Module 5

Applications of Fuzzy Systems

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 1

Contents
Fuzzy AHP
Fuzzy TOPSIS
Mamdani Fuzzy Inference Model
Tagaki Sugeno Fuzzy Inference Model

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 2

Fuzzy AHP
• The Fuzzy Analytic Hierarchy Process (Fuzzy AHP) is an extension of
the traditional Analytic Hierarchy Process (AHP), incorporating fuzzy
logic to handle the uncertainty and vagueness often present in
decision-making processes.

• Fuzzy AHP is a versatile tool that can be adapted to various complex

decision-making scenarios, providing a structured and systematic
approach to handle the inherent uncertainty in human judgments.

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 3

Applications of Fuzzy AHP
• Supply Chain Management
Supplier Selection: Evaluating and selecting the best suppliers based
on multiple criteria such as cost, quality, delivery performance, and
flexibility.

• Project Management
Project Selection: Prioritizing and selecting projects based on criteria
like cost, time, risk, and strategic alignment.

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 4

• Healthcare
Medical Decision Making: Assisting in medical diagnosis and
treatment selection by evaluating multiple diagnostic criteria and
treatment options.
Healthcare Facility Location: Choosing optimal locations for
healthcare facilities based on accessibility, population density, and
resource availability.

• Transportation
Route Selection: Optimizing transportation routes by considering
factors like distance, cost, traffic, and safety.

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 5

• Real Estate
Property Valuation: Evaluating and ranking real estate properties
based on location, price, amenities, and future value prospects.

• Education
Curriculum Development: Designing and prioritizing educational
programs and courses based on relevance, demand, and resource
availability.
Institutional Evaluation: Assessing and ranking educational
institutions based on criteria like academic performance,
infrastructure, and student satisfaction.

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 6

AHP (Analytical Hierarchy Process) Example

Criteria Price or Cost Storage space Camera Looks

Mobile 1 15000 16 GB 12 MP 5 Excellent
Mobile 2 11000 20 GB 8 MP 3 Poor
Mobile 3 16000 32 GB 20 MP 4 Better

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 7

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 8
Step 2: Pair-wise comparison matrix: It is created with the
help of scale of relative importance.
This will give relative importance of different attributes or criteria
with respect to goal. 1 Equal
Importance
3 Moderate
Importance
5 Strong
Importance
7 Very Strong
Importance
9 Extreme
Importance
2,4,6,8 Intermediate
Values
1/3,1/5, Values for
Fundamental scale [Saaty, 1980] 1/7,1/9 inverse
comparison
19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 9
 Price or Storage Camera Looks
Cost Space
Price or Cost is of a strong
Price or 1 5 4 7 importance than storage
Cost space
Storage 1/5 1 1/2 3
Space
Camera 1/4 2 1 3
Looks 1/7 1/3 1/3 1

1 Equal Importance
Price or Storage Camera Looks
Cost Space 3 Moderate Importance
Price or 1 5 4 7 5 Strong Importance
Cost 7 Very Strong
Storage 0.2 1 0.5 3 Importance
Space 9 Extreme Importance
Camera 0.25 2 1 3 2,4,6,8 Intermediate Values
Looks 0.14 0.33 0.33 1 1/3,1/5,1/7, Values for inverse
Sum 1.59 8.33 5.83 14 1/9 comparison
19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 10
 Price or Storage Camera Looks
Cost Space
Divide each with the respective
Price or 1/1.59 5/8.33 4/5.83 7/14
sum
Cost
Storage 0.2/1.59 1/8.33 0.5/5.83 3/14
Space
Camera 0.25/1.59 2/8.33 1/5.83 3/14
Looks 0.14/1.59 0.33/8.33 0.33/5.83 1/14
Sum 1.59 8.33 5.83 14

Price or Storage Camera Looks

Cost Space
Normalized Pair-wise
Price or 0.6289 0.6002 0.6861 0.5000
comparison matrix:
Cost
To bring the values to common
Storage 0.1258 0.1200 0.0858 0.2143 range
Space
Camera 0.1572 0.2401 0.1715 0.2143
Looks 0.0898 0.0400 0.0572 0.0714
19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 11
 Price or Storage Camera Looks Criteria
Cost Space Weights
Price or 0.6289 0.6002 0.6861 0.5000 0.6038 Average of Row
Cost
Storage 0.1258 0.1200 0.0858 0.2143 0.1365
Space
Camera 0.1572 0.2401 0.1715 0.2143 0.1958
Looks 0.0898 0.0400 0.0572 0.0714 0.0646
Step 3: Calculating the Consistency: To check the calculated value is consistent or not
Criteria 0.6038 0.1365 0.1958 0.0646
weights
Price or Cost Storage Space Camera Looks Consider the same
Price or 1 5 4 7 pairwise matrix which is
Cost not normalized

Storage 0.2 1 0.5 3

Space
Camera 0.25 2 1 3
Looks 0.14 0.33 0.33 1
19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 12
Criteria 0.6038 0.1365 0.1958 0.0646
Weights
Price or Cost Storage Space Camera Looks Multiply each value in
the column with the
Price or 1 * 0.6038 5 * 0.1365 4 * 0.1958 7 * 0.0646
criteria value
Cost
Storage 0.2 * 0.6038 1 * 0.1365 0.5 * 0.1958 3 * 0.0646
Space
Camera 0.25 * 0.6038 2 * 0.1365 1 * 0.1958 3 * 0.0646
Looks 0.14 * 0.6038 0.33 * 0.1365 0.33 * 0.1958 1 * 0.0646

Criteria 0.6038 0.1365 0.1958 0.0646

Weights
Price or Cost Storage Space Camera Looks
Price or 0.6038 0.6825 0.7832 0.4522
Cost
Storage 0.1208 0.1365 0.0979 0.1938
Space
Camera 0.1510 0.2730 0.1958 0.1938
Looks
19-07-2024
0.0863 0.0455 0.0653
Prof. Trupthi Rao, Dept. of AI&DS, GAT
0.0646 13
 Price or Storage Camera Looks Weighted
Cost Space sum value
Price or 0.6038 0.6825 0.7832 0.4522 2.5217
Cost
Storage 0.1208 0.1365 0.0979 0.1938 0.5490 Addition of
Space entire row
Camera 0.1510 0.2730 0.1958 0.1938 0.8136
Looks 0.0863 0.0455 0.0653 0.0646 0.2616

Price or Storage Camera Looks Weighted Criteria Weighted sum

Cost Space sum value weights value/criteria
weights
Price or 0.6038 0.6825 0.7832 0.4522 2.5217 0.6038 4.1763
Cost
Storage 0.1208 0.1365 0.0979 0.1938 0.5490 0.1365 4.0219
Space
Camera 0.1510 0.2730 0.1958 0.1938 0.8136 0.1958 4.1552
Looks 0.0863 0.0455 0.0653 0.0646 0.2616 0.0646 4.0495

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 14

 Price or Storage Camera Looks Weighted Criteria Weighted sum
Cost Space sum value weights value/criteria
weights
Price or 0.6038 0.6825 0.7832 0.4522 2.5217 0.6038 4.1763
Cost
Storage 0.1208 0.1365 0.0979 0.1938 0.5490 0.1365 4.0219
Space
Camera 0.1510 0.2730 0.1958 0.1938 0.8136 0.1958 4.1552
Looks 0.0863 0.0455 0.0653 0.0646 0.2616 0.0646 4.0495

Lamda = (4.1763+4.0219+4.1552+4.0495)/4 = 4.1007

Here n=4, as we have 4 criteria

Consistency Index (C.I) = (Lamda – n)/n-1

= (4.1007 – 4)/4-1

= 0.1007/3

19-07-2024 = 0.0335 Prof. Trupthi Rao, Dept. of AI&DS, GAT 15

Consistency Ratio (C.R) = C.I/R.I
where C.I = Consistency Ratio, RI= Random Index

R.I is the consistency index of randomly generated pairwise matrix.

C.R = 0.0335 / 0.9

= 0.037311

0.10 is the standard value we compare.

If C.R value is less than 0.10 then criteria weights are

acceptable.

If C.R value is greater than 0.10, then re-evaluate the pairwise

comparison.

Since 0.037311 < 0.10, the pair-wise comparison matrix is

reasonably Consistent.

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 16

The criteria with the highest weight will be assigned the highest rank.

Criteria Criteria Rank

weights
Price or 0.6038 1
Cost
Storage 0.1365 3
Space
Camera 0.1958 2
Looks 0.0646 4

The criteria with the highest weight is Price or Cost. Therefore, the mobile with lowest Price will
be chosen. Hence Mobile 2 with Price 11000 is chosen.

19-07-2024 17
Fuzzy AHP – Using Geometric Mean

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 18

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 19
Step 2: Pair-wise comparison matrix: It is created with the help of scale of
relative importance.
This will give relative importance of different attributes or criteria with respect
to the goal.

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 20

 Price or Storage Camera Looks
Cost Space
Price or (1,1,1) (4,5,6) (3,4,5) (6,7,8)
Cost
Storage 1/5 (1,1,1) 1/2 (2,3,4)
Space
Camera 1/4 (1,2,3) (1,1,1) (2,3,4)
Looks 1/7 1/3 1/3 (1,1,1)

Fuzzified pairwise
For reciprocals comparison matrix

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 21

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 22
19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 23
19-07-2024 24
De-Fuzzification process

(0.428+0.610+0.859)/3 = 0.633

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 25

 Criteria Rank
weights
Price or 0.601 1
Cost
Storage 0.138 3
Space
Camera 0.197 2
Looks 0.065 4

Since Price is less ($200) for Mobile 2, it is chosen.

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 26

Fuzzy TOPSIS
• The FUZZY TOPSIS (Technique for Order Preference by Similarity to Ideal Solution)
method is an extension of the classical TOPSIS method, incorporating fuzzy set
theory to handle the uncertainty and vagueness in the decision-making process.
• This method is particularly useful when dealing with qualitative and subjective
criteria that are difficult to quantify precisely.

Applications:
• Supply Chain Management: Supplier selection and evaluation.
• Engineering: Material selection and process optimization.
• Healthcare: Medical diagnosis and treatment planning.
• Environmental Management: Assessment of environmental impact.
• Marketing: Product development and customer satisfaction analysis.
19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 1
Steps:
• Construct the Fuzzy Decision Matrix
• Normalize the Fuzzy Decision Matrix
• Determine FPIS (Fuzzy Positive Ideal Solution) and FNIS (Fuzzy Negative Ideal
Solution)
• Calculate the Distance of every alternative from FPIS and FNIS
• Calculate the Closeness Coefficient
• Rank the Alternatives (alternative with the highest closeness coefficient is the
best alternative)

19-07-2024 Prof. Trupthi Rao, Dept. of AI&DS, GAT 2

 YOUYA
Date.

Fuy TOPSTS \Techotque fon Cndesn

Pederente by Similasity. to Tdeal
Ogolution:
This nethad aaks using fuzzs Jogic and
helps in chooring the be't alternative ina
mticaíteGata "system.
The steps învolved in thit method ane as
ollowr:
G(onstruct the &uzzy decitton mat six [e DM
@Nomalfze the FM.
9 Detesmine FPIS fuzzy forttive Tceal
Solution & FNTSy
Tdeal solution)
@(alulate the dirtance f6n ol eveRy alBeunciv
forn FIS_and FNTS
O Calculate the clore ngr coe icient Cccl
Rank the alternatives Khe alteinative
Wth highest cc ehauá be auiancá the
highert 'nank
0 Find the best altenative don the qtven éu74
deciston motix ntng TO PsIS 'method
(atteial Aeinive A3
(3,4,s) (4,s,6) (,3,4)
|(I,23) (3,4,5) (4,5,6)
(O,34)3,4,s) (s,6,t)
 W

Page No.:

Date: YOUVA

Shepa Nbrmali~e FDM.

IRvtde exexs kuzy value in the FoM by
its maximom value. Moimum value is to
be charen auith mepact to every ceiteta
Maximum value unden'S:(4,s,6) e
Maximum yalue unet (y4,s )
MaximumNolue unde (a: (s, 6,1)
Nomalied FDM.
(oittia A4

Conteaia A

(01S, 0:3,43333) ( ) (os, G-6 0-661

2o-25, 0-4, 0-s) 0-45, 0:30B) (, 1, )
(o.4,oso-54140-60-6663, J,,1)
0.4143l
 08364 dt-
02649 03908t O.0438+
Qs (o4-)²+
3
3-1)
(og-12+Q 4s-D'
1s-)
dt
b,) , 5) a
dirtancc ucltdcian uring. Calcule
FNTS. &
alalate Gom Stiopa
a Hane
of dis FPTS
O.s) O-f, (O:4, fNI
o4,0s) (o-2, FNTS=
lI,1,) FPIS=
66) 10- o-6, (O-s, FNTS
(l,o1,) FfIS 3 C Fo
Coiteata. yey value
în
in wog to.
FNIS andcsiteria. each in
e PIScO1esponde
best to DtStep3:
YOUVA
Date:
No.: Page
S
dz a7s-0) fo3-0s03-0
3

(o4-0.6+los
o 66)

3
(os3 -0 -66)2
d,- 0- 3094
Fon Aa:

t
3 o-t-ifo-s-tifo
NO+o: O438 0- 1218
204070
 T
Paga No.:
Date: YOUNA
do

tIfot5-o21 4(o-4-o 4-4(0g2

3

tfos-oas)+fo: 66-oafo1
d 0:8003.

(o-6-0'+ 66 -)

dgt o- 4185
dz
3
4:o4)'+l-0s+(-0.9
3

-0-4) +(i-0r)? -o1a)

d3 o-13 43
 Date:

Sttpr alculateroruses o-tAciert (cc)

di +d;t

Fon A

Q3096tE
diitdit 0-3096 t08364

0-2003
datd,t 0-8003+0. 40o
CGz 0 66 28

O-1343
d td,t O-7343+ 014185
 M WT
Paga No.:
YOUYA
Date:

: Rank the altenatives.

AlHeinaive Rank
A 1

A ir the best altennative.