byjusexamprep.

com

2
byjusexamprep.com

CHAPTER INDUSTRIAL ENGINEERING

6 LINEAR PROGRAMMING

1. LINEAR PROGRAMMING (GEORGE B. DANTZING)

• L.P is used for optimization of our limited resources when there are no of alternate solution

possible for the problem.

• It is a mathematical technique and the turn linear used for the variable and it simply means

that the relationship between different variable can be represented in the form of straight

line.

1.1. Requirement of L.P

• Objective fun- It is a main fun which we optimize and it should be clearly identifiable

and measurable in quantitative term like maximum of profit, sell or min of cost.

• Constraints/Conditions: These are the limitation or restriction with in which we

need to optimize our objective function.

• All the variables used in the objective fun and constraint should be linear and non-

negative.

1.2. Laws/Rules in L.P

1.2.1. Law of certainty: In LP model the various parameter like objective function

coefficient, constraints & resources are known exactly, and their value does not change

with time.

1.2.2. Law of Proportionality:

e.g. 1 Product → R1 5/- profit

5 Product → R5 5 × 5 profit

1.2.3. Law of continuity (Divisibility): In LP model decision variables are continuous

and they are permitted to take any non- negative value that satisfy all the constraints.

3
byjusexamprep.com

1.3. General Statement of LP:

Max. Z = C1x1 + C2x2 + C3x3 + C4x4 + ............ + Cnxn – objective fun

Non-negative condition → x1, x2 ........... n > 0

Where aij, bi, cj are constant & xj is variable

i = 1, 2, 3 ....... n

j = 1, 2, ........... n

aij : technological constant or substitution

bi : resource value

cj : Profit coefficient

xj : decision or choice variable

1.4. Graphical Method:

1.4.1. Steps

(i) identify the problem, define decision variables, objective fun and constraints.

(ii) Draw a graph that include all the constraints and identify the common feasible region.

(iii) find out the point within the feasible region which optimizes the objective fun these

point gives the final optimum soln.

One of the corner point of the feasible region gives the final solution because objective

fun is straight line with constant slope (Isocost line) and as it moves away from origin it

value increases and the optimum value will be at one the corner extreme point.

Objective fun will be tangent to that point and give the optimum solution.
Example: Find the solution for the following LPP.
Zmax = 8 x1 + 7 x2
Subject to 3x1 + x2 ≤ 66
x1 +x2 ≤ 45
x1 ≤ 20
x 1 , x2 ≥ 0

4
byjusexamprep.com

Solution:

x1 x2 ………… (1)
+ 1
22 66

x1 x2 ..……… (2)
+ 1
45 45
x1
1 ………… (3)
20

x2 ………… (4)
1
40

Corner Points are:

O (0,0); A(20,0); B(20,15.33); C(10.5,34.5); D(5,40); E(0,40).
Zmax = 8x1 + 7x2
Zo = 0
ZA = 8 × 20 + 7 × 0 = 160
ZB = 8 × 20 + 7 × 15.33 = 267.33
ZC = 8 × 10.5 + 7 × 34.5 = 325 ← optimal
ZD = 8 × 5 + 7 × 40 = 320
ZE = 8 × 0 + 7 × 40 = 280
Hence, Z is Optimal at ‘C’
x1 = 10.5; x2 = 34.5; Zmax = 325
Example: Find the solution for the following LPP.
Zmax = 28 x1 + 30x2
Subjected to
6x1 + 3x2 ≤ 18
3x1 + x2 ≤ 8

4x1 + 5x2 ≤ 30
x 1 , x2 ≥ 0

5
byjusexamprep.com

x1 x2
+ 1
3 6

x1 x2
+ 1
8/3 8

x1 x
+ 21
7.5 6

Corner points O(0, 0), A(8/3, 0), B(2, 2), C(0, 6)

Zmax = 28 x1 + 30 x2
Zo = 0
8 2
ZA = 28  + 30  0 = 74
3 3

ZB = 28 × 2 + 30 × 2 = 16

ZC = 28 × 0 + 30 × 6 = 180

Optimal at ‘C’ and x1 = 0 at optimality.

i.e. one of the basic variables has a zero value hence, the solution is degenerate.

Example: Find the solution for the following LPP.

Zmin = 1.5 x1 + 2.5x2
Subject to
x1 + 3x2 ≥ 3
x1 + x2 ≥ 2
x 1 , x2 ≥ 0

6
byjusexamprep.com
x1 x2
+ 1 ……………….. (1)
3 1
x1 x2
+ 1 ………………… (2)
2 2

Corner points are:

A(3, 0); B(0, 2); C(3/2, 1/2);
Zmin = 1.5 x1 + 2.5x2
ZA = 1.5 × 3 + 2.5 × 0 = 4.5
ZB = 1.5 × 0 + 2.5 × 2 = 5
ZC = 1.5 × 1.5 + 2.5 × 1/2 = 3.5 ← Optimal
Zmin = 3.5; x1 = 1.5, x2 = 1/2 is the solution.

1.4.2. Binding and non-binding constraints

[x1 = 1000/3. x2 = 400/3]

1. 2x1 + 2.5 x2 < 1000 → 1000 = 1000 Binding

2. 3x1 + 1.5 x2 < 1200 → 1000 = 1000 Binding

3. 1.5x1 + 4 x2 < 1200 → 1033.33 ≠ 1000 Non-binding

 When we put the value of optimum solution in the constant and L.H.S. = R.H.S

the constant is termed as binding otherwise non-binding

 the final solution is always obtained from the binding constraint.

1.4.3. Redundant Constraints: Constant which does not become part of boundary

making the feasible region is termed as redundant or unnecessary constraint.

Inclusion or Exclusion of such constraint does not have any effect on final optimum

solution of the problem.

7
byjusexamprep.com

1.4.4. Special cases:

1. Infinite or multi optimal solution: Infinite solution means we get the same optimum

value of objective function for different varying variables.

We always get a unique solution when the slope of objective function is different from

constraints.

We get infinite solutions if same value of objective function is obtained at two or more

points and these points lie on same straight line.

2. No solution or infeasibility: In some condition constraint may in inconsistent in such

a manner that it is not possible to find a feasible region where all the constraints are

satisfied, it is termed as no solution or infeasibility.

Example: Find the solution for the following LPP.
Zmax = 3x1 – 2x2
Subject to
x1 + x2 ≤ 1
x1 + x2 ≥ 2
x 1 , x2 ≥ 0
Solution:
x1 x2
+ 1 …………………... (1)
1 1
x1 x2
+ 1 ……………...….. (2)
2 2

There is no common solution space satisfying both the equations hence, the solution is
infeasible.

8
byjusexamprep.com

3. Unbounded solution: In some condition the highest value of objective fun goes upto
infinite and it simply means that the common feasible region is not bounded by the limit
on the constraints it is termed as unbounded solution.
Example: Find the solution for the following LPP.
Zmax = 3 x1 + 2x2
Subject to
x1 – x2 ≤ 1
x1 + x2 ≥ 3
x 1 , x2 ≥ 0
Solution:
x1 x2 ……………………… (1)
+ 1
1 1
……………………… (2)
x1 x2
+ 1
3 3

Unbounded solution space and unbounded optimal solution to problem.

1.5. SIMPLEX METHOD
It is the step by step procedure in which we proceed in a systematic manner from an
initial feasible solution with an improvement upon initial solution until in certain no. of
steps we reach the final optimum solution.
This method also checks the corner points of the feasible region but in multi dimensions
depending upon the no. of variables.
1.5.1. Standard form for Simplex method
• In standard form the objective function is the maximization or minimization type.
All the constraints are extent to non-negative restrictions.
• The right-hand side of each constrain is not negative. All decision variable are ≥ 0.

9
byjusexamprep.com

Standard form:
Zmax = 2x1 + 3x2 + 0S1 + 0S2
Subjected to
x1+ 2x2 + S1 = 10
– 2x1 – x2 = 10
x1 + 5x2 – S2 = 8
x 1 , x2 ≥ 0
1.5.2. Simplex Procedure:
1. Setup the objective function.
2. Setup the constrained equations.
3. Convert the inequalities to equalities.
Example:
Zmax = 2x + 5y

Subjected to x + 4y ≤ 24
3x + y ≤ 21
x+y≤9
x, y ≥ 0
Zmax = 2x + 5y + 0S1 + 0S2 + 0S3, where S1, S2 and S3 are slack variables.
x + 4y + 0S1 = 24
3x + y + 0S2 = 21
x + y + 0S3 = 9
1.5.3. Basic simplex table:

SV = solution vector
Example:
Zmax = 24a + 48b
Subjected to:
a + 2b ≤ 48
2a + b ≤ 60
a, b ≥ 0

10
byjusexamprep.com

Solution:
Zmax = 24a + 48b + 0S1 + 0S2.
a + 2b + S1 = 48
2a + b + S2 = 60.
where, S1, S2 → slack variables and they have no effect in profit.
Setup the standard simplex table with Cj as coefficient of objective function.

Where, P.E = Pivot element,

E.V = Entering vector
L.V = Leaving vector
• The coefficient of the variables in the constraint equation form the body of the matrix.
• The variable corresponding to unit matrix are called the initial solution to the problem.
• Determine Zj which is the matrix multiplication of the solution vector column with the
individual columns of matrices.
• Determine Ci – Zj
• The most positive Cj – Zj is the entering vector for maximization and most negative (–
ve) Cj – Zj is the entering vector for minimization.
• The elements which are entering vector are also called entering vector elements.
• They are also called intersection elements for the respective rows.
• Obtain minimum ratio by dividing Bo column with the respective entering vector
elements.
• For both maximization and minimization, minimum positive ratio is called as the leaving
vector.
• The intersection of the entering vector and leaving vector is called as the pivot element.
• For this problem the present variable 0S1 is leaving the solution and in its place 48b is
entering the solution.

11
byjusexamprep.com

• Drive the pivot element by dividing the row containing the pivot element with the pivot
element and write the replaced row.
• As a change is made in one of the rows of matrix a corresponding change should be
made in the remaining rows by using the formula
New element = (old element) – (intersection element) × (corresponding element in
replaced row)
3 1
∴ New elements are = , 0, − , 1, 36
22
• Repeat the above procedure from the calculation of Zj onwards till the solution obtained
is optimal and the optimally is given by all the zeros or negative values in Cj – Zj row for
maximization, Cj – Zj row is called as net evaluation row.
• All zeros or positive values in Cj – Zj row for minimization.
• From the above problem,
b = 24, S2 = 36, a = S1= 0.
∴ Zmax = 24 × 0 + 48 × 24 = 1152
• Variables in this solution are called basic variables.
• Variables which are not in the solution are called non basic variables.
• For any basic variable at optimality in (Cj – Zj) the values will definitely be zero.

1.6. BIG. M method

This method is used for ≥ or = type constraints or mix constraints ≤ or ≥
e.g.
5x1 – x2 ≥ 24

5x1 – x2 – S1 = 24

Put x1 = 0
x2 = 0
S1 = –24, Which is not possible because it should be positive (Basic variable)
For this solution, introduce an artificial variable (A1)
5x1 – x2 – S1 + A1 = 24

Non
Basic

Put x1 = 0
x2 = 0 and S1 = 0
∴ A1 + 24
• for maximisation problems, in Z = ------------ MA1
• for maximisation problems in z = ----------- + MA2

where, M is Higher than any finite number

12
byjusexamprep.com

• Big M method is a modified form of simplex method, and it is always used whenever
the constraints are of (≥ or =) type irrespective of whether the problem is for
maximisation or for minimisation.
•In this condition, we introduce artificial variables in the current solution to get an initial
working matrix. These artificial variables must not appear in the final solution and this is
ensured by providing an extremely negative value (M) to their profit coefficients in the
objection function.

1.6.1. Special cases

1. Unique solution: If number of basic variables are equal to the number of zeros,
then the solution is unique.

2. Infinite or multi-optimum solution: When a non-basic variable in an optimum

solution has zero value for Δj row then the solution is not unique and it indicates that the

problem has infinite no. of solution.

3. Unbounded solution: If in a case all the value in the replacement ratio column are

either –ve or infinite then the solution terminates, and it indicates that the problem has

unbounded solution.

4. No-solution/infeasibility: When in the final solution artificial variable remains in the

basis then there is no feasible solution to the problem.

5. Degenerate solution:

13
byjusexamprep.com

m → no. of equations
• When one or more of the basic variables becomes equal to zero during calculation then
the solution is called degenerate & the condition is known as degeneracy. In a
degenerate solution the no. of basic variables becomes less than equality constraint.
1.7. Shadow Price:
The shadow price of a resource is the unit price that is equal to the increase in profit to
be realized by one additional unit of that resource. Thus, the dual variable is also referred
to as the shadow price or imputed price of a resource. This is highest price the
manufacturer would be willing to pay for the resources.
Assessment of this nature is post-optimal or sensitivity analysis.
The main results to be drawn from the sensitivity analysis presentation are:
• Changes in the different coefficients of the model can affect either the optimality or
feasibility of the current solution and will lead to one of three situations regarding the
optimal solution.
(i) The variables and their optimal values remain essentially unchanged.
(ii) The variables remain the same, but their optimal values change.
(iii) The variables and their values change completely.
In cases (i) and (ii), sensitivity analysis methods are very efficient. Even in the third case,
the new optimal solution is normally recovered in a few additional iterations.
• The addition of a new constraint can never improve the value of the objective function.
• The additional of a new variable can never worsen the value of the objective function.
1.8. Dual Problem:
Corresponding to a linear programming problem is another linear programming problem
formulated from the Parameters of the original problem.
• A new LP, derived from the primal, according to a set of transformation rules.
Dual Theorem of Linear Programming: States a theoretic relationship between the
primal and dual problems
Dual Variables: are the variables of the dual linear programming problem
Primal Problem: is the original linear programming problem.
Post-Optimality (Sensitivity) Analysis: of a linear programming problem is a study
of the effect of changes of the profit of resource level on the solution.
• The study of how sensitive an optimal solution is to the model assumptions and to data
changes. Range of Optimality: The range of values over which a basic variables
coefficient can change without causing a change in the optimal solution mix.
Canonical form: An LP problem is said to be in canonical form if

14
byjusexamprep.com

i. all the decision variables are non – negative

ii. the objective function is of maximization type
iii. all constraints are less than or equal to type except for non – negative conditions.
• The initial given problem is termed as primal and problem obtained by transposing row
and column but having the same optimum value of the objective fun is termed as dual.
Primal Dual
max im 
→ Minimum
n → m

m → n

bi → cj

cj 
→ bi

< type constraint: will change to ≥ constraint

≥ constraint: will change to ≤ constraints
Primal constraint i is “=” type: yj unrestricted in sign
Primal variable unrestricted in sign: Dual constraint j is “=” type

Example:
Find the dual for the following L.P.P:
Min Z = 10x1 + 6x2 – 9x3
4x1 + 3x2 – x3 ≥ 8
3x1 – 4x3 ≤ 7
x2 – 3x3 ≤ 4
5x1 + 2x2 ≥ 3
2x1 – 2x2 – 3x3 ≤ 14
x 1 , x2 , x 3 ≥ 0
Solution:
Because problem is of minimisation. All the constraints should be of ≥ Type.
i.e. 4x1 + 3x2 – x3 ≥ 8
–3x1 + 4x3 ≥
–x2 + 3x3 ≥ –4
5x1 + 2x2 ≥ 3
–2x1 + 2x2 + 3x3 ≥ –14; x 1 , x2 x 3 ≥ 0

15
byjusexamprep.com

∴ Dual is
Max W = 8y1 – 7y2 – 4y3 + 3y4 – 14y5
Constraints are,
4y1 – 3y2 + 5y4 – 2y5 ≤ 10
3y1 + 0y2 – y3 + 2y4 + 2y5 ≤ 6
–y1 + 4y2 + 3y3 + 0y4 + 3y5 ≤ – 9
Where, y1, y2, y3, y4, y5 ≥ 0 are dual variables.

16
byjusexamprep.com

PRACTICE QUESTIONS

1. A company making cold drinks has two bottling plants located at towns T1 and T2 Each
plant produces three drinks A, B and C and their production capacity per day is shown
below:

Plant at
Cold drinks
T1 T2

A 6,000 2,000

B 1,000 2,500

C 3,000 3,000

The marketing department of the company forecasts a demand of 80,000 bottles of A,

22,000 bottles of B and 40,000 bottles of C during the month of June. The operating costs
per day of plants at T1 and T2 are Rs. 6,000 and Rs. 4000 respectively. Find (graphically)
the number of days for which each plant must be run in June so as to minimize the
operating costs while meeting the market demand.
2. A firm uses milling machines, grinding machines and lathes to produce two motor parts.
The machining times required for each part, the machining times available on different
machines and the profit on motor part are given below:

Machining time reqd. for

Type of the motor part (minutes) Max. time available per
machine I II week (minutes)

Milling 10 4 2,000
machines
Grinding 3 2 900
machines
Lathes 6 12 3,000

Profit/unit (Rs) 100 40

Determine the number of parts I and II to be manufactured per week to maximize the
profit.

17
3. Using simplex method
Maximize Z = 5x1 + 3x2
Subject to
x1 + x2 ≤ 2, 5x1
+ 2x2 ≤ 10,
3x1 + 8x2 ≤ 12,
x1, x2 ≥ 0.
4. Maximize Z = 107x1 + x2 + 2x3,
Subject to the constraints:
14x1 + x2 – 6x3 + 3x4 = 7,

16x1 + 1 x2 – 6x3  5,
2
3x1 – x2 – x3  0,
x1, x2, x3, x4  0.
5. Use Big-M method to
Minimize Z = 2x1 + x2
Subject to 3x1 + x2 = 3, 4x1 + 3x2 ≥ 6, x1 + 2x2 ≤ 3, x1, x2 ≥ 0.
6. Construct the dual of the L.P.P.
Maximize Z = 4x1 + 9x2 + 2x3.
Subject to 2x1 + 3x2 + 2x3 ≤ 7, 3x1
– 2x2 + 4x3 = 5,
x1, x2, x3 ≥ 0.
ANSWERS:
1. x1 = 12 days, x2 = 4 days, Zmin = Rs 88,000.
2. The maximum value of Z occurs at two vertices (200, 0) and (125, 187.5) and the
maximum value of Z will be 20,000.
3. The optimum solution is x1 = 2, x2= 0 and maximum Z = 10.
4. As Cj is positive under some columns, the solution is not optimal. Here 113/3 being the
largest positive value of Cj, x3 is the incoming variable. But all the values of θ being
≤0, x3 will not enter the basis. This indicates that the solution to this problem is
unbounded.
5. x1 = 3/5, x2 = 6/5, max Z = -12/5 Or
Min Z = -max Z = 12/5
6. Let y1 and y2 be the dual variables associated with the first and second constraints. Then
the dual problem is
Minimize W = 7y1 + 5y2,
Subject to 2y1 + 3y2 ≤ 4, 3y1 – 2y2 ≤ 9, 2y1 + 4y2 ≤ 2,
y1 ≥ 0,
y2 is unrestricted in sign. (Because 2nd constraint in primal has equality sign)

****