Scribd
Upload
8 views39 pages

MCDM

The document outlines the Multicriteria Decision Making (MCDM) process, detailing its steps, techniques like SAW and AHP, and applications in personal and professional life. It emphasizes the importance of evaluating alternatives based on multiple criteria rather than seeking an optimal solution. Additionally, it provides examples of decision matrices and the AHP process for prioritizing options based on expert judgment and relative importance.

Uploaded by

jay lakhani
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
8 views39 pages

MCDM

The document outlines the Multicriteria Decision Making (MCDM) process, detailing its steps, techniques like SAW and AHP, and applications in personal and professional life. It emphasizes the importance of evaluating alternatives based on multiple criteria rather than seeking an optimal solution. Additionally, it provides examples of decision matrices and the AHP process for prioritizing options based on expert judgment and relative importance.

Uploaded by

jay lakhani
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 39

MCDM (SAW & AHP)

Decision Making
• Four steps:
• identification of the problem
• deriving the preferences
• evaluation of alternatives;
• identification of the best alternatives.
• Multicriteria decision making primarily do not look at the optimal
solution but intends to analyse various alternatives
• Also called MCDA (Multicriteria decision analysis)
• MCDM can be divided into two groups:
• Multiple attribute decision making (MADM)
• limited number of predetermined alternatives and
discrete preference ratings
• multiple objective decision making (MODM)
• optimal solution of set of goals prevailing among the
constrains.
MCDM process:
• Identification of the relevant criteria and alternatives from the
existing theory and practice.
• Assignment of numerical values to criteria to indicate their
relative importance and to quantify the impacts of the
alternatives on these criteria.
• Using a formal mathematical procedure for analysing
numerical values to determine the ranking (priorities) of the
alternatives.
Single-criteria versus multi-criteria decision
making
Y MCDM?
• Decision makers do not look for an optimal solution but wish to investigate the set of non-dominated
solutions.
• A non-dominated solution is the one which does not permit the deviation from it to any other solution without
sacrificing in at least one criterion.
Applications of MCDM in personal life
Applications of MCDM in professional life
Classification of MCDM techniques
• SAW
• AHP
• ANP
• TOPSIS
• ELECTRE
• VIKOR
• MOORA
• COPRAS….
Real-Life Applications
• Bridge Construction
• reduce traffic congestion
• reduce total distance to travel
• add to the elegance of the city.
Why MCDM Techniques Are More Popular in
Industry Practitioners?
• its potential to evaluate various conflicting alternatives (may be
called choice, strategy, policy, scenario, etc.) on set of criteria.
• can handle different units of measurement among the criteria
• advantage in evaluating the intangible factors like brand, image,
risk, difficulty level, etc.
Simple Additive Weightage (SAW)
• Also known as weighted linear combination or scoring method.
• Based on the weighted average
Problem:
• The customer satisfaction is derived from the perceived fairness
of exchange situations.
• Customers perceive higher equity when they receive distinct
advantages from the service provider, and good conflict handling
policy, characterized by distributive, procedural justice and
interactional justice (fairness).
• The problem considered here attempts to prioritize various hotels
for set of key criteria
Alternatives:
• Types of hotels.
• P1: Ellaa hotel in Gachibowli.
• P2: The Manohar hotel in Begumpet.
• P3: Eaglewood hotel in Gachibowli.
• P4: Radisson Blu Plaza hotel in Banjara hills.
• P5: Dream Valley Resorts in Shamshabad.
• P6: Aalankrita Resorts and spa in Ameerpet.
Criteria
• C1: Location of the hotel in Hyderabad.
• C2: Price in Rs.
• C3: Rating of hotel.
• C4: Type of hotel.
• C5: Amenities.

• C2 is cost criteria and remaining are benefit criteria.


Step 1: Obtaining the decision matrix

C1 C2 C3 C4 C5
P1 3.2 4700 4 5 5
P2 4.2 6000 3.4 5 5
P3 3.2 1375 4 3 3
P4 4.5 5680 4.3 5 5
P5 4.1 2600 4 3 3
P6 3.5 4100 4.4 4 5
Step 2: Obtain the normalized decision matrix
• rij = xij /xj*, xj*= Maximum value if the jth criterion is a benefit criterion.
• rij = xj*/ xij, xj*= Minimum value if the jth criterion is a cost criterion

C1 C2 C3 C4 C5
P1 0.711111 0.292553 0.909091 1 1
P2 0.933333 0.229167 0.772727 1 1
P3 0.711111 1 0.909091 0.6 0.6
P4 1 0.242077 0.977273 1 1
P5 0.911111 0.528846 0.909091 0.6 0.6
P6 0.777778 0.335366 1 0.8 1
Obtain the weighted score for each alternative
using the normalized decision matrix
• Sum Product of weights and normalized decision matrix scores
• Weightages: C1 = 0.1, C2 = 0.3, C3 = 0.1, C4 = 0.2, C5 = 0.3 (given)

Hotels Score Rank


P1 0.749786 3
P2 0.739356 4
P3 0.76202 2
P4 0.770351 1
P5 0.640674 6
P6 0.738388 5
Problem 2: Prioritization of manufacturing
systems
• Different types of manufacturing systems.
P1: Mass production.
P2: Batch production.
P3: Job shop.
P4: Project.

• The factors in manufacturing systems.


X1: Volume of units produced in year.
X2: No of products to be produced.
X3: Flexibility.
X4: Variable cost.
X5: Quality of the product

Weightages: X1 = 0.3, X2 = 0.2, X3 = 0.1, X4 = 0.1, X5 = 0.3


Criteria scores
X1 X2 X3 X4 X5
P1 Very high Medium Very low Very low Very low
P2 Medium High Medium Medium Medium
P3 Low Very high High High High
P4 Very low Very Low Very high Very high Very high
Decision matrix with numerical values

X1 X2 X3 X4 X5
P1 9 5 1 9 1
P2 5 7 5 5 5
P3 3 9 7 3 7
P4 1 1 9 1 9
Normalized decision matrix
C1 C2 C3 C4 C5

P1 1 0.555556 0.111111 1 0.111111


P2 0.555556 0.777778 0.555556 0.555556 0.555556
P3 0.333333 1 0.777778 0.333333 0.777778
P4 0.111111 0.111111 1 0.111111 1
Obtain the weighted score for each alternative
using the normalized decision matrix
Weightages: C1 = 0.1, C2 = 0.3, C3 = 0.1, C4 = 0.2, C5 = 0.3 (given)

Alts scores rank


P1 0.555556 3
P2 0.6 2
P3 0.644444 1
P4 0.466667 4
AHP (Analytic Hierarchy Process )
• developed in 1970 by Thomas L. Saaty.
• accommodates linear hierarchical structure of the problems
including criteria, sub-criteria and alternatives.
• include the inputs in both the forms
• actual measurement of criteria such as cost, profit
• subjective assessment by experts that captures their satisfaction,
opinions, beliefs and priorities related to any decision.
• ease in converting its subjective opinion into objective.
The Saaty rating scale
AHP process
• AHP begins with the construction of a decision matrix
representing relative significance of various “attributes” relative to
each other.
• If attribute A is “very much important” than attribute B then, it is
rated at 7, and the importance of B to A is rated as 1/7.
• Relative weights of the factors are calculated for the problem
under consideration.
• The matrix formed using the relative weights is called reciprocal
matrix or judgement matrix.
Consistency Ratio (CR)
• A measure of the consistency of the judgement considered through
AHP compared to that of completely random samples of judgement
• If the value of CR significantly exceeds 0.1, the judgement is deemed to
be very close to being purely random which may require the entire
process to be repeated or abandoned as futile.
𝐶𝐼 (𝐶𝑂𝑁𝑆𝐼𝑆𝑇𝐸𝑁𝐶𝑌 𝐼𝑁𝐷𝐸𝑋)
• 𝐶𝑅 =
𝑅𝐼 (𝑅𝐴𝑁𝐷𝑂𝑀 𝐼𝑁𝐷𝐸𝑋)
• In practical world, CR more than 0.1 is also acceptable sometimes.
• Higher the CR, higher is the inconsistency of judgement, because it
gets closer to random judgement.
Step-by-Step Procedure of AHP
• Step-1 : formation of reciprocal matrix (n criteria * n criteria).
• Get expert opinion
• Use Saaty rating scale to convert qualitative data to quantitative weights
• Step-2: calculation of eigen vector values
𝑛𝑡ℎ 𝑟𝑜𝑜𝑡 𝑜𝑓 𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑜𝑓 𝑣𝑎𝑙𝑢𝑒𝑠 𝑖𝑛 𝑎 𝑟𝑜𝑤
• 𝐶𝑙𝑜𝑠𝑒 𝑎𝑝𝑝𝑟𝑜𝑥𝑖𝑚𝑎𝑡𝑖𝑜𝑛 =
𝑠𝑢𝑚 𝑜𝑓 𝑎𝑙𝑙 𝑡ℎ𝑒 𝑟𝑜𝑜𝑡𝑠
Calculating eigenvector
nth root of
Eigen
A B C D Product of
vector
values
A 1.00 0.33 0.11 0.20 0.2934 0.06
B 3.00 1.00 1.00 1.00 1.3161 0.26
C 9.00 1.00 1.00 3.00 2.2795 0.45
D 5.00 1.00 0.33 1.00 1.1362 0.23
Totals 5.03 1.00

Eigen vectors represent the weights or importance of the criteria,


higher the eigen vector, higher the weight
Then we calculate the consistency ratio
𝐶𝐼 (𝐶𝑂𝑁𝑆𝐼𝑆𝑇𝐸𝑁𝐶𝑌 𝐼𝑁𝐷𝐸𝑋)
• 𝐶𝑅 =
𝑅𝐼 (𝑅𝐴𝑁𝐷𝑂𝑀 𝐼𝑁𝐷𝐸𝑋)
𝜆𝑚𝑎𝑥 −𝑛
• 𝐶𝑂𝑁𝑆𝐼𝑆𝑇𝐸𝑁𝐶𝑌 𝐼𝑁𝐷𝐸𝑋 𝐶𝐼 = ’
𝑛−1
• 𝜆𝑚𝑎𝑥 of a criteria calculation:
• sum product of ( row value of the criteria, eigen vector column)
• Divide it by the respective eigen vector, you get 𝜆𝑚𝑎𝑥
• All 𝜆𝑚𝑎𝑥 must satisfy 𝜆𝑚𝑎𝑥 ≥ 𝑛
• Take their mean. Use it to calculate the CI.
• For perfect data : 𝜆𝑚𝑎𝑥 = 𝑛
• For our data based on human judgement : 𝜆𝑚𝑎𝑥 ≥ 𝑛
𝑅𝐴𝑁𝐷𝑂𝑀 𝐼𝑁𝐷𝐸𝑋
• Get the value from saaty’s table as per the order of the matrix.
• Order of the matrix = no of criteria
Steps for an actual problem:
• AHP comes under the family of pairwise comparison
• That means we have to carry the previous steps for
• Between all the criteria
• Between the alternatives for each criteria separately

• First one will give us the weights, the second one will give us
normalized score of alternatives in each criteria.
• Take their sum product and you will get the final score. Rank it in
descending order.
two levels AHP
‘more’ levels AHP
File:AHPJones01.png

Chevrole
Audi BMW Dodge Fiat Hyundai
t
Structure the Hierarchy

You might also like