Operations Analytics (Class 23-24)

© 2015 McGraw-Hill Education. All rights reserved.
1
© 2015 McGraw-Hill Education. All rights reserved. 2
Sensitivity Analysis
 Change in the value of R.H.S of a

constraints
 Change in the values of objective

function coefficient
 Addition of a new variable
 Addition of a new constraint 13

Change in the value of R.H.S of a
constraints
• Change feasible region of the LPP
• May change the values of the objective
function & optimal solution
• Entire quantity of resource at plant 2 was used
in the optimal solution
• Binding Constraint: L.H.S = R.H.S
• Quantity by which the objective function value
changes whenever the R.H.S of a constraint is
changed by one unit is called as Shadow
Price
constraints
• Implies with increase in 1 hour for plant 2
capacity, the increase in profit is INR 1.5
• It is a marginal value of one hour of plant 2 capacity
• For plant 1, out of 4 hours we could utilize only 2
hours. i.e. 2 hours are unused
• Thus adding more resources at plant 1 will not
help manager to reap more profit
• And shadow price thus will be zero
• They are called as non-binding constraint (L.H.S ≠
R.H.S)
constraints
• Implies shadow price will be non-zero only for binding
constraint
• In binding constraint, the resource is completely
used & thus additional resource is likely to change
the optimal solution & objective function value
• On the other hand, in a non-binding constraint there
is already excess resource. Additional resource
will not have any impact on the optimal solution &
thus on the objective function value
• Shadow price is the change in objective function
value for unit change in R.H.S of a constraint
constraints
• Shadow price value is valid within an interval
• As R.H.S. increase continuously, it is likely to
become non-binding constraints (excess capacity)
• Similarly when R.H.S. decreases continuously,
resource may become scarce & its marginal value is
likely to increase
• Thus, the current shadow price is valid only in an interval
of the R.H.S. value – Allowable increase & Allowable
decrease; e.g. 1.5 is valid for range: 6-18 (Plant 2)
• If R.H.S is 19, shadow price will become zero & if
R.H.S. is 5, shadow price increases to 2.5
Change in the coefficient values in
objective function
• Can change due to increase in material cost,
increase in production cost etc.
• If objective coefficient will change, it is likely to
change the optimal solution
• As long as the coefficient (profit in this case) for
product 1 or door lies between allowable
increase & allowable decrease range; the
current optimal solution will remain optimal;
i.e. X1 = 2 & x2 = 6
• This range is called as the range of optimality
Addition of a New Constraint
• May change the feasible region of

the problem & may change the
current optimal solution
• If newly added constraint is

redundant than there will be no
change in the current optimal
solution
Addition of a New Variable
• Adding a new decision

variable is likely to change
many constraints of the
problem & thus may change
the optimal solution

The Linear Programming Model
• General problem terminology and

examples
– Resources: money, particular types of
machines, vehicles, or personnel
– Activities: investing in particular projects,
advertising in particular media, or shipping
from a particular source
• Problem involves choosing levels of
activities to maximize overall measure of
performance

• Standard form

Example 4: Diet Problem
• A person wants to decide the constituents

of a diet
• To fulfil his daily requirements of proteins,
fats and carbohydrates at the minimum
cost
• Choice is to be made from four different
types of food
• The yields per unit of these foods are
given in the table
• Formulate LP model for the problem

Food Type Yield per unit Cost
per
unit
Protein Fats Carbohydrates
1. 3 2 6 45
2. 4 2 4 40
3. 8 7 7 85
4. 6 5 4 65
Minimum 800 200 700
Requirement
• Xj: Quantity of food type ‘j’ Decision Variables
Objective Function: Minimize Total Diet Cost

Minimize Z = 45 X1 + 40 X2 + 85 X3 + 65 X4
St. 3 X1 + 4 X2 + 8 X3 + 6 X4 ≥ 800
Constraints
2 X1 + 2 X2 + 7 X3 + 5 X4 ≥ 200
6 X1 + 4 X2+ 7 X3 + 4 X4 ≥ 700
Xj ≥ 0 Non-negativity Constraints

Example 5: Advertising Problem
• An advertising company wishes to plan its

advertising strategy in 3 different media:
TV, radio and magazines
• The purpose is to reach a large number of
potential customers
• Following data has been obtained from
market survey

• Formulate LP model for the problem

Strategy Types of Media
TV Radio Magazine 1 Magazine 2
Cost of an 30,000 20,000 15,000 10,000

advertising unit
Number of 2,00,000 6,00,000 1,50,000 1,00,000
Potential
Customers
Reached per unit
Number of Female 1,50,000 4,00,000 70,000 50,000
Customers
Reached per unit

• The company wants to spend not more than

Rs. 4,50,000 on advertising
• At least 1 million exposures should take place
among female customers
• Advertising on magazines must be limited to
Rs. 1,50,000
• At least 3 advertising units to be bought of
magazine 1 and 2 units of magazine 2
• The number of advertising units on TV and
radio should each be between 5 and 10
• Formulate a LP model for the problem
• Xj: Number of units bought using media type ‘j’

Decision Variables
• Objective Function: Maximize Potential
Customers Reach
Maximize Z = 2,00,000 X1 + 6,00,000 X2 +
1,50,000 X3 + 1,00,000 X4
St. 30,000 X1 + 20,000 X2 + 15,000 X3 +
10,000 X4 ≤ 4,50,000
1,50,000 X1 + 4,00,000 X2 + 70,000 X3
+ 50,000 X4 ≥ 10,00,000 Constraints
15,000 X3 + 10,000 X4 ≤ 1,50,000

X3 ≥ 3
Constraints
X4 ≥ 2
5 ≤ X1 ≤ 10
5 ≤ X2 ≤ 10
Xj ≥ 0 Non-negativity Constraints

Assumptions of Linear Programming
• Additivity
– Every function in a linear programming model
is the sum of the individual contributions of
the activities
– The contribution to the objective function

for any variable is independent of the other
decision variables
– This may not be satisfied in few cases

– Assume that a marketing team is trying to
find an optimal advertisement mix that will
have maximum reach at minimum cost
– Various channels are: TV, Radio,
Newspaper, Social media etc.
– All these channels are not independent
– If we reach 5 million with TV and 3 million
with radio we cannot say that if we use
both channels then we can reach to 8
million customers
• Divisibility
– Decision variables in a linear programming
model may have any values
• Including noninteger values
– Assumes activities can be run at fractional
values
– In such situations, one can round the optimal
decision variables up or down to the nearest
integer and get an answer that is reasonably
close to the optimal integer solution
• Proportionality assumption
– The contribution of each activity to the value of the
objective function (or left-hand side of a functional
constraint) is proportional to the level of the activity
– The contribution of any decision variable to the
objective function is proportional to its value
– If assumption does not hold, one must use
nonlinear programming techniques example:
manufacturer may provide discounts based on
quantity of purchase (contribution may follow
step function)
• Certainty
– Value assigned to each parameter of a linear
programming model is assumed to be a
known constant
– All the constant terms, objective function

and constraint coefficients as well as the
right hand sides, are know with absolute
certainty and will not change

Product Mix Problem
Maximize 𝑍 = 6 𝑥1 + 5𝑥2
Subject to
𝑥1 + 𝑥2 ≤ 5
3𝑥1 + 2𝑥2 ≤ 12
𝑥1 ≥ 0
𝑥2 ≥ 0

Steps in Graphical Solution Method

• Plot the constraints on a graph by eliminating
inequality sign and assuming arbitrary values for DV’s
• Restrict yourself to the first quadrant where both the

decision variables (𝒙𝟏, 𝒙𝟐 ) are ≥ zero to fulfil the non-
negativity requirement
• Identify the feasible region that contain the set of

points satisfying all the constraints

Steps in Graphical Solution Method

• Identify corner points
• Evaluate the objective function at all the corner

points
• The corner point that has the best value of the

objective function (max/min) is optimal

Product Mix Problem
1.) Graphical Solution Method
𝑥1 0 5
𝑥2 Constraint 1
𝑥2 5 0
(0, 5) A
𝑥1 + 𝑥2 = 5
𝑥1
(0, 0) O E
(5, 0)

Product Mix Problem
𝑥1 0 5
𝑥2 Constraint 1
𝑥2 5 0
𝑥1 0 4
Constraint 2
𝑥2 6 0
(0, 6) B 3𝑥1 + 2𝑥2 = 12

(0, 5) A
𝑥1 + 𝑥2 = 5
𝑥1
(0, 0) O D E
(4, 0) (5, 0)

Product Mix Problem
𝑥1 0 5
𝑥2 Constraint 1
𝑥2 5 0
𝑥1 0 4
Constraint 2
𝑥2 6 0
(0, 6) B 3𝑥1 + 2𝑥2 = 12

(0, 5) A
𝑭𝒆𝒂𝒔𝒊𝒃𝒍𝒆 𝑹𝒆𝒈𝒊𝒐𝒏 𝑥1 + 𝑥2 = 5
𝑥1
(0, 0) O D E
(4, 0) (5, 0)

Product Mix Problem
𝑥1 0 5
𝑥2 Constraint 1
𝑥2 5 0
𝑥1 0 4
Constraint 2
𝑥2 6 0
(0, 6) B 3𝑥1 + 2𝑥2 = 12

(0, 5) A
C (2, 3)
𝑥1
(0, 0) O D E
(4, 0) (5, 0)

Product Mix Problem
𝑥1 0 5
𝑥2 Constraint 1
𝑥2 5 0
𝑥1 0 4
Constraint 2
𝑥2 6 0
(0, 6) B 3𝑥1 + 2𝑥2 = 12

(0, 5) A
C (2, 3) 𝑶𝒑𝒕𝒊𝒎𝒖𝒎/𝑶𝒑𝒕𝒊𝒎𝒂𝒍 Solution
𝑥1
(0, 0) O D E
(4, 0) (5, 0)

Product Mix Problem
𝑥1 0 5
𝑥2 Constraint 1
𝑥2 5 0
𝑥1 0 4
Constraint 2
𝑥2 6 0
(0, 6) B 3𝑥1 + 2𝑥2 = 12

(0, 5) A
𝑥1
(0, 0) O D E
(4, 0) (5, 0)

Product Mix Problem
𝑥1 0 5
𝑥2 Constraint 1
𝑥2 5 0
𝑥1 0 4
Constraint 2
𝑥2 6 0
(0, 6) B 3𝑥1 + 2𝑥2 = 12

(0, 5) A
𝑥1
(0, 0) O D E
(4, 0) (5, 0)
𝑰𝒔𝒐 𝑷𝒓𝒐𝒇𝒊𝒕 𝑳𝒊𝒏𝒆𝒔
A LP problem involving only two decision

variables can be solved using a graphical solution
method
• Feasible Region
– It includes every solution point that satisfies
all the constraints
• Corner Points/Extreme Points
– Corner points of the feasible region is referred
to as the extreme points

• Feasible solution
– Solution for which all constraints are satisfied
– Might not exist for a given problem
• Infeasible solution
– Solution for which at least one constraint is
violated
• Optimum/Optimal solution
– Has most favorable value of objective function
– Might not exist for a given problem
• No need to evaluate all feasible solution points.

In fact, consider only the feasible solutions
that occur at the extreme points of the
feasible region.
• The optimal solution is found by evaluating

the four extreme-point solutions for the
problem represented by ‘OACD’ and selecting
the one that provides the largest profit
contribution.

• Corner-point feasible (CPF) solution

– Solution that lies at the corner of the feasible
region
• Linear programming problem (LPP) with
feasible solution and bounded feasible
region
– Must have CPF solutions and optimal
solution(s)
– Best CPF solution must be an optimal
solution
Linear Programming Model: Special Case
• Unbounded: solution to a LP problem is unbounded if the
value of the solution may be made infinitely large or
infinitely small depending on maximization/minimization
objective function without violating any of the
constraints.
• This condition might be termed managerial utopia; for
example, if this condition were to occur in a profit
maximization problem, the manager could achieve an
unlimited profit.
• In real-life LP problems, the occurrence of unbounded
solution means that the problem has been improperly
formulated.
• The occurrence of an unbounded solution is often the
result of a missing constraint during problem formulation.
Wyndor Glass Company Problem
Maximize 𝑍 = 3 𝑥1 + 5𝑥2
Subject to
𝑥1 ≤ 4
2𝑥2 ≤ 12
3𝑥1 + 2𝑥2 ≤ 18
𝑥1 ≥ 0
𝑥2 ≥ 0




The Essence of an Algebraic Method
• Surplus variables are also known as

negative slack variables
• In the objective function, it carries a zero

coefficient

One to One Mapping of Algebraic and
Graphical Solution Method
The four feasible solutions (i.e. 1, 3, 5 & 6) that we have got
with the algebraic method basically represents the four
corner points (i.e. O, A, C and D) obtained in the graphical
solution method 𝑥2
(0, 6) B 3𝑥1 + 2𝑥2 = 12

(0, 5) A
𝑥1
(0, 0) O D E
(4, 0) (5, 0)

One to One Mapping of Algebraic and
Graphical Solution Method
• The solution number 2 and 4 are not
represented here because they are
infeasible solutions
• Any point inside the feasible region is
dominated by the corner points except for
origin
• Thus, it is reasonable to assume the value
of two variables equal to zero in the
algebraic method
• In the algebraic method, fixing the two

variables equal to zero provides the basic
solution
• Out of six basic solutions, only four solutions
are feasible and they are called as basic
feasible solution
• In general for n variables and m equations,
we have nCm basic solutions
• All basic feasible solution will represent
the four corner points
• The variables that we have fixed equal to zero in

the algebraic method are called as non-basic
variables and the variables that we are trying to
solve for are called as basic variables
• Every basic solution is a set of non-basic and
basic variables
• In general the number of basic variables are
equal to the number of equations or
constraints that we have in the LP model and
the remaining variable would become non-basic
variable
Disadvantage with an Algebraic Method
• With algebraic method, we would end up evaluating
4C or nC
2 m basic solutions and it is a very large
number of solutions
• In total six basic solutions are evaluated and two out of
six are found to be infeasible and thus it is useless to
the objective function
• Also, it is not necessary to get the better value of
objective function with each basic feasible solution
• There is a need of method that can avoid getting
infeasible solution, can provide better value of
objective function with each iteration and can
terminate when it reaches to the optimum value 77
Setting Up the Algebraic Method
• First step: convert functional inequality

constraints into equality constraints
– Done by introducing slack or surplus
variables
– Resulting form known as augmented form
Example: constraint 𝑥1 ≤ 4 is equivalent to
𝑥1 + 𝑥3 = 4 and 𝑥3 ≥ 0

Setting Up the Algebraic Method for

Wyndor Glass Company Problem Solution
Using Algebraic Method
Maximize 𝑍 = 3𝑥1 + 5𝑥2+0𝑥3+0𝑥4+0𝑥5
Subject to
𝑥1 + 𝑥3 = 4
2𝑥2 + 𝑥4 = 12
3𝑥1 + 2𝑥2 + 𝑥5 = 18
𝑥1 ≥ 0
𝑥2 ≥ 0
𝑥3 ≥ 0
𝑥4 ≥ 0
𝑥5 ≥ 0

• The number of basic variable is equal to 3 so we can fix
the value of two variables equal to zero
1.) If 𝑥1 and 𝑥2 = 0; then 𝑥3 = 4, 𝑥4 = 12, and 𝑥5 = 18; Z = 0

4.) If 𝑥5 and 𝑥1 = 0; then 𝑥2 = 9, 𝑥3 = 4, and 𝑥4 = -6; no need*
8.) If 𝑥4 and 𝑥5 = 0; then 𝑥1 = 2, 𝑥2 = 6 and 𝑥3 = 2; Z = 36

The Essence of the Algebraic Method

Setting Up the Algebraic Method
• Properties of a basic solution

– Each variable designated basic or nonbasic
– Number of basic variables equals number of
functional constraints
– The nonbasic variables are set equal to zero
– Values of basic variables obtained as
simultaneous solution of system of equations
– If basic variables satisfy nonnegativity
constraints, basic solution is a (BF) solution

The Graphical and Algebraic Method
Mapping

Operations Analytics (Class 23-24)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Operations Analytics (Class 23-24)

Uploaded by

Copyright:

Available Formats

© 2015 McGraw-Hill Education. All rights reserved.

 Change in the value of R.H.S of a

 Change in the values of objective

 Addition of a new variable

 Addition of a new constraint 13

• May change the feasible region of

• If newly added constraint is

• Adding a new decision

© 2015 McGraw-Hill Education. All rights reserved. 20

• General problem terminology and

© 2015 McGraw-Hill Education. All rights reserved. 22

© 2015 McGraw-Hill Education. All rights reserved. 23

• A person wants to decide the constituents

• Formulate LP model for the problem

• Xj: Quantity of food type ‘j’ Decision Variables

Objective Function: Minimize Total Diet Cost

© 2015 McGraw-Hill Education. All rights reserved. 26

• An advertising company wishes to plan its

© 2015 McGraw-Hill Education. All rights reserved. 28

• Formulate LP model for the problem

Cost of an 30,000 20,000 15,000 10,000

© 2015 McGraw-Hill Education. All rights reserved. 29

• The company wants to spend not more than

• Xj: Number of units bought using media type ‘j’

15,000 X3 + 10,000 X4 ≤ 1,50,000

© 2015 McGraw-Hill Education. All rights reserved. 32

– The contribution to the objective function

– This may not be satisfied in few cases

– All the constant terms, objective function

© 2015 McGraw-Hill Education. All rights reserved. 38

© 2015 McGraw-Hill Education. All rights reserved. 39

Steps in Graphical Solution Method

• Restrict yourself to the first quadrant where both the

• Identify the feasible region that contain the set of

© 2015 McGraw-Hill Education. All rights reserved. 40

Steps in Graphical Solution Method

• Evaluate the objective function at all the corner

• The corner point that has the best value of the

© 2015 McGraw-Hill Education. All rights reserved. 41

© 2015 McGraw-Hill Education. All rights reserved. 46

(0, 6) B 3𝑥1 + 2𝑥2 = 12

© 2015 McGraw-Hill Education. All rights reserved. 47

(0, 6) B 3𝑥1 + 2𝑥2 = 12

© 2015 McGraw-Hill Education. All rights reserved. 48

(0, 6) B 3𝑥1 + 2𝑥2 = 12

© 2015 McGraw-Hill Education. All rights reserved. 49

(0, 6) B 3𝑥1 + 2𝑥2 = 12

© 2015 McGraw-Hill Education. All rights reserved. 50

(0, 6) B 3𝑥1 + 2𝑥2 = 12

© 2015 McGraw-Hill Education. All rights reserved. 51

(0, 6) B 3𝑥1 + 2𝑥2 = 12

A LP problem involving only two decision

© 2015 McGraw-Hill Education. All rights reserved. 53

• No need to evaluate all feasible solution points.

• The optimal solution is found by evaluating

© 2015 McGraw-Hill Education. All rights reserved. 55

• Corner-point feasible (CPF) solution

© 2015 McGraw-Hill Education. All rights reserved. 58

© 2015 McGraw-Hill Education. All rights reserved. 59

© 2015 McGraw-Hill Education. All rights reserved. 60

© 2015 McGraw-Hill Education. All rights reserved. 61

• Surplus variables are also known as

• In the objective function, it carries a zero

© 2015 McGraw-Hill Education. All rights reserved. 72