Sustainable Manufacturing

MMZG534/DMZG534

Lecture14
BITS Pilani
Work Integrated Learning
Pavan Kumar Potdar
Mechanical Engineering
Programmes Division
 BITS Pilani
Work Integrated Learning
Programmes Division

Multi Criteria Decision Making

MCDM
• consists of constructing a global preference relation for a
set of alternatives evaluated using several criteria
• selection of the best actions from a set of alternatives,
each of which is evaluated against multiple, and often
conflicting criteria.

• MADM (Multi Attribute Decision Making)

• these problems are assumed to have a predetermined , limited
number of decision alternatives.

• MODM (Multi Objective Decision Making)

• the decision alternatives are not given, instead the set of decision
alternatives is explicitly defined by constraints using multiple
objective programming, the number of potential decision
alternatives may be large.
BITS Pilani, WILPD
MCDM: Examples
• In determining an electric route for power transmission in a
city, several criteria could be considered:
• Cost
• Health
• Reliability
• Importance of areas

• Selection of best supplier

• Cost
• Quality
• Flexibility
• Delivery Time

BITS Pilani, WILPD

Example of MCDM

• Objective is to buy a smart phone

• Factors / Criteria we use could be Quality, Price,
Camera, Looks, Safety, Ease of use.
• We may have few alternatives in mind (Samsung,
Motorola, Iphone etc)
• Taking a decision when we have many criteria will be
difficult as there are many things to think about.
• Let us analyse this step by step for better understanding

BITS Pilani, WILPD

We can make a table of the criteria and the alternative as
below. C1-6 are the criteria. A1-3 are the alternatives we
have.
A1 A2 A3
Samsung I Phon Motorola
C1 Quality
C2 Camera
C3 Comfort
C4 Safety
C5 Looks
C6 Price

BITS Pilani, WILPD

We need to give ratings for each criteria for each
alternative. The rating scale can be taken as (1 – low to
9 – high).

A1 A2 A3
Samsung I Phone Motorola
C1 Quality 7 8 6
C2 Camera 8 9 6
C3 Comfort 7 9 7
C4 Safety 7 8 4
C5 Looks 7 9 6
C6 Price 8 5 6

BITS Pilani, WILPD

For the ratings given, the selection is obviously the one
alternative which has the max average value.
A1 A2 A3
Samsung I Phone Motorola
C1 Quality 7 8 6
C2 Camera 8 9 6
C3 Comfort 7 9 7
C4 Safety 7 8 4
C5 Looks 7 9 6
C6 Price 8 5 6

Avg 7.33 8.00 5.83

BITS Pilani, WILPD

The problem in this case is that we assumed all the criteria
are of equal weightage, which is not true in real sense.
Some criteria are more important than others. Hence if
you give weightage to criteria, the same table looks like
this.
A1 A2 A3 A1 A2 A3
Weight Samsung I Phone Motorola Samsung I Phone Motorola
C1 Quality 0.25 7 8 6 1.75 2 1.5
C2 Camera 0.20 8 9 6 1.6 1.8 1.2
C3 Comfort 0.05 7 9 7 0.35 0.45 0.35
C4 Safety 0.05 7 8 4 0.35 0.4 0.2
C5 Looks 0.05 7 9 6 0.35 0.45 0.3
C6 Price 0.40 8 5 6 3.2 2 2.4

Avg 1.00 7.33 8.00 5.83 1.27 1.18 0.99

BITS Pilani, WILPD

When the criteria are given weights, you can see the
decision has changed from I Phone to Samsung.

The criteria weights are normally calculated using pair-wise

comparison method to identify the relative relation and
influence of one criteria over the other. (There are other
methods too to calculate the weights.)

We will see how it is done in AHP

BITS Pilani, WILPD

 BITS Pilani
Work Integrated Learning
Programmes Division

Multi Criteria Decision Making: AHP

 Analytical Hierarchy Process
• AHP uses a hierarchical structure and pairwise
comparisons.
• An AHP hierarchy has at least three levels:
1. the main objective of the problem at the top
2. multiple criteria that define alternatives in the middle (m)
3. competing alternatives at the bottom (n)
Hierarchical value tree Buy a Car L1: Goal

Performance Style Costs L2: Criteria

Braking Acceleration Maintenance Initial L3: Sub-criteria

Ford Chevy BMW Mazda L4: Alternatives

BITS Pilani, WILPD

AHP Steps
Thomas L. Saaty developed AHP containing following steps:

1. Define the problem and determine its goal.

2. Structure the hierarchy from the top (the objectives from a
decision-maker's viewpoint) through the intermediate levels
(criteria on which sub-sequent levels depend) to the lowest
level which usually contains the list of alternatives.
3. Construct a set of pair-wise comparison matrices (size nXn)
for each of the lower levels with one matrix for each element
in the level immediately above by using the relative scale
measurement shown in Table. The pair-wise comparisons are
done in terms of which element dominates the other.
4. There are n(n-1) = judgments required to develop the set of
matrices in step3. Reciprocals are automatically assigned in
each pair-wise comparison.
BITS Pilani, WILPD
AHP Steps
5. Hierarchical synthesis is now used to weight the
eigenvectors by the weights of the criteria and the sum is
taken over all weighted eigenvector entries corresponding
to those in the next lower level of the hierarchy.
6. Having made all the pair-wise comparisons, the
consistency is determined by using the Eigen value, λmax,
to calculate the consistency index, CI as follows: CI=(λmax
– n)/(n-1), where n is the matrix size. Judgment
consistency can be checked by taking the consistency
ratio (CR) of CI with the appropriate value. The CR is
acceptable, if it does not exceed 0.10.
7. Steps 3-6 are performed for all levels in the hierarchy

BITS Pilani, WILPD

Example to understand the
AHP process.
The objective is to buy a mobile phone and we have
identified few criteria which are relevant. The hierarchy is
given below.

BITS Pilani, WILPD

Level 1 – the objective of study – To buy a mobile phone

Level 2 – list the various criteria that must be considered

like – price, storage, camera and looks.

Level 3 – gives the various alternatives under

consideration. Mobiles 1 – 5

This is a case of MCDM but to do this correctly we need

the criteria weights, which we will find using the Pair-
wise – Comparison method.

BITS Pilani, WILPD

Pair wise comparison matrix

• It gives the relative importance of one criteria with

respect to the goal.
• A standard scale of relative importance is used for this.

BITS Pilani, WILPD

Pair wise comparison matrix template

Price Storage Camera Looks

Price
Storage

Camera
Looks

BITS Pilani, WILPD

To populate the matrix, you need to answer question such
as
“How important is the Price compared to Storage” when
you want to buy a phone?

If the row criteria is having more importance than the

column criteria, the value of relative importance as per
the case is assigned in the row and 1/<value> assigned
for the column for that criteria.

BITS Pilani, WILPD

If the price has “Strong importance” to the storage, the
value assigned will look like
Price Storage Camera Looks
Price 5
Storage 1/5
Camera
Looks

The diagonal will have 1 as the value. Based on the

responses, the other cells are populated.

BITS Pilani, WILPD

Assuming the resposes are populated, the pair wise
comparison matrix will look like
Price Storage Camera Looks
Price 1 5 4 7
Storage 1/5 1 1/2 3
Camera 1/4 2 1 3
Looks 1/7 1/3 1/3 1

Price Storage Camera Looks

Price 1.00 5.00 4 7
Storage 0.20 1 0.50 3
Camera 0.25 2 1 3
Looks 0.14 0.33 0.33 1
SUM 1.59 8.33 5.83 14.00
BITS Pilani, WILPD
Normalized pair wise comparison matrix is calculated by
dividing each cell in column by its column sum.

Price Storage Camera Looks

Price 0.63 0.60 0.685714 0.5
Storage 0.13 0.12 0.085714 0.214286
Camera 0.16 0.24 0.171429 0.214286
Looks 0.09 0.04 0.057143 0.071429

BITS Pilani, WILPD

Next calculate the Criteria Weights by averaging the values
in each row.

Price Storage Camera Looks Criteria Wt

Price 0.63 0.60 0.685714 0.5 0.60338
Storage 0.13 0.12 0.085714 0.214286 0.13639
Camera 0.16 0.24 0.171429 0.214286 0.19567
Looks 0.09 0.04 0.057143 0.071429 0.06456

BITS Pilani, WILPD

Calculate the consistency of the calculations. Use the non-
normalized matrix for this purpose.
Criteria Wt 0.60338 0.13639 0.19567 0.06456
Price Storage Camera Looks
Price 1.00 5.00 4 7
Storage 0.20 1 0.50 3
Camera 0.25 2 1 3
Looks 0.14 0.33 0.33 1

Criteria Wt 0.60338 0.13639 0.19567 0.06456

Price Storage Camera Looks
Price 0.60 0.68 0.782665 0.451951
Storage 0.12 0.14 0.097833 0.193693
Camera 0.15 0.27 0.195666 0.193693
Looks 0.09 0.05 0.065222 0.064564

BITS Pilani, WILPD

Weighted Sum value is calculated by summing up each row
values.

Criteria Wt 0.60338 0.13639 0.19567 0.06456 Weighted

Price Storage Camera Looks Sum Value
Price 0.60 0.68 0.782665 0.451951 2.52
Storage 0.12 0.14 0.097833 0.193693 0.55
Camera 0.15 0.27 0.195666 0.193693 0.81
Looks 0.09 0.05 0.065222 0.064564 0.26

BITS Pilani, WILPD

Calculate the ratio of Weighted sum value and the criteria
weights. This is done by dividing the WSV for each row
with its corresponding Criteria Wt.
Weighted Criteria
Price Storage Camera Looks Sum Value Wt Ratio
Price 0.60 0.68 0.782665 0.451951 2.52 0.60338 4.176388
Storage 0.12 0.14 0.097833 0.193693 0.55 0.13639 4.022228
Camera 0.15 0.27 0.195666 0.193693 0.81 0.19567 4.154955
Looks 0.09 0.05 0.065222 0.064564 0.26 0.06456 4.049398

BITS Pilani, WILPD

Calculate the Consistency Index

Where n = number of criteria.

Calculate the Consistency Ratio:

RI is the Random Index – a standard value based on “n”

BITS Pilani, WILPD

Since the value of CR is less than 0.10, we can
assume that the consistency of data and proceed
with the decision making.

We now have the weights of criteria and we can

proceed with the MCDM method to take a decision.

BITS Pilani, WILPD

BITS Pilani, WILPD
Fundamental Scale of
absolute numbers

BITS Pilani, WILPD

 Random Index table

N 1 2 3 4 5 6 7 8 9
Random Index 0 0 0.58 0.9 1.12 1.24 1.32 1.41 1.45

Use AHP method for decision making and providing ranking to the factors

• Design factor is “moderately preferred” over production factor. (3)

• Design factor is “strongly preferred” over Use factor. (5)
• Design factor is “moderately to strongly preferred” over Disposal factor. (4)
• Production factor is “moderately to strongly preferred” over Use factor. (4)
• Production factor is “moderately preferred” over Disposal factor. (3)
• Use factor is “equally to “moderately preferred” over Disposal factor. (2)

BITS Pilani, WILPD

Comparison matrix: Criteria

Design Factor Production Use Factor Disposal Factor

Factor
Design Factor 1 3 5 4
Production 1/3 1 4 3
Factor
Use Factor 1/5 1/4 1 2
Disposal Factor 1/4 1/3 1/2 1

BITS Pilani, WILPD

Practice Problem (Purchasing
Land) - Suggest best location

Loc 1 Loc 2 Loc 3 Loc 4 Loc 5

Cost of land 9 6 9 3 6

Size 5 2 3 6 5

Distance to city 2 6 2 6 6

Loan from Banks 9 7 7 6 6

Ownership Status 3 6 3 6 6

Facilities 4 6 4 6 6

BITS Pilani, WILPD

Pair wise comparison matrix:

Ownersh
CoL Size Distance Loan ip Facilities
Cost of land 1 5 5 4 4 5
Size 0.2 1 3 3 4 2
0.33333
Distance to city 0.2 3 1 4 4 1
0.33333
Loan from Banks 0.25 3 0.25 1 1 1
Ownership Status 0.25 0.25 0.25 1 1 1
Facilities 0.2 0.5 1 1 1 1
BITS Pilani, WILPD
Problem with multiple
responses – Car selection

MS MAH Skoda VW Tata

M1 M2 M3 M4 M5
Mileage 8 6 4 5 6
Features 6 7 8 8 6
Comfort 4 7 7 8 4
Cost 6 7 2 3 5
Storage Space 3 5 7 6 9
Brand 9 5 9 2 7
Resale Value 9 5 2 3 4

BITS Pilani, WILPD

Response 1

Storage Resale
Mileage Features Comfort Cost Space Brand Value
Mileage 1 2 2 2 8 2 6
Features 1/2 1 2 2 8 3 5
Comfort 1/2 1/2 1 5 8 3 5
Cost 1/2 1/2 1/5 1 2 2 4
Storage Space 1/8 1/8 1/8 1/2 1 1/7 1/5
Brand 1/2 1/3 1/3 1/2 5 1 2
Resale Value 1/6 1/5 1/5 1/4 5 1/2 1

BITS Pilani, WILPD

Response 2

Storage Resale
Mileage Features Comfort Cost Space Brand Value
Mileage 1 3 2 3 8 3 5
Features 1/3 1 3 2 8 5 5
Comfort 1/2 1/3 1 6 8 3 5
Cost 1/3 1/2 1/6 1 2 3 4
Storage
Space 1/8 1/8 1/8 1/2 1 1/7 1/2
Brand 1/3 1/5 1/3 1/3 5 1 2
Resale
Value 1/5 1/5 1/5 1/4 2 1/2 1

BITS Pilani, WILPD

Response 3

Storage Resale
Mileage Features Comfort Cost Space Brand Value
Mileage 1 2 2 2 6 4 6
Features 1/2 1 2 3 7 3 4
Comfort 1/2 1/2 1 6 8 3 5
Cost 1/2 1/3 1/6 1 3 1/3 3
Storage
Space 1/6 1/7 1/8 1/3 1 1/7 1/2
Brand 1/4 1/3 1/3 3 5 1 2
Resale
Value 1/6 1/4 1/5 1/3 2 1/2 1

BITS Pilani, WILPD

BITS Pilani, WILPD