Vous êtes sur la page 1sur 16

Analytic Hierarchy Process

Multiple-criteria decision-making
Real world decision problems
multiple, diverse criteria
qualitative as well as quantitative information

Comparing apples and oranges?


Spend on defence or agriculture?
Open the refrigerator - apple or orange?

AHP
Information is decomposed into a hierarchy of
alternatives and criteria
Information is then synthesized to determine
relative ranking of alternatives
Both qualitative and quantitative information
can be compared using informed judgements
to derive weights and priorities

Example: Car Selection


Objective
Selecting a car

Criteria
Style, Reliability, Fuel-economy

Cost?

Alternatives
Civic Coupe, Saturn Coupe, Ford Escort,
Mazda Miata

Hierarchical tree
S e le c t in g
a N ew C ar
S t y le
- Civic
- Saturn
- Escort
- Miata

R e lia b ilit y
- Civic
- Saturn
- Escort
- Miata

FuelE conom y
- Civic
- Saturn
- Escort
- Miata

Ranking of criteria
Weights?
AHP
pair-wise relative importance
[1:Equal, 3:Moderate, 5:Strong, 7:Very strong, 9:Extreme]

Style

Reliability

Fuel Economy

Style

1/1

1/2

3/1

Reliability

2/1

1/1

4/1

Fuel Economy

1/3

1/4

1/1

Ranking of priorities
Eigenvector

[Ax = x]

Iterate
1. Take successive squared powers of matrix
2. Normalize the row sums
Until difference between successive row sums is
less than a pre-specified value

1
0.5
2
1
0.333 0.25
Row sums
12.75
22.3332
4.8333
39.9165

3
4
1.0

squared

3.0
1.75 8.0
5.3332 3.0
14.0
1.1666 0.6667 3.0

Normalized
Row sums
0.3194
0.5595
0.1211
1.0

0.3196
New iteration gives normalized row sum 0.5584
0.1220
Difference is:

0.3194
0.5595 0.1211

0.3196
0.5584
0.1220

- 0.0002
= 0.0011
- 0.0009

Preference
Style
.3196
Reliability
.5584
Fuel Economy .1220
S e le c t in g
a N ew C ar
1 .0
S t y le
.3 1 9 6

R e lia b ilit y
.5 5 8 4

FuelE conom y
.1 2 2 0

Ranking alternatives
Style
Civic

Civic
1/1

Saturn
Escort
Miata

4/1
1/4
6/1

Reliability Civic
Civic
1/1
Saturn
Escort
Miata

1/2
1/5
1/1

Saturn
1/4
1/1
1/4
4/1
Saturn
2/1
1/1
1/3
1/2

Escort Miata
4/1
1/6
4/1
1/1
5/1

1/4
1/5
1/1

Escort Miata
5/1
1/1
3/1
1/1
4/1

2/1
1/4
1/1

Eigenvector
.1160
.2470
.0600
.5770

.3790
.2900
.0740
.2570

Miles/gallon

Fuel Economy
(quantitative
information)

Normalized

Civic

34

.3010

Saturn
Escort
Miata

27
24
28

.2390
.2120
.2480

113

1.0

S e le c t in g
a N ew C ar
1 .0
S t y le
.3 1 9 6

- Civic .1160
- Saturn .2470
- Escort .0600
- Miata .5770

R e lia b ilit y
.5 5 8 4

- Civic
- Saturn
- Escort
- Miata

.3790
.2900
.0740
.2570

Fuel E conom y
.1 2 2 0

- Civic
- Saturn
- Escort
- Miata

.3010
.2390
.2120
.2480

Ranking of alternatives
Style Reliability
Civic
Saturn
Escort
Miata

.1160

Fuel
Economy
.3196

.3790 .3010

.2470
.0600

.2900 .2390
.0740 .2120

.5770

.2570

.2480

.5584
.1220

.3060
.2720
.0940
.3280

Handling Costs
Dangers of including Cost as another criterion
political, emotional responses?

Separate Benefits and Costs hierarchical trees


Costs vs. Benefits evaluation
Alternative with best benefits/costs ratio

Cost vs. Benefits

MIATA
CIVIC
SATURN
ESCORT

Cost

Normalized
Cost

Cost/Benefits
Ratio

$18K
$12K
$15K
$9K

.333
.222
.2778
.1667

.9840
1.3771
.9791
.5639

Complex decisions
Many levels of criteria and sub-criteria

Application areas

strategic planning
resource allocation
source selection, program selection
business policy
etc., etc., etc..

AHP software (ExpertChoice)


computations
sensitivity analysis
graphs, tables

Group AHP