Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Analytic Hierarchy Process LOGO Contents

(Multi-criteria decision making)

Introduction

How does AHP work?

Applications

Q&A

LOGO
25/01/2021
Speaker: Ha Thi Xuan Chi 1 25/01/2021 2

LOGO Introduction LOGO Introduction

IBM NASA Harvard University Colorado State University


Several approaches for MCDM
exist: Goodyear IRS Yale University University of Cambridge
Ford Motor Co. FBI A few of the MIT Duke University
vANP many American University Purdue University
vAHP Citibank Department of Universities
vTOPSIS Defense using/teaching
vWeighted score method AHP Naval War College Katz School of Business
Xerox World Bank George Washington Wharton School of Business
Boeing Texaco University
a few of the thousands of Michigan State University Johns Hopkins University
AT&T Eastman
organizations using Kodak Stanford University University of Maryland
AHP
General Motors Inter-American
Bank
25/01/2021 3 25/01/2021 4
LOGO Introduction LOGO Example
Which one I
should I want to
choose? buy a car

Maybe AHP
can help you

25/01/2021 5 25/01/2021 6

LOGO LOGO

What is AHP?
What characteristics of the
It sounds to car are important to you?
AHP (Analytic complicate to me.
Hierarchy But, give it a try Well, I consider Style,
Process) is a Reliability, Fuel economy
multicriteria
decision-making
system

Solve complex Dr Thomas L.Saaty


decision-making developed
problems the process in
the early 1970’s
and...

25/01/2021
http://www.123ahp.com/Primjer.aspx?idPrimjer=26 7 25/01/2021 8
LOGO Hierarchy tree LOGO

L0 Goal Buy a car School Selection

Fuel
L1 Criteria Style Reliability Selecting the most appropriate school for the kid
economy

L2 Alternatives Mercedes Audi BMW Lexus Learning Friends


School Vocational College Music
life training prep. classes

v Why hierarchy tree?


v How many criteria in one level? School A School B School C
v How many levels?
www.ilkertopcu.net

25/01/2021 9 25/01/2021 10

LOGO LOGO

R&D Project Selection Whom to marry - a compatible spouse

Flexibity Independence Growth Challenge Commitment Humor Intelligence


Goal Future of the Firm
Psychological Physical Socio-cultural Philosophical Aesthetic

Communication Food Sociability World View Housekeeping


Criteria
Technical Marketing Financial Manufacture & Problem Solving

Family & Children Shelter Finance Theology Sense of Beauty


& Intelligence
Regulatory Compliance Capability to Market NPV Capability to Manufacture
Sub- Development Cost Market Growth Capital Invest Facility/Equp. Req.
Temper Sex Understanding
Criteria Prob. of Tech. Success Market Share ROI Safety
R&D and Eng. Resources Market Potential Unit Cost Security
Development Time Customer Acceptance
Patent Position
Affection
Ratings
(for each Outstanding Outstanding Outstanding Outstanding
Above Average Above Average Above Average Above Average Loyalty
Sub- Average Average Average Average
Criterion) Below Average Below Average Below Average Below Average
Case 1
marry don’t marry

Case 2
p1 p2 p99 Campbell Graham McGuire Faucet
25/01/2021 11 25/01/2021 12
www.ilkertopcu.net www.ilkertopcu.net
LOGO LOGO Analytic Hierarchy Process
Whom to marry - a compatible spouse
§ A powerful and understandable methodology that
allows groups or individuals to combine qualitative
Flexibity Independence Growth Challenge Commitment Humor Intelligence
and quantitative factors in decision making process.
Psychological Physical Socio-cultural Philosophical Aesthetic § A method for complicated and unstructured problems.
Communication
& Problem Solving
Food Sociability World View Housekeeping § Uses a hierarchical model having levels of goal,
Family & Children Shelter Finance Theology Sense of Beauty
& Intelligence criteria, possible sub-criteria, and alternatives.
Temper Sex Understanding

Security

Affection

Loyalty

Case 1
marry don’t marry

Case 2
Campbell Graham McGuire Faucet
www.ilkertopcu.net

25/01/2021 13 25/01/2021 14

LOGO Priorities: Importance -Preference LOGO

v AHP captures priorities from paired comparison


Well, the Fuel economy is for me more
judgments of the elements of the decision with important than Reliability...and Reliability a little
respect to each of their parent criteria: more important than Style.
– Paired comparison judgments are arranged in a matrix I would not like to have an ugly car with high
– Priorities are derived from the matrix as its principal reliability, and then I do not want it to have low
eigenvector fuel economy even if it is well designed car.
What criteria of Uh, it is hard for me to define the rank of
– It also allows for the measurement of inconsistency in criteria.
the car are the
judgment most important
to you?

25/01/2021 15 25/01/2021 16
LOGO Pair wise comparison LOGO Pair wise comparison

Reliability 1/3
Style
Style
Reliability
Fuel
economy 1/7

Scale of reciprocals for inverse


Scale of important level: comparisons
9 7 5 3 1 1/3 1/5 1/7 1/9
8 6 4 2 ½ ¼ 1/6 1/8
Reliability Fuel
1/5
economy
Extremely Important Equal Important Extremely

Moderate Moderate Very important


Very important
important important

25/01/2021 17 25/01/2021 18

Normalized principal Eigen


LOGO Structure a matrix LOGO
vector
Reliability

Reliability
economy

economy

Average
Style

Style
Fuel

Fuel
Priority
vector
What happen if
0.084 using different
Style 1 1/3 1/7 Style 1/11 1/19 5/47 0.084 normalization
methods ?
0.193
Reliability 3 1 1/5 Reliability 3/11 3/19 7/47 0.193

Fuel Fuel 0.723


7 5 1 7/11 15/19 35/47 0.723
economy economy

S = [ S1 S2 ... Sn ] 1
Sum 11 19/3 47/35

25/01/2021 19 25/01/2021 20
LOGO Calculate Eigen value LOGO
Priority
Eigen value lmax - n
Consitency Index: CI =
vector
Sum lmax
0.084
n -1
11 19/3 47/35 x = 3.11 CI
0.193 Consitency Ratio: CR =
RI
0.723
Random Consitency Index : RI

lmax = S ´ W Size of comparison matrix : n

25/01/2021 21 25/01/2021 22

LOGO LOGO Results

lmax - n 3.11 - 3
CI = = = 0.055
n -1 3 -1
0.055
CR = = 0.094 < 10% (acceptable)
0.58
v Why CR< 10% ?

The size of comparison matrix: n=3


n 1 2 3 4 5 6 7 8 9 10
RI 0 0 0.58 0.9 1.12 1.24 1.32 1.41 1.45 1.49

25/01/2021 23 25/01/2021 24
LOGO Pair wise comparison for alternatives LOGO Results
Style Reliability

Mercedes Audi 1/3 1/5

BMW 5 5

Lexus 9 3 Mercedes: 11.6% Mercedes: 37.9%

Audi BMW 3 1/5 Audi: 24.7% Audi: 29.%

Lexus 1/5 3 BMW: 6% BMW: 74%

BMW Lexus 5 7 Lexus: 57.8% Lexus: 25.7%

25/01/2021 25 25/01/2021 26

LOGO LOGO Results

Miles/gallon Normalized

Mercedes 34 .3010

Fuel Audi 27 .2390


economy
BMW 24 .2120 Mercedes: 11.6% Mercedes: 37.9% Mercedes: 30.1% 30% 1
Lexus 28 .2480
113 1.0
Audi: 24.7% Audi: 29.% Audi: 23.9% 25% 3

BMW: 6% BMW: 74% BMW: 21.2%


17% 4

Lexus: 57.8% Lexus: 25.7% Lexus: 24.8%


28% 2
25/01/2021 27 25/01/2021 28
LOGO LOGO
Priority
Priority vector of
vector of Weights WOW, it’s sort
alternatives
criteria of amazing, this
is the car that
1 Mercedes best fits my
0.12 0.38 0.30 0.30 preferences and
0.084
= criteria
3 Audi
0.25 0.29 0.24 0.25
x 0.193
4 BMW
0.06 0.07 0.21 0.17
0.723
2 Lexus
0.58 0.26 0.25 0.28

25/01/2021 29 25/01/2021 30

LOGO AHP applications LOGO Software


v Super Decisions: www.superdecisions.com
question please, I can
Sure!
select in the same v Download: http://www.superdecisions.com/downloa
method for cars, house,
suppliers,
ds/
projects and etc. Right?

It’s simply
changed What if I change
my mind in some
preference or
criteria?

Expert
choice, What software
Excel do support for
computing AHP
Wow, so process?
25/01/2021 convenient 31 25/01/2021 32
Comparison of criteria w.r.t. goal
LOGO Example LOGO

School life

Vocational
School Selection

Learning

training
Friends
Priorities

classes
College

Music
prep.
Selecting the most appropriate school for the kid Learning 1 4 3 1 3 4 .3047
Friends 1/4 1 7 3 1/5 1 .1486
School Vocational College Music
Learning Friends
life training prep. classes
School life 1/3 1/7 1 1/5 1/5 1/6 .0382
Vocational
training 1 1/3 5 1 1 1/3 .1414
School A School B School C College prep. 1/3 5 5 1 1 3 .2208
www.ilkertopcu.net Music classes 1/4 1 6 3 1/3 1 .1463
25/01/2021 33 25/01/2021 34
www.ilkertopcu.net

Consistency
LOGO LOGO

.3202 .9667 3.0186


Comparison of schools w.r.t. criteria
w .5571 Aw 1.6881 Aw 3.0300
.1226
Learning
.3687
Friends w School life
A B C Priorities A B C 3.0065
A
Priorities B C Priorities

A 1 1/3 1/2 .16 A 1 1


average
1 .33 3.0183
A 1 5
λ1 max
.46

B 3 1 3 .59 B 1 1 1 .33 B 1/5 1 1/5 .09

C 2 1/3 1 .25 C 1 1 1 .33 C 1 5 1 .46


Thank you for your attention!
n RI
λmax - n
3 0.52 CI= =0.0092
Vocational training n - 1prep.
College Music classes
4 0.89 A B C Priorities A B C Priorities A B CCIPriorities
CR1 =6
Consistency index
5 1.11 A 1 9 7 .75 A 1 1/2 1 .25 A 4 .69 =1.76%
6 1.25 RI
B 1/9 1 1/5 .06 B 2 1 2 .50 B 1/6 1 1/3 .09
7 1.35
RI= SEE Consistency ratio
.19 1/2 =0.52
.25 .22
8 1.40
C 1/7 5 1
TABLE
C 1 1 C 1/4 3 1

www.ilkertopcu.net
9 1.45 Random consistency index
25/01/2021 35 25/01/2021 36
www.ilkertopcu.net
LOGO LOGO Questions
v How to dealing with high CR values?
(suggested 10%.If >10%, reconsider input: make question again and reconsider
the answers)
v Can we change the AHP scale from -9 to +9?
(No, must be positive matrix, need rescale the negative value into positive follow
AHP scales )
v I have more than 20 criteria, why RI table is only until n=10?
(don’t need RI for more than n=7, because the number of pair comparison n*(n-
1)/2 too large make people confuse to answer correctly)
v Is it possible to obtain a negative consistency ratio AHP?
(No, may be happen due to error in the approximation of Eigen value,
MS excel just can give the approximation of Eigen value, not the real one,
Maxima, Expert choice are good)

25/01/2021 37 25/01/2021 38

You might also like