CS6704-Resource Management Techniques Department of CSE 2016-2017

UNIT I- LINEAR PROGRAMMING

Principal components of decision problem – Modeling phases – LP Formulation and graphic solution – Resource
allocation problems – Simplex method – Sensitivity analysis.
UNIT-I / PART-A
1 What is operations research?
Operations research is the study of optimization techniques. OR is the application of scientific methods,
techniques and tools to problems involving the operations of systems to find optimal solutions. It is applied in
decision theory.
2 List some applications of OR.
 To find optimal assignment of various jobs to different machines.
 To find the waiting time and number of customers waiting in the queue and system in queuing model.
 To find the minimum transportation cost with optimal allocation of goods from different origins to
various destinations.
 To find decision in problems of marketing, finance, production planning and control.
3 How is an OR model organized?
The first crucial step in any OR model is the definition of the alternatives or the decision variables of the
problem. Next, the decision variables are used to construct the objective function and the constraints of the
model. With the three steps completed, the OR model is organized in the following general format: maximize
or minimize objective function subject to constraints.
4 List the different techniques used to solve the OR model.
 Linear programming: it is designed for models with strict linear objective and constraint functions.
 Integer programming: in which the variables assume integer values.
 Dynamic programming: in which the original model can be decomposed into smaller sub-problems.
 Network programming: in which the problem can be modelled as a network.
 Nonlinear programming: in which the functions of the model are non-linear.
5 What are the phases of an OR study?
The principal phases for implementing OR in practice include:
 Definition or formulation of the problem.
 Construction of the mathematical model.
 Derive the solution of the model.
 Validation of the model.
 Implementation of the solution.
6 What is linear programming?
Linear programming (LP) applies to optimization models in which the objective and constraint functions
are strictly linear. It is a technique used for determining optimum utilization of limited resources to meet
out the given objectives. The objective is to maximize the profit or minimize the resources (men, machine,
materials and money).
7 What are the basic components for LP model?
The LP model, as in any OR model has three basic components:
 Decision variables that we seek to determine
 Objective (goal) that we aim to optimize
 Constraints that we need to satisfy
8 What are the requirements for employing LPP technique?
 There must be a well-defined objective function.
 There must be alternative courses of action to choose.
 At least some of the resources must be in limited supply, which give rise to constraints.
 Both the objective function and the constraints must be linear equations or inequalities.
9 What are the properties that LP must satisfy?
Linearity implies that LP must satisfy two properties:
 Proportionality: requires the contribution of each decision variable in the objective function and its
requirements in the constraints to be directly proportional to the value of the variable.
 Additivity: stipulates that the total contribution of all the variables in the objective function and their
requirements in the constraints are the direct sum of the individual contribution or requirements of
 CS6704-Resource Management Techniques Department of CSE 2016-2017
each variable.
10 Mention the steps involved in graphical LP solution.
The graphical procedure includes two steps:
 Determination of the solution space that defines all feasible solutions of the model.
 Determination of the optimum solution from among all the feasible points in the solution space.
11 What is the mathematical model of linear programming?
The linear programming involving more than two variables may be expressed as follows:
Maximize or Minimize Z = c1 x1 + c2 x2+ …..+ cnxn
Subject to constraints
a11x1+a12x2+…………+ a1nxn ≤ or = or ≥ b1
a21x1+a22x2+…………+ a2nxn ≤ or = or ≥ b2
……………………………………………….
……………………………………………….
am1x1+am2x2+…………+ amnxn ≤ or = or ≥ bm
and the non-negativity restrictions x1,x2,….xn≥ 0
12 Define solution, feasible solution and optimal solution of LPP
A set of variables x1,x2….xn which satisfies the constraints of LPP is called a solution. Any solution to a
Linear Programming Problem which satisfies the non negativity restrictions of LPP’s called the feasible
solution. Any feasible solution which optimizes the objective function of the LPP’s called the optimal
solution.
12 Define slack and slack variable.
In ≤ constraints, the right hand side can be thought of as representing the limit on the availability of a
resource, the left hand side represents the usage of this limited resource. The difference between the RHS and
LHS of the ≤ constraint yields the unused or slack amount of the resource.
To convert the ≤ inequality to an equation, a non-negative slack variable is added to the LHS of the
constraint.
For eg: 6x1 + 4x2 ≤ 24 is converted to 6x1 + 4x2+ s1 = 24, s1 ≥ 0, s1 slack variable
14 Define surplus and surplus variable.
A ≥ constraint normally sets a lower limit on the activities of the LP model. The amount by which the LHS
exceeds the minimum limit represents a surplus.
The conversion from ≥ to = is achieved by subtracting a non-negative surplus variable from the LHS of the
inequality.
For eg: x1 + x2 ≥ 800 is converted to x1 + x2– S1 = 800, S1 ≥ 0, S1 surplus variable.
15 What are the characteristics of LPP in canonical form?
In matrix notation the canonical form of LPP can be expressed as:
Maximize Z = CX (objective function)
Subject to AX ≤ b (constraints)
And X ≥ 0 (non-negativity restrictions)
Where C = (c1, c2, ……..,cn)

Charcteristics:
 The objective function is of maximization type
 All constraints are of ≤ type
 All variables xiare non-negative
16 What is sensitivity analysis?
Sensitivity analysis investigates the changes in the optimal solution resulting from making changes in
 CS6704-Resource Management Techniques Department of CSE 2016-2017
parameters of the LP model. Sensitivity analysis is also called post optimality analysis.
17 What are the characteristics of LPP in standard form?
The general linear programming problem in the form
Maximize or Minimize Z = c1 x1 + c2 x2 + …..+ cnxn
Subject to constraints
a11x1 + a12 x2 +…………+ a1nxn = b1
a21x1 + a22 x2 +…………+ a2nxn = b2
……………………………………………….
……………………………………………….
am1x1 + am2 x2 +…………+ amnxn = bm
and the non-negativity restrictions x1, x2,….xn ≥ 0 is known as standard form.
In matrix notation the standard form of LPP can be expressed as:
Maximize Z = CX (objective function)
Subject to constraints AX = b and X ≥ 0
Characteristics:
 The objective function is of maximization type.
 All the constraint equation must be of equal type by adding slack or surplus variables
 RHS of each constraint must be positive type.
 All variables are non-negative.
18 List all possible cases that can arise in sensitivity analysis with their actions to obtain new solution.
Condition resulting from the changes Recommended action
Current solution remains optimal and feasible No further action is necessary
Current solution becomes infeasible Use dual simplex to recover feasibility
Current solution becomes non-optimal Use primal simplex to recover optimality
Current solution becomes both non-optimal and Use the generalized simplex method to obtain a new
infeasible solution
19 What are the changes that affect feasibility?
There are two types of changes that could affect feasibility of the current solution:
 Changes in resources availability (or right side of the constraints or requirement vector b) and
 Addition of new constraints.
20 What are the changes that affect optimality?
There are two particular situations that could affect optimality of the current solution:
 Changes in the original objective coefficients.
 Addition of new economic activity (variable) to the model.
UNIT-I / PART-B
1 Solve the following LPP by graphical method:
Max Z = 3x1 + 2x2 subject to -2x1 + x2 ≤ 1, x1 ≤ 2, x1 + x2 ≤ 3 and x1, x2 ≥ 0.
Solution: First consider the inequality constraints as equalities.
-2x1 + x2 = 1 (1)
x1 = 2 (2)
x1 + x2 = 3 (3)
x1 = 0 (4)
x2 = 0 (5)
From the line -2x1 + x2 = 1
Put x1 = 0  x2 =1  (0,1)
Put x2 =0  -2x1 = 1  x1 = -0.5  (-0.5, 0)

So the line (1) passes through the points (0,1) and (-0.5, 0). The points on this line satisfy the equation
-2x1 + x2 = 1. Now origin (0, 0) on substitution, gives -0 + 0 = 0 < 1; hence it also satisfies the inequality -2x 1 +
x2 ≤ 1. Thus all points on the origin side and on this line satisfy the inequality -2x 1 + x2 ≤ 1. Similarly
interpreting the other constraints, the feasible solution is OABCD. The feasible region is also known as
solution space of the LPP. Every point in this area satisfies all the constraints.
Now the aim is to find vertices of the solution space. B is the point of intersection of x1 = 2 and x1 +
 CS6704-Resource Management Techniques Department of CSE 2016-2017
x2 = 3. Solving these two equations, x1 = 2 and x2 = 1. Therefore the vertex B( 2,1). Similarly C is the

intersection of -2x1 + x2 = 1 and x1 + x2 = 3. Solving these C( , ).

Therefore the vertices solution space are O(0,0), A(2,0), B(2,1), C( , ) and D(0,1). The value of

Z at these vertices are given by (Z = 3x1 + 2x2).

Vertex Value of Z
O(0,0) 0
A(2,0) 6
B(2,1) 8

C( , )

D(0,1) 2
Since the problem is of maximization type, the optimum solution to the LPP is
Maximum Z = 8, x1 = 2 and x2 = 1
2 A company manufactures two types of printed circuits. The requirements of transistors, resistors
and capacitors for each type of printed circuits along with other data are given below:
Circuit Stock
A B Available
Transistor 15 10 180
Resistor 10 20 200
Capacitor 15 20 210
Profit Rs. 5 Rs. 8
How many circuits of each type should the company produce from the stock to earn maximum profit?
Solution: Let x1 be the number of type A circuits and x2 be the number of type B circuits to be produced. To
produce these units of type A and type B the company requires
Transistors = 15x1 +10x2
Resistor = 10x1 + 20x2
Capacitor = 15x1 + 20x2
Since the availability of these transistors, resistors and capacitors are 180, 200 and 210 respectively, the
constraints are
15x1 +10x2 ≤ 180
10x1 + 20x2 ≤ 200
15x1 + 20x2 ≤ 210
and x1 ≥ 0, x2 ≥ 0
Since the profit from type A is Rs. 5 and from type B is Rs. 8 per units, the total profit is 5x1 + 8x2.
Therefore the complete formulation of the LPP is
Maximize Z = 5x1 + 8x2
Subject to
15x1 +10x2 ≤ 180
10x1 + 20x2 ≤ 200
15x1 + 20x2 ≤ 210
and x1 ≥ 0, x2 ≥ 0

By using graphical method the solution space is given with the shaded area OABCD with vertices
O(0,0), A(12,0), B(10,3), C(2,9) and D(0,10). The values of Z of these vertices are given by Z=5x1+8x2
Vertex Value of Z
O(0,0) 0
A(12,0) 60
B(10,3) 74
C(2,9) 82
D(0,10) 80
 CS6704-Resource Management Techniques Department of CSE 2016-2017

Since the problem is of maximization type, the optimal solution is

Maximize Z = 82, x1 = 2, x2 = 9
3 Obtain all the basic solutions to the following system of linear equations:
2x1 + 6x2 + 2x3 + x4 = 3, 6x1 + 4x2 + 4x3 + 6x4 = 2 Which of them are basic feasible solutions and which are
non-degenerate basic solutions? Is the non-degenerate solution feasible?
Solution: Since there are 4 variables with 2 equations there are 4C 2 = 6 different basic solutions, which
are given in the following table:
Is the solution
Values of the basic Is the Is the solution
No Basic Non-basic feasible and
variables given by the solution non-
. variables variables non-
constraint equations feasible? degenerate?
degenerate?
2x1 + 6x2 = 3
1 x1, x2 x3 = x4 = 0 6x1 + 4x2 = 2 Yes No No
 x1 = 0, x2 = 1/2
2x1+ 2x3 = 3
2 x1, x3 x2 = x4 = 0 6x1 + 4x3 = 2 No Yes No
 x1 = -2, x3 = 7/2
2x1 + x4 = 3
3 x1, x4 x2 = x3 = 0 6x1 + 6x4 = 2 No Yes No
 x1 = 8/3, x4 = -7/3
6x2 + 2x3 = 3
4 x2, x3 x1 = x4 = 0 4x2 + 4x3 = 2 Yes No No
 x2 = ½, x3 = 0
6x2 + x4 = 3
5 x2, x4 x1 = x3 = 0 4x2 + 6x4 = 2 Yes No No
 x2 = ½, x4 = 0
2x3 + x4 = 3
6 x3, x4 x1 = x2 = 0 4x3 + 6x4 = 2 No Yes No
 x3 = 2, x4 = -1
4 Use simplex method to solve the LPP: Maximize Z = 4x1 + 10x2
subject to 2x1 + x2 ≤ 50, 2x1 + 5x2 ≤ 100, 2x1 + 3x2 ≤ 90 and x1 , x2 ≥ 0
Solution: By introducing the slack variables s1, s2 and s3, the problem in standard form becomes

Maximize Z = 4x1 + 10x2 + 0s1 + 0s2 + 0s3

Subject to
2x1 + x2 + s1 + 0s2 + 0s3 = 50
2x1 + 5x2 + 0s1 + s2+ 0s3 = 100
2x1 + 3x2 + 0s1 + 0s2 + s3 = 90
and x1 , x2, s1, s2, s3 ≥ 0

Since there are 3 equations with 5 variables, the initial basic feasible solution is obtained by equating (5-3)=2
variables to 0.
Therefore the initial basic feasible solution is s1 = 50, s2 = 100, s3 = 90.
(x1 = 0, x2 = 0, non-basic)

The initial simplex table is given by

Cj (4 10 0 0 0)
CB YB XB x1 x2 s1 s2 s3  = min XBi/air
0 s1 50 2 1 1 0 0 50
0 s2 100 5 5 0 1 0 20*
0 s3 90 2 3 0 0 1 30
Zj - Cj 0 -4 -10 0 0 0
 CS6704-Resource Management Techniques Department of CSE 2016-2017
The improved basic feasible solution is given in the following simplex table

First Iteration:
Cj (4 10 0 0 0)
CB YB XB x1 x2 s1 s2 s3
0 s1 30 8/5 0 1 -1/5 0
10 x2 20 2/5 1 0 1/5 0
0 s3 30 4/5 0 0 -3/5 1
Zj - Cj 200 0 0 0 2 0

Since all Zj - Cj ≥ 0, the current basic feasible solution is optimal. Therefore the optimal solution is
Max Z = 200, x1 = 0, x2 = 20
5 Use simplex method to Minimize Z = x2 - 3x3 + 2x5
Subject to 3x2 - x3 + 2x5 ≤ 7, -2x2 + 4x3 ≤ 12, -4x2 + 3x3 + 8x5 ≤ 10 and x2 , x3 , x5 ≥ 0
Solution: Since the objective function is of minimization type, convert it into mazimization type as follows:
Maximize (-Z) = Maximize Z* = -x2 + 3x3 - 2x5
Subject to
3x2 - x3 + 2x5 ≤ 7
-2x2 + 4x3 ≤ 12
-4x2 + 3x3 + 8x5 ≤ 10
x2 , x3 , x5 ≥ 0

By introducing non-negative slack variables s1, s2 and s3, the standard form of the LPP becomes
Maximize Z* = -x2 + 3x3 - 2x5 + 0s1 + 0s2 + 0s3
Subject to constraints
3x2 - x3 + 2x5 + s1 + 0s2 + 0s3 = 7
-2x2 + 4x3 +0s1 + s2 + 0s3 = 12
-4x2 + 3x3 + 8x5 +0s1 + 0s2 + s3 = 10
x2 , x3 , x5, s1, s2, s3 ≥ 0

Therefore the initial basic feasible solution is given by s1 = 7, s2 = 12, s3 = 10

(x2 = x3 = x5 = 0, non-basic)

Initial Iteration:
Cj (-1 3 -2 0 0 0)
CB YB XB x2 x3 x5 s1 s2 s3 
0 s1 7 3 -1 2 1 0 0 -
0 s2 12 -2 (4) 0 0 1 0 12/4 = 3*
0 s3 10 -4 3 8 0 0 1 10/3 = 3.33
*
Zj - Cj 0 1 -3 2 0 0 0

Since Z2* - C2 < 0, the current basic feasible solution is not optimal. The non-basic variable s2 leaves the basis.

Second Iteration:
Cj (-1 3 2 0 0 0)
CB YB XB x2 x3 x5 s1 s2 s3
-1 x2 4 1 0 4/5 2/5 1/10 0
3 x3 5 0 1 2/5 1/5 3/10 0
0 s3 11 0 0 10 1 -1/2 1
Zj* - Cj 11 0 0 12/5 1/5 4/5 0

Since all (Zj* - Cj) ≥ 0, the current basic feasible solution is optimal. Therefore the solution is given by
 CS6704-Resource Management Techniques Department of CSE 2016-2017
Maximize Z* = 11, x2 = 4, x3 = 5, x5 = 0
But Minimize Z = -Maximize Z* = -11
Minimize Z = -11, x2 = 4, x3 = 5, x5 = 0
6 A gear manufacturing company received an order for three specific types of gear for regular supply.
The management is considering to devote the available excess capacity to one or more of the
three types A, B and C. The available capacity on the machines which might limit output and the
number of machine hours required for each unit of the respective gear is also given below:
Productivity in machine hours/ unit
Machine Type Available machine hours/week
Gear A Gear B Gear C
Gear Hobbing m/c 250 8 2 3
Gear Shaping m/c 150 4 3 0
Gear Grinding m/c 50 2 - 1
The unit profit would be Rs. 20, Rs. 6 and Rs. 8 respectively for the gears A, B and C. Find how much of
each gear the company should produce in order to maximize profit?
Solution: Let x1, x2 and x3 be the number of units of gear A, B and C produced respectively to maximize the
profit. The mathematical formulation of the LPP is given by
Maximize Z = 20x1 + 6x2 + 8x3
Subject to
8x1 + 2x2 + 3x3 ≤ 250
4x1 + 3x2 ≤ 150
2x1 + x3 ≤ 50
And x1, x2, x3 ≥ 0
By introducing non-negative slack variables s1, s2 and s3 the standard form of the LPP becomes
Maximize Z* = 20x1 + 6x2 + 8x3 + 0s1 + 0s2 + 0s3
Subject to
8x1 + 2x2 + 3x3 s1 + 0s2 + 0s3 = 250
4x1 + 3x2 0s1 + s2 + 0s3 = 150
2x1 + x3 +0s1 + 0s2 + s3 = 50
And x1, x2, x3, s1, s2, s3 ≥ 0

Therefore the initial basic feasible solution is given by s1 = 250, s2 = 150, s3 = 50

(x1 = x2 = x3 = 0, non-basic)

Initial Iteration:
Cj (20 6 8 0 0 0)
CB YB XB x1 x2 x3 s1 s2 s3 
0 s1 250 8 2 3 1 0 0 250/8
0 s2 150 4 3 0 0 1 0 150/4
0 s3 50 (2) 0 1 0 0 1 50/2*
Zj - Cj 0 -20 -6 -8 0 0 0

Third Iteration:
Introduce x3 and drop x1
Cj (20 6 8 0 0 0)
CB YB XB x1 x2 x3 s1 s2 s3
0 s1 0 -2/3 0 0 1 -2/3 -3
6 x2 50 4/3 1 0 0 1/3 0
8 x3 50 2 0 1 0 0 1
Zj - Cj 700 4 0 0 0 2 8

Since all Zj - Cj ≥ 0, the current feasible solution is optimal. Therefore the optimal solution is
Maximize Z = 700, x1 = 0, x2 = 50, x3 = 50
 CS6704-Resource Management Techniques Department of CSE 2016-2017

Therefore the company should produce 50 units of gear B, 50 units of gear C and none of gear A in order to
have maximum profit of Rs. 700.
7 Consider the LPP Max Z = 2x1 + x2 + 4x3 - x4 subject to x1 + 2x2 + x3 - 3x4 ≤ 8, - x2 + x3 + 2x4 ≤ 0, 2x1 + 7x2 -
5x3 - 10x4 ≤ 21 and x1 , x2 , x3 , x4 ≥ 0.
a) Solve the LPP.
b) Discuss the effect of change of b2 to 11.
c) Discuss the effect of change of b to [3-2,4].

Solution: a) By using regular simplex method, the optimal simplex table is displayed below:
Cj (2 1 4 -1 0 0 0)
CB YB XB x1 x2 x3 x4 s1 s2 s3
1 x2 0 0 1 -1 -2 0 -1 0
2 x1 8 1 0 3 1 1 2 0
0 s3 5 0 0 -4 2 -2 3 1
Zj - Cj 16 0 0 1 1 1 3 0

Therefore the optimal solution is

Max Z = 16, x1 =8, x2 = x3 = x4 = 0.

b) If b2 is changed to 11 , then b = [8, 11, 21]

The solution is infeasible. That is the change from b2 from 0 to 11 affects the feasibility of the solution. The
changes in the final table reduce to

Cj (2 1 4 -1 0 0 0)
CB YB XB x1 x2 x3 x4 s1 s2 s3
1 x2 -11 0 1 -1 (-2) 0 -1 0
2 x1 30 1 0 3 1 1 2 0
0 s3 38 0 0 -4 2 -2 3 1
Zj - Cj 49 0 0 1 1 2 3 0

Since the current solution is infeasible, use dual simplex method.

The optimal solution to the new problem is
Max Z = 87/2, x1 = 49/2, x2 = 0 = x3, x4 = 11/2

c) If b is changed from [8, 0, 21] to [3, -2, 4]

The solution is infeasible. The change of b from [8, 0, 21] to [3, -2, 4] affects the feasibility of the solution

Cj (2 1 4 -1 0 0 0)
CB YB XB x1 x2 x3 x4 s1 s2 s3
1 x2 -3 1 1 0 0 0 5/2 ½
0 s1 14 -2 0 0 -5 1 -17/2 -3/2
4 x3 -5 1 0 1 2 0 7/2 ½
Zj - Cj -23 3 0 0 9 0 33/2 5/2

Since x3 leaves the basis and all a3j ≥ 0, the new problem poses no feasible solution.
8 Consider the LPP Max Z = 3x1 + 4x2 + x3 + 7x4 subject to 8x1 + 3x2 + 4x3 + x4 ≤ 7, 2x1 + 6x2 + x3 + 5x4 ≤ 3,
x1 + 4x2 + 5x3 + 2x4 ≤ 8 and x1 , x2 , x3 , x4 ≥ 0
a) Solve the LPP.
b) If a new constraint 2x1 + 3x2 + x3 + 5x4 ≤ 4 is added to the above LPP, discuss the effect of change
in the optimal solution of the original problem.
c) If a new constraint 2x1 + 3x2 + x3 + 5x4 ≤ 2 is added (or) if the upper limit of the above constraint
is reduced to 2, discuss the effect of change in the optimum solution of the original problem.
 CS6704-Resource Management Techniques Department of CSE 2016-2017
Solution: a) By introducing the non-negative slack variables s1, s2 and s3 and then solving the problem by
simplex method, the optimal simplex table is given by:
Cj (3 4 1 7 0 0 0)
CB YB XB x1 x2 x3 x4 s1 s2 s3
3 x1 16/19 1 9/38 1/2 0 5/38 -1/38 0
7 x4 5/19 0 21/19 0 1 -1/19 4/19 0
0 s3 126/19 0 59/38 9/2 0 -1/38 -15/38 1
Zj - Cj 83/19 0 169/38 1/2 0 1/38 53/38 0

The optimal solution is

Min Z = 83/19, x1 = 16/19, x2 = 0, x3 = 0, x4 = 5/19

b) Now consider the new constraint 2x1 + 3x2 + x3 + 5x4 ≤ 4

As this constraint is satisfied by the optimal solution, the solution remains feasible and optimal for the
modified problem also. Here the constraint 2x1 + 3x2 + x3 + 5x4 ≤ 4 is redundant.

c) Consider the new constraint 2x1 + 3x2 + x3 + 5x4 ≤ 2

As this constraint is not satisfied by the current optimal solution, the current solution is no longer optimal for
the modified problem. In order to find the new optimal solution, a new row is added. By adding the non-
negative slack variable s4 the constraint becomes 2x1 + 3x2 + x3 + 5x4 + s4 = 2
Cj (3 4 1 7 0 0 0 0)
CB YB XB x1 x2 x3 x4 s1 s2 s3 s4
3 x1 16/19 1 9/38 1/2 0 5/38 -1/38 0 0
7 x4 5/19 0 21/19 0 1 -1/19 4/19 0 0
0 s3 126/19 0 59/38 9/2 0 -1/38 -15/38 1 0
0 s4 -1 0 -3 0 0 0 -1 0 1
Zj - Cj 83/19 0 169/38 1/2 0 1/38 53/38 0 0

First iteration:
Drop s4 and introduce s2

Cj (3 4 1 7 0 0 0 0)
CB YB XB x1 x2 x3 x4 s1 s2 s3 s4
3 x1 33/38 1 12/38 1/2 0 5/38 0 0 -1/38
7 x4 1/19 0 9/19 0 1 -1/19 0 0 4/19
0 s3 267/38 0 52/19 9/2 0 -1/38 0 1 -15/38
0 s2 1 0 3 0 0 0 1 0 -1
Zj - Cj 113/38 0 5/19 1/2 0 1/38 0 0 53/38

The new optimal solution is

Max Z = 113/38, x1 = 33/38, x2 =0, x3 = 0, x4 = 1/19
The additional constraint has decreased the optimum value of the objective function.
9 Consider the LPP Max Z = 5x1 + 3x2 subject to 3x1 + 5x2 ≤ 15, 5x1 + 6x2 ≤ 10 and x1 , x2 ≥ 0.
a) Solve the LPP.
b) Find how far the component C1 of C can be increased without affecting the optimality of the
solution.

Solution: a) By introducing the non-negative slack variables s1 and s2 and using regular simplex method, the
optimal simplex table is displayed below:
Cj (5 3 0 0)
CB YB XB x1 x2 s1 s2
0 s1 9 0 7/5 1 -3/5
 CS6704-Resource Management Techniques Department of CSE 2016-2017
5 x1 2 1 6/5 0 1/5
Zj - Cj 10 0 3 0 1

The optimal solution is

Max Z = 10, x1 = 2, x2 = 0

b) Here C1  CB i.e., C1 is the coefficient of the basic variable x1 and C1 =5

 Max {0, -5/2, -5} ≤ ∆C1 ≤ ∞
 0 ≤ ∆C1 ≤ ∞
 C 1 – 0 ≤ C1 ≤ C1 + ∞
 5 – 0 ≤ C1 ≤ 5 + ∞
 5 ≤ C1 ≤ ∞, which is the range over C 1 can increase without destroying the optimality of the
solution.
10 Consider the LPP Max Z= 3x1 + 5x2 subject to x1 ≤ 4, 3x1 + 2x2 ≤ 18 and x1 , x2 ≥ 0.
a) Solve this LPP.

b) If a new variable x5 is added to this problem with a column and C5 = 7, find the change in

the optimal solution.

Solution: a) By using simplex method, the optimal solution of the original problem is
Cj (3 5 0 0)
CB YB XB x1 x2 x3 x4
0 x3 4 1 0 1 0
5 x2 9 3/2 1 0 1/2
Zj - Cj 45 9/2 0 0 5/2

Max Z = 45, x1 =0, x2 = 9

b) If a new variable x5 is introduced with a5 = and C5 =7, then

ā5 = = = ā5

and Z5 – C5 = CB ā5 – C5 = (0, 5) – 7 = -2. The optimality of the solution is affected. The change in the

optimal table is reduced to

Cj (3 5 0 0 7)
CB YB XB x1 x2 x3 x4 x5 
0 x3 4 1 0 1 0 (1) 4
5 x2 9 3/2 1 0 1/2 1 9
Zj - Cj 45 9/2 0 0 5/2 -2

First iteration: Introduce x5 and drop x3

Cj (3 5 0 0)
CB YB XB x1 x2 x3 x4 x5
7 X5 4 1 0 1 0 1
5 x2 5 1/2 1 -1 1/2 0
Zj - Cj 53 13/2 0 2 5/2 0

Since all Zj - Cj ≥ 0, the current solution is optimal. The optimal solution to the new problem is
Max Z = 53, x1 = 0, x2 = 5, x5 = 4.
 CS6704-Resource Management Techniques Department of CSE 2016-2017

UNIT II DUALITY AND NETWORKS

Definition of dual problem – Primal – Dual relationships – Dual simplex methods – Post optimality analysis –
Transportation and assignment model – Shortest route problem.
UNIT-II / PART-A
1 What is a dual problem?
For every linear programming problem there is a unique linear programming problem associated with it,
involving the same data and closely related optimal solutions. The original problem is called the Primal
problem while the other is called its dual problem. In general the two problems are said to be duals of each
other.
2 Mention the importance of duality.
The importance of duality concept is due to two main reasons:
 If the primal contains a large number of constraints and a smaller number of variables, the
computation can be reduced by converting into the dual problem and solving it.
 The interpretation of the dual variables from the cost or economic point of view proves extremely
useful in making future decisions in the activities being programmed.
3 How is a dual problem constructed from the primal?
 A dual variable is defined for each primal (constraint) equation.
 A dual constraint is defined for each primal variable.
 The constraint (column) coefficients of a primal variable define the LHS coefficients of the dual
constraint and its objective coefficient define the RHS.
 The objective coefficients of the dual equal the RHS of the primal constraint equations.
4 How can optimal dual solution be obtained?
The primal and dual solutions are so closely related that the optimal solution of the primal problem directly
yields the optimal solution of the dual.
Method 1:
(Optimal values of dual variables) = (Row vector of original objective coefficients of optimal primal basic
variables) x (optimal primal inverse)
Method 2:
(Optimal primal z-coefficient of any variable xj) = (LHS of the jth dual constraint) – (RHS of the jth dual
constraint)
5 How are the primal and dual objective values related?
In the primal-dual problems, if one is maximization the other must be minimization. From this standpoint,
the objective values in the two problems are related in the following manner:
For any feasible primal and dual solutions,
(Objective value in the maximization problem) ≤ (Objective value in the minimization problem)
6 What is dual simplex method?
In dual simplex method, LP starts at a basic solution that is optimal but infeasible, and successive iterations
remain basic and optimal as they move towards feasibility. At the last iteration the feasible optimal solution
is found.
7 Define transportation model.
There are m sources and n destinations, each represented by a node. The arcs represent the routes linking the
sources and destinations. Arc(i,j) joining source i to destination j, carries cij the transportation cost per unit and
the amount shipped xij. The amount of supply at source i is ai, and the amount of demand at destination j is bj.
the objective of the model is to determine the unknowns xij that will minimize the total transportation cost
while satisfying all the supply and demand restrictions.
8 What are the methods used in transportation problem to obtain the initial basic feasible solution.
 North-west corner rule: The first assignment is made in the cell occupying the upper left-hand corner
of the transportation table.
 Least cost method: Identify the cell with smallest cost and assign the minimum value.
 Vogel’s approximation method: Find the difference between the smallest and next smallest cost in
 CS6704-Resource Management Techniques Department of CSE 2016-2017
each row (column) and write them in brackets against the row (column).
9 How will you overcome degeneracy in a transportation problem?
If the number of occupied cells in a m x n transportation problem is less than (m+n-1) then the problem is said
to be degenerate where m is the number of rows and n is the number of columns in the transportation table. To
resolve degeneracy, allocate an extremely small amount (close to zero) to one or more empty cells of the
transportation table, so that the total number of occupied cells becomes (m+n-1) at independent positions.
The small amount is denoted by .
10 What is balanced transportation problem & unbalanced transportation problem?
When the sum of supply is equal to demands, then the problem is said to be balanced transportation
problem. A transportation problem is said to be unbalanced if the total supply is not equal to the total
demand.
11 How do you convert an unbalanced transportation problem into a balanced one?
The unbalanced transportation problem is converted into a balanced one by adding a dummy row (source) or
dummy column (destination) with zero cost vectors whichever is necessary. The unit transportation cost of
the dummy row/ column elements are assigned to zero. Then the problem is solved by the usual procedure.
12 What is assignment problem?
An assignment problem is a particular case of a transportation problem in which the objective is to assign a
number of tasks (jobs, origins or sources) to an equal number of facilities (machines, persons or
destinations) where each operator performs only one operation, the overall objective is to maximize the total
profit or minimize the overall cost of the given assignment.
13 What do you understand by restricted assignment? Explain how you should overcome it?
In the assignment technique, it may not be possible to assign a particular task to a particular facility due to
technical difficulties or other restrictions. This can be overcome by assigning a very high processing time or
cost (it can be ∞) to the corresponding cell.

14 Give the mathematical formulation of an assignment problem.

The assignment problem can be expressed as
Maximize Z =

Where cij is the cost of assigning ith machine to the jth job subject to the constraints

xij =

i.e) xij = 1 or 0 xij (xij – 1) = 0  xij2 = xij

= 1, i = 1, 2, …, n and

=1, j = 1, 2, … , n
15 Give the difference between transportation problem and assignment problem.
Transportation Problems Assignment Problems
Supply at any source may be any positive Supply at any source will be 1. ie) ai = 1
quantity ai.
Demand at any destination maybe a positive Demand at any destination will be 1.
quantity bj. ie) bj = 1
One or more source to any number of One source to one destination.
destinations.
16 How do you identify alternative solution in assignment problem?
If the final cost matrix contains more than required number of zeroes at the independent position. This
implies that there is more than one optimal solution with some optimum assignment cost.
17 Define unbounded assignment problem and what are the rules to recognize it?
In some LP models, the values of the variables may be increased indefinitely without violating any of the
 CS6704-Resource Management Techniques Department of CSE 2016-2017
constraints, meaning that the solution space is unbounded in at least one direction. As a result, the objective
value may increase (maximization case) or decrease (minimization case) indefinitely.
The rule for recognizing unboundedness is that if at any iteration all the constraint coefficients of any non
basic variable are zero or negative, then the solution space is unbounded in that direction. The objective
coefficient of that variable is negative in the case of maximization or positive in the case of minimization,
then the objective value is unbounded as well.
18 What is a travelling salesman problem?
A salesman normally must visit a number of cities starting from his head quarters. The distance between every
pair of cities are assumed to be known. The problem of finding the shortest distance if the salesman starts
from his head quarters and passes through each city exactly once and returns to the headquarters is
called Travelling Salesman problem.
19 Define transshipment model.
A problem in which available commodity frequently moves from one or more number of sources to any
number of destinations before reaching its actual destination is called as transshipment problems. The
transshipment model recognizes that it may be cheaper to ship through intermediate or transient nodes
before reaching the final destination.
20 What are the algorithms that are used to solve shortest-route problem?
The two algorithms for solving both cyclic and acyclic networks:
 Dijikstra’s algorithm
 Floyd’s algorithm
UNIT-II / PART-B
1 Use duality to solve the following LPP. Minimize Z = 2x1 + 2x2 subject to 2x1 + 4x2 ≥ 1, -x1 - 2x2 ≤ -1, 2x1 +
x2 ≥ 1 and x1 , x2 ≥ 0.
Solution: Given primal LPP is Minimize Z = 2x1 + 2x2
Subject to
2x1 + 4x2 ≥ 1
x1 + 2x2 ≥ 1
2x1 + x2 ≥ 1
and x1 , x2 ≥ 0
Its dual problem is Max W = y1 + y2 + y3
Subject to
2y1 + y2 + 2y3 ≤ 2
4y1 + 2y2 + y3 ≤ 2
and y1, y2, y3 ≥ 0
By introducing the non-negative slack variables s1 and s2, the standard form of the dual LPP becomes
Max W = y1 + y2 + y3 + 0s1 + 0s2
Subject to
2y1 + y2 + 2y3 + s1 + 0s2 = 2
4y1 + 2y2 + y3 +0s1 + s2 = 2
and y1, y2, y3, s1, s2 ≥ 0
The initial basic feasible solution is s1 = 2, s2 = 2.

First iteration:
bj (1 1 1 0 0)
CB YB XB y1 y2 y3 s1 s2 
0 s1 2 2 1 2 1 0 2/2
0 s2 2 (4) 2 1 0 1 2/4
Wj - bj 0 -1 -1 -1 0 0

Third iteration: Introduce y2 and drop y1

bj (1 1 1 0 0)
CB YB XB y1 y2 y3 s1 s2
1 y3 2/3 0 0 1 2/3 -1/3
 CS6704-Resource Management Techniques Department of CSE 2016-2017
1 y2 2/3 2 1 0 -1/3 2/3
Wj - bj 4/3 1 0 0 1/3 1/3

Since all Wj - bj ≥ 0, the current basic solution is optimal. The optimal solution to the dual LPP is
Max W = 4/3, y1 =0, y2= 2/3, y3 = 2/3
Here it is observed that the primal variables x1 and x2 respectively correspond to the slack variables s1 and s2
of the dual problem. As seen from the above table, the net evaluations W j - bj corresponding to the slack
variables s1 and s2 are 1/3 and 1/3 respectively.

Therefore the optimal solution to the original primal LPP is

Min Z = 4/3, x1 = 1/3, x2 = 1/3
2 Prove using duality theory that the following linear program is feasible but has no optimal solution.
Minimize Z = x1 - x2 + x3 subject to x1 – x3 ≥ 4, x1 - x2 + 2x3 ≥ 3 and x1 , x2 , x3 ≥ 0.
Solution: Given primal LPP is
Min Z = x1 - x2 + x3
Subject to
x1 – x3 ≥ 4
x1 - x2 + 2x3 ≥ 3
and x1 , x2 , x3 ≥ 0

Its dual problem is

Max W = 4y1 + 3y2
Subject to
y1 + y2 ≤ 1
0y1 + y2 ≥ 1
-y1 + 2y2 ≤ 1
and y1, y2 ≥ 0

By introducing the slack variables s1, s3 and surplus variable s2 and an artificial variable R1, the standard form
of the dual LPP is
Max W = 4y1 + 3y2 + 0s1 + 0s2 + 0s3 – MR1
Subject to
y1 + y2 + s1 + 0s2 + 0s3 = 1
0y1 + y2 + 0s1 – s2 + 0s3 + R1 = 1
-y1 + 2y2 + 0s1 + 0s2 + s3 = 1
and y1, y2, s1, s2, s3, R1 ≥ 0

The initial basic feasible solution is given by s1 =1, R1 = 1, s3 = 1 (basic)

(y1 = y2 = s2 =0, non-basic)

Initial iteration
bj (4 3 0 0 -M 0)
1 2 s3
CB YB X s1 R1 
s2
0 s1 1 1 1 0 1 0 0 1
-M R1 1 0 1 -1 0 1 0 1
0 s3 1 -1 (2) 0 0 0 1 ½*
Wj - bj -M -4 -M-3 M 0 0 0

Second iteration: Introduce y1 and drop s1

bj (4 3 0 0 -M 0)
CB YB XB y1 y2 s2 s1 R1 s3
4 y1 1/(3) 1 0 0 2/3 0 -1/3
 CS6704-Resource Management Techniques Department of CSE 2016-2017
-M R1 1/3 0 0 -1 -1/3 1 -1/3
3 y2 2/3 0 1 0 1/3 0 1/3
Wj - bj 0 0 M 0
Since all Wj - bj ≥ 0, and an artificial variable R1 appears in the basis at non-zero level, the dual problem has
no optimal basic feasible solution. Therefore there exists no finite optimum solution to the given primal LPP.
3 Using dual simplex method solve the LPP. Minimize Z = 2x1 + x2 subject to 3x1 + x2 ≥ 3, 4x1 + 3x2 ≥ 6, x1 +
2x2 ≥ 3 and x1 , x2 ≥ 0.
Solution: After converting the objective function in to maximization type and all the constraints in ≤ type, the
given LPP becomes
Max Z* = -2x1 - x2
Subject to
-3x1 - x2 ≤ -3
-4x1 - 3x2 ≤ -6
-x1 - 2x2 ≤ -3
and x1 , x2 ≥ 0.

By introducing the non-negative slack variables s1, s2 and s3, the LPP becomes
Max Z* = -2x1 - x2 + 0s1 + 0s2 + 0s3
Subject to
-3x1 - x2 + s1 = -3
-4x1 - 3x2 + s2 = -6
-x1 - 2x2 + s3 = -3
and x1 , x2, s1, s2, s3 ≥ 0.

The initial basic solution is given by s1 = -3, s2 = -6, s3 = -3 (basic)

(x1 = x2 = 0, non-basic)

Initial iteration
Cj (-2 -1 0 0 0)
CB YB XB x1 x2 s1 s2 s3
0 s1 -3 -3 -1 1 0 0
0 s2 -6 -4 (-3) 0 1 0
0 s3 -3 -1 -2 0 0 1
Zj* - Cj 0 2 1 0 0 0

Since Zj* - Cj ≥ 0 and all X Bi < 0, the current solution is not an optimum basic feasible solution. Since X B2=-6
is the most negative, the corresponding basic variable s2 leaves the basis.

Second iteration: Drop s1 and introduce x1

Cj (-2 -1 0 0 0)
CB YB XB x1 x2 s1 s2 s3
-2 x1 3/5 1 0 -3/5 1/5 0
-1 x2 6/5 0 1 4/5 -3/5 0
0 s3 0 0 0 1 -1 1
Zj* - Cj -12/5 0 0 2/5 1/5 0

Since all Zj* - Cj ≥ 0, and all X Bi ≥ 0, the current solution is an optimum basic feasible solution. The optimal
solution is
Max Z* = -12/5, x1 = 3/5, x2 = 6/5
But Min Z = -Max Z* = -(-12/5) = 12/5
Min Z = 12/5, x1 = 3/5, x2 = 6/5
 CS6704-Resource Management Techniques Department of CSE 2016-2017

4 Determine basic feasible solution to the following transportation problem using North West
Corner rule:

Sink
A B C D E Supply
P 2 11 10 3 7 4
Origin Q 1 4 7 2 1 8
R 3 9 4 8 12 9
Demand 11 10 3 7
3 3 4 5 6 1 1

4 7 2 1
Solution: Since ai = bj = 21, the given problem is 8
balanced. Therefore there exists a feasible solution to the
transportation problem. 9 4 8 12
9
2 11 10 3 7
3 4 3 4 5 6
1 4 7 2 1
8

3 9 4 8 12
9

3 3 4 5 6

Following North West Corner rule, the first allocation is made in the cell (1,1)
Here x11 =min {a1, b1} = min {4,3} = 3
Therefore allocate 3 to cell (1,1) and decrease 4 by 3 ie., 4-3 = 1

As the first column is satisfied cross out the first column and the resulting reduced transportation table is

Finally the initial basic feasible solution is shown in the following table.
2 11 10 3 7
3 1

1 4 7 2 1
2 4 2

3 9 4 8 12
3 6

The initial transportation cost = Rs. 2 x 3 + 11 x 1 + 4 x 2 + 7 x 4 + 2 x 2 + 8 x 3 + 12 x 6

 CS6704-Resource Management Techniques Department of CSE 2016-2017
Rs. 153/-
5 Find the optimal transportation cost of the following matrix using least cost method for finding
the critical solution.
Market
A B C D E Available
P 4 1 2 6 9 100
Factory Q 6 4 3 5 7 120
R 5 2 6 4 8 120
Demand 40 50 70 90 90

Solution: Since ∑ai = ∑bj = 340, the given transportation problem is balanced. By using least cost method the
initial solution is shown in the table:

4 1 2 6 9
50 50

6 4 3 5 7
10 20 90

5 2 6 4 8
30 90

The initial transportation cost

= Rs. 1410

For optimality: Since the number of non-negative independent allocations is (m+n-1) apply MODI method:
The new basic feasible solution is displayed in the following table:
4 1 2 6 9
10 50 40

6 4 3 5 7
30 90

5 2 6 4 8
30 90

The above table satisfies the rim conditions so apply the MODI method.
4 1 2 6 3 9 6
10 50 40 u1=0
3 3
6 5 4 2 3 5 4 7
30 90 u2=1
1 2 1
5 2 2 6 3 4 8 7
30 90 u3=1
0 3 1

v1=4 v2=1 v3=2 v4=3 v5=6

The optimum allocation schedule is given by x11 = 10, x12 = 50, x13 = 40, x23 = 30, x25 = 90, x31 = 30, x34
 CS6704-Resource Management Techniques Department of CSE 2016-2017
= 90 and the optimum transportation cost Rs. 1400/-
6 Find the non-degenerate basic feasible solution for the following transportation problem using:
a) North west corner approach
b) Least cost method
c) Vogel’s approximation method.

To Supply

10 20 5 7 10
13 9 12 8 20
From 4 5 7 9 30
14 7 1 0 40
Demand 3 12 5 19 50
60 60 20 10

Solution: Since ∑ai = ∑bj = 150, the transportation problem is balanced. There exist a basic feasible solution to
this problem.
a) The starting solution by NWC rule is shown in the table
10 20 5 7
10

13 9 12 8
20

4 5 7 9
30

14 7 1 0
40

3 12 5 19
20 20 10

The initial transportation cost = Rs. 1290/-, as 0

b) Least Cost Method:

10 20 5 7
10

13 9 12 8
20

4 5 7 9
10 20

14 7 1 0
10 20 10

3 12 5 19
50

Since the number of non-negative allocations at independent positions is (m+n-1)=8, the solution is non-
degenerate basic feasible.
The initial transportation cost = Rs. 760/-

c) Vogel’s Approximation Method:

 CS6704-Resource Management Techniques Department of CSE 2016-2017
10 20 5 7
10

13 9 12 8
20

4 5 7 9
30

14 7 1 0
10 20 10

3 12 5 19
50 

The initial transportation cost = Rs. 670/- as 0

7 The processing time in hours, for the jobs when allocated to the different machines are indicated
below. Assign the machines for the jobs so that the total processing time is minimum.
Machines
M1 M2 M3 M4 M5

J1 9 22 58 11 19
J2 43 78 72 50 63
Jobs J3 41 28 91 37 45
J4 74 42 27 49 39
J5 36 11 57 22 25

Solution: The cost matrix of the given problem is

9 22 58 11 19

43 78 72 50 63

41 28 91 37 45

74 42 27 49 39

36 11 57 22 25

Since the number of rows is equal to the number of columns in the cost matrix, the given assignment problem
is balanced.
Select the smallest cost element in each row and subtract from all elements in corresponding row
and get a reduced matrix.
0 13 49 2 10

0 35 29 7 20

13 0 63 9 17

47 15 0 22 12

25 0 46 11 14

Select the smallest cost element in each column and subtract from all elements in corresponding
 CS6704-Resource Management Techniques Department of CSE 2016-2017
column. The reduced matrix is
0 13 49 0 0

0 35 29 5 10

13 0 63 7 7

47 15 0 20 2

25 0 46 9 4

Cover all zeros by drawing a minimum number of straight lines

0 13 49 0 0

0 35 29 5 10

13 0 63 7 7

47 15 0 20 2

25 0 46 9 (4)

The optimal assignment schedule is J1  M4, J2  M1, J3  M2, J4  M3, J5  M5 and the optimum
processing time is
= 11 + 43 + 28 + 27 + 25 hours
= 134 hours
8 A company has four machines to do three jobs. Each job can be assigned to one and only one
machine. The cost of each job on each machine is given in the following table. What are the job
assignments which will minimize the cost?
Machines
1 2 3 4
A 18 24 28 32
Jobs B 8 13 17 19
C 10 15 19 22

Solution: The cost matrix of the given assignment problem is

18 24 28 32

8 13 17 19

10 15 19 22

Since the number of rows is less than the number of columns in the cost matrix, the given assignment problem
is unbalanced. To make it balanced, add a dummy job D (row) with 0 cost elements. The balanced cost matrix
is given by:
 CS6704-Resource Management Techniques Department of CSE 2016-2017

18 24 28 32

8 13 17 19

10 15 19 22

0 0 0 0

Now select the smallest cost element in each row (column) and subtract this from all the elements of
corresponding row (column), the reduced matrix is given as:

0 6 10 14

0 5 9 11

0 5 9 12

0 0 0 0

Cover the all zeros by drawing minimum number of straight lines. Choose the smallest cost
element not covered by these straight lines.
0 6 10 14

0 5 9 11

0 (5) 9 12

0 0 0 0

(0) 1 1 5

0 (0) 0 2

0 0 (0) 3

9 4 0 (0)

The optimum assignment schedule is given by A  1, B  2, C  3, D  4 and the optimum assignment cost
= 18 + 13 + 19 + 0 = 50/- units of cost.
9 A company has a team of four salesmen and there are four districts where the company wants to
start its business. After taking into account the capabilities of salesman and the nature of districts,
the company estimates that the profit per day in rupees for each salesman in each district is as
below. Find the assignment of salesmen to various districts which will yield maximum profit.
Districts
 CS6704-Resource Management Techniques Department of CSE 2016-2017
1 2 3 4
A 16 10 14 11
B 14 11 15 15
Salesmen C 15 15 13 12
D 13 12 14 15
Solution: The cost matrix of the given assignment problem is
(16) 10 14 11

14 11 15 15

15 15 13 12

13 12 14 15

Convert to equivalent minimization problem by subtracting all cost elements in the cost matrix
by the highest cost element 16.
0 6 2 5

2 5 1 1

1 1 3 4

3 4 2 1

Get the reduced cost matrix

0 6 2 5

1 4 0 0

0 0 2 3

2 3 1 0

Since each row and each column contains at least one zero, make the assignment in rows and
columns having single zero.
(0) 6 2 5

1 4 (0) 0

0 (0) 2 3

2 3 1 (0)

The optimum assignment schedule is given by A  1, B  3, C  2, D  4 and the optimum profit

= Rs. (16 + 15 + 15 + 15)
= Rs. 61/-.
10 Solve the following travelling salesman problem so as to minimize the cost per cycle.
To
A B C D E
A - 3 6 2 3
CS6704-Resource Management Techniques Department of CSE 2016-2017
B 3 - 5 2 3
From C 6 5 - 6 4
D 2 2 6 - 6
E 3 3 4 6 -
Solution: The cost matrix of the given travelling salesman problem is
∞ 3 6 2 3

3 ∞ 5 2 3

6 5 ∞ 6 4

2 2 6 ∞ 6

3 3 4 6 ∞

Subtract the smallest cost element in each row (column) from all the elements of the
corresponding row (column) we get:
∞ 1 3 0 1

1 ∞ 2 0 1

2 1 ∞ 2 0

0 0 3 ∞ 4

0 0 0 3 ∞

Now make assignment in rows and columns having single zeros

∞ 1 3 (0) 1

1 ∞ 2 0 1

2 1 ∞ 2 (0)

0 (0) 3 ∞ 4

0 0 (0) 3 ∞

Since some rows and columns are without assignments, the current assignment is not optimal. Cover all the
zeros by drawing minimum number of straight lines.

∞ (1) 3 0 1

1 ∞ 2 0 1

2 1 ∞ 2 0
 CS6704-Resource Management Techniques Department of CSE 2016-2017
0 0 3 ∞ 4

0 0 0 3 ∞

The optimum assignment is given by A  D, B  C, C  E, D  B, E  A

ADBCEA
UNIT III INTEGER PROGRAMMING
Cutting plan algorithm – Branch and bound methods, Multistage (Dynamic) programming.
UNIT-III/ PART-A
1 What is integer programming?
A linear programming problem in which some or all of the variables in the optimal solution are restricted to
assume non-negative integer (or discrete) values is called an Integer Programming Problem (IPP) or Integer
Linear Programming (ILP).
2 State the general form of an integer programming problem?
The general Integer Programming Problem is given by Maximize Z = CX, Subject to the constraints AX ≤ b,
X ≥ 0 and some or all variables are integers.
3 Define All IPP and Mixed IPP.
All (pure) IPP: In a linear programming problem, if all the variables in the optimal solution are restricted to
assume non-negative integer values, then it is called the pure (all) IPP.

Mixed IPP: In a linear programming problem, if only some of the variables in the optimal solution are
restricted to assume non-negative integer values, while the remaining variables are free to take any non-
negative values, then it is called mixed IPP.
4 List out some of the applications of IPP?
 IPP occur quite frequently in business and industry.
 All transportation, assignment and travelling salesman problems are IPP, since the decision
variables are either zero or one.
 All sequencing and routing decisions are IPP as it requires the integer values of the decision
variables.
 Capital budgeting and production scheduling problem are IPP. In fact, any situation involving
decisions of the type either to do a job or not to do can be treated as an IPP.
 All allocation problems involving the allocation of goods, men, machines, give rise to IPP since such
commodities can be assigned only integer and not fractional values.
5 What is the importance of Integer Programming?
In linear programming problem, all the decision variables allowed to take any non-negative real values, as it is
quite possible and appropriate to have fractional values in many situations. In many situations, like business
and industry, these decision variables make sense only if they have integer values in the optimal solution.
Hence a new procedure has been developed in this direction for the case of LPP subjected to additional
restriction that the decision variables must have integer values.
6 What is the strategy used in integer programming algorithms?
The ILP algorithms are based on exploiting the computational success of LP. The strategy of these algorithms
involves three steps.
Step 1: Relax the solution space of ILP by deleting the integer restriction on all integer variables and replacing
any binary variable y with the continuous range 0 ≤ y 1. The result of the relaxation is a regular LP.
Step 2: Solve the LP and identify its continuous optimum.
Step 3: Starting from the continuous optimum point, add special constraints that iteratively modify the LP
solution space in a manner that will eventually render an optimum extreme point satisfying the integer
requirements.
7 What are the methods for generating the special constraints in ILP?
Two general methods have been developed generating the special constraints:
 Branch and Bound (B&B) Method: It starts with the continuous optimum, but systematically
 CS6704-Resource Management Techniques Department of CSE 2016-2017
partitions the solution space into sub problems that eliminate parts that contain no feasible integer
solution.

 Cutting Plane Method: This method solves the IPP as ordinary LPP by ignoring the integer
restriction and then introducing additional constraints one after the other to cut (eliminate)
certain part of the solution space until an integral solution is obtained.
8 Write an algorithm for Gomory’s Fractional Cut algorithm?
Step 1: Convert the minimization IPP into an equivalent maximization IPP and all the coefficients and
constraints should be integers.
Step 2: Find the optimum solution of the resulting maximization LPP by using simplex method.
Step 3: Test the integrity of the optimum solution.
Step 4: Rewrite each XBi
Step 5: Express each of the negative fractions if any, in the kth row of the optimum simplex table as the sum of
a negative integer and a non-negative fraction.
Step 6: Find the fractional cut constraint.
Step 7: Add fractional cut constraint at the bottom of optimum simplex table obtained in step 2.
Step 8: Go to step 3 and repeat the procedure until an optimum integer solution is obtained.
9 Mention some suggestions that are helpful in computation in ILP.
The most important factor affecting computation in ILP is the number of integer variables and the feasible
range in which they apply. It may be advantageous to reduce the number of integer variables in the ILP
model as much as possible. The following suggestions may provide helpful:
 Approximate the integer variables by continuous ones whenever possible.
 For the integer variables, restrict their feasible ranges as much as possible.
 Avoid the use of nonlinearity in the model.
10 What is a fractional cut?
In the cutting plane method, the fractional cut constraints cut the unused area of the feasible region in the
graphical solution of the problem. i.e. cut that area which has no integer-valued feasible solution. Thus these
constraints eliminate all the non-integral solutions without loosing any integer-valued solution. A desired cut
which represents a necessary condition for obtaining an integer solution is referred to as the fractional cut
because all its coefficients are fractions.
11 What is mixed integer problem?
In the mixed integer programming problem only some of the variables are integer constrained, while other
variables may take integer or other real values. The problem is first solved as a continuous LPP by ignoring the
integer condition. If the values of the integer constrained variables are integers then the current solution is an
optimal solution to the given mixed IPP. Otherwise select the source row which corresponds to the largest
fractional part fk among those basic variables which are constrained to be integers. Then construct Gomorian
constraint from the source row.
12 What is dynamic programming?
Dynamic programming is the mathematical technique of optimization using multistage decision process. It is
a process in which a sequence of interrelated decisions has to be made. It provides a systematic procedure for
determining the combination of decisions which maximize overall effectiveness.
13 Specify the need for dynamic programming.
Decision making process consists of selecting a combination of plans from a large number of alternative
combinations. This involves lot of computational work and time. Dynamic programming deals with such
situations by dividing the given problem into sub problems or stages. Only one stage is considered at a time
and the various infeasible combinations are eliminated with the objective of reducing the volume of
computations. The solution is obtained by moving from one stage to the next and is completed when the final
stage is reached.
14 What is forward and backward recursion?
Forward recursion is one in which the computation proceed from stage 1 to stage n. The same can be solved
by backward recursion starting at stage n and ending at stage 1. Both yield the same solution. Although
forward appears more logical, DP invariably uses backward recursion. The reason for this preference is that, in
general backward recursion may be more efficient computationally.
15 State Bellman’s principle of optimality.
 CS6704-Resource Management Techniques Department of CSE 2016-2017
An optimal policy (set of decisions) has the property that whatever be the initial state and initial decisions, the
remaining decisions must constitute an optimal policy for the state resulting from the first decision.
16 List some characteristics of dynamic programming problems.
The characteristics of dynamic programming problems may be outlined as:
 Each problem can be divided into stages, with a policy decision required at each stage.
 Each stage has number of states associated with it.
 The effect of the policy decision at each stage is to transform the current state into a state associated
with the next stage.
 The current state of the system is described by state variables.
17 State Markovian property.
For dynamic programming problem, in general, the knowledge of the current state of the system conveys all
of the information about its previous behaviour necessary for determining the optimal policy hence forth.
This is the Markovian property.
18 Define stage. A stage may be defined as the portion of the problem that possesses a set of mutually exclusive
alternatives from which the best alternative is to be selected. If we are to take six sequential decisions then we
have six stages.
19 Mention some applications of dynamic programming.
 Used for production, scheduling and employment smoothening problems.
 Used to determine the inventory level and formulating inventory recording.
 Applied for allocating scarce resource to different alternative uses.
 Used to determine the optimal combination of advertising media.
 Applied in replacement theory to determine at which age the equipment is to be replaced for optimal
return from the facility.
20 What are the steps involved in dynamic programming algorithm?
Step 1: Identify the decision variables and specify objective function to be optimized.
Step 2: Decompose the given problem into smaller sub problems. Identify state variables a t each stage.
Step 3: Write the general recursive relationship for computing the optimal policy.
Step 4: Write the relation giving the optimal decision function for one stage sub-problem and solve it.
Step 5: Solve the optimal decision function for 2-stage, 3-stage, ....(n-1)-stage and n-stage problem.
UNIT-III / PART-B
1 Find the optimum integer solution to the following LPP. Maximize Z = x1 + x2 subject to constraints 3x1
+ 2x2 ≤ 5, x2 ≤ 2 and x1 > 0, x2 ≥ 0 and are integers.
Solution: Ignoring the integrality condition and introducing non-negative slack variables x3 and x4, the
standard form of the continuous LPP becomes
Maximize Z = x1 + x2 + 0x3 + 0x4
Subject to
3x1 + 2x2 + x3 + 0x4 = 5
0x1 + x2 + 0x3 + x4 = 2
and x1, x2, x3, x4 ≥ 0
The initial basic feasible solution is given by x3 = 5, x4 = 2, (x1 = x2 = 0, non-basic)
Initial iteration:
Cj (1 1 0 0)
CB YB XB x1 x2 x3 x4 
0 x3 5 (3) 2 1 0 5/3
0 x4 2 0 1 0 1 -
Zj - Cj 0 -1 -1 0 0
Third iteration:
Cj (1 1 0 0 0)
CB YB XB x1 x2 x3 x4 s1
1 x1 0 1 0 0 -1 1
1 x2 2 0 1 0 1 1
0 x3 1 0 0 1 1 -3
Zj - Cj 2 0 0 0 0 1
 CS6704-Resource Management Techniques Department of CSE 2016-2017
For the above LPP the feasible regions OABC, the optimum solution is
Max Z = 7/3, x1 = 1/3, x2 = 2
Since this solution is not an integer optimum solution, introduce the secondary Gomorian constant.

+ ≥

Express this in terms of x1 and x2

3x1 + 2x2 + x3 = 5  x3 = 5 – 3x1 -2x2
and x2 + x4 = 2  x4 = 2 – x2
Substituting in the Gomory constants

(5 – 3x1 -2x2) + (2 – x2) ≥

 5 – 3x1 – 2x2 + 2 – x2 ≥ 1
 -3x1 – 3x2 + 7 ≥ 1  -3x1 – 3x2 ≥ -6
 3x1 + 3x2 ≤ 6  x1 + x2 ≤ 2
The required optimal integer valued solution is
Max Z = 2, x1 = 0, x2 = 2 (or)
Max Z = 2, x1 = 1, x2 = 1
2 Using Gomory’s cutting plane method maximize Z = 2x1 + 2x2 subject to 5x1 + 3x2 ≤ 8, 2x1 + 4x2 ≤ 8 and
x1 , x2 ≥ 0 and are all integers.
Solution: Ignore the integrality condition and introducing the non-negative slack variables x3 and x4 the
standard form of the continuous LPP becomes
Maximize
Z = 2x1 + 2x2 + 0x3 + 0x4
Subject to
5x1 + 3x2 + x3 + 0x4 = 8
2x1 + 4x2 + 0x3 + x4 = 8
and x1 , x2, x3, x4 ≥ 0

The initial basic feasible solution is given by

x3 =8, x4 = 8 (basic)
(x1 = x2 = 0, non-basic)

Initial iteration:
Cj (2 2 0 0)
CB YB XB x1 x2 x3 x4 
0 x3 8 (5) 3 1 0 8/5*
0 x4 8 2 4 0 1 8/2
Zj - Cj 0 -2 -2 0 0

First Iteration: Introduce x1 drop x3

Cj (2 2 0 0)
CB YB XB x1 x2 x3 x4 
2 x1 8/5 1 3/5 1/5 0 8/3
24/14=
0 x4 24/5 0 (14/5) -2/5 1
12/7*
Zj - Cj 16/5 0 -4/5 2/5 0

Fourth iteration: Introduce x4 and drop s2

Cj (2 2 0 0 0 0)
CB YB XB x1 x2 x3 x4 s1 s2
2 x1 1 1 0 0 0 1 -4/5
2 x2 1 0 1 0 0 -1 1
 CS6704-Resource Management Techniques Department of CSE 2016-2017
0 x3 0 0 0 1 0 -2 1
0 x4 2 0 0 0 1 2 -12/5
Zj - Cj 4 0 0 0 0 0 2/5

The optimal solution to the new problem is

Max Z = 4, x1 = 1, x2 = 1
3 Solve the following IPP. Minimize Z= -2x1 - 3x2 subject to 2x1 + 2x2 ≤ 7, x1 ≤ 2, x2 ≤ 2 and x1 , x2 ≥ 0 and
are integers.
Solution: Given IPP is
Minimize Z= -2x1 - 3x2
Subject to
2x1 + 2x2 ≤ 7
x1 ≤ 2
x2 ≤ 2
x1 , x2 ≥ 0 and integers.

Maximize
Z*= 2x1 + 3x2
Subject to
2x1 + 2x2 ≤ 7
x1 ≤ 2
x2 ≤ 2
x1 , x2 ≥ 0 and integers.

Ignoring the integrality condition and introducing the non-negative slack variables x3, x4, x5 the standard form
of the continuous LPP becomes.

Maximize
Z*= 2x1 + 3x2 +0x3 + 0x4 + 0x5

Subject to
2x1 + 2x2 + x3 + 0x4 + 0x5 = 7
x1 + 0x2 + 0x3 + x4 + 0x5 = 2
0x1 + x2 + 0x3 + 0x4 + x5 = 2
x1 , x2, x3, x4, x5 ≥ 0 and integers.

The basic feasible solution is given by x3=7, x4=2, x5=2 (basic)

(x1=x2=0, non-basic)

Initial iteration:

Cj (2 3 0 0 0)
CB YB XB x1 x2 x3 x4 x5 
0 x3 7 2 2 1 0 0 7/2
0 x4 2 1 0 0 1 0 -
0 x5 2 0 (1) 0 0 1 2*
Zj*- Cj 0 -2 -3 0 0 0

First Iteration: Introduce x2 and drop x5

Cj (2 3 0 0 0)
CB YB XB x1 x2 x3 x4 x5 
0 x3 3 (2) 0 1 0 -2 3/2*
 CS6704-Resource Management Techniques Department of CSE 2016-2017
0 x4 2 1 0 0 1 0 2
3 x2 2 0 1 0 0 1 -
Zj*- Cj 6 -2 0 0 0 3

Third iteration: Drop s1 and introduce x3

Cj (2 3 0 0 0 0)
CB YB XB x1 x2 x3 x4 x5 s1
2 x1 1 1 0 0 0 -1 1
0 x4 1 0 0 0 1 1 -1
3 x2 2 0 1 0 0 1 0
0 x3 1 0 0 1 0 0 -2
Zj*- Cj 8 0 0 0 0 1 2

The optimal integer solution is

Max Z* = 8, x1=1, x2=2
Min Z = -Max(-Z) = -Max Z* = -8
Min Z = -8, x1=1, x2=2
4 Solve the following mixed integer programming problem by Gomory’s cutting plane algorithm:
Maximize Z = x1 + x2 subject to 3x1 + 2x2 ≤ 5, x2 ≤ 2 and x1 , x2 ≥ 0 and x1 an integers.
Solution: Ignoring the integrality condition and introducing the non-negative slack variables x3, x4, the
standard form of the continuous LPP becomes
Maximize
Z = x1 + x2 + 0x3 + 0x4
Subject to
3x1 + 2x2 + x3 + 0x4 = 5
0x1 + x2 + 0x3 + x4 = 2
and x1 , x2 , x3, x4 ≥ 0

The initial basic feasible solution is given by

x3=5, x4=2 (basic) (x1=x2=0, non-basic)

Initial iteration:
Cj (1 1 0 0)
CB YB XB x1 x2 x3 x4 
0 x3 5 (3) 2 1 0 5/3
0 x4 2 0 1 0 1 -
Zj - Cj 0 -1 -1 0 0

First iteration: Introduce x1 and drop x3

Cj (1 1 0 0)
CB YB XB x1 x2 x3 x4 
1 x1 5/3 1 2/3 1/3 0 5/2
0 x4 2 0 (1) 0 1 2
Zj - Cj 5/3 0 -1/3 1/3 0

Third iteration: Drop s1 and introduce x3

Cj (1 1 0 0 0)
CB YB XB x1 x2 x3 x4 
1 x1 0 1 0 0 -1 1
1 x2 2 0 1 0 1 0
0 x3 1 0 0 1 1 -3
 CS6704-Resource Management Techniques Department of CSE 2016-2017
Zj - Cj 2 0 0 0 0 1

The required solution is

Max Z = 2, x1=0, x2=2
5 Solve the following mixed integer programming problem minimize Z = x1 - 3x2 subject to x1 + x2 ≤ 5, -2x1
+ 4x2 ≤ 11 and x1 , x2 ≥ 0 and x2 is an integer.
Solution: Given mixed IPP be
Minimize
Z = x1 - 3x2
Subject to
x1 + x2 ≤ 5
-2x1 + 4x2 ≤ 11
x1 , x2 ≥ 0 and x2 is an integer.
Maximize
Z* = -x1 + 3x2
Subject to
x1 + x2 ≤ 5
-2x1 + 4x2 ≤ 11
x1 , x2 ≥ 0 and x2 is an integer.
Ignoring the integrality condition and introducing the non-negative slack variables x3 and x4, the standard
form of the continuous LPP becomes
Maximize
Z* = -x1 + 3x2 + 0x3 + 0x4
Subject to
x1 + x2 + x3 + 0x4 = 5
-2x1 + 4x2 +0x3 + x4 = 11
x1 , x2 , x3, x4

The initial basic feasible solution is given by

x3=5, x4=11 (basic)
(x1=x2=0, non-basic)

Initial iteration:
Cj (-1 3 0 0)
CB YB XB x1 x2 x3 x4 
0 x3 5 1 1 1 0 5/1
0 x4 11 -2 (4) 0 1 11/4
Zj - Cj 0 1 -3 0 0

First Iteration: Introduce x2 and drop x4

Initial iteration:
Cj (-1 3 0 0)
CB YB XB x1 x2 x3 x4 
0 x3 9/4 (3/2) 0 1 -1/4 3/2
3 x2 11/4 -1/2 1 0 1/4 -
Zj - Cj 33/4 -1/2 0 0 3/4

Third iteration: Drop s1 and introduce x3

Cj (-1 3 0 0 0)
CB YB XB x1 x2 x3 x4 s1
-1 x1 1/2 1 0 0 -1/2 2
3 x2 3 0 1 0 0 1
0 x3 3/2 0 0 1 1/2 -3
 CS6704-Resource Management Techniques Department of CSE 2016-2017
Zj - Cj 17/2 0 0 0 1/2 1

The optimal integer solution is

Max Z* = 17/2, x1=1/2, x2=3
Min Z = -Max Z* = -17/2
Min Z = -17/2, x1=1/2, x2=3
6 Use branch and bound technique to solve the following: Maximize Z = x1 + 4x2 subject to constraints 2x1
+ 4x2 ≤ 7, 5x1 + 3x2 ≤ 15 and x1 , x2 ≥ 0 and integers.
Solution: Ignoring the integrality condition the continuous LPP becomes
Maximize
Z = x1 + 4x2
Subject to
2x1 + 4x2 ≤ 7
5x1 + 3x2 ≤ 15
and x1 , x2 ≥ 0 and integers.

By using the graphical method the solution space in given by the region OABC. The optimum solution of this
problem is
Max Z = 7, x1=0, x2=7/2

Since x2=7/2, this problem should be branched into two problems.

Sub-problem (1)
Maximize
Z = x1 + 4x2
Subject to
2x1 + 4x2 ≤ 7
5x1 + 3x2 ≤ 15
x2 ≤ 1
and x1 , x2 ≥ 0
Its solution space is given by the region OABDE and its optimal solution is
Max Z = 11/2, x1=3/2, x2=1

Since x1=3/2, this sub-problem is branched again.

Sub-problem (2)
Maximize
Z = x1 + 4x2
Subject to
2x1 + 4x2 ≤ 7
5x1 + 3x2 ≤ 15
x2 ≥ 2
and x1 , x2 ≥ 0
Its solution space is given by region OABC and FGH and it has no feasible solution.

Sub-problem (3)
Maximize
Z = x1 + 4x2
Subject to
2x1 + 4x2 ≤ 7
5x1 + 3x2 ≤ 15
x2 ≤ 1
x1 ≤ 1
and x1 , x2 ≥ 0
Its solution space is given and its optimal solution is
Max Z = 5, x1=1, x2=1
 CS6704-Resource Management Techniques Department of CSE 2016-2017
7 Use branch and bound to solve the following IPP. Max Z = 3x1 + 4x2 subject to 7x1 + 16x2 ≤ 52, 3x1 - 2x2 ≤
18 and x1 , x2 ≥ 0 and are all integers.
Solution: Ignoring the integrality condition and introducing the non-negative slack variables x3, x4, the
standard form of the continuous LPP becomes
Max
Z = 3x1 + 4x2 + 0x3 + 0x4
Subject to
7x1 + 16x2 + x3 + 0x4 = 52
3x1 - 2x2 + 0x3 + x4 = 18
x1 , x2 , x3, x4 ≥ 0
The initial basic feasible solution is given by x3 = 52, x4 = 18 (basic)
(x1= x2= 0, non-basic)
Initial Iteration:
Cj (3 4 0 0)
CB YB XB x1 x2 x3 x4 
0 x3 52 7 (16) 1 0 52/16
-
0 x4 18 3 -2 0 1
Zj - Cj 0 -3 -4 0 0

First Iteration: Introduce x2 and drop x3

Cj (3 4 0 0)
CB YB XB x1 x2 x3 x4 
4 x2 13/4 7/16 1 1/16 0 52/7
0 x4 49/2 (31/8) 0 1/8 1 196/31*
Zj - Cj 13 -5/4 0 1/4 0

Second Iteration: Introduce x1 and drop x4

Cj (3 4 0 0)
CB YB XB x1 x2 x3 x4
4 x2 15/31 0 1 -1/12 7/62
3 x1 196/31 1 0 1/3 8/31
Zj - Cj 648/31 0 0 2/3 19/43

Since all (Zj –Cj)  0, the current basic solution is optimal, but non-integer.
x2 = 15/31 = 0 + 15/31
x1 = 196/31 = 6 + 10/31

The maximum fractional part is 15/31 which corresponds to x2

0 < x2 < 1 ⇒ x2  0 (or) x2  1

Applying these two conditions separately in the continuous LPP, the two sub-problems are:
Sub-problem 1
Max
Z = 3x1 + 4x2
Subject to
7x1 + 16x2 ≤ 52
3x1 - 2x2 ≤ 18
x2 ≤ 0
x1 , x2 ≥ 0
The optimal solution is given by Max Z= 18, x1=6, x2=0. Since the solution is in integers this sub-problem is
fathomed. The lower limit of the objective function is 18.

Sub-problem 2
 CS6704-Resource Management Techniques Department of CSE 2016-2017
Max
Z = 3x1 + 4x2
Subject to
7x1 + 16x2 ≤ 52
3x1 - 2x2 ≤ 18
x2 ≥ 1
x1 , x2 ≥ 0
The optimal solution is given by Max Z=19.43, x1=5.14, x2=1
Also in sub-problem (2) since x1=5.14
 5 < x1 < 6
 x1 ≤ 5 or x1 ≥ 6

Applying these two conditions separately in sub-problem (2), solve sub-problem (3) and sub-problem (4).
Sub-problem 3
The optimal solution is given by Max Z= 19.25, x1 = 5, x2 = 1.063
Sub-problem 4
This sub-problem has no feasible solution. So this sub-problem is fathomed.
From the available integer optimal solutions, the best optimal solution is Max Z=19, x1=5, x2=1, which is the
required optimal solution.
8 Solve the following by dynamic programming. Min Z = y1 + y2 + ......+ yn subject to constraints y1 y2...... yn
= b and y1 y2...... yn ≥ 0 (or) factorize a positive quantity b into n factors in such a way so that their sum
is a minimum.
Solution: To develop the recursive equation:
Let fn(b) be the minimum attainable sum y1 + y2 + ......+ yn when the positive quantity b is factorized into n
factors y1, y2, …, yn.

For n=1 (One stage problem)

Here b is factorized into one factor y1=b only

f1(b) = {y1}=b (trivial case)

For n=2 (Two stage problem)

Here b is factorized into two factors y1=x and y2=b/x so that y1.y2=b

f2(b) = {y1+y2}

= {x + }

= {x + f1 (b/x)}

For n=3 (Three stage problem)

Here b is factorized into three factors y1=x and y2.y3=b/x so that y1.y2.y3=b

f3(b) = {y1+y2+y3}

= {x + f2 (b/x)}

In general the recursive equation for the n-stage problem is

 CS6704-Resource Management Techniques Department of CSE 2016-2017

fn(b) = {x + fn-1 (b/x)}

9 By dynamic programming technique, solve the problem. Min Z = x12 + x22 + x32 subject to constraints x1 +
x2 + x3 ≥ 15 and x1 , x2, x3 ≥ 0.
Solution: To develop the recursive equation
It is a three stage problem. The decision variables are x1, x2, x3 and the state variables are s1, s2 and s3 are
defined as
s3 = x1+x2+x3 ≥15
s2 = x1+x2 = s3-x3
s1 = x1 = s2-x2

Let fi(si) be the minimum value of Z at the i-th stage where

Si = x1+x2+….+xi (i=1,2,3)

Now the recursive equations are

f1(s1) = {x12} = s12 = (s2-x2)2 → (1)

f2(s2) = {x12 + x22}

= {x22 + f1(s1)} → (2)

f3(s3) = {x12 + x22 + x32}

= {x32 + f2(s2)} → (2)

To solve the recursive equations:

From (1)
f1(s1) = (s2-x2)2
From (2)

f2(s2) = {x22 + f1(s1)}

= {x22 + f1(s2-x2)}

= {2x22 – 2s2x2 + s22}

The function {2x22 – 2s2x2 + s22} will attain minimum when x2 = s2/2

f2(s2) = {2x22 – 2s2x2 + s22}

= 2(s2/2)2 -2s2(s2/2) + s22 = s22/2

From (3)

f3(s3) = {x32 + f2(s2)}

 CS6704-Resource Management Techniques Department of CSE 2016-2017

= {x32 + f2(s3-x3)}

= {x32 + (s3-x3)2/2}

The function x32 + (s3-x3)2/2 will attain its minimum when x3 = s3/3

f3(s3) = {x32 + (s3-x3)2/2}

= (s3/3)2 + (s3-s3/3)2/2
= s32/3

But s3  15, i.e., minimum s3 =15

Therefore Z is minimum when x3 = s3/3 = 15/3 = 5

But s2 = s3 – x3 = 15 – 5 = 10
Therefore x2 = s2/2 = 10/2 = 5

Also s1 = s2 –x2 = 10 -5 =5
Therefore x1 = s1 = 5

f3(s3) = s32/3 = 152/3 =75 = f3(15)

Therefore the optimal policy is (5, 5, 5) and Min Z = 75 = f3(15).

10 Solve the following LPP using dynamic programming principles. Max Z = 2x 1 + 5x2 subject to 2x1 + x2 ≤
43, 2x2 ≤ 46 and x1 , x2 ≥ 0.
Solution: The problem consists of two decision variables and two resources (constraints). Hence the problem
has two stages and two state variables: Let (B1j, B2j) be the state of the system at stage j (j=1,2) and
fj(B1j,B2j) be the optimal (maximum) value of the objective function for stage j=1,2 given the state (B1j,B2j)

Using backward computation procedure

F2(B12,B22)= {5x2} =

Since Max{x2} which satisfies 0≤x2≤B12, 0≤x2≤B22/2 is the minimum of B12, B22/2
x2*=Max{x2} = Min{B12,B22/2} → (1)
f2 (B12,B22) = 5Min{B12,B22/2} → (2)

Also f1 (B11,B21) = {2x1+f2(B11-2x1,B21-0)}

From (2)
F2(B11-2x1,B21)= 5 Min {B11-2x1,B21/2}

F1(B11.B21) = {2x1+ 5 Min(B11-2x1,B21/2)}

Since it is a two stage problem, at the first stage B11=43, b21=46

F1(B11,B21) = f1(43,46)

`` = {2x1+ 5 Min(43-2x1, 46/2)}

 CS6704-Resource Management Techniques Department of CSE 2016-2017

= {2x1+ 5 Min(43-2x1, 23)} → (3)

From (3)

F1(B11,B21) = Max {2x1 + 5 }

= Max {

Since Max of 2x1 +115, 0≤x1≤10 occurs at x1 = 10 and

Max of 215 -8x1, 10≤x1≤43/2 also occurs at x1 =10
F1(B11,B21) = 2(10) +115 = 135
Now
x2 = Min {B11-2x1,B21/2}
= Min {43-2x1,23}
= Min {23,23} = 23
The optimal solution is Max Z = 135, x1=10, x2=23
UNIT IV CLASSICAL OPTIMISATION THEORY
Unconstrained external problems, Newton – Ralphson method – Equality constraints – Jacobean methods –
Lagrangian method – Kuhn – Tucker conditions – Simple problems.
UNIT-IV / PART-A
1 What is nonlinear programming?
Nonlinear programming is an extension of linear programming. In real life problems the objective function
may be nonlinear but the set of constraints may be linear or nonlinear. Such problems are called as nonlinear
problems. A set of nonlinear programming problems include:
 Nonlinear programming problems of general nature.
 Quadratic programming problems.
 Separable programming problems
2 How do classical optimization problems determine points of maxima and minima?
Classical optimization theory uses differential calculus to determine points of maxima and minima extrema) for
unconstrained and constrained functions. The methods may not be suitable for efficient numerical
computations, but the underlying theory provides the basis for most nonlinear programming algorithms.
3 What is an unconstraint problem?
An extreme point of a function f(X) defines either maximum or minimum of the function. Mathematically, a
point X0= (x10,….., xj0,……., xn0) is a maximum if f(X0 + h) ≤ f (X0) for all h = (h1,...., hj,…., hn) and |hj| is
sufficiently small for all j.
4 What is the necessary condition for an n variable function to have extrema?
A necessary condition for X0 to be an extreme point of f(x) is that f(X0) = 0.
5 What is the sufficient condition for a function to have extrema?
A sufficient condition for a stationary point X 0 to be an extremum is that Hessian matrix H evaluated at X 0
satisfy the following conditions:
 H is positive definite if X0 is minimum point.
 H is negative definite if X0 is maximum point.
6 What are the various aspects of non-linear optimization?
 Unconstrained extremum points.
 Constrained problems with equality constraints – Lagrangean method.
 Constrained problems with inequalities – Kuhn Tucker conditions.
 Quadratic programming.
7 Define Newton-Raphson method
The Newton-Raphson method is an iterative procedure for solving simultaneous nonlinear equations. It is part
of gradient methods for optimizing unconstrained functions numerically.
 CS6704-Resource Management Techniques Department of CSE 2016-2017
8 List the types of constrained problems.
There are 2 types of constrained problem
 Equality constraints
 Inequality constraints
9 List the methods to solve equality constraints
Equality constraints present two methods.
 Jacobian method
 Lagrangean method
The Lagrangean method can be developed logically from the Jacobian method. It provides an economic
interpretation of Lagrangean method.
10 What is the mathematical form of Lagrangean method?
The form for nonlinear programming: Maximize or minimize Z = f(X 1, X2, ….., Xj,……, Xn) subject to Gi(X1,
X2, ….., Xj,……, Xn) = bi, i = 1, 2, ….., m, Xj  0, j = 1, 2, ….., n.
11 Define Lagrangean method.
The Lagrangean method for identifying the stationary points of optimization problem with equality constraints
are defined by the procedure which is developed formally as L(X, ) = f (X) - g(X). The function L is called
the Lagrangean function and the parameters  the Lagrange multipliers.
12 Mention the steps involved in Lagrangean method.
Step 1: Form the Lagrangean function.
Step 2: The first partial derivative of L with respect to X j is obtained, where j varies from 1 to n, and also with
respect to i, where i varies from 1 to m. then equate them to 0.
Step 3: Solution to equations in step 2 are found.
Step 4: The bordered Hessian square matrix [HB] of size n + m is formed.
Step 5: The stationary points (X1*, X2*, ….., Xj*,……, Xn*) are tested for maximization/minimization objective
function.
13 What is the condition to be checked for maximization type objective function?
The stationary point will be given the maximum objective function value if the sign of each of the last (n – m)
principal minor determinants of the bordered Hessian matrix is the same as that of (-1) m+1, ending with the
(2m+1)th principal minor determinant.
14 What is the condition to be checked for minimization type objective function?
The stationary point will be given the minimum objective function value if the sign of each of the last (n – m)
principal minor determinants of the bordered Hessian matrix is the same as that of (-1) m, ending with the
(2m+1)th principal minor determinant.
15 What is the mathematical form of constrained optimization problem with inequality constraints?
The inequality constraints can be converted into equations and use the Lagrangean method. The problem is of
the form: Maximize Z = f(X) subject to g(X)  0, where X = (x1, x2, …., xn) and g = (g1, g2, …., gm)T. The
functions f(X) and gi(Xi) are assumed to be twice differentiable.
16 Define Jacobian matrix
The Jacobian matrix is the matrix of all first order partial derivatives of a vector valued function. When the
matrix is a square matrix, both the matrix and its determinant are referred to as Jacobian matrix. J m x m is
called the Jacobian matrix. The Jacobian J is assumed nonsingular. This is possible because m equations are
independent by definition.
17 Mention about sensitivity analysis in the Jacobian method.
The Jacobian method can be used to study the effect of small changes in the right hand side of the constraints
on the optimal value of f. This type of investigation is called sensitivity analysis and is similar to what is
carried out in linear programming.
18 What is sensitive coefficient?
The effect of the small change g on the optimum value of f can be studied by evaluating the rate of change of
f with respect to g. These rates are usually referred to as sensitivity coefficients.
19 Define KKT
The Karush–Kuhn–Tucker (KKT) conditions (also known as the Kuhn–Tucker conditions) are first order
necessary conditions for a solution in nonlinear programming to be optimal, provided that some regularity
 CS6704-Resource Management Techniques Department of CSE 2016-2017
conditions are satisfied. Allowing inequality constraints, the KKT approach to nonlinear programming
generalizes the method of Lagrange multipliers, which allows only equality constraints. The system of
equations corresponding to the KKT conditions is usually not solved directly, except in the few special cases
where a closed-form solution can be derived analytically.
20 What is the mathematical form of Kuhn Tucker conditions?
The form for nonlinear programming which is having a maximization objective function with all less than or
equal to type constraints: Maximize Z = f(X 1, X2, ….., Xj,……, Xn) subject to Gi(X1, X2, ….., Xj,……, Xn)
 bi, i = 1, 2, ….., m, Xj  0, j = 1, 2, ….., n.
UNIT-IV / PART-B
1 Explain Classical optimization theory concept in detail.
2 What is unconstrained external problem. Explain in detail.
3 Explain in detail about Newton-Ralphson method in detail.
4 By using the Newton-Raphson’s method find the positive root of the quadratic equation 5x2 + 11x – 17 = 0
correct to 3 significant figures.
5 What is Equality constraints.
6 Explain Jacobean Method in detail.
7 Explain Lagrangian method with example.
8 Use the Lagrangian method to maximize the function f (x; y) = xy subject to the constraint x + 2y ≤200
9 Explain Kuhn-Tucker conditions with example.
10 Explain the Kuhn-Tucker conditions and solve them in the following Kuhn-Tucker problems
a) Max f(x,y)= xy subject to x+4Y ≤ 16
b) Max f(x,y)=x2y subject to 2x2+y2 ≤ 3
c) Max f(x,y,z)=xyz subject to x2 + y2 ≤ 1 and x + z ≥ 1.
UNIT V OBJECT SCHEDULING
Network diagram representation – Critical path method – Time charts and resource leveling – PERT.
UNIT-V / PART-A
1 What is network?
A network is defined as a graphical representation of a project’s operation and is composed of all the events
and activities in sequence along with their inter relationship and inter dependencies. Project networks can be
used to plan, record, and document a schedule and to track actual results against the schedule.
2 What is Event in a network diagram?
Each activity in the project network begins and ends with an event (a circle or a bubble). The events are
numbered (the numbers do not necessarily imply sequence), and a given activity is identified by the numbers
associated with its beginning and ending events.
3 What are the three main phases of project?
Planning – This phase involves a listing of tasks or jobs that must be performed to complete a project under
considerations. Scheduling – This phase involves the laying out of the actual activities of the projects in a
logical sequence of time in which they have to be performed.
Control – This phase consists of reviewing the progress of the project whether the actual performance is
according to the planned schedule and finding the reasons for difference, if any, between the schedule and
performance.
4 What are the two basic planning and controlling techniques in a network analysis?
Critical Path Method (CPM) - This is the unit cost that is only for your reference when comparing one plan to
another. Programme Evaluation and Review Technique (PERT) – It takes into account the uncertainty project
durations into account.
5 Define activity & Critical Activities?
An activity is a task or an item of work to be done in a project. It is the element of the project and it may be a
process, material handling, procurement cycle etc. In a Network diagram all the activities in any critical path
are called critical activities.
6 Define non critical activities?
Activities which have a provision such that the event if they consume a specified time over and above the
estimated time the project will not be delayed are termed as non-critical activities.
 CS6704-Resource Management Techniques Department of CSE 2016-2017
7 Define Dummy Activities & duration?
Dummy activities - When two activities start at a same time, the head event are joined by a dotted arrow
known as dummy activity which may be critical or non-critical. Duration - It is the estimated or the actual time
required to complete a trade or an activity.
8 Define total project time & Critical path?
Total project time is the time taken to complete a project and just found from the sequence of critical activities.
In other words it is the duration of the critical path. Critical path is the sequence of activities which decides the
total project duration. It is formed by critical activities and consumes maximum resources and time.
9 Define Optimistic?
Optimistic time estimate or least time estimate (t o or a) is the duration of any activity when everything goes on
very well during the project. i.e labours are available and come in time, machines are working properly, money
is available whenever needed, there is no scarcity of raw material needed etc.,
10 Define Pessimistic?
Pessimistic time estimate or greatest time estimate is denoted by t p or b is the duration of any activity when
almost everything goes against our will and a lot of difficulties are faced while doing a project.
11 Define most likely time estimation?
Most likely time estimate is the duration of any activity when sometimes thing go on very well, sometimes
things go on very bad while doing the project. Two assumptions are, (i) the activity durations are independent.
It follows the β – distribution.
12 What is a parallel critical path?
When critical activities are crashed and the duration is reduced other paths may also become critical such
critical paths are called as parallel critical path. When there is more than one critical path in a network, project
duration can be reduced only when the duration of critical activity is reduced or the duration of different
activities on different critical paths are reduced
13 What is standard deviation and variance in PERT network?
The expected time of an activity in actual execution is not completely reliable and is likely to vary. If the
variability is known we can measure the reliability of the expected time as determined from three estimates.
The measure of the variability of possible activity time is given by standard deviation, their probability
distribution. Variance of the activity is the square of the standard deviation
14 What is meant by resource analysis?
Resources are required to carry out the project tasks. They can be equipment, facilities, funding which are
required for the completion of a project activity. The lack of resource will therefore be a constraint on the
completion of a project activity. Resource scheduling, availability and optimization are considered key to
successful project management.
15 What are the three time estimates used in the context of PERT? How are the expected duration of an
activity and its standard deviation calculated?
Optimistic time estimate or least time estimate (to or a)
Pessimistic time estimate or greatest time estimate (tp or b)
Most likely time estimate (tm or b)
Expected Duration = (te+4tm+tp)/6
Standard deviation = (tp-to)/6
16 Explain the term: Resource smoothing.
Resource smoothing means the uniform distribution of resource allocate on over the project. That is scheduling
of activities, within the limits of their float, so that fluctuations in individual resource requirements are
minimized.
17 Define a dummy arrow used in a network and state two purposes for which it is used.
Dummy activity is a hypothetical activity which requires zero time and zero resources for completion. Dummy
arrow represents an activity with zero duration. It is represented by dotted line and is inserted in the network to
clarify activity pattern under the following situations:
i. It is created to make activities with common starting and finishing events distinguishable, and
ii. To identify and maintain the proper precedence relationship between activities those are not connected
by events.
18 What are the advantages of CPM and PERT techniques?
 CS6704-Resource Management Techniques Department of CSE 2016-2017
It encourages a logical discipline in planning, scheduling and control of projects. It helps to effect considerable
reduction of project times and the cost. It helps better utilization of resources like men, machines, materials and
money with reference to time. It measures the effect of delays on the project and procedural changes on the
overall schedule.
19 Differentiate between PERT and CPM?
CPM-Network is built on the basis of activity, Deterministic nature, one time estimation
PERT- An event oriented network, Probabilistic nature, three time estimation (i.e optimistic, pessimistic and
most likely time).
20 Mention some advantages of PERT.
 It compels managers to plan their projects critically and analyse all factors affecting the progress of the
plan. The process of the network analysis requires that the project planning be conducted on
considerable detail from the start to the finish.
 It provides the management a tool for forecasting the impact of schedule changes and be prepared to
correct such situations. The likely trouble spots are located early enough so as to apply some preventive
measures or corrective actions.
 A lot of data can be presented in a highly ordered fashion. The task relationships are graphically
represented for easier evaluation and individuals in different locations can easily determine their role in
the total task requirements.
 The PERT time (Te) is based upon 3-way estimate and hence is the most objective time in the light of
uncertainties and results in greater degree of accuracy in time forecasting.
UNIT-V / PART-B
1 Explain the following:
a) Critical Path Method
b) PERT
2 Draw the network and determine the critical path and also calculate the floats:

Jobs 1-2 1-3 2-4 3-4 3-5 4-5 4-6 5-6

Duration
6 5 10 3 4 6 2 9
(Days)
3 Construct PERT networks and find the critical path and find all floats.

Activity 1-2 1-3 2-4 3-4 3-5 4-9 5-6 5-7 6-8 7-8 8-10 9-10

Time 4 1 1 1 6 5 4 8 1 2 5 7

4 Explain the following

a) Explain the three rime estimates used in PERT
b) What are the main assumptions under in PERT computations
c) Distinguish between PERT and CPM
5 Explain PERT procedure to determine
a) The expected critical path
b) Expected variance of the project length and
c) The probability of completing the project within a specified time.
6 The following table shows the jobs of a project with their duration in days. Draw the network and determine
the critical path. Also calculate all the floats.
 CS6704-Resource Management Techniques Department of CSE 2016-2017

Jobs 1-2 1-3 1-4 2-5 3-7 4-6 5-7 5-8

Duration 10 8 9 8 16 7 7 7

Jobs 6-7 6-9 7-10 8-10 9-10 10-11 11-12

Duration 8 5 12 10 15 8 5

7 Draw the network for the following project and compute the earliest and latest times for each event and
also find the critical path.
Activity Imme.predecessor Time (days)
1-2 - 5
1-3 - 4
2-4 1-2 6
3-4 1-3 2
4-5 2-4 1
4-6 2-4 & 3-4 7
5-7 4-5 8
6-7 4-6 4
7-8 6-7&5-7 3
8 Draw the network model of the problem and carry out the critical path computations.
Immediate
A A B,C D D D G H D
Predecessor
Duration(days) 21 14 14 3 70 70 14 1 1
Immediate
J D L D N O E,F,P Q,I,K M,R
Predecessor
Duration(days) 1 1 7 7 7 14 1 1 1
9 A small project consists of jobs as given in the table below. Each job is listed with its normal time and a
minimum or crash time (in days). The cost (in Rs.per day) for each job is also given:

Job Normal Duration (in Minimum (crash) Cost of Crashing

(i-j) days) Duration (in days) (Rs. per day)
1-2 9 5 20
1-3 8 5 25
1-4 15 10 30
2-4 5 3 10
3-4 10 6 15
4-5 2 1 40
a) What is the normal project length and the minimum project length?
b) Determine the minimum crashing cost of schedules ranging from normal length down to, and
including, the minimum length schedule. That is, if L = Length of the schedule, find the costs of
schedules which are L, L-1, L-2 and so on.
c) Overhead costs total Rs.60 per day. What is the optimum length schedule in terms of both crashing and
overhead cost? List the schedule duration of each job for your solution.
10 The following information is available.
Activity No. of Days No. of men reqd. per day
A 1-2 4 2
B 1-3 2 3
C 1-4 8 5
D 2-6 6 3
CS6704-Resource Management Techniques Department of CSE 2016-2017
E 3-5 4 2
F 5-6 1 3
G 4-6 1 8
a) Draw the network and find the critical path.
b) What is the peak requirement of Manpower? On which day(s) will this occur?
c) If the minimum labour available on any day is only 10, when can the project be completed.