. a = 9, b = 0, p = 0, q = 0, r = 0, A1 = 0, A2 = 12, A3 = 21 and Z = Rs. 366/–

Cost in Cj 4 2 0 0 0 10 10 10 Replace–
Program Rs. Require– a b p q r A1 A2 A3 ment ratio
ment

a 4 9 1 0.33 – 0.33 0 0 0.33 0 0 27.27

A2 10 12 0 0.67 0.33 –1 0 – 0.33 1 0 17.91
A3 10 21 0 1.67 0 0 –1 0 0 1 12.51
Net Evaluation 0 – 22.72 – 1.98 10 10 8.22 0 0

Table: III. a = 4.84, b = 12.51, p = 0, q = 0, r = 0, A1 = 0, A2 = 3.6, A3 = 0, Z = Rs. 80.50.

Cost in Cj 4 2 0 0 0 10 10 10 Replace–
Program Rs. Require– a b p q r A1 A2 A3 ment ratio
ment

a 4 4.84 1 0 – 0.33 0 – 0.198 0.33 0 0.198 Neglect

A2 10 3.6 0 0 0.33 –1 – 0.4 – 0.33 1 0.4 10.9
b 2 12.57 0 1 0 0 – 0.6 0 0 0.6 Infinity
Net Evaluation 0 0 – 1.98 10 3.592 8.02 0 4.008 —
76 Operations Research

Table: IV. a = 3, b = 18.05, p = 9, q = 0, r = 0, A1 = 0, A2 = 0, A3 =0, Cost Z = Rs. 48.10

Cost in Cj 4 2 0 0 0 10 10 10 Replace–
Program Rs. require– x y p q r A1 A2 A3 ment ratio
ment

a 4 3 1 0 – 0.485 0.5 0 0.485 – 0.5 0

p 0 9 0 0 0.425 – 2.5 1 – 0.425 0.25 – 1
b 2 18.05 0 1 0.445 – 1.5 0 – 0.445 1.5 0
Net Evaluation 0 0 1.05 1 0 8.95 9 0

As the elements of net evaluation row are either zeros or positive elements, the solution is optimal.
A = 3 and B = 12.57 and the optima cost Z = Rs. 48.10. The shadow price = 1.05 × 27 + 1 × 21
= Rs. 49.35.
The difference is due to decimal calculations.
Problem 3. 13: Solve the Minimization L.P.P. given below:
Min. Z = 1x – 3y + 2z S.t.
3x – 1y – + 3z ≤ 7
– 2x + 4y + 0z ≤ 12
– 4x + 3y + 8z ≤ 10 and x, y, and z all ≥ 0.
Solution: As the objective function is of minimization type and the constraints are of ≤ type, we
can rewrite the problem in simplex format as:
Maximize Z = –1x + 3y – 2z + 0S1 + 0S2 + 0S3 S.t.
3x – 1y + 3z + 1S1 + 0S2 + 0 S3 = 7
– 2x + 4y + 0z + 0S1 + 1S2 + 0S3 = 12
– 4x + 3y + 8z + 0S1 + 0S2 + 1S3 = 10 and x, y, z, S1, S2 and S3 all ≥ 0.

Table: I. x = 0, y = 0, z = 0, S1 = 7, S2 = 0, S3= 0 and Z = Rs.0./–

Problem Profit Capacity,= –1 3 –2 0 0 0 Replacement

variable Rs. Z units x y z S1 S2 S3 ratio
S1 0 7 3 –1 3 1 0 0 —
S2 0 12 –2 4 0 0 1 0 12/4 = 3
S3 0 10 –4 3 8 0 0 1 10/3 = 3.3.

Net Evaluation. – 1 3 –2 0 0 0
Linear Programming Models : Solution by Simplex Method 77

Table: II. x = 0, y = 3, z = 0, S1 = 10, S2 = 0, S3 =1 and Z = 3 × 3 = Rs.9.00.

Problem Profit Capacity, –1 3 –2 0 0 0 Replacement

variable Rs. Z = units x y z S1 S2 S3 ratio
S1 0 10 3.5 0 3 1 0.25 0 2.86
y 3 3 – 0.5 1 0 0 0.25 0 —
S3 0 1 – 2.5 0 8 0 – 0.75 1 —
Net Evaluation. 0.5 0 –2 0 – 0.75 0

Table III. x = 2.86, y = 4.43, z = 0, S1 = 0, S2 = 0 and S3 = 8.14, Z = Rs. 10.43.

Problem Profit Capacity, –1 3 –2 0 0 0 Replacement

variable Rs. Z = units x y z S1 S2 S3 ratio
x –1 2.86 1 0 0.86 0.29 0.07 0 —
y 3 4.43 0 1 0.43 0.143 0.285 0 —
S3 0 8.14 0 0 10.14 0.174 – 0.57 1 —
Net Evaluation. 0 0 – 2.43 – 0.14 – 0.68 0

Answer: X = 2.86, Y = 4.43, Z = 0 and Profit Z = Rs. 10.43.

3.8. MIXED PROBLEMS

As a mathematical interest, we may deal with some problems which have the characteristics of both
maximization and minimization problems. These problems may not exist in real world, but they are
significantly important as far as mathematical interest. These problems are generally known as Mixed
problems. Let us work out some problems of this nature. (by using big – M method).
Problem 3.14: Solve the following L.P.P.:
Minimize Z = 4a + 2b S.t.
3a + 1b ≥ 27
–1a –1b ≤ –21
1a + 2b ≥ 30 and both a and b are ≥ 0
The right hand side of any inequality or equation should not be negative. Hence we have to
multiply the second inequality by –1. Then the given problem becomes:
Minimize Z = 4a + 2b s.t. OR Maximize Z = – 4a – 2b s.t
3a + 1b ≥ 27 3a + 1b ≥ 27
1a + 1b ≥ 21 1a + 1b ≥ 21
1a + 2b ≥ 30 and both a and b ≥ 0 1a + 2b ≥ 30 and both a and b ≥ 0.
78 Operations Research

The simplex version of the problem is:

Minimize Z = 4a + 2b +0p +0q + 0r + MA1 + MA2 + MA3 s.t.
3a + 1b –1p + 0q +0r + 1A1 + 0A2 + 0A3 = 27
1a + 1b + 0p – 1q + 0r + 0A1 + 1A2 + 0A3 = 21
1a + 2b + 0p + 0q – 1r + 0A1 + 0A2 + 1A3 = 30 and a, b, p, q, r, A1, A2, A3 all ≥ 0
The simplex format of Maximization version is: In maximization version we use negative sign
for big –M.
Maximize Z = –4a – 2b + 0p + 0q + 0r – MA1 – MA2 – MA3 s.t.
3a + 1b – 1p + 0q + 0r + 1A1 + 0A2 + 0A3 = 27
1a + 1b + 0p – 1q + 0r + 0A1 + 1A2 + 0A3 = 21
1a + 2b + 0p + 0q – 1r + 0A1 + 0A2 + 1A3 = 30 and a, b, p, q, r, A1, A2 , A3 all ≥ 0
Let us solve the maximization version.

Table: I. a = 0, b = 0 p = 0 q = 0 r = 0 A1 = 27 A2 = 21 A3 = 30 and Z = Rs. 78 M.

Problem Profit Capacity –4 –2 0 0 0 –M –M –M Replace–

variable Rs. C = Units a b p q r A1 A2 A3 ment ratio

A1 –M 27 3 1 –1 0 0 1 0 0 27/3 = 9

A2 –M 21 1 1 0 –1 0 0 1 0 21/1 = 21

A3 –M 30 1 2 0 0 –1 0 0 1 30/1 = 30
Net – 4 + 5M – 2 + 4M –M –M –M 0 0 0
evaluation

Table: II. a = 9, b = 0, p = 0, q = 0, r = 0, A1 = 0, A2 = 12, A3 = 21, Z = –33M –36

Problem Profit Capacity –4 –2 0 0 0 –M –M – M Replace–

variable Rs. C = Units a b p q r A1 A2 A 3 ment ratio
a –4 9 1 1/3 – 1/3 0 0 0 0
A2 –M 12 0 2/3 1/3 –1 0 1 0
A3 –M 21 0 5/3 1/3 0 –1 0 1
Net 0 –2 + 7/3M 4/3 + 2/3M – M –M 0 0
evaluation

Note: Artificial variable removed is not entered.

Linear Programming Models : Solution by Simplex Method 79

Table: III. a = 24/5, b = 63/5, p = 0, q = 0, r = 0, A1 = 0, A2 = 18/5, A3 = 0

Problem Profit Capacity –4 –2 0 0 0 –M – M – M Replace–

variable Rs. C = Units a b p q r A1 A2 A 3 ment ratio

a –4 24/5 1 0 – 2/5 0 1/5 0

A2 –M 18/5 0 0 1/5 –1 2/5 1

b –2 63/5 0 1 1/5 0 – 3/5 0

Net 0 0 – 6/5 + 1/5M – M – 2/5 + 2/5M 0

evaluation

Table: IV. a = 3, b = 18, p = 0, q = 0, r = 9 and Z = Rs. 48.00

Problem Profit Capacity –4 –2 0 0 0 –M –M –M Replace–

variable Rs. C = Units a b p q r A1 A2 A3 ment ratio

a –4 3 1 0 – 1/2 1/2 0

r 0 9 0 0 1/2 – 5/2 1

b –2 18 0 1 1/2 – 3/2 0

Net 0 0 –1 –1 0
evaluation

A = 3, B = 18 and Z = Rs. 48/–. That is for minimization version; the total minimum cost is Rs.
48/–
Problem 3.15: Solve the following L.P.P.
Maximize Z = 1a + 2b + 3c – 1d S.t.
1a + 2b + 3c = 15
2a + 1b + 5c = 20
1a + 2b + 1c + 1f = 10 and a, b, c, f all are ≥ 0.
In this problem given constraints are equations rather than inequalities. Also by careful examination,
we can see that in the third equation variable ‘f ’ exists and it also exists in objective function. Hence we
consider it as a surplus variable and we add two more slack variables ‘d’ and ‘e’ is added to first and
second equations. But the cost coefficient in objective function for variables d and e will be –M. When
we use big M in maximization problem, we have to use –M in objective function. While solving the
problem, all the rules related to solving maximization problem will apply. Hence now the simplex format
of the problem is as follows:
Maximize Z = 1a + 2b + 3c – 1d – Me – Mf s.t.
1a + 2b + 3c + 1d + 0e + 0f = 15
2a + 1b + 5c + 0d + 1e + 0f = 20
1a + 2b + 1c + 0d + 0e + 1f = 10 and a, b, c, d, e, f all = 0.
80 Operations Research

Table I. a = 0, b = 0, c = 0, d = 15, e = 20, f = 10 and Z = Rs. 0.

Problem Profit Cj Capacity, 1 2 3 –1 – M – M Replacement

variable Rs. units a b c d e f ratio

d –1 15 1 2 3 1 0 0 5

e –M 20 2 1 5 0 1 0 4

f –M 10 1 2 1 0 0 1 10

Net 2 + 3M 4 + 3M 4 + 8M 0 0 0
evaluation.

Table II. A = 0, b = 0, c = 4, d = 0, e = 0, f = 0., Z = Rs. 3 × 4 = Rs. 12.

Problem Profit C j Capacity, 1 2 3 –1 –M –M Replacement

Variable Rs. units a b c d e f Ratio

d –1 6 3/5 9/5 0 1 0 0 30/9

c 3 4 2/5 1/5 1 0 0 0 20

f –M 3 – 1/5 7/5 0 0 1 1 15/7

Net 2/5 + M /5 16/5 + 7/5M 0 M –M 0

evaluation.

TableL: III. a = 0, b = 15/7, c = 25/7, d = 15/7, e = 0, f = 0, Z = Rs. 107 – 15/7 = Rs.92.

Problem Profit Profit 1 2 3 –1 –M –M

variable Rs. capacity a b c d e f

d –1 15/7 6/7 0 0 1 0 0

c 3 25/7 3/7 0 1 0 – 5/7 – 5/7

b 2 15/7 – 1/7 1 0 0 5/7 5/7

Net 6/7 0 0 0 – M + 5/7 – M + 5/7

evaluation
Linear Programming Models : Solution by Simplex Method 81

Table: IV. a =5/2, b = 5/2, c = 5/2, d = 0, e = 0, f = 0, Z = Rs. 15/–

Problem Profit Profit 1 2 3 –1 –M –M

Variable Rs. capacity a b c d e f

a 1 5/2 1 0 0 7/6 0 0

c 3 5/2 0 0 1 – 1/2 – 5/7 – 5/7

b 2 5/2 0 1 0 1/6 5/7 5/7

0 0 0 – 4/12 – M + 5/7 – M = 5/7

As all the elements of net evaluation row are either zeros or negative elements, the solution is
optimal.
Z = Rs. 15/–. And a = b = c = 5/2.
Problem 3.16: Solve the given l.p.p: Simplex format is:
Maximize 4x + 3y s.t. Maximize Z = 4x + 3y + 0S + 0p + 0q – MA1 – MA2 s.t.
1x + 1y ≤ 50 1x +1y+ 1S + 0p + 0q + 0A1 + 0A2 = 50
1x + 2y ≥ 80 1x + 2y +0S – 1p + 0q + 1A1 + 0A2 = 80
3x + 2y ≥ 140 3x + 2y + 0S + 0p – 1q + 0A1 + 1A2 = 140
And both x and y ≥ 0 and x, y, S, p, q, A1 and A2 all ≥ 0

Table: I. X = 0, y, = 0, S = 50, p = 0 q = 0, A1 = 80, A2 = 140 and Z = Rs. 220 M.

Problem Profit Profit: 4 3 0 0 0 –M –M Replace–

variable Rs. capacity units x y S p q A1 A2 ment ratio

S 0 50 1 1 1 0 0 0 0 50

A1 –M 80 1 2 0 –1 0 1 0 40

A2 –M 140 3 2 0 0 –1 0 1 70

Net 4 + 4M 3 + 4M 0 M M 0 0
evaluation.

Now in the net evaluation row the element under variable ‘x’ is 4 + 4M is greater than the element
3 + 3M. But if we take ‘x’ as the incoming variable we cannot send artificial variable out first. Hence we
take ‘y’ as the incoming variable, so that A1 go out first.
82 Operations Research

Table: II. X = 0, y = 40, S = 10, A1 = 0, A2 = 60, p = 0, q = 0 and Z = Rs. 120 – 60 M.

Problem Profit Profit: 4 3 0 0 0 –M – M Replace–

variable Rs. capacity units x y S p q A1 A 2 ment ratio

S 0 10 0.5 0 1 0.5 0 – 0.5 0 20

y 3 40 0.5 1 0 – 0.5 0 0.5 0 80

A2 –M 60 2 0 0 2 –1 –2 1 30

Net 2.5 + 8M 0 0 1.5 – 2M – M – 1.5 – 3M 0

evaluation.

Table: III. x = 20, y = 30, p = 0, q = 0, A1 = 0, A2 = 20, Z = Rs. 170/–

Problem Profit Profit: 4 3 0 0 0 –M –M Replace–

variable Rs. capacity units x y S p q A1 A2 ment ratio

x 4 20 1 0 2 1 0 –1 0

y 3 30 0 1 –1 –1 0 1 0

A2 –M 20 0 0 –4 0 –1 0 1

Net 0 0 – 5 –4M –1 – M –M=1 0

evaluation.

In the last table though basis variables have the opportunity cost as 0, still artificial variable A2
exists in the problem, hence the original problem has no feasible solution.
X = 20, Y = 30.
Problem 3.17: Solve the given l.p.p. Simplex version of the problem is:
Maximize Z = 1a + 1.5b + 5c + 2d s.t. Maximize Z = 1a + 1.5b + 5c + 2d + 0S1 + 0S2 – MA1 –
MA2 s.t.
3a + 2b + 1c + 4d ≤ 6 3a + 2b +1c + 4d + 1S1 + 0S2 + 0A1 + 0A2 = 6
2a + 1b + 5c + 1d ≤ 4 2a +1b + 5c + 1d + 0S1 +1S2 + 0A1 + 0A2 = 4
2a + 6b – 4c + 8d = 0 2a + 6b – 4c + 8d + 0S1 + 0S2 +1A1 + 0A2 = 0
1a + 3b – 2c + 4d = 0 1a +3b – 2c + 4d + 0S1 + 0S2 + 0A1 + 1A2 = 0
And a, b, c, d all ≥ 0 and a, b, c, d, S1, S2, A1, A2 all ≥ 0
Linear Programming Models : Solution by Simplex Method 83

Table I. A = 0, b = 0, c = 0, d = 0, S1 = 6, S2 = 0, A1 = 0, A2 = 0 and Z = Rs. 0.00

Problem Profit C j Capacity 1 1.5 5 2 0 0 –M –M Replace–

variable Rs. units a b c d S1 S2 A1 A2 ment ratio
S1 0 6 3 2 1 4 1 0 0 0 6/4 = 1.5
S2 0 4 2 1 5 1 0 1 0 0 4/1 = 4
A1 –M 0 2 6 –4 8 0 0 1 0 0
A2 –M 0 1 3 –2 4 0 0 0 1 0
Net 1 + 3M 1.5 + 9M 5 – 6M 2 + 12M 0 0 0 0
evaluation

Note: Both A1 and A2 have the same replacement ratio. That is to say there is a tie in
outgoing variables. This type of situation in Linear Programming Problem is known as
DEGENERACY. To Solve degeneracy refer to the rules stated earlier.

Now let us remove as and when a surplus variable goes out, which will ease our calculation and
also save time.
Table: II. a = 0, b = 0, c = 0, d = 0, S1 = 6, S2 = 4, A1 = 0, A2 = 0, Z = Rs. 0/–

Problem Profit Cj Capacity 1 1.5 5 2 0 0 –M –M Replace–

variable Rs. units a b c d S1 S2 A1 A2 ment ratio

S1 0 6 2 –1 3 0 1 0 0

S2 0 4 7/4 1/4 11/2 0 0 1 0

A1 –M 0 0 0 0 0 0 0 1

d 2 0 1/4 3/4 – 1/2 1 0 0 0

Net 1/2 0 6 0 0 0 0
evaluation

In the given problem the inequalities number 3 and 4 i.e., 2a + 6b – 4c + 8d = 0 and

1a + 3b – 2c + 4d = 0 appears to be similar. If you carefully examine, we see that
2a + 6b – 4c + 8d = 0 is 2 × (1a + 3b – 2c + 4d = 0). Hence one of them may be considered as
redundant and cancelled. i.e., third constraint is double the fourth constraint hence we can say it is not
independent. Hence we can eliminate the third row and the column under A1 from the tableau. For
identifying the redundancy, one need not wait for final optimal table.
84 Operations Research

Table: III. (Second reduced second table) S1 = 6, S2 = 4, d = 0, a = 0, b = 0, c = 0, Z = Rs. 0

Problem Profit Cj Capacity 1 1.5 5 2 0 0 –M –M Replace–

variable Rs. units a b c d S1 S2 A1 A2 ment ratio

S1 0 6 2 –1 3 0 1 0 2

S2 0 4 7/4 1/4 11/2 0 0 1 8/11

d 2 0 1/4 3/4 – 1/2 1 0 0 —

Net 1/2 0 6 0 0 0
evaluation.

Table: IV. S1 = 42/11, C = 5, d = 4/11 and Z = Rs. 4.36.

Problem Profit Cj Capacity 1 1.5 5 2 0 0 –M –M Replace–

variable Rs. units a b c d S1 S2 A1 A2 ment ratio

S1 0 42/11 23/22 – 25/22 0 0 1 – 6/11

c 5 8/11 7/22 1/22 1 0 0 2/11

d 2 4/11 9/22 17/22 0 1 0 1/11

Net – 31/22 – 3/11 0 0 0 – 12/11

evaluation

As all the elements of net evaluation row are either zeros or negative elements the solution at this
stage is optimal. Hence c = 8/11, d = 4/11 and Z = Rs. 48/11 = Rs. 4.36.

3.10. ARTIFICIAL VARIABLE METHOD OR TWO PHASE METHOD

In linear programming problems sometimes we see that the constraints may have ≥ , ≤ or = signs. In
such problems, basis matrix is not obtained as an identity matrix in the first simplex table; therefore, we
introduce a new type of variable called, the artificial variable. These variables are fictitious and cannot
have any physical meaning. The introduction of artificial variable is merely to get starting basic feasible
solution, so that simplex procedure may be used as usual until the optimal solution is obtained. Artificial
variable can be eliminated from the simplex table as and when they become zero i.e, non–basic. This
process of eliminating artificial variable is performed in PHASE I of the solution. PHASE II is then
used for getting optimal solution. Here the solution of the linear programming problem is completed in
two phases, this method is known as TWO PHASE SIMPLEX METHOD. Hence, the two–phase
method deals with removal of artificial variable in the fist phase and work for optimal solution in the
second phase. If at the end of the first stage, there still remains artificial variable in the basic at a
positive value, it means there is no feasible solution for the problem given. In that case, it is not
Linear Programming Models : Solution by Simplex Method 85

necessary to work on phase II. If a feasible solution exists for the given problem, the value of objective
function at the end of phase I will be zero and artificial variable will be non-basic. In phase II original
objective coefficients are introduced in the final tableau of phase I and the objective function is optimized.
Problem 3.18: By using two phase method find whether the following problem has a feasible
solution or not?
Maximize Z = 4a + 5b s.t. Simplex version is: Max. Z = 4a + 5b + 0S1 + 0S2 – MA s.t.
2a + 4b ≤ 8 2a + 4b + 1S1 + 0S2 + 0A = 8
1a + 3b ≥ 9 and both a and b are ≥ 0. 1a + 3b + 0S1 – 1S2 + 1A = 9 and
a, b, S1, S2, A all are ≥ 0
Phase I
Maximize Z = 0a + 0b + 0S1 + 0S2 – 1A s.t.
2a + 4b + 1S1 + 0S2 + 0A = 8
1a + 3b + 0S1 – 1S2 + 1A = 9 and a, b, S1, S2 and A all ≥ 0.
Table: I. a = 0, b = 0, S1 = 8, S2 = 0, A = 9 and Z = – Rs 9

Problem Profit Cj Capacity, 0 0 0 0 –1 Replacement

variable Rs. units a b S1 S2 A ratio.
S1 0 8 2 4 1 0 0 8/4 = 2
A –1 9 1 3 0 –1 1 9/3 = 3
Net evaluation 1 3 0 –1 0

Table: II. a = 0, b = 2, S1 = 0, S2 =0, A = 3 and Z = Rs. – 3/–

Problem Profit Cj Capacity, 0 0 0 0 –1 Replacement

variable Rs. units a b S1 S2 A ratio.
b 0 2 0.5 1 0.25 0 0 4
A –1 3 0.5 0 – 0.75 –1 1 8
Net evaluation 0.5 0 – 0.75 –1 0

Table: III. a = 4, b = 0, S1 =0, S2 = 0, A = 1 and z = Rs. –1/–

Problem Profit Cj Capacity, 0 0 0 0 –1 Replacement

variable Rs. units a b S1 S2 A ratio.
a 0 4 1 2 0.5 0 0
A –1 1 0 –1 –1 –1 1
Net evaluation 0 –1 –1 –1 0
86 Operations Research

As the artificial variable still remains as the basic variable and has a positive value, the given
problem has no feasible solution.
If we examine the same by graphical means, we can see that the problem has no feasible region.

Problem 3.18: Simplex version:

Maximize Z = 4x + 3y s.t. Maximize Z = 4x + 3y + 0S1 + 0S2 – MA s.t.
2x + 3y ≤ 6 2x + 3y + 1S1 + 0S2 + 0A = 6
3x + 1y ≥ 3 3x + 1y + 0S1 – 1S2 +1A = 3
Both x and y all ≥ 0 x, y, S1, S2, and A all = 0.
Phase I
0x + 0y +0S1 + 0S2 – 1A s.t.
2x + 3y + 0S1 + 0S2 +0A = 6
3x + 1y + 0S1 – 1S2 + 1A3, x, y, S1, S2, and A all ≥ 0

Table: I. x = 0, y = 0, S1 = 6, S2 = 0, A = 3 and Z = 3 × –1 = – Rs. 3/–

Problem Profit Cj Capacity, 0 0 0 0 –1 Replacement

variable Rs. units x y S1 S2 A ratio.
S1 0 6 2 3 1 0 0 3
A –1 3 3 1 0 –1 1 1

Net evaluation 3 1 0 –1 0

Table: II. x = 1, y = 0, S1 = , S2 = 0, A = 0. Z = Rs. 0.

Problem Profit Cj Capacity, 0 0 0 0 –1 Replacement

variable Rs. units x y S1 S2 A ratio.
S1 0 0 0 7/3 1 2/3 – 2/3
x 0 1 1 1/3 0 – 1/3 1/3
Net evaluation 0 0 0 0 1
Linear Programming Models : Solution by Simplex Method 87

As there is no artificial variable in the programme, we can get the optimal solution for the given
problem.
Hence Phase II is:
Phase II

Table: III. S1 = 4, x = 1, y = 0, S2 = 0, A = 0 and Z = Rs. 1 × 4 = Rs.4 /–

Problem Profit Cj Capacity, 0 0 0 0 –1 Replacement

variable Rs. units x y S1 S2 A ratio.
S1 0 4 0 7/3 1 2/3 – 2/3 6
x 4 1 1 1/3 0 – 1/3 1/3 – 3 (neglect)

Net evaluation 0 1/3 0 4/3 4/3 – M

Table: IV. x = 3, y = 0, S1 = 6, S2 = 0, A = 0 and Z = Rs. 3x4 = Rs. 12/–

Problem Profit Cj Capacity, 0 0 0 0 –1 Replacement

variable Rs. units x y S1 S2 A ratio.
S2 0 6 0 7/2 3/3 1 –1
x 4 3 1 3/2 ½ 0 0
Net evaluation 0 –1 –2 0 –M

Optimal solution is x = 4 and Z = Rs. 12/–. Graphically also student can work to find optimal
solution.
Problem 3.19:
Solve the following L.P.P. by two–phase method: Simplex version:
Maximize Z = 2a – 1b + 1c s.t. Maximize Z = 2a – 1b + 1c + 0S1 + 0S2 + 0S3 – 1A1 – 1A2 s.t
1a + 1b – 3c ≤ 8 1a + 1b –3c + 1S1 + 0S2 + 0S3 + 0A1 + 0A2 = 8
4a – 1b + 1c ≥ 2 4a – 1b + 1c + 0S1 – 1S2 + 0S3 + 1A1 + 0A2 = 2
2a + 3b – 1c ≥ 4 2a + 3b – 1c + 0S1 + 0S2 – 1S3 + 0A1 + 1A2 = 4
And a, b, c all ≥ 0. a, b, c, S1, S2, S2 A1 and A2 all are ≥ 0.
Phase I
The objective function is Maximize Z = 0a + 0b + 0c + 0S1 + 0S2 + 0S3 + (– 1) A1 + (–1) A2
The structural constraints will remain same as shown in simplex version.
88 Operations Research

Table: I. a = 0, b = 0, c = 0, S1 =8, S2 =0, S3 = 0, A1 = 2, A2 = 4, Z = – Rs. 6/–

Problem Profit Cj 0 0 0 0 0 0 –1 –1 Replace–

variable Rs. Capacity units a b c S1 S2 S3 A1 A2 ment ratio

S1 0 8 1 1 –3 1 0 0 0 0 8

A1 –1 2 4 –1 1 0 –1 0 1 0 1/2

A2 –1 4 2 3 –1 0 0 –1 0 1 2

Net 6 2 0 0 –1 –1 0 0
evaluation

Table: II. a = 1/2, b = 0, c = 0, S1 = 15/2, S2 = 0, S3 = 0, A1 = 0, A2 = 3, Z = – Rs.3/–

Problem Profit Cj 0 0 0 0 0 0 –1 –1 Replace–

variable Rs. Capacity units a b c S1 S2 S3 A1 A2 ment ratio

S1 0 15/2 0 3/4 – 11/4 1 1/4 0 0

a 0 1/2 1 – 1/4 1/4 0 – 1/4 0 0

A2 – 1 3 0 7/2 – 5/2 0 1/2 –1 1

Net 0 7/2 – 3/2 0 1/2 –1 0

evaluation

Table: III. a = 5/7, b = 6/7, c = 0, S1 = 45/7, S2 = 0, S3 = 0, A1 = 0, A2 = 0, Z = Rs.0

Problem Profit Cj 0 0 0 0 0 0 –1 –1 Replace–

variable Rs. Capacity units a b c S1 S2 S3 A1 A2 ment ratio

S1 0 45/7 0 0 – 19/7 1 1/14 5/14

a 0 5/7 1 0 1/2 0 – 3/14 – 1/14

b 0 6/7 0 1 – 3/7 0 1/7 – 2/7

Net 0 0 0 0 0 0
evaluation

As there are no artificial variables, we can go for second phase.

Linear Programming Models : Solution by Simplex Method 89

Phase II.
Table: I. a = 5/7, b = 6/7, c = 0, S1 = 45/7, S2 = 0, S3 = 0, A1 = 0, A2 = 0, Z = Rs. 4/7.

Problem Profit Cj 2 –1 1 0 0 0 –1 – 1 Replace–

variable Rs. Capacity units a b c S1 S2 S3 A1 A 2 ment ratio

S1 0 45/7 0 0 – 19/7 1 1/14 5/14

a 2 5/7 1 0 1/7 0 – 3/14 – 1/14

b –1 6/7 0 1 – 3/7 0 1/7 3/7

Net 0 0 2/7 0 4/7 – 1/7

evaluation

Table: II.

Problem Profit Cj 2 –1 1 0 0 0 –1 –1 Replace–

variable Rs. Capacity units a b c S1 S2 S3 A1 A2 ment ratio

S1 0 6 0 – 7/2 – 5/2 1 0 1/2

a 2 2 1 3/2 – 1/2 0 0 – 7/2

S2 0 6 0 7 –3 1 1 –2

Net 0 –4 2 0 0 7
evaluation

Highest positive element under S3 in net evaluation row shows that the problem has unbound
solution.
Problem 3.20: This can be written as:
Minimize Z = 15/2a – 3b + 0c s.t. Maximize Z = –15/2 a + 3b – 0c s.t
3a – 1b –1c ≥ 3 3a – 1b – 1c ≥ 3
1a – 1b +1c ≥ 2 1a – 1b + 1c ≥ 2 and a, b, c all ≥ 0
a, b, c all ≥ 0
Simplex version is:
Maximize Z = 15/2 a – 3b – 0c + 0S1 + 0S2 – 1A1 – 1A2 s.t.
3a – 1b – 1c – 1S1 + 0S2 + 1A1 + 0A2 = 3
1a – 1b + 1c + 0S1 – 1S2 + 0A1 + 1A2 = 2 and a, b, c, S1, S2, A1 and A2 all ≥ 0
In Phase I we give profit coefficients of variables as zero.
Maximize: Z = 0a + 0b + 0c +0S1 + 0S2 –1A1 – 1 A2 s.t.
The constraints remain same.
90 Operations Research

Phase I
Table: I. a = 0, b = 0, c = 0, S1 = 0, S2 = 0, A1 = 3, A2 = 2 and Z = – Rs.5/–

Problem Profit Cj 0 0 0 0 0 –1 –1 Replace–

variable Rs. Capacity a b c S1 S2 A1 A2 ment ratio

A1 –1 3 3 1 1 1 0 1 0 3/3 = 1

A2 –1 2 1 –1 1 0 –1 0 1 2/1 = 2

Net 4 0 2 1 1 0 0
evaluation

Table: II. a = 1, b = 0, c = 0, S1 = 0, S2 = 0, A1 = 0, A2 = 1, and Z = – Rs.1/–

Problem Profit Cj 0 0 0 0 0 –1 –1 Replace–

variable Rs. Capacity a b c S1 S2 A1 A2 ment ratio

a 0 1 1 – 1/3 – 1/3 – 1/3 0 1/3 0 —

A2 –1 1 0 – 2/3 4/3 1/3 –1 – 4/3 1 3/4

N.E. 0 – 2/3 4/3 1/3 –1 – 4/3 0

Table: III. a = 5/4, b = 0, c = ¾, S1 = 0, S2 =0, A1 = 0, A2 = 0 and Z = Rs.0/–

Problem Profit Cj 0 0 0 0 0 –1 –1 Replace–

variable Rs. Capacity a b c S1 S2 A1 A2 ment ratio

a 0 5/4 1 – 1/2 0 – 1/4 – 1/4 1/4 1/4

c 0 3/4 0 – 1/2 1 1/4 – 3/4 – 1/4 3/4

Net 0 0 0 0 0 –1 –1
evaluation

Phase II
Table: I. a = 5/4, b = 0, c = ¾, S1= 0, S2 = 0, A1 = 0, S2 = 0, Z = Rs. 75/8.

Problem Profit Cj – 15/2 3 0 0 0 –1 –1 Replace–

variable Rs. Capacity a b c S1 S2 A1 A2 ment ratio

a – 15/2 5/4 1 – 1/2 0 – 1/4 – 1/4

c 0 3/4 0 – 1/2 1 1/4 – 3/4

Net 0 – 3/4 0 – 15/8 – 15/8

evaluation
Linear Programming Models : Solution by Simplex Method 91

As all the net evaluation row elements are negative or zeros, the solution is optimal.
a = 5/4, c = 3/5, and minimum Z = Rs. 75/8.
The Disadvantages of Big M method over Two–phase method:
1. Big M method can be used to find the existence of feasible solution. But it is difficult
and many a time one gets confused during computation because of manipulation of
constant M. In two–phase method big M is eliminated and calculations will become
easy.
2. The existence of big M avoids the use of digital computer for calculations.

3.11. DEGENERACY IN LINEAR PROGRAMMING PROBLEMS

The degeneracy in linear programming problems and the methods of solving degeneracy, if it exists,
are discussed earlier in the chapter. To recollect the same a brief discussion is given below:
While improving the basic feasible solution to achieve optimal solution, we have to find the key
column and key row. While doing so, we may come across two situations. One is Tie and the other is
Degeneracy.
The tie occurs when two or more net evaluation row elements of variables are equal. In maximization
problem, we select the highest positive element to indicate incoming variable and in minimization we
select lowest element to indicate incoming variable (or highest numerical value with negative sign).
When two or more net evaluation row elements are same, to break the tie, we select any one of them
to indicate incoming variable and in the next iteration the problem of tie will be solved.
To select the out going variable, we have to select the lowest ratio or limiting ratio in the replacement
ratio column. Here also, some times during the phases of solution, the ratios may be equal. This
situation in linear programming problem is known as degeneracy. To solve degeneracy, the following
methods are used:
1. Select any one row as you please. If you are lucky, you may get optimal solution, otherwise
the problem cycles.
OR
2. Identify the rows, which are having same ratios. Say for example, S1 and S3 rows having
equal ratio. In such case select the row, which contains the variable with smaller subscript.
That is select row containing S1 as the key row. Suppose the rows of variable x and z are
having same ratio, then select the row-containing x as the key row.
3. (a) Divide the elements of unit matrix by corresponding elements of key column. Verify the
ratios column-wise in unit matrix starting from left to right. Once the ratios are unequal, the
degeneracy is solved. Select the minimum ratio and the row containing that element is the
key row. (This should be done to the rows that are in tie).
(b) If the degeneracy is not solved by 3 (a), then divide the elements of the main matrix by the
corresponding element in the key column, and verify the ratios. Once the ratios are unequal,
select the lowest ratios. (This should be done only to rows that are in tie).
Problem 3.21: A company manufactures two product A and B. These are machined on machines
X and Y. A takes one hour on machine X and one hour on Machine Y. Similarly product B takes 4 hours
on Machine X and 2 hours on Machine Y. Machine X and Y have 8 hours and 4 hours as idle capacity.
92 Operations Research

The planning manager wants to avail the idle time to manufacture A and B. The profit contribution of
A is Rs. 3/– per unit and that of B is Rs.9/– per unit. Find the optimal product mix.
Solution: Simplex format is:
Maximize Z = 3a + 9b s.t. Maximize Z = 3a + 9b + 0S1 + 0S2 s.t.
1a + 4b ≤ 8 1a + 4b + 1S1 + 0S2 = 8
1a + 2b ≤ 4 both a and b are ≥ 0 1a + 2b + 0S1 + 1S2 = 4 and
a, b, S1, S2 all ≥ 0.
Table: I. A = 0, b = 0, S1 = 8, S2 = 4 and Z = Rs.0/–

Problem Profit Cj 3 9 0 0 Replacement

variable Rs. Capacity a b S1 S2 ratio

S1 0 8 1 4 1 0 8/4 = 2

S2 0 4 1 2 0 1 4/2 = 2

Net evaluation 3 9 0 0

Now to select the out going variable, we have to take limiting ratio in the replacement ratio
column. But both the ratios are same i.e. = 2. Hence there exists a tie as an indication of degeneracy
in the problem. To solve degeneracy follow the steps mentioned below:
(i) Divide the elements of identity column by column from left to right by the corresponding key
column element.
Once the ratios are unequal select the lowest ratio and the row containing that ratio is the key
row.
In this problem, for the first column of the identity (i.e. the S1 column) the ratios are: 1/4, and
0/2. The lowest ratio comes in row of S2. Hence S2 is the outgoing variable. In case ratios are equal go
to the second column and try.
Table: II. a = 0, b = 2, S1 = 0, S2 = 0, Z = Rs. 18/–

Problem Profit Cj 3 9 0 0 Replacement

variable Rs. Capacity a b S1 S2 ratio

S1 0 0 1 –2 –1 0

b 9 2 0 1/2 1/2 1

Net evaluation 0 – 9/2 – 3/2 0

Optimal solution is b = 2 and Profit is 2 × 9 = Rs. 18/–

Linear Programming Models : Solution by Simplex Method 93

Problem 3.22: Simplex version is:

Maximize Z = 2x + 1y s.t Maximize Z = 2x + 1y + 0S1 + 0S2 + 0S3 s.t.
4x + 3y ≤ 12 4x + 3y + 1S1 + 0S2 + 0S3 = 12
4x + 1y ≤ 8 4x + 1y + 0S1 + 1S2 + 0S3 = 8
4x – 1y ≤ 8 4x – 1y + 0S1 + 0S2 + 1S3 = 8
Both x and y are ≥ 0 x, y, S1, S2, S3 all ≥ 0

Table: I. x = 0, y = 0, S1 = 12, S2 = 8, S3 = 8 and Z = Rs. 0/–

Problem Profit Cj 2 1 0 0 0 Replacement

variable Rs. Capacity units x y S1 S2 S3 ratio
S1 0 12 4 3 1 0 0 12/4 = 3
S2 0 8 4 1 0 1 0 8/4 = 2
S3 0 8 4 –1 0 0 1 8/4 = 2

Net evaluation 2 1 0 0 0

As the lowest ratio (= 2) is not unique, degeneracy occurs. Hence divide the elements of the
identity column by column from left to right and verify the ratios.
In this problem, the elements of the first column of identity under S1 (for 2nd and 3rd row) are
0,0. If we divide them by respective key column element, the ratios are 0/4 and 0/4 which are equal,
hence we cannot break the tie.
Now try with the second column i.e., column under S2. The elements of 2nd and 3rd row are 1
and 0. If we divide them by respective elements of key column, we get 1/4 and 0/4. The ratios are
unequal and the lowest being zero for the third row. Hence S3 is the outgoing variable.

Table: II. x = 2, y = 0, S1 = 4, S2 = 0, S3 = 0, and Z = Rs. 4/–

Problem Profit Cj 2 1 0 0 0 Replacement

variable Rs. Capacity units x y S1 S2 S3 ratio
S1 0 4 0 4 1 0 –1 1/4
S2 0 0 0 2 0 1 –1 0/2
X 2 2 1 – 1/4 0 0 1/4 —
Net evaluation 0 3/2 0 0 – 1/2
94 Operations Research

Table: III. x = 2, y = 0, S1 = 4, S2 = 0, S3 = 0, Z = Rs. 4/–

Problem Profit Cj 2 1 0 0 0 Replacement

variable Rs. Capacity units x y S1 S2 S3 ratio
S1 0 4 0 0 1 –2 1 4/1
y 1 0 0 1 0 1/2 – 1/2 —
x 2 2 1 0 0 1/8 1/8 2 × 8/1 = 16

Net evaluation 0 0 0 – 3/4 1/4

Table: IV. x = 3/2, y = 2, S1 = 0, S2 = 0, S3 = 4 and Z = Rs. 3 + 2 = Rs. 5/–

Problem Profit Cj 2 1 0 0 0 Replacement

variable Rs. Capacity units x y S1 S2 S3 ratio
S3 0 4 0 0 1 –2 1
y 1 2 0 1 1/2 – 1/2 0
x 2 3/2 1 0 – 1/8 3/8 0
Net evaluation 0 0 – 1/4 – 1/4 0

As the elements of net evaluation row are either zeros or negative elements, the solution is optimal.
x = 3/2 and y = 2, and Z = Rs. 2 × 3/2 + 2 × 1 = Rs. 5/–
Shadow price = ¼ × 12 + ¼ × 8 = Rs. 5/–.

3.12.1. Special Cases

Some times we come across difficulties while solving a linear programming problem, such as
alternate solutions and unbound solutions. Let us solve some problems of special nature.
Problem. 3.23:
Solve the given l.p.p. by big – M method. Simplex version is:
Maximize Z = 6a + 4b s.t. Maximize Z = 6a + 4b + 0S1 + 0S2 + 0S3 – MA
2a + 3b ≤ 30 2a + 3b + 1S1 + 0S2 + 0S3 + 0A = 30
3a + 2b ≤ 24 3a + 2b + 0S1 + 1S2 +0S3 + 0A = 24
1a + 1b ≥ 3 and x, y both ≥ 0. 1a + 1b + 0S1 + 0S2 – 1S3 + 1A = 3
a, b, S1, S2, S3, A all ≥ 0.
Linear Programming Models : Solution by Simplex Method 95

Table: I. a = 0, b = 0, S1 = 30, S2 = 24, S3 = 0, A = 3 and Z = Rs. 3M

Problem Profit Cj 6 4 0 0 0 –M Replacement

variable Rs. Capacity units a b S1 S2 S3 A ratio.

S1 0 30 2 3 1 0 0 0 30/2 = 15

S2 0 24 3 2 0 1 0 0 24/3 = 8

A –M 3 1 1 0 0 –1 1 3/1 = 3

Net 6+M 4+M 0 0 –M 0

evaluation

Table II. a = 3, b = 0, S1 = 24, S2 = 15, S3 = 0, A = 0, Z = Rs. 18/– (out going column eliminated)

Problem Profit Cj 6 4 0 0 0 –M Replacement

variable Rs. Capacity units a b S1 S2 S3 A ratio

S1 0 24 0 1 1 0 2 24/2 = 12

S2 0 15 0 –1 0 1 3 15/3 = 5

A 6 3 1 1 0 0 –1 —

Net 0 –2 0 0 6
evaluation

Table: III. a = 8, b = 0, S1 = 14, S2 = 5, S3 = 0, A = 0, Z = Rs. 48

Problem Profit Cj 6 4 0 0 0 –M Replacement

variable Rs. Capacity units a b S1 S2 S3 A ratio
S3 0 14 0 5/3 1 – 2/3 0 42/5
S2 0 5 0 – 1/3 0 1/3 0 —
a 6 8 1 2/3 0 1/3 0 24/2 = 12
Net 0 0 0 –3 0
evaluation

In the above table, as all the net evaluation elements are either zeros or negative elements the
optimal solution is obtained. Hence the answer is: a = 8 and the profit Z = Rs. 48/–. But the net
evaluation element in the column under 'b' is zero. This indicates that the problem has alternate solution.
If we modify the solution we can get the values of basis variables, but the profit Z will be the same.
This type of situation is very helpful to the production manager and marketing manager to arrange the
96 Operations Research

production schedules and satisfy the market demands of different segments of the market. As the
profits of all alternate solutions are equal, the manager can select the solution, which is more needed by
him. Now let us workout the alternate solution.
Table: IV. a = 12 /5, b = 42 / 5 , S1 = 0, S2 = 0 , S3 = 39/5, A = 0, Z = Rs. 48/–

Problem Profit Cj 6 4 0 0 0 –M Replacement

variable Rs. Capacity units a b S1 S2 S3 A ratio

b 4 42/5 0 1 3/5 – 2/5 0

S3 0 39/5 0 0 1/5 1/5 1

a 6 12/5 1 0 – 2/5 3/5 0

Net 0 0 0 –2 0
evaluation

All the elements of net evaluation rows are either zeros or negative elements so the solution is
Optimal. The answer is : a = 12/5, b = 42/5 and Z = 12/5 × 6 + 42/5 × 4 = Rs. 48/– . Shadow price =
element in column under S2 multiplied by element on the R.H.S of second inequality i.e., 24 × 2 = Rs.
48/–. Once we get one alternate solution then any number of solutions can be written by using the
following rule.
(a) One alternate value of the basis variable is: First value × d + second value × (1–d). In the
given example, the first value of ‘a’ is 8 and second value of ‘a’ is 12/5. Hence next value is 8 d + (1
– d) × 12/5. Like this we can get any number of values. Here ‘d’ is any positive fraction number: for
example 2/5, 3/5 etc. It is better to take d = 1/2, so that next value is 1/2 × first value + 1/2 × second
value. Similarly the values of other variables can be obtained.

3.12.2. Unbound Solutions

In linear programming problem, we come across certain problem, where feasible region is
unbounded i.e., the value of objective function can be increased indefinitely. Then we say that the
solution is UNBOUND. But it is not necessary, however, that an unbounded feasible region should yield
an unbounded value of the objective function. Let us see some examples.
Problem 3.24: Maximize Z = 107a + 1b + 2c s.t
14a +1b – 6c + 3d = 7
16a + ½ b + 6c ≤ 5
3a – 1b – 6c ≤ 0 and a, b, c and d all ≥ 0.
We find that there is variable ‘d’ in the first constraint, with coefficient as 3. Second and third
constraints do not have variable ‘d’. Hence we can divide the first constraint by 3 we can write as: 14/
3 a + 1/3 b – 6/3 c + 3/3 d = 7/3. If we write like this, we can consider the variable 'd' as slack variable.
Hence the given l.p.p. becomes as:
Maximize Z = 107 a + 1b + 2c s.t
14/3a + 1/3b – 2c + 1d = 7/3
16a + ½ b – 6c ≤ 5
3a – 1b – 1c ≤ 0
and a, b,c, and d all ≥ 0
Linear Programming Models : Solution by Simplex Method 97

Simplex version is
Maximize Z = 107a + 1b + 2c + 0d + 0S1 + 0S2 s.t.
14/3 a + 1/3b – 2c + 1d + 0S1 + 0S2 = 7/3
16a + ½ b – 6c + 0d + 1S1 + 0S2 = 5
3a – 1b – 1c + 0d + 0S1 + 1S2 = 0
and a, b, c, d, S1, S2 all ≥ 0.
TableL: I. a = 0, b = 0, c = 0, d = 7/3, S1 = 5, S2 = 0 and Z = Rs. 0.

Problem Cost Cj 107 1 2 0 0 0 Replacement

variable Rs. Capacity units a b c d S1 S2 ratio

d 0 7/3 14/3 1/3 –2 1 0 0 7/14

S1 0 5 16 1/2 6 0 1 0 5/16

S2 0 0 3 –1 –1 0 0 1 0/3

Net evaluation 107 1 2 0 0 0

Table: II. a = 0, b = 0, c = 0, d = 7/3, S1 = 5, S2 = 0 and Z = Rs.0/–

Problem Cost Cj 107 1 2 0 0 0 Replacement

variable Rs. Capacity units a b c d S1 S2 ratio

d 0 7/3 0 17/9 – 4/9 1 0 – 14/9 —

S1 0 5 0 35/6 – 2/3 0 1 – 16/3 —

a 107 0 1 – 1/3 – 1/3 0 0 1/3 —

Net evaluation 0 110/3 113/3 0 0 107/3

As the elements of incoming variable column are negative we cannot workout replacement ratios
and hence the problem cannot be solved. This is an indication of existence of unbound solution to
the given problem.
Problem 3.25: Simplex version is:
Maximize Z = 6x – 2y s.t. Maximize Z = 6x – 2y + 0S1 + 0S2 s.t.
2x – 1y ≤ 2 2x – 1y + 1S1 + 0S2 = 2
1x + 0y ≤ 4 and both x and y are ≥ 0. 1x + 0y + 0S1 + 1S2 = 4
And x, y, S1 and S2 all ≥ 0.
98 Operations Research

Table: I. x = 0, y = 0, S1 = 2, S2 = 0 and Z = Rs. 0

Problem Profit Cj 6 –2 0 0 Replacement

variable Rs. Capacity units x y S1 S2 ratio

S1 0 2 2 1 1 0 2/2 = 1

S2 0 4 1 0 0 1 4/1 = 4

Net evaluation 6 –2 0 0

Table: II. x = 1, y = 0, S1 = 0, S2 = 3, Z = Rs. 6/–

Problem Profit Cj 6 –2 0 0 Replacement

variable Rs. Capacity units x y S1 S2 ratio

x 6 1 1 – 1/2 1/2 0 —

S2 0 3 0 1/2 – 1/2 1 7/2

Net evaluation 0 1 3 0

Table: III. x = 4, y = 6, S1 = 0, S2 = 0, Z = 24 – 12 = Rs. 12/–

Problem Profit Cj 6 –2 0 0 Replacement

variable Rs. Capacity units x y S1 S2 ratio

x 6 4 1 0 0 1

y –2 6 0 1 –1 2

Net evaluation 0 0 –2 –2

X = 4, Y = 6 and Z = Rs. 12/–

(Note: Students can verify the answer by graphical method).
(Note that in the first table the elements in column under ‘y’ are –1 and 0. This indicates
that the feasible region is unbound. But the problem has solution. This clearly states that
the problem may have unbound region but still it will have solution).

Problem 3.26: Simplex version is:

Maximize Z = 3a + 2b S.t. Maximize Z = 3a + 2b + 0S1 + 0S2 – MA s.t
2a + 1b ≤ 2 3a + 1b + 1S1 + 0S2 + 0A = 2
3a + 4b ≥ 12 3a + 4b + 0S1 – 1S2 + 1A = 12
Both a and b are > 0 a, b, S1, S2 and A all ≥ 0.
Linear Programming Models : Solution by Simplex Method 99

Table: I. a = 0, b = 0, S1 = 2, S2 = 0, A = 12 and Z = Rs. – 12 M

Problem Profit Cj 3 2 0 0 –M Replacement

variable Rs. Capacity units a b S1 S2 A ratio

S1 0 2 3 1 1 0 0 2/1 = 2

A –M 12 3 4 0 –1 1 12/4 = 3

Net evaluation 3 + 3M 2 + 4M 0 –M 0

Table: II. b = 2, a = 0, S1 = 0, S2 = 0, A = 4, Z = Rs. 4 – 4M

Problem Profit Cj 3 2 0 0 –M Replacement

variable Rs. Capacity units a b S1 S2 A ratio

b 2 2 2 1 1 0 0

A –M 4 –5 0 –4 –1 0

Net evaluation – 1 –5M 0 – 2 – 4M – M 0

In the above solution, though the elements of net evaluation row are either negative or
zeros, the presence of artificial variable ‘A’ as problem variable with value 4 shows that the
problem has no feasible solution because the positive value of A violates the second constraint
of the problem.
(Students can verify the solution by graphical method).

3.12.3. Problem with Unrestricted Variables

The non–negativity constraint of a linear programming problem restricts that the values of all
variables in the problem say for example x, y and z or a, b, c and d etc must be = 0. Some times we may
come across a situation that the values of the variables are unrestricted, i.e., may assume any value ( 0
or > 1 or < 1) i.e., to say that the ≥ sign is not required. In such cases to maintain non-negativity
restriction for all variables, each variable is replaced by two non-negative variables, say for example: x,
y and z are replaced by x' and x", y' and y", z' and z" respectively. If x' > x", then x is positive, while
x' < x" then x is negative.
Problem 3.27: Maximize Z = 2x + 3y s.t.
–1x + 2y ≤ 4
1x + 1y ≤ 6
1x + 3y ≤ 9 and x and y are unrestricted.
As it is given that both x and y are unrestricted in sign they are replaced by non–negative variables,
x' and x", y' and y" respectively. This is subjected to x = x' – x" and y = y' – y". By introducing slack
variables and non–negative variables the simplex format is:
Maximize Z = 2 (x' – x") + 3 (y' – y") + 0S1 + 0S2 + 0S3 s.t.
100 Operations Research

– 1 (x' – x") + 2 (y' – y") + 1S1 + 0S2 + 0S3 = 4

1 (x' – x") + 1 (y' – y") + 0S1 + 1S2 + 0S = 6
1 (x' – x") + 3 (y' – y") + 0S1 + 0S2 + 1S3 = 9
x', x", y', y" and S1, S2 and S3 all ≥ 0.
Table: I. x' = x = y' = y" = 0, S1 = 4, S2 =6, S3 9 and Z = Rs. 0/–

Problem Profit Cj 2 –2 3 –3 0 0 0 Replace–

variable Rs. Capacity units x' x" y' y" S1 S2 S3 ment ratio

S1 0 4 –1 1 2 –2 1 0 0 2

S2 0 6 1 –1 1 –1 0 1 0 6

S3 0 9 1 –1 3 –3 0 0 1 3

Net evaluation 2 –2 3 –3 0 0 0

Table: II. x' = 0, x" = 0, y' = 3, y" = 0, S1 = 0, S2 = 4, S3 = 3, Z = Rs. 6/–

Problem Profit Cj 2 –2 3 –3 0 0 0 Replace–

variable Rs. Capacity units x' x" y' y" S1 S2 S3 ment ratio

y' 3 2 – 1/2 ½ 1 –1 1/2 0 0 –4

S2 0 4 3/2 – 3/2 0 0 – 1/2 1 0 8/3

S3 0 3 5/2 – 5/2 0 0 – 3/2 0 1 6/5

Net evaluation 7/2 – 7/2 0 0 – 3/2 0 0

Table:III. x' = 6/5, x" = 0, y' = 13/3, y" = 0, S1 = 0, S2 = 11/5, S3 = 0, Z = Rs. 10.20

Problem Profit Cj 2 –2 3 –3 0 0 0 Replace–

variable Rs. Capacity units x' x" y' y" S1 S2 S3 ment ratio

y' 3 13/5 0 0 1 –1 1/5 0 1/5 13

S2 0 11/5 0 0 0 0 2/5 1 – 3/5 11

x' 2 6/5 1 –1 0 0 –3/5 0 2/5 —

Net evaluation 0 0 0 0 3/5 0 – 7/5

Linear Programming Models : Solution by Simplex Method 101

Table: IV. x' = 9/2, x" = 0, y' = 3/2, y" = 0, S1 = 0, S2 = 11/2, S2 = 0, S2 = 0, Z = Rs. 13.50

Problem Profit Cj 2 –2 3 –3 0 0 0 Replace–

variable Rs. Capacity units x' x" y' y" S1 S2 S3 ment ratio

y' 3 3/2 0 0 1 –1 0 – 1/2 1/2

S1 0 11/2 0 0 0 0 1 5/2 – 3/2

x’ 2 9/2 1 –1 0 0 0 3/2 – 1/2

Net evaluation 0 0 0 0 0 – 3/2 – 1/2

Shadow price = 3/2 × 6 + ½ × 9 = 9 + 4.5 = Rs. 13.5

x = x' – x" = 9/2 – 0 = 9/2
y = y' – y" = 3/2 – 0 = 3/2 and Z = 2 × 9/2 + 3 × 3/2 = Rs. 13.50.
Problem 3.28: Minimize Z = 1a + 1b + 1c s.t
1a – 3b + 4c = 5
1a – 2b ≤ 3
2a – 1c ≥ 4 and
a and b are ≥ 0 and c is unrestricted.
(Note: Whenever the range of a variable is not given in the problem, it should be understood
that such variable is unrestricted in sign).
Solution: Replace variable 'c' by c' – c" and introducing slack variable for inequality 2 and
surplus variable and artificial variable for inequalities 1 and 3 respectively, we can write simplex version
of the problem.
Simplex version:
Maximize Z = – 1a – 1b – 1c' + 1c" + 0A1 + 0S2 – MA1 – MA2 s.t.
1a – 3b + 4c' – 4c" + 1A1 + 0S1 + 0S2 + 0A2 = 5
1a – 2b + 0c' – 0c" + + 0A1 + 1S1 + 0S2 + 0A2 = 3
0a + 2b – 1c' + 1c" + 0A1 + 0S1 – 1S2 + 1A2 = 4
And a, b, c', c", S1, S2, S3 and A1, A2 all ≥ 0.
Table: I. A = 0, b = 0, c' = 0, c" = o, S1= 3, S2 = 0, A1 = 5, A2 = 4 and Z = Rs. – 9M

Problem Profit Cj –1 –1 –1 1 0 0 – M – M Replace–

variable Rs. Capacity Units a b c' c" S1 S2 A 1 A 2 ment ratio

A1 –M 5 1 –3 4 –4 0 0 1 0 5/4

S1 0 3 1 –2 0 0 1 0 0 0 3/0

A2 –M 4 0 2 –1 1 0 –1 0 1 negative

Net evaluation – 1 + M – 1 – M – 1 + 3M 1 – 3M 0 –M 0 0
102 Operations Research

Table: II. A = 0, b = 0, c' = 5/4, c" = 0, S1 = 3, S2 = 0, A1 = 0, A2 = 21/4, Z = Rs. –5/4 – 21/4 M

Problem Profit Cj –1 –1 –1 1 0 0 – M – M Replace–

variable Rs. Capacity units a b c' c" S1 S2 A 1 A 2 ment ratio

c' –1 5/4 1/4 – 3/4 1 –1 0 0 0 Negative

S1 0 3 1 –2 0 0 1 0 0 Negative

A2 –M 21/4 1/4 5/4 0 0 0 –1 1 21/5

Net evaluation – 3/4 + M/4 – 7/4 + 5/4 M 0 0 0 –M 0

Table: III. a = 0, b = 21/5, c' = 22/5, c" = 0, S1 = 57/5, S2 = 0, A1 = 0, A2 = 0, Z = Rs. 43/5

Problem Profit Cj –1 –1 –1 1 0 0 –M –M Replace–

variable Rs. Capacity units a b c' c" S1 S2 A1 A2 ment ratio

c' –1 22/5 2/5 0 1 –1 0 – 3/5

S1 0 57/5 7/5 0 0 0 1 – 8/5

b –1 21/5 1/5 1 0 0 0 – 4/5

Net evaluation – 2/5 0 0 0 0 – 7/5

As the elements of net evaluation row are either negative or zeros, the solution is optimal. The
answer is a = 0, b = 21/5, c = 22/5 – 0 = 22/5, and Z = 21/5 + 22/5 = Rs. 43/5.
(In this example, the columns of artificial variable is eliminated whenever it goes out of the
programme)
Example 3.29: Solve the given l.p.p
Maximize Z = 0a + 8b s.t.
a–b ≥ 0
2a + 3b ≤ – 6 and both a and b are unrestricted.
As a and b are unrestricted, they may be + ve or –ve or zero. As non–negativity constraint is a
condition for simplex method, this can be solved by writing a = a' – a" and b = b' – b" so that a', a",
b' and b" all ≥ 0.
Now the revised problem is: Simplex version is:
Maximize Z = 0a' + 0a" + 8b' – 8b" s.t. Maximize Z = 0a' + 0a" + 8b' – 8b" + 0S1 + 0S2 – MA1
– MA2 s.t.
(1a' – 1a") – 1 (b' – b") ≥ 0 a' – a" – b' + b" – 1S1 + 0S2 + 1A1 + 0A2 = 0
– 2 (a' – a") – 3 (b' – b") > 6 – 2a' + 2a" – 3b' + 3b" + 0S1 – 1S2 + 0A1 + 1A2 = 6
And a', a", b', b" all ≥ 0 a', a", b', b", S1, S2, A1 and A2 all ≥ 0.
Linear Programming Models : Solution by Simplex Method 103

Table: I. a' = 0, a" = 0, b' = 0, b" = 0, S1 = 0, S2 = 0, A1 = 0 and A2 = 6 and Z = Rs.–6M

Problem Profit Cj 0 0 8 –8 0 0 –M –M Replace–

variable Rs. Capacity units a' a" b' b" S1 S2 A1 A2 ment ratio

A1 –M 0 1 –1 –1 1 1 0 0 1 0

A2 –M 6 –2 2 –3 3 0 –1 0 1 6/3 = 2

Net evaluation –M M 8 – 4M 4M – 8 – M –M 0 1

Table: II. a' = 0, a" = 0, b' = 0, b" = 0, S1 = 0, S2 = 0, A1 = 0, A2 = 6 and Z = Rs. –6M

Problem Profit Cj 0 0 8 –8 0 0 – M – M Replace–

variable Rs. Capacity units a' a" b' b" S1 S2 A1 A 2 ment ratio

b" –8 0 1 –1 –1 1 –1 0 0 Negative

A2 –M 6 –5 5 0 0 3 –1 1 6/5

Net evaluation 8 – 5M –8 + 5M 0 0 –8 + 3M –M 0

Table: III. a" = 6/5, b" = 6/5, a' = 0, b' = 0, S1 = 0, S2 = 0, A1 = 0, A2 = 0, Z = Rs. –8/5

Problem Profit Cj 0 0 8 –8 0 0 –M –M Replace–

variable Rs. Capacity units a' a" b' b" S1 S2 A1 A2 ment ratio

b" –8 6/5 0 0 –1 1 –2/5 1/5

a" 0 6/5 –1 1 0 0 3/5 –1/5

Net Evaluation. 0 0 0 0 –16/5 –8/5

Answer: a" = 6/5, b" = 6/5 and Z = Rs.– 48/5. The shadow price = 8/5 × 6 = Rs. 48/5. = Rs.
9.60
Problem 3.30: Solve the given l.p.p.
Maximize Z = 4x + 5y – 3z s.t.
1x + 1y + 1z = 10 ...(1)
1x – 1y ≥ 1 ...(2)
104 Operations Research

2x + 3y + 1z ≤ 40 ...(3)
And x, y, z all ≥ 0.
Solution: (Note: In case there is equality in the constraints, then one variable can be
eliminated from all inequalities with ≤ or ≥ sign by subtracting the equality from the inequality.
Then the variable, which is eliminated from equality, is treated as slack variable for further
calculations).
Now, subtracting (1) from (3) to eliminate z from (3) and retaining z to work as slack variable for
the equality (1), the given problem becomes:
Given Problem: Simplex version:
Maximize Z = 4x + 5y – 3z s.t. Maximize: Z = 4x + 5y – 3z + 0S1 – MA + 0S2 s.t.
1x + 1y + 1z = 10 1x + 1y + 1z + 0S1 + 0A + 0S2 = 10
1x – 1y + 0z ≥1 1x – 1y + 0z –1S1 + 1A + 0S2 = 1
1x + 2y + 0z ≤ 30 1x + 2y + 0z + 0S1 + 0A + 1S2 = 30
x, y, and z all ≥ 0. x, y, z, S1, S2 A all ≥ 0.

Table: I. X = 0, y = 0, z = 10, S1 = 0, S2 30, A = 1 and Z = Rs. –30 – M

Problem Profit Cj 4 5 –3 0 –M 0 Replacement

variable Rs. Capacity units x y z S1 A S2 ratio

z –3 10 1 1 1 0 0 0 10/1 = 10

A –M 1 1 –1 0 –1 1 0 1/1 = 1

S2 0 30 1 2 0 0 0 1 30/1 = 30

Net evaluation 7+M 8–M 0 –M 0 0

Table: II. x = 1, y = 0, z = 9, S1 = 0, S2 = 29, A = 0, Z = Rs. – 23/–.

Problem Cost Cj 4 5 –3 0 –M 0 Replacement

variable Rs. Capacity units x y z S1 A S2 ratio

z –3 9 0 2 1 1 0 9/2 = 4.5

x 4 1 1 –1 0 –1 0 Negative

S2 0 29 0 3 0 1 1 29/3 = 9.33

Net evaluation 0 9 0 4 0
Linear Programming Models : Solution by Simplex Method 105

Table: III. x = 11/2, y = 9/2, z = 0, S1 = 0, S2 = 31/2, A = 0, Z = Rs. 45/2 + 44/2 = 89/2 =

Rs. 44.50.

Problem Cost Cj 4 5 –3 0 –M 0 Replacement

variable Rs. Capacity units x y z S1 A S2 ratio

y 5 9/2 0 1 1/2 1/2 0

x 4 11/2 1 0 1/2 – 1/2 0

S2 0 31/2 0 0 – 3/2 – 1/2 1

Net evaluation 0 0 – 15/2 – 1/2 0

As all the net evaluation row elements are either negative or zeros, the solution is optimal.
x = 11/2, y = 9/2 and Z = Rs. 44.50.
Problem 3.31: Maximize Z = 4x + 5y – 3z + 50 s.t
1x + 1y + 1z = 10,
1x – 1y ≥ 1
2x + 3y + 1z ≤ 40
x, y, z all ≥ 0.
Note: If any constant is included in the objective function, it should be deleted in the
beginning and finally adjusted in optimum value of Z and if there is equality in the constraints,
then one variable can be eliminated from the inequalities with ≥ or ≤ sign. In this example
constant 50 presents in the objective function. Also one equality is present in the constraints.
As done earlier, variable ‘z’ is eliminated from third constraint and is considered as slack
variable in first equality.
The problem 3.30 and 3.31 are one and the same except a constant present in the objective
function of problem 3.31.
The solution for the problem 3.30 is x = 11/2, y = 9/2 and Z = 89/2.
While solving the problem 3.31, neglect 50 from the objective function, and after getting the final
solution, add 50 to that solution to get the answer. That is the solution for the problem 3.31 is Z = Rs.
89/2 + 50 = Rs. 189/2.

3.13. DUALITY IN LINEAR PROGRAMMING PROBLEMS

Most important finding in the development of Linear Programming Problems is the existence of duality
in linear programming problems. Linear programming problems exist in pairs. That is in linear
programming problem, every maximization problem is associated with a minimization problem.
Conversely, associated with every minimization problem is a maximization problem. Once we have a
problem with its objective function as maximization, we can write by using duality relationship of linear
programming problems, its minimization version. The original linear programming problem is known
as primal problem, and the derived problem is known as dual problem.
The concept of the dual problem is important for several reasons. Most important are (i) the
variables of dual problem can convey important information to managers in terms of formulating their
future plans and (ii) in some cases the dual problem can be instrumental in arriving at the optimal
solution to the original problem in many fewer iterations, which reduces the labour of computation.
106 Operations Research

Whenever, we solve the primal problem, may be maximization or minimization, we get the solution
for the dual automatically. That is, the solution of the dual can be read from the final table of the primal
and vice versa. Let us try to understand the concept of dual problem by means of an example. Let us
consider the diet problem, which we have discussed while discussing the minimization case of the
linear programming problem.
Example: The doctor advises a patient visited him that the patient is weak in his health due to shortage
of two vitamins, i.e., vitamin X and vitamin Y. He advises him to take at least 40 units of vitamin X and
50 units of Vitamin Y everyday. He also advises that these vitamins are available in two tonics A and B.
Each unit of tonic A consists of 2 units of vitamin X and 3 units of vitamin Y. Each unit of tonic B
consists of 4 units of vitamin X and 2 units of vitamin Y. Tonic A and B are available in the medical shop
at a cost of Rs. 3 per unit of A and Rs. 2,50 per unit of B. The patient has to fulfill the need of vitamin
by consuming A and B at a minimum cost.
The problem of patient is the primal problem. His problem is to minimize the cost. The tonics
are available in the medical shop. The medical shop man wants to maximize the sales of vitamins A and
B; hence he wants to maximize his returns by fixing the competitive prices to vitamins. The problem of
medical shop person is the dual problem. Note that the primal problem is minimization problem
and the dual problem is the maximization problem.
If we solve and get the solution of the primal problem, we can read the answer of dual problem
from the primal solution.
Primal problem: Dual Problem:
Minimize Z = 3a + 2.5b. s.t. Maximize Z = 40x + 50y s.t.
2a + 4b ≥ 40 2x + 3y ≤ 3
3a + 2b ≥ 50 4x + 2y ≤ 2.50
both a and b are ≥ 0. both x and y are ≥ 0.
Solution to Primal: (Minimization problem i.e., patient’s problem)
Problem Cost Cj 3 2.50 0 0 M M
variable Rs. Requirement a b p q A1 A2
b 2.50 5/2 0 1 –3/8 1/4 3/8 –1/4
a 3 15 1 0 1/4 –1/2 –1/4 1/2
Net evaluations 0 0 3/16 7/8 M – 3/16 M – 7/8

Answer: a = 15 units, b = 2.5 units and total minimum cost is Rs. 51.25
Solution to Dual: (Maximization problem i.e medical shop man’s problem)
Problem Profit Cj 40 50 0 0
variable Rs. Capacity units x y S1 S2
y 50 7/8 0 1 1/2 –1/4
x 40 3/16 1 0 –1/4 3/8
Net evaluation 0 0 –15 –5/8
Linear Programming Models : Solution by Simplex Method 107

Answer: x = 3/16, y = 7/8, and maximum profit is Rs. 51.25

The patient has to minimize the cost by purchasing vitamin X and Y and the shopkeeper has to
increase his returns by fixing competitive prices for vitamin X and Y. Minimum cost for patient is
Rs. 51.25 and the maximum returns for the shopkeeper is Rs. 51.25. The competitive price for tonics
is Rs.3 and Rs.2.50. Here we can understand the concept of shadow price or economic worth of
resources clearly. If we multiply the original elements on the right hand side of the constraints with the
net evaluation elements under slack or surplus variables we get the values equal to the minimum cost of
minimization problem or maximum profit of the maximization problem. The concept of shadow price
is similar to the economist’s concept of the worth of a marginal resource. In other words, we can see
for a manufacturing unit it is machine hour rate. It is also known as imputed value of the resources.
One cannot earn more than the economic worth of the resources he has on his hand. The fact that the
value of the objective function in the optimal program equals to the imputed value of the
available resources has been called the FUNDAMENTAL THEOREM OF LINEAR
PROGRAMMING.
By changing the rows of the primal problem (dual problem) into columns we get the dual
problem (primal problem) and vice versa.
To understand the rationale of dual problem and primal problem, let us consider another example.
Problem 3.32: A company manufactures two products X and Y on three machines Turning,
Milling and finishing machines. Each unit of X takes, 10 hours of turning machine capacity, 5 hours of
milling machine capacity and 1 hour of finishing machine capacity. One unit of Y takes 6 hours of
turning machine capacity, 10 hours of milling machine capacity and 2 hours of finishing machine
capacity. The company has 2500 hours of turning machine capacity, 2000 hours of milling machine
capacity and 500 hours of finishing machine capacity in the coming planning period. The profit
contribution of product X and Y are Rs. 23 per unit and Rs. 32 per unit respectively. Formulate the
linear programming problems and write the dual.
Solution:

Pro duct
Department X Y Available Capacity

Turning 10 6 2500
Milling 5 10 2000
Finishing 1 2 500
Profit per unit in Rs. 23 32

Let us take the maximization problem stated to be primal problem. Associated with this maximization
problem is the minimization problem that is the dual of the given primal problem. Let us try to formulate
the dual by logical argument.
The primal problem is the seller’s maximization problem, as the seller wants to maximize his
profit. Now the technology, i.e., the machinery required are with the seller and they are his available
resources. Hence he has to prepare the plans to produce the products to derive certain profit and he
wants to know what will be his profit he can get by using the available resources. Hence the buyer’s
problem is:
Maximize 23 x + 32y s.t
108 Operations Research

10x + 6y ≤ 2500
5x + 10y ≤ 2000
1x + 2y ≤ 500 and both x and y ≥ 0 (This we shall consider as primal problem).
Associated with seller’s maximization problem is a buyer’s minimization problem. Let us assume
seller will pretend to be the buyer. The rationale is the buyer, it is assumed, will consider the purchase
of the resources in full knowledge of the technical specifications as given in the problem. If the buyer
wishes to get an idea of his total outlay, he will have to determine how much must he pay to buy all the
resources. Assume that he designates variables by a, b, c, to represent per unit price or value that he
will assign to turning, milling and finishing capacities, respectively, while making his purchase plans.
The total outlay, which the buyer wishes to minimize, will be determined by the function 2500a +
2000b + 500c, which will be the objective function of the buyer. The linear function of the objective
function mentioned, must be minimized in view of the knowledge that the current technology yields a
profit of Rs. 23 by spending 10 machine hours of turning department, 5 machine hours of milling
department and 1 man-hour of finishing department. Similarly we can interpret other constraints also.
Now the buyer’s minimization problem will be:
Minimize. Z = 2500a + 2000b + 500c s.t.
10a + 5b + 1c ≥ 23
6a + 10b + 2c ≥ 32 and all a, b, c, are ≥ 0 (This minimization version is the dual of seller’s
primal problem given above), where a, b, and c are dual variables and x, y, z are primal variables.
The values assigned to the dual variables in the optimal tableau of the dual problem, represent
artificial accounting prices, or implicit prices or shadow prices, or marginal worth or machine
hour rate of various resources. Because of this, we can read the values of dual variables from the net
evaluation row of final tableau of primal problem. The values will be under slack variable column in net
evaluation row.
The units of the constraint to which the dual variable corresponds determine the dimension of any
dual variable.
In this problem the dimension of variable ‘a’ as well as ‘b’ is rupees per machine hour and that of
variable 'c' is rupees per man-hour.
Another important observation is by definition, the entire profit in the maximization must be
traced to the given resources, the buyer’s total outlay, at the equilibrium point, must equal to the total
profit. That is, optimal of the objective function of the primal equals to the optimal value of the objective
function of the dual. This observation will enable the problem solver to check whether his answer is
correct or not. The total profit (cost) of maximization problem (minimization problem) must be equal
to the shadow price (or economic worth) of resources.
The given primal problem will have symmetrical dual. The symmetrical dual means all given
structural constraints are inequalities. All variables are restricted to nonnegative values.
Now let us write primal and dual side by side to have a clear idea about both.
Primal problem: Dual Problem:
Maximize Z = 23x + 32y s.t. Minimize Z = 2500a + 2000b + 5000c s.t.
10x + 6y ≤ 2500 10a + 5b + 1c ≥ 23
5x + 10y ≤ 2000 6a + 10b + 2c ≥ 32
1x + 2y ≤ 500
Both x and y are ≥ 0. All a, b, c, are ≥ 0.
Linear Programming Models : Solution by Simplex Method 109

Now let us discuss some of the important points that are to be remembered while dealing with
primal and dual problem. Hence the characteristics are:
Note:
1. If in the primal, the objective function is to be maximized, then in the dual it is to be
minimized.
Conversely, if in the primal the objective function is to be minimized, then in the
dual it is to be maximized.
2. The objective function coefficients of the prima appear as right-hand side numbers in
the dual and vice versa.
3. The right hand side elements of the primal appear as objective function coefficients
in the dual and vice versa.
4. The input - output coefficient matrix of the dual is the transpose of the input – output
coefficient matrix of the primal and vice versa.
5. If the inequalities in the primal are of the “less than or equal to” type then in the
dual they are of the “greater than or equal to” type. Conversely, if the inequalities
in the primal are of the “greater than or equal to” type; then in the dual they are of
the “less than or equal to” type.
6. The necessary and sufficient condition for any linear programming problem and its
dual to have optimum solution is that both have feasible solution. Moreover if one of
them has a finite optimum solution, the other also has a finite optimum solution.
The solution of the other (dual or primal) can be read from the net evaluation row
(elements under slack/surplus variable column in net evaluation row). Then the values
of dual variables are called shadow prices.
7. If the primal (either) problem has an unbound solution, then the dual has no solution.
8. If the i th dual constraints are multiplied by –1, then i th primal variable computed
from net evaluation row of the dual problem must be multiplied by –1.
9. If the dual has no feasible solution, then the primal also admits no feasible solution.
10. If k th constraint of the primal is equality, then the k th dual variable is unrestricted
in sign.
11. If p th variable of the primal is unrestricted in sign, then the p th constraint of the
dual is a strict equality.
Summary:
Primal Dual
(a) Maximize. Minimize
(b) Objective Function. Right hand side.
(c) Right hand side. Objective function.
(d) i th row of input-output coefficients. i th column of input output coefficients.
(e) j th column of input-output coefficients. j the row of input-output coefficients.
(f) i th relation of inequality ( ≤ ). i th variable non-negative.
(g) i th relation is an equality (=). i the variable is unrestricted in sign.
(h) j th variable non-negative. j relation an inequality ( ≥ ).
(i) j th variable unrestricted in sign. j th relation an equality.
110 Operations Research

Note:
1. Primal of a Prima is Primal
2. Dual of a Dual is Primal.
3. Primal of a Dual is Primal.
4. Dual of a Primal is Dual.
5. Dual of a Dual of a Dual is Primal.

3.13.1. Procedure for converting a primal into a dual and vice versa
Case 1: When the given problem is maximization one:
The objective function of primal is of maximization type and the structural constraints are of ≤
type. Now if the basis variables are x, y and z, give different name for variables of dual. Let them be a,
b, and c etc. Now write the structural constraints of dual reading column wise. The coefficients of
variables in objective function will now become the left hand side constants of structural constraints.
And the left hand side constants of primal will now become the coefficients of variables of objective
function of dual. Consider the example given below:
Primal Dual
Maximize: Z = 2x + 3y s.t. Minimize: Z = 10a + 12b s.t.
1x + 3y ≤ 10 1a + 2b ≥ 2
2x + 4y ≤ 12 and 3a + 4b ≥ 3 and
Both x and y are ≥ 0 both a and b are ≥ 0
Case 2: When the problem is of Minimization type:
The procedure is very much similar to that explained in case 1. Consider the example below:
Primal Dual
Minimize Z = 10x + 12y s.t Maximize Z = 2a + 3b s.t.
1x + 2 y ≥ 2 1a + 3b ≤ 10
3x + 4y ≥ 3 and 2a + 4b ≤ 12 and
Both x and y are ≥ 0 Both a and b are ≥ 0.
Case 3: When the problem has got both ≥ and ≤ constraints, then depending upon the objective
function convert all the constraints to either ≥ or ≤ type. That is, if the objective function is of
minimization type, then see that all constraints are of ≥ type and if the objective function is of
maximization type, then see that all the constraints are of ≤ type. To convert ≥ to ≤ or ≤ to ≥, simply
multiply the constraint by –1. Once you convert the constraints, then write the dual as explained in
case 1 and 2. Consider the example:
Primal: The primal can be written as:
Maximize Z = 2a + 3b s.t. Maximize Z = 2a + 3b s.t.
1a + 4b ≤ 10 1a + 4b ≤ 10
2a + 3b ≥ 12 and –2a – 3b ≤ – 12 and both a and b are ≥ 0.
Both a and b are ≥ 0
Linear Programming Models : Solution by Simplex Method 111

Now dual is:

Minimize Z = 10x – 12y s.t.
1x – 2y ≥ 2
4x – 3y ≥ 3 and both x and y are ≥ 0.
Similarly when the objective function of primal is of minimization type, then same procedure is
adopted. See that all the constraints are of ≥ type.
Primal: Primal can be written as:
Minimise Z = 4a + 5b s.t. Minimise Z = 4a + 5b s.t.
3a + 2b ≤ 10 – 3a – 2b ≥ – 10
2a + 4b ≥ 12 and 2a + 4b ≥ 12 and both a and b are ≥ 0.
Both a and b are ≥ 0.
The dual is:
Maximize Z = – 10x + 12y s.t.
– 3x + 2y ≤ 4
– 2x + 4y ≤ 5 and
Both x and y are ≥ 0.
Case 4: When one of the constraint is an equation, then we have to write two versions of the equation,
that is remove equality sign and write ≥ and ≤ sign for each one of them respectively and then write
the dual as usual. Consider the example given below:
Primal: Primal can be written as:
Maximize Z = 2x + 3y s.t. Maximise Z = 2x + 3y s.t.
1x + 2y ≥ 10 1x + 2y ≥ 10
2x + 2y = 20 and 2x + 2y ≥ 20
Both x and y are ≥ 0. 2x + 2y ≤ 20 and both x and y are ≥ 0
This can be written as: The dual is:
Maximise Z = 2x + 3y s.t. Minimise Z = –10a – 20b + 20c s.t.
–1x – 2y ≤ – 10 –1a – 2b + 2c ≥ 2
–2x – 2y ≤ – 20 –2a – 2b + 2c ≥ 3 and a, b, c all ≥ 0.
2x + 2y ≤ 20 and both x and y are ≥ 0.
Problem 3.32:
Write dual of the given l.p.p. Dual of the given problem is:
Minimize Z = 3x + 1y s.t. Maximize Z = 2a + 1b s.t.
2x + 3y ≥ 2 2a + 1b ≤ 3
1x + 1y ≥ 1 and 3a + 1b ≤ 1 and
Both x and y are ≥ 0. Both a and b are ≥ 0
Simplex version (Primal) Simplex version: (Dual)
Minimize Z = 3x + 1y + 0p + 0q + MA1 + MA2 s.t. Maximize Z = 2a + 1b + 0S1 + 0S2 s.t.
2x + 3y – 1p + 0q + 1A1 + 0A2 = 2 2a + 1b + 1S1 + 0S2 = 3
112 Operations Research

1x + 1y + 0p – 1q + 0A1 + 1A2 = 1 and 3a + 1b + 0S1 + 1S2 = 1 and

x, y, p, q, A1 and A2 all ≥ 0 a, b, S1, S2 all ≥ 0
Problem 3.33: Write the dual of the primal problem given and solve the both and interpret the
results.
Primal Problem: Simplex version:
Maximize Z = 5a + 20b s.t. Maximize Z = 5a + 20b + 0S1 + 0S2 + 0S3 s.t.
5a + 2b ≤ 20 5a + 2b + 1S1 + 0S2 + 0S3 = 20
1a + 2b ≤ 8 1a + 2b + 0S1 + 1S2 + 0S3 = 8
1a + 6b ≤ 12 and 1a + 6b + 0S1 + 0S2 + 1S3 = 12
Both a and b ≥ 0 and a, b. S1, S2, and S3 all ≥ 0
First let us solve the Primal problem by using simplex method and then write the dual and solve
the same.
Table: I. a = 0, b = 0, S1 =20, S2 = 8, S3 = 12 and Z = Rs. 0

Problem Profit Cj 5 20 0 0 0 Replacement

variable Rs. Capacity units a b S1 S2 S3 ratio
S1 0 20 5 2 1 0 0 10
S2 0 8 1 2 0 1 0 4
S3 0 12 1 6 0 0 1 2

Net evaluation 5 20 0 0 0

Table: II. a = 0, b = 2, S1 16, S2 = 4, S3 = 0, and Z = Rs. 40/–

Problem Profit Cj 5 20 0 0 0 Replacement

variable Rs. Capacity units a b S1 S2 S3 ratio
S1 0 16 14/4 0 1 0 – 1/3 24/7 = 21 3/7
S2 0 4 2/3 0 0 1 – 1/3 12/2 = 6
B 20 2 1/6 1 0 0 – 1/6 12
Net evaluation 5/3 0 0 0 – 10/3

Table: III. a = 24/7, b = 10/7, S1 = 0, S2 = 12/7, S3 = 0, and Z = Rs. 45.75.

Problem Profit Cj 5 20 0 0 0 Replacement

variable Rs. Capacity units a b S1 S2 S3 ratio
a 5 24/7 1 0 3/14 0 – 1/14
S2 0 12/7 0 0 – 1/7 1 – 2/7
b 20 10/7 0 1 – 1/28 0 5/28
Net evaluation 0 0 – 5/14 0 – 45/14
Linear Programming Models : Solution by Simplex Method 113

As all the values in net evaluation row are either zeros or negative elements, the solution is
optimal.
Hence, a = 24/7, b = 10/7 and optimal profit is Rs. 45.75.
Now let us solve the dual of the above.
Dual of the given problem: Simplex version:
Minimize Z = 20x + 8y + 12z s.t.
i.e., Maximize Z = –20x – 8y – 12z s.t. Maximize Z = –20x – 8y – 12z + 0S1 + 0S2 –
MA1 – M A s.t.
5x + 1y + 1z ≥ 5 5x + 1y + 1z – 1S1 + 0S2 + 1A1 + 0A2 = 5
2x + 2y + 6z ≥ 20 2x + 2y + 6z + 0S1 – 1S2 + 0A1 + 1A2 = 20
and x, y, and z all ≥ 0 and x, y, z, S1, S2, A1, A2 all ≥ 0.
Two Phase version is: Maximize: Z = 0x + 0y + 0z + 0S1 + 0S2 – 1A1 – 1A2 s.t.
5x + 1y + 1z – 1S1 + 0S2 + 1A1 + 0A2 = 5
2x + 2y + 6z + 0S1 – 1S2 + 0A1 + 1A2 = 20
And x, y, z, S1, S2, A1, A2 all ≥ 0.

Table: I. x = 0, y = 0, z = 0, S1 = 0, S2 = 0, A1 = 5, A2 = 20, Z = – Rs 25/–

Problem Profit Cj 0 0 0 0 0 –1 –1 Replace–

variable Rs. Capacity units x y z S1 S2 A1 A2 ment ratio

A1 –1 5 5 1 1 –1 0 1 0 1

A2 –1 20 2 2 6 0 –1 0 1 10

Net evaluation 7 3 7 –1 –1 0 0

Table: II. a = 1, y = 0, z = 0, S1 = 0, S2 = 0, A1 = 0, A = 20, Z = – Rs.18/–

Problem Profit Cj 0 0 0 0 0 –1 –1 Replace–

variable Rs. Capacity units x y z S1 S2 A1 A2 ment ratio

x 0 1 1 1/5 1/5 –1/5 0 1/5 0 5

A2 –1 18 0 8/5 28/5 2/5 –1 –2/5 1 45/14

Net evaluation 0 8/5 28/5 2/5 –1 –2/5 0

114 Operations Research

Table: III. x = 5/14, y = 0, z = 45/14, S1 = 0, S2 = 0, A1 = 0, A2 = 0, Z = Rs. 0

Problem Profit Cj 0 0 0 0 0 –1 –1 Replace–

variable Rs. Capacity units x y z S1 S2 A1 A2 ment ratio

x 0 5/14 1 1/7 0 –3/14 1/28 3/14 –1/28

z 0 45/14 0 2/7 1 1/14 –5/28 –1/14 5/28

Net evaluation 0 0 0 0 0 –1 –1

Table: IV. x = 5/14, y = 0, z = 45/14, S1 = 0, S2 = 0, A1 = 0, A2 = 0 and Z = – Rs. 45.75.

Problem Profit Cj 0 0 0 0 0 –1 –1 Replace–

variable Rs. Capacity units x y z S1 S2 A1 A2 ment ratio

x –20 5/14 1 1/7 0 –3/14 1/28 3/14 –1/28

z –12 45/14 0 2/7 1 1/14 –5/28 –1/14 5/28

Net evaluation 0 –12/7 0 –24/7 –10/7 24/7 – M 10/7 – M

x = 5/14, z = 45/14 and Z = Rs. 45.75.

Now let us compare both the final (optimal solution) table of primal and dual.
Optimal solution table of Primal
Table: III.a = 24/7, b = 10/7, S1 = 0, S2 = 12/7, S3 = 0, and Z = Rs. 45.75.

Problem Profit Cj 5 20 0 0 0 Replacement

variable Rs. Capacity units a b S1 S2 S3 ratio

a 5 24/7 1 0 3/14 0 –1/14

S2 0 12/7 0 0 –1/7 1 –2/7

b 20 10/7 0 1 –1/28 0 5/28

Net evaluation 0 0 –5/14 0 –45/14

Linear Programming Models : Solution by Simplex Method 115

Optimal Solution table of Dual:

Table: IV. x = 5/14, y = 0, z = 45/14, S1 = 0, S2 = 0, A1 = 0, A2 = 0 and Z = – Rs. 45.75.

Problem Profit Cj 0 0 0 0 0 –1 –1 Replace–

variable Rs. Capacity units x y z S1 S2 A1 A2 ment ratio

x –20 5/14 1 1/7 0 –3/14 1/28 3/14 –1/28

z –12 45/14 0 2/7 1 1/14 –5/28 –1/14 5/28

Net evaluation 0 –12/7 0 –24/7 –10/7 24/7 – M 10/7 – M

Construction of Dual problem:

Problem 3.34: Construct the dual of the given l.p.p.
Maximize Z = 5w + 2x + 6y + 4z s.t
1w + 1x + 1y + 1z ≤ 140
2w + 5x + 6y + 1z ≥ 200
1w + 3x + 1y + 2z ≤ 150
And w, x, y, z all are ≥ 0.
Solution: Dual: As the constraint 2 is of ≥ type and other two are ≤ type and the objective
function is maximization, we have to write the constraint 2 also as ≤ type. Hence multiply the constraint
by –1 and write the dual problem.
Hence the given problem is:
Maximize Z = 5w + 2x + 6y + 4z s.t.
1w + 1x + 1y + 1z ≤ 140
–2w – 5x – 6y – 1z ≤ – 200
1w + 3x + 1y + 2z ≤ 150
And w, x, y, z all are ≥ 0.
Dual is: Minimize Z = 140 a – 200 b + 150c s.t.
1a – 2b + 1c ≥ 5
1a – 5b + 3c ≥ 2
1a – 6b + 1z ≥ 6
1a – 1b + 2z ≥ 4
And a, b, c all ≥ 0.
Problem 3.35: (Primal): Maximize Z = 4a + 3b s.t.
2a + 9b ≤ 100
3a + 6b ≤ 120
1a + 1b = 50 and both a and b are ≥ 0.
When there is equality in the constraints, we have to write two versions of the same, i.e., both ≤
and ≥ version and then we have to write the dual. As the objective function is maximization type, we
116 Operations Research

must take care that all the inequalities are of ≤ type. In case the objective function is minimization type,
then all the inequalities must be of ≥ type.
Dual of the given problem:
The given problem is written as: This can be written as:
Maximize 4a + 3b s.t. Maximize Z = 4a + 3b s.t.
2a + 9b ≤ 100 2a + 9b ≤ 100
3a + 6b ≤ 120 3a + 6b ≤ 120
1a + 1b ≤ 50 1a + 1b ≤ 50
1a + 1b ≥ 50 and both a and b are ≥ 0 –1a – 1b ≤ –50 and both a and b are ≥ 0.
Dual of the primal:
Minimize 100w + 120x + 50y – 50z s.t
2w + 3x + 1y – 1z ≥ 4
9w + 6x + 1y – 1z ≥ 3 and w, x, y, and z all are > 0.
As the number of rows is reduced, the time for calculation will be reduced. Some times while
solving the l.p.p it will be better to write the dual and solve so that one can save time. And after getting
the optimal solution, we can read the answer of the primal from the net evaluation row of the dual.
Problem 3.36: Write the dual of the given primal problem:
Minimize: 1a + 2b + 3c s.t.
2a + 3b – 1c ≥ 20
1a + 2b + 3c ≤ 15
0a + 1b + 2c = 10 and
a, b, c all ≥ 0
As the objective function is minimization type all the inequalities must be of ≥ type. Hence
convert the second constraint by multiplying by –1 into ≥ type. Rewrite the third constraint into ≤ and
≥ type and multiply the ≤ type inequality by –1 and convert it into ≥ type. And then wire the dual.
Given problem can be written as: Dual is:
Minimize 1a + 2b + 3c s.t Maximize Z = 20w – 15x + 10y – 10z s.t
2a + 3b – 1c ≥ 20 2w – 1x + 0y – 0z ≤ 1
–1a – 2b – 3c ≥ – 15 3w – 2x+ 1y – 1z ≤ 2
0a + 1b + 2c ≥ 10 – 1w – 3x + 2y – 2z ≤ 3 and
–0a – 1b – 2c ≥ – 10 and w, x, y and z all ≥ 0.
a, b, and c all ≥ 0.

Points to Remember
1. While writing dual see that all constraints agree with the objective function. That
is, if the objective function is maximization, then all the inequalities must be ≤
type. In case the objective function is Minimization, then the inequalities must be
of ≥ type.
2. If the objective function is maximization and any one or more constraints are of
≥ type then multiply that constraint by –1 to convert it into ≤ type. Similarly, if
Linear Programming Models : Solution by Simplex Method 117

the objective function is minimization type and one or more constraints are of ≤
type, then multiply them by –1 to convert them into ≥ type.
3. In case any one of the constraint is an equation, then write the two versions of the
same i.e., ≤ and ≥ versions, then depending on the objective function, convert
the inequalities to agree with objective function by multiplying by –1.

3.14. SENSITIVITY ANALYSIS

While solving a linear programming problem for optimal solution, we assume that:
(a). Technology is fixed, (b). Fixed prices, (c). Fixed levels of resources or requirements, (d).
The coefficients of variables in structural constraints (i.e. time required by a product on a particular
resource) are fixed, and (e) profit contribution of the product will not vary during the planning period.
These assumptions, implying certainty, complete knowledge, and static conditions, permit us to design
an optimal programme. The condition in the real world however, might be different from those that are
assumed by the model. It is, therefore, desirable to determine how sensitive the optimal solution is to
different types of changes in the problem data and parameters. The changes, which have effect on the
optimal solution are: (a) Change in objective function coefficients (aij), (b) Resource or requirement
levels (bi), (c) Possible addition or deletion of products or methods of production. The process of
checking the sensitivity of the optimal solution for changes in resources and other components of the
problem, is given various names such as: Sensitivity Analysis, Parametric Programming and Post
optimality analysis or what if analysis.
Post optimality test is an important analysis for a manager in their planning process, when they
come across certain uncertainties, say for example, shortage of resources due to absenteeism, breakdown
of machinery, power cut off etc. They may have to ask question ‘what if’, a double–edged sword.
They are designed to project the consequences of possible changes in the future, as well as the impact
of the possible errors of estimation of the past. The need for sensitivity analysis arises due to (i) To
know the effect of and hence be prepared for, possible future changes in various parameters and
components of the problem, (ii) To know the degree of error in estimating certain parameters that
could be absorbed by the current optimal solution. Or to put in other way, sensitivity analysis answers
questions regarding what errors of estimation could have been committed, or what possible
future changes can occur, without disturbing the optimality of the current optimal solution.
The outcome of sensitivity analysis fixes ranges i.e., upper limits and lower limits of parameters
like Cj, aij, bi etc. within which the current optimal programme will remain optimal. Hence, we can say
that the sensitivity analysis is a major guide to managerial planning and control. Also sensitivity analysis
arises the need for reworking of the entire problem from the very beginning each time a change
is investigated or incorporated. The present optimal solution can be used to study the changes with
minimum computational effort. By adding or deleting a new column (product) or adding or deleting a
new row (new process) we can analyze the changes with respect to Cj, aij, and bi.
To summarize the Sensitivity analysis include:
1. Coefficients (Cj) of the objective function, which include:
(a) Coefficients of basic variables (Cj).
(b) Confidents of non–basic variables.
2. Changes in the right hand side of the constraints (bi).
118 Operations Research

3. Changes in aij, the components of the matrix, which include:

(a) Coefficients of the basic variables, aij.
(b) Coefficients of non-basic variables.
4. Addition of new variables to the problem.
5. Addition of new or secondary constraint.
The above changes may results in one of the following three cases:
Case [Link] optimal solution remains unchanged, that is the basic variables and their values remain
essentially unchanged.
Case II. The basic variables remain the same but their values are changed.
Case III. The basic solution changes completely.

3.14.1. Change in the Objective Coefficient

(a) Non-basic Variables
Consider a change in the objective coefficient of the non-basic variable in the optimal solution.
Any change in the objective coefficient of the non-basic variable will affect only its index row coefficient
and not others.
Problem 3.37: Maximize Z = 2a + 2b + 5c + 4d s.t
1a + 3b + 4c + 3d ≤ 10
4a + 2b +6c + 8d ≤ 25 and a, b, c, and d all are ≥ 0.
Table: I. a = 0, b = 0, c = 0, d = 0, S1 = 10, S2 = 25, Z = Rs.0 /–

Problem Profit Cj 2 2 5 4 0 0 Replacement

variable Rs. Capacity units a b c d S1 S2 ratio

S1 0 10 1 3 4 3 1 0 10/4

S2 0 25 4 2 6 8 0 1 25/6

Net evaluation 2 2 5 4 0 0

Table: II. a = 0, b = 0, c = 5/2, d = 0, S1 = 0, S2 = 10, Z = Rs. 12.50

Problem Profit Replacement

variable Rs. Capacity units a b c d S1 S2 ratio

c 5 5/2 1/4 3/4 1 3/4 1/4 0 10

S2 0 10 5/2 – 5/2 0 7/5 3/2 1 4

Net evaluation 3/4 – 1.75 0 1/4 – 5/4 0

Linear Programming Models : Solution by Simplex Method 119

Table: III. a = 4, b = 0, c = 1.5, d = 0, S1 = 0, S2 = 0, Z = Rs. 15.50 /–

Problem Profit Cj 2 2 5 4 0 0
variable Rs. Capacity units a b c d S1 S2

c 5 3/2 0 1 1 2/5 2/5 – 1/10

a 2 4 1 –1 0 7/5 – 3/5 2/5

Net evaluation 0 –1 0 – 4/5 – 4/5 – 3/10

Here ‘a’ and ‘c’ are basic variables and ‘b’ and ‘d’ are non–basic variables. Consider a small
change x1 in the objective coefficient of the variable ‘b’, then its index row (net evaluation row)
element becomes:
(2 + x1) – (–2 + 5) = x1 –1. If variable ‘b’ wants to be an incoming variable x1 – 1 must be
positive. Then the value of x1 should be > 1. Hence when the value of the increment is > 1 then the
present optimal solution changes.
Similarly for D, if x2 is the increment, then (4 + x2) – (14/5 + 10/5) = (4 + x2) – 24/5, hence if 'd'
wants to become incoming variable then the value of x2 > 4/5. To generalize, one can easily conclude
that for non-basic variables when its objective coefficient just exceeds its index row coefficient in the
optimal solution, the present solution ceases to be optimal.
(b) Basic variables:
Now let us consider a change in the objective coefficient of the basic variable in the optimal
solution. Here, it affects the net evaluation row coefficients of all the variables. Hence, as soon as the
net evaluation row coefficients of basic variables become negative, it leaves the solution, and that of
non-basic variable becomes positive, it becomes an incoming variable. In either case, the present
optimal solution changes.
Consider the above example. Let us say that there is a small reduction ‘x1’ in the objective
coefficient of variable ‘a’ i.e., (2– x1) then the net evaluation row coefficients of variables are:
Variable Corresponding change in net evaluation row element.
a (2–x1) – 1 ( 2–x1) = 0
b 2 – {–1 (2–x1) + 5) = – (x1 + 1)
c 0
d – ( 4/5 + 7/5 x1)
S1 –4/5 + 3/5 x1
S2 – 3/10 + 2/5 x1
Results:
(a) For any value of x1 variable 'b' cannot enter the solution.
(b) As soon as x1 is > 4/7, variable 'd' enters the solution.
(c) For any value of x1, S1 will not enter into solution.
(d) As soon as x1 > 3/4 , S2 claims eligibility to enter into solution.
A reduction in objective coefficient of variable 'a' by more than 4/7, the present optimal solution
change. i.e., the value is 2 – 4/7 = 10/7.
120 Operations Research

When the objective coefficient of variable ‘a’ increases by a value x2, the changes are:
Variable Corresponding change.
a (2 + x2) – (2 + x2) = 0
b 2 – (–1) {( 2+x2) + 5)} = –1 + x2
c 0
d –4/5 – 7/5 x2
S1 –4/5 + 3/5 x2
S2 –3/10–2/5 x2
If x2 is ≥ 1 the variable ‘b’ claims the entry into solution and the optimal solution changes.
For any value of x2 variables ‘d’ and S1 are not affected.
If x2 is > 4/3, S1 claims the entry into solution, and the optimal changes.
Hence, as soon as objective coefficient of variable ‘a’ increases by more than 1 the present
optimal solution changes. Hence the maximum permissible value of objective coefficient of ‘a’ is 2 + 1
= 3 for the present optimal solution to remain. That is the range for objective coefficient of variable ‘a’
is 10/7 to 3.

3.14.2. Change in the right – hand side of the constraint

The right hand side of the constraint denotes present level of availability of resources (or requirement
in minimization problems). When this is increased or decreased, it will have effect on the objective
function and it may also change the basic variable in the optimal solution.
Example 3.38: A company manufactures three products: X, Y and Z by using three resources. Each
unit of product X takes three man hours and 10 hours of machine capacity and 1 cubic meter of
storage place. Similarly, one unit of product Y takes 5 man-hours and 2 machine hours on 1cubic meter
of storage place and that of each unit of products Z is 5 man-hours, 6 machine hours and 1 cubic meter
of storage place. The profit contribution of products X, Y and Z are Rs. 4/–, Rs.5/– and Rs. 6/–
respectively. Formulate the linear programming problem and conduct sensitivity analysis when
Maximize Z = 4x + 5y + 6z s.t.
3x + 5y + 5z ≤ 900
10x + 2y + 6z ≤ 1400
1x + 1y + 1z ≤ 250 and all x, y, and z are ≥ 0
The final table of the solution is:
x= 50, y = 0, z = 150, S1 = 0, S2 = 0, S3 = 50 and Z = Rs. 1100/–
Problem Profit Cj 4 2 6 0 0 0
variable Rs. Capacity units x y z S1 S2 S3

z 6 150 0 11/8 1 5/16 – 3/32 0

x 4 50 1 – 5/8 0 – 3/16 5/32 0

S3 0 50 0 1/4 0 – 1/8 – 1/16 1

Net evaluation 0 – 23/4 0 – 9/8 –1/16 0

Linear Programming Models : Solution by Simplex Method 121

The solution is x = 50, y = 0, z = 150, S1 = 0, S2 = 0, S3 = 50 and profit Z = Rs. 1100/–

Here man- hours are completely utilized hence S1 = 0, Machine hours are completely utilized,
hence S3 = 0 but the storage capacity is not completely utilized hence still we are having a balance of 50
cubic meters of storage place i.e., S3 = 50.
Value of dual variable under S1 in net evaluation row is 9/8. This is the shadow price or per unit
price of the resource. The resource is man- hours. Hence it means to say that as we go on increasing
one hour of man-hour resource, the objective function will go on increasing by Rs. 9/8 per hour.
Similarly the shadow price of machine hour is Rs. 1/16 and that of storage space is Rs.0. Similar
reasoning can be given. That is every unit increase in machine hour resource will increase the objective
function by Rs. 1/16 and that of storage space is Rs.0/–
Now let us ask ourselves what the management wants to do:
Question No.1. If the management considers to increase man-hours by 100 hours i.e., from 900
hours to 1000 hours and machine hours by 200 hours i.e., 1400 hours to 1600 hours will the optimal
solution remain unchanged?
Now let us consider the elements in the identity matrix and discuss the answer to the above
question.

Problem variable S1 S2 S3 Capacity B–1 × b =

z 5/16 – 3/32 0 1000 5000/16 – 4800/32 + 0 = 325/2

x – 3/16 5/32 0 1600 – 3000/16 + 8000/32 + 0 = 125/2

S3 – 1/8 – 1/16 1 250 – 1000/8 – 1600/16 + 250 = 25

Now x = 125/2, and z = 325/2 and S3 = 25. As x and z have positive values the current optimal
solution will hold well. Note that the units of x and z have been increased from 50 and 150, to 125/2 and
325/2. These extra units need the third resource, the storage space. Hence storage space has been
reduced from 50 to 25.
Shadow price indicates that resource of man-hours can be increased to increase objective function.
A solution to question No.1 above, showed that with increase of man- hours by 100 (i.e., from 900 to
1000 hours), the basic variables remain the same (i.e., x and z and S3) with different values at the
optimal stage.
Question No. 2: Up to what values the resource No.1, i.e., man-hours can be augmented without
affecting the basic variables? And up to what value the resource man-hours can be without affecting
the basic variables?
Let α be the increment in man- hour resource, then:
Problem
S1 S2 S3 Capacity B–1 × b =
variable
z 5/16 –3/32 0 900 + α 5/16(900 + α) – 3/32 × 1400 + 0 × 250 = 150 + 5/16α

x –3/16 5/32 0 1400 – 3/16 (900 + α) + 5/32 × 1400 + 0 × 250 = 50 – 3/16α

S3 –1/8 –1/16 1 250 – 1/8 (900 + α) – 1/16 × 1400 + 1 × 250 = 50 – 1/8 α

122 Operations Research

As z = 150 + 5/16α, z remains positive for any value of α ≥ 0.

As x = 50 – 3/16α, x remains positive for any value of α ≤ 800/3.
As S3 = 50 – α/8, S3 remains positive for any value of α ≤ 400.
Therefore, the present basis remains feasible for the increment of resource 1, i.e., man-hours by
800/3 only. For further increase, basis will change. Similarly if the resource 1 being contemplated, the
solution would be:

Problem
S1 S2 S3 Capacity B–1 × b =
variable
z 5/16 – 3/32 0 900 – α 5/16 (900 – α) – 3/32 × 1400 + 0 × 250 = 150 – 5/16α

x – 3/16 5/32 0 1400 – 3/16 (900 – α) + 5/32 × 1400 + 0 × 250 = 50 + 3/16α

S3 – 1/8 – 1/16 1 250 – 1/8 (900 – α) – 1/16 × 1400 + 1 × 250 = 50 + 1/8α

Thus z will be negative, when 5/16α > 150 i.e., α > 480. Hence when resource No.1 i.e. Man-
hour is reduced to the level below 420 = (900 – 480) the present basis will change.
Hence the range for resource No. 1, i.e., man-hours to retain the present basic variables is:
(900 + 800/3) to (900 – 480) = 3500 /3 to 420.

3.14.3. Dual Simplex Method

We remember that while getting optimal solution for a linear programming problem, by using
simplex method, we start with initial basic feasible solution (with slack variables in the programme for
maximization problem and artificial variables in the programme for minimization problem) and through
stages of iteration along simplex algorithm we improve the solution till we get optimal solution. An
optimal solution is one, in terms of algorithm for maximization, in which, the net evaluation row are
either negative elements are zeros, i.e., the dual variables are feasible. Now with the knowledge of
Primal and Dual relationship, we know that the net evaluation row elements are the values of dual
variables and hence it suggests dual feasible solution. There are situations, where the primal solution
may be infeasible, but corresponding variables indicate that dual is feasible. Thus, solution of primal is
infeasible but optimum. In a dual simplex method initial solution is infeasible but optimum, and through
iteration it reaches feasibility at which stages it also reaches true optimum.
Problem 3.39: The problem can be written as:
Minimize Z = 2a + 1b s.t. Maximize Z = –2a – 1b s.t.
3a + 1b ≥ 3 –3a – 1b ≤ –3
4a + 3b ≥ 6 –4a – 3b ≤ –6
1a + 2b ≤ 3 and both a and b are ≥ 0. 1a + 2b ≤ 3 and both a and b are ≥ 0.
The linear programming version is:
Maximize Z = –2a – 1b + 0S1 + 0S2 + 0S3 s.t
–3a – 1b + 1S1 + 0S2 + 0S3 = –3
–4a – 3b +0S1 + 1S2 + 0S3 = –6
1a + 2b + 0S1 + 0S2 + 1S3 = 3
And a, b, S1, S2, and S3 all ≥ 0.
Linear Programming Models : Solution by Simplex Method 123

Table: I. a = 0, b = 0, S1 = –3, S2 = –6, S3 =3 and Z = Rs.0/–

Problem Profit Cj –2 –1 0 0 0 Replacement

variable Rs. Capacity units a b S1 S2 S3 ratio

S1 0 –3 –3 –1 1 0 0 3

S2 0 –6 –4 –3 0 1 0 1/2

S3 0 3 1 2 0 0 1 3/2

Net evaluation –2 –1 0 0 0

Quotient Row: –4/–2 –1/–3

Now observe here, both S1 = –3 and S2 = –6 are negative and S3 = 3 positive. This shows that the
basic variables of primal (S1 and S3) are infeasible. Moreover the net evaluation row elements are
negative or zeros, the solution is optimal. Now let us change the solution towards feasibility without
disturbing optimality. That is deciding the incoming variable and outgoing variable. In regular simplex
method, we first decide incoming variable. In dual simplex we first decide outgoing variable,
i.e., key row.
(a) Criterion for out going variable:
The row, which has got the largest negative value (highest number with negative sign) in the
capacity column, becomes the key row and the variable having a solution in that row becomes out
going variable. If all the values in the capacity column are non-negative and if all net evaluation row
elements are negative or zeros, the solution is optimal basic feasible solution. In the given example, the
row containing S2 is having highest number with negative sign; hence S2 is the out going variable.
(b) Criterion for incoming variable:
Divide the net evaluation row elements by corresponding coefficients (if negative) of the key
row. The column for which the coefficient is smallest becomes key column and the variable in that
column becomes entering variable. (If all the matrix coefficients in the key row are positive, the
problem has no feasible solution).
In the problem given, variable satisfies the condition, hence variable 'b' is the incoming variable.
The rest of the operation is similar to regular simplex method.
Table: II. a = 0, b = 2, S1 = –1, S2 = 0, S3 = –1 and Z = Rs. –2

Problem Profit Cj –2 –1 0 0 0 Replacement

variable Rs. Capacity units a b S1 S2 S3 ratio

S1 0 –1 – 5/3 0 1 – 1/3 0 3/5

b –1 2 4/3 1 0 – 1/3 0 6/4

S3 0 –1 – 5/3 0 0 2/3 1 3/5

Net evaluation – 2/3 0 0 – 1/3

quotient: – 1/5 – – –1 – –
124 Operations Research

The solution is infeasible. As net evaluation row elements are negative the solution remains optimal
and as basic variables are negative, it is infeasible.
Now both rows of S1 and S3 are having –1 in capacity column, any one of them becomes key
row. Let us select first row as key row. The quotient row shows that ‘a’ as the incoming variable.

Table: III. a = 3/5, b = 6/5, S1 = 0, S2 = 0, S3 = 0, Z = –Rs. 12/5

Problem Profit Cj –2 –1 0 0 0 Replacement

variable Rs. Capacity units a b S1 S2 S3 ratio

a –2 3/5 1 0 – 3/5 1/5 0

b –1 6/5 0 1 4/5 – 3/5 0

S3 0 0 0 0 –1 1 1

Net evaluation 0 0 – 2/5 – 1/5 0

As the net evaluation row elements are negatives or zeros, the solution is optimal and feasible.
Answer is a = 3/5 and b = 6/5 and profit Z = – Rs. 12/5. That is for minimization problem the minimum
cost is Rs. 12/5.
To Summarize:
(i) The solution associated with a basis is optimal if all basic variables are ≥ 0.
(ii) The basic variable having the largest negative value (highest number with negative
sign), is the outgoing variable and the row containing it is the key row.
(iii) If α is the matrix coefficient of the key row and is < 0, the variable for which (index
row coefficient/αα) is numerically smallest will indicate incoming variable.
(iv) If α > 0, for all the variables in the key row the problem is infeasible.

Problem 3.40: The problem can be written as:

Minimize Z = 20x + 16y s.t. Maximize Z = – 20x – 16y + 0S1 + 0S2 + 0S3 + 0S4 s.t.
1x + 1y ≥ 12 –1x – 1y + 1S1 + 0S2 + 0S3 + 0S4 = –12
2x + 1y ≥ 17 –2x –1y + 0S1 + 1S2 + 0S3 + 0S4 = –17
2x + 0y ≥ 5 –2x +0y + 0S1 + 0S2 + 1S3 + 0S4 = –5
0x + 1y ≥ 6 and 0x – 1y + 0S1 + 0S2 + 0S3 + 1S4 = – 6
Both x and y are ≥ 0 and x, y, S1, S2, S3 and S4 all ≥ 0.
Linear Programming Models : Solution by Simplex Method 125

Solution:
Table: I. x = 0, y = 0, S1 = –12, S2 = –17, S3 = –6, S4 –6, Z = Rs.0/–

Problem Profit Cj – 20 – 16 0 0 0 0 Replacement

variable Rs. Capacity units x y S1 S2 S3 S4 ratio

S1 0 – 12 –1 –1 1 0 0 0 12

S2 0 – 17 –2 –1 0 1 0 0 8.5

S3 0 –5 –2 0 0 0 1 0 2.5

S4 0 –6 0 –1 0 0 0 1 0

Net evaluation – 20 – 16 0 0 0 0

Quotient 10 16

As all the elements of net evaluation row are negative, the solution is optimal but as slack variables
have –ve values, the solution is infeasible. The row with highest number with negative sign becomes
outgoing variable. Here S2 is out going variable i.e., it becomes key row. By dividing net evaluation row
elements by corresponding key row elements, quotient row elements are obtained, which show that x
is the incoming variable (lowest number).
Table: II. x = 17/2, y = 0, S1 = –7/2, S2 = 0, S3 = 12, S4 = –6 and Z = – Rs. 170/–

Problem Profit Cj – 20 – 16 0 0 0 0 Replacement

variable Rs. Capacity units x y S1 S2 S3 S4 ratio

S1 0 – 7/2 0 – 1/2 1 –1/2 0 0 7

x – 20 17/2 1 1/2 0 –1/2 0 0 Negative

S3 0 12 0 1 0 –1 1 0 12

S4 0 –6 0 –1 0 0 0 1 6

Net evaluation 0 –6 0 – 10 0 0

Quotient – 6 – – – –

Solution is optimal and infeasible as net evaluation row elements are negative and slack variables
are negative.
126 Operations Research

Table: III. x = 11/2, y = 6, S1 = –1/2, S2 = 0, S3 = 6, S4 = 0, Z = – Rs.206/–

Problem Profit Cj – 20 – 16 0 0 0 0 Replacement

variable Rs. Capacity units x y S1 S2 S3 S4 ratio

S1 0 – 1/2 0 0 1 – 1/2 0 – 1/2 Negative

x – 20 11/2 1 0 0 – 1/2 0 1/2

S3 0 6 0 0 0 –1 1 1

y – 16 6 0 1 0 0 0 –1

Net evaluation 0 0 0 – 10 0 –6

Quotient 20 12

Table: IV. X = 5, y = 7, and Z = – Rs. 212/–

Problem Profit Cj – 20 – 16 0 0 0 0 Replacement

variable Rs. Capacity units x y S1 S2 S3 S4 ratio

S4 0 1 0 0 –2 1 0 1

x – 20 5 1 0 1 –1 0 0

S3 0 5 0 0 2 –2 1 0

y – 16 7 0 1 –2 1 0 0

Net evaluation 0 0 – 12 –4 0 0

Quotient

As all the elements of net evaluation row are either negative or zeros the solution is optimal and as
all the slack variables and basic variables have positive values the solution is feasible.
x = 5, y = 7, S1 = 0, S2 = 0. S3 = 5, S4 = 1 and Z = – Rs. 212/– i.e minimum optimal is Rs.212/
Problem 3.41: The problem can be written as:
Minimize Z = 10a + 6b + 2c s.t. Maximize Z = –10a – 6b – 2c + 0S1 + 0S2
–1a + 1b + 1c ≥ 1 1a – 1b –1c + 1S1 + 0S2 = –1
3a + 1b – 1c ≥ 2 –3a – 1b + 1c + 0S1 + 1S2 = –2
And all a, b, c, ≥ 0. And a, b, c, S1, S2 all ≥ 0
Linear Programming Models : Solution by Simplex Method 127

Solution:
Table: I. a = 0, b = 0, X = 0, S1 = –1, S2 = –2, Z = Rs. 0/–

Problem Profit Cj – 10 –6 –2 0 0 Replacement

variable Rs. Capacity units a b c S1 S2 ratio

S1 0 –1 1 –1 –1 1 0 –1

S2 0 –2 –3 –1 1 0 1 2/3

Net evaluation –10 –6 –2 0 0

Quotient 10/3 6

Net evaluation row elements are negative hence solution is optimal but slack variables are negative,
hence the solution is infeasible. Variable ‘a’ has lowest quotient hence incoming variable and S2 has got
highest element with negative sign, it is out going variable and the row having S2 is the key row.
Table: II. a = 2/3, b = 0, z = 0, S1 = –5/3, S2 = 0, Z = – Rs. 20/3.

Problem Profit Cj – 10 –6 –2 0 0 Replacement

variable Rs. Capacity units a b c S1 S2 ratio

S1 0 –5/3 0 –4/3 –2/3 1 1/3 Negative

a –10 2/3 1 1/3 –1/3 0 –1/3

Net evaluation 0 –8/3 –16/3 0 –10/3

Quotient 2 8

Variable ‘b’ has lowest positive quotient, it is incoming variable, S1 has highest number with
negative sign, it is the out going variable. The solution is infeasible optimal.
Table: III. a = 1/4, b = 5/4, c = 0, S1 = 0, S2 = 0, Z = – Rs. 10/–

Problem Profit Cj – 10 –6 –2 0 0 Replacement

Variable Rs. Capacity units a b c S1 S2 ratio

b –6 5/4 0 1 1/2 –3/4 –1/4

a – 10 1/4 1 0 –1/2 1/4 –1/4

Net evaluation 0 0 –4 –2 –4

The solution is optimal and feasible. a = ¼, b = 5/4, c = 0 and Z = – Rs. 10/–. Hence the minimum
cost is Rs. 10/–
128 Operations Research

3.14.4. Addition of a New Constraint

Whenever a linear programming problem is formulated, the constraints, which are considered
important and significant, are considered and the problem is solved to find the optimal solution. After
obtaining the optimal solution for the problem, the optimal solution is checked to see whether it satisfies
the remaining constraints of secondary importance. This approach reduces the size of the problem to
be handled in the first instance and reduces the calculation part. We may come across a situation that
at a later stage, newly identified significant constraints have to be introduced into the problem. The
situation in either case amounts addition of one or several constraints. With the values of optimal basic
variables it is first checked whether they satisfy the new constraint (s). If so, the solution remains
optimal. If not, the constraints are introduced in the optimal tableau, and by elimination of coefficients
of basic variables in the new constraints are reduced to zero. With that, if solution is feasible, and non–
optimal, regular simplex method is used for optimization. On the other hand, if solution is infeasible but
optimal, dual simplex method is used for optimization. New constraint, which is not satisfied by the
previous optimal solution, is called tighter constraint, because it changes the solution.
Problem 3.42: (Repetition of problem 3.39)
Minimize Z = 2a + 1b s.t.
3a + 1b ≥ 3
4a + 3b ≥ 6
1a + 2b ≤ 3 and both a and b are ≥ 0.
Optima solution obtained by dual simplex method is: a = 3/5 and b = 6/5.
Let us suppose that new constraint added is 5a + 5b ≥ 10. The simplex version of this inequality
is:
5a + 5b – S4 = 10.
Let us substitute the values of ‘a’ and ‘b’ in the above.
5 × 3/5 + 5 × 6/5 – S4 = 10. This gives a value of –1 to S4, which violates the non-negativity
constraint and hence the solution is not optimal. Hence by introducing the new constraint in the final
table, we have to get a new optimal solution.
Table: I.

Problem Profit Cj –2 –1 0 0 0 0 Replacement

variable Rs. Capacity units a b S1 S2 S3 S4 ratio

a –2 3/5 1 0 –3/5 1/5 0 0 Row 1

b –1 6/5 0 1 4/5 –3/5 0 0 Row 2

S3 0 0 0 0 –1 1 1 0 Row 3

S4 0 –10 –5 –5 0 0 0 1 Row 4

Net evaluation 0 0 –2/5 –1/5 0 10

Multiplying the row 1 by 5 and adding it to row 4 and multiplying row 2 by 5 and adding it to row
4, we get
Linear Programming Models : Solution by Simplex Method 129

Table: II.

Problem Profit Cj –2 –1 0 0 0 0 Replacement

vriable Rs. Capacity units a b S1 S2 S3 S4 ratio

a –2 3/5 1 0 –3/5 1/5 0 0

b –1 6/5 0 1 4/5 –3/5 0 0
S3 0 0 0 0 –1 1 1 0
S4 0 –1 0 0 1 –2 0 1
Net evaluation 0 0 – 2/5 – 1/5 0 0
Quotient 1/10

a = 3/5, b = 6/5, S3 = 0, S4 = –1, Hence the solution is infeasible. As the net evaluation row
elements are negative or zeros the solution is optimal. Hence by using dual simplex method, we get S3
as incoming variable and S4 as the out going variable.
Table: III.

Problem Profit Cj –2 –1 0 0 0 0 Replacement

variable Rs. Capacity units a b S1 S2 S3 S4 ratio
a –2 1/2 1 0 –1/2 0 0 1/10
b –1 3/2 0 1 1/2 0 0 –3/10
S3 0 –1/2 0 0 –1/2 0 1 1/2 Negative
S2 0 1/2 0 0 –1/2 1 0 –1/2
Net evaluation 0 0 –1/2 0 0 1/10
Quotient 1

Solution is infeasible and optimal. S1 is the incoming variable and S3 is the outgoing variable.
Table: IV.

Problem Profit Cj –2 –1 0 0 0 0 Replacement

variable Rs. Capacity units a b S1 S2 S3 S4 ratio
a –2 1 1 0 0 –1 0 –2/5
b –1 1 0 1 0 1 1 1/5
S1 0 1 0 0 1 0 –2 –1
S2 0 1 0 0 0 1 –1 –1

Net evaluation 0 0 0 –1 0 –3/5

130 Operations Research

As the net evaluation row elements are either negatives or zeros the solution is optimal. As all
variables have positive values the solution is feasible.
a = 1, b = 1, and Z min is Rs. 3/–. Here the basic variables remain same but the optimal value of
cost is changed.
Problem 3.42: (Extension of problem 3.40)
Add a new constraint 4x + 3y ≥ 40 to problem 3.40 and examine whether basic variables change
and if so what are the new values of basic variables?
The optimal solution obtained is x = 5, y = 7 and Z = Rs. 212/–
Substituting the values in the new constraint, 4 × 5 + 3 × 7 = 41 which is ≥ 40. Hence the
condition given in the new constraint is satisfied. Therefore it is not tighter constraint. This will not
have any effect on the present basis. Let us verify the same.
The simplex version of the new constraint is: –4x – 3y + 1S5 = 40. The optimal table is:
Table: I.

Problem Profit Cj – 20 – 16 0 0 0 0 0 Replace–

variable Rs. Capacity units x y S1 S2 S3 S4 S5 ment ratio
S4 0 1 0 0 –2 1 0 1 0 Row 1
x –20 5 1 0 1 –1 0 0 0 Row 2
S3 0 5 0 0 2 –2 1 0 0 Row 3
y –16 7 0 1 –2 1 0 0 0 Row 4
S5 0 –40 –4 –3 0 0 0 0 1 Row 5
Net evaluation 0 0 –12 –4 0 0 0
Quotient

Now to convert matrix coefficients of basic variables in row 5 to zero, multiply row 4 by 3 and
adding it to row 5 and multiplying row 2 by 4 and adding it to row 5, we get:
Table: II.

Problem Profit Cj – 20 – 16 0 0 0 0 0 Replace–

variable Rs. Capacity units x y S1 S2 S3 S4 S5 ment ratio
S4 0 1 0 0 –2 1 0 1 0
x –20 5 1 0 1 –1 0 0 0
S3 0 5 0 0 2 –2 1 0 0
y –16 7 0 1 –2 1 0 0 0
S5 0 1 0 0 –2 –1 0 0 1

Net evaluation 0 0 –12 –4 0 0 0

Linear Programming Models : Solution by Simplex Method 131

x = 5, y = 7 and minimum Z = Rs. 212/– As the net evaluation row is negative or zeros, and all
problem variables have positive value, the solution is feasible and optimal. Values of basic variables have
not changed.
Problem 3.43: (Repetition of Problem 3.41).
Add the following two new constraints to problem No. 3.41 and find the optimal solution.
Minimize Z = 10a + 6b + 2c s.t
–1a + 1b + 1c ≥ 1
3a + 1b – 1c ≥ 2
New constraints are:
4a + 2b + 3c ≤ 5
8a – 1b + 1c ≥ 5
The simplex format of new constraints is:
4a + 2b + 3c + 1S3 = 5
8a – 1b + 1c – 1S4 = 4
Earlier values of a = 1/4 and b = 5/4 and z = 0
4 × ¼ + 2 × 5/4 + 3 × 0 + 1S3 = 5, gives S3 = 3/2, which is feasible.
8 × ¼ – 5/4 × 1+ 0 – S4 = 4 gives S4 = –13/4 is not feasible. Hence the earlier basic solution is
infeasible.
We have seen that the first additional constraint has not influenced the solution and only second
additional constraint will influence the solution, by introducing both the constraints in optimal table we
get:
Table: I.

Problem Profit Cj – 16 –6 –2 0 0 0 0 Replace–

variable Rs. Capacity units a b c S1 S2 S3 S4 ment ratio

b –6 5/4 0 1 ½ –3/4 –1/4 0 0 Row 1

a –16 1/4 1 0 –1/2 1/4 –1/4 0 0 Row 2

S3 0 5 4 2 3 0 0 1 0 Row 3

S4 0 –4 –8 1 –1 0 0 0 1 Row 4

Net evaluation

Multiply row 1 by 2 and subtract it from row 3.

Subtract row 1 from row 4.
Multiply row 2 by 4 and subtract from row 3.
Multiply row 2 by 8 and add it to row 4
132 Operations Research

Table II.

Problem Profit Cj – 10 –6 –2 0 0 0 0 Replace–

variable Rs. Capacity units a b c S1 S2 S3 S4 ment ratio

b –6 5/4 0 1 1/2 –3/4 –1/4 0 0

a –10 1/4 1 0 –1/2 1/4 –1/4 0 0

S3 0 3/2 0 0 4 1/2 3/2 1 0

S4 0 –13/4 0 0 –11/2 11/4 –7/4 0 1

Net evaluation 0 0 –4 –2 –4 0 0

Quotient 8/11 16/7

As S4 = – 13/4, the basic solution is not feasible. As net evaluation row elements are either
negative or zeros the solution is optimal. Students can see that by using dual simplex method, variable
‘c’ will enter the solution and S4 will leave the solution.

FLOW CHART FOR SIMPLEX METHOD