DESIGN AND IMPLEMENTATION OF A FUZZY EXPERT

DECISION SUPPORT SYSTEM FOR VENDOR SELECTION

Case Study in OIEC IRAN(Oil Industerial Engineering and Construction)

Maryam Ramezani, G. A.Montazer

Information Technoldy Department,Tarbayat Modares University, Ale Ahmad Street,Tehran, Iran

Keywords: Vendor Selection, Fuzzy logic theory, Decision Support System.

Abstract: Supplier selection and evaluation is a complicated multi objective process with many uncertain factors.

Sealed bid evaluation is the most common approach for supplier selection purpose in Iran. In this paper, a

fuzzy expert decision support system is developed for solving the vendor selection problem with multiple

objectives, in which some of the parameters are fuzzy in nature. Basic important factors considered for

supplier selection are price, quality and delivery time. The designed system has been designed and

implemented and evaluated in a lead famous company and the results are discussed.

1 INTRODUCTION

The explosive growth in business to business

commerce is expected to revolutionize the

transaction process between buyers and sellers.

Effective purchasing and supply management can

contribute significantly to the success of most

organizations (Johnson, 2002). Procurement and

supply management are one of the most significant

parts in EPC (Engineering, Procurement, and

Construction

) contracts. When a supplier selection

decision needs to be made, the buyer generally

establishes a set of evaluation criteria that can be

used to compare potential sources. The basic criteria

typically utilized for this purpose are pricing

structure, delivery (lead-time and reliability),

product quality, and service (i.e., personnel,

facilities, research and development, capability,

etc.). Frequently, these evaluation criteria conflict

with one another. In addition, the importance of each

criterion varies from one purchase to the next. This

situation can be more complicated further by the fact

that some of the criteria are primarily quantitative

(price, quality, etc.) and some are qualitative

(service, etc.).(Garfamy, 2004) The literature on

supplier selection spans over three decades and

covers virtually all the aspects of business.

Researchers have long sought to understand and

model the relationships between suppliers and

buyers (Bhutta, 2001).

In this paper, vendor

selection process is simulated by use of expert

systems and fuzzy theory. The paper is organized as

follows. In Section 2, we present the importance of

the mentioned subject and a brief review of the

literature on vendor selection process and the main

approaches being executed in Iran. Section 3 gives

an introduction to fuzzy and expert DSS system as

well as fuzzy expert system architecture. Section 4

describes the design and development of the system.

Section 5 describes the implementation of the

system in one of the famous oil companies in Iran

and finally Section 6 concludes the paper.

2 LITERATURE REVIEW

The source-selection decision is highly complex and

purchaser’s most difficult responsibility. First, such

a decision involves more than one selection criterion

when choosing among the available suppliers.

Second, criteria included in the supplier selection

process may frequently contradict each other (lowest

price against a poor quality).

Third complication surrounding the supplier

selection decision arises from internal policy

constraints and externally imposed system

constraints placed on the buying process. Fourth, as

organizational requirements and market conditions

change, the importance of the analysis of tradeoffs

among the selection criteria may be increased

(Weber, 2000; Weber,2000a). Garfamy classifies the

main Supplier selection criteria as cost, quality,

cycle time, service, relationship, organization

(Garfamy, 2004) which every criterion is composed

243

Ramezani M. and A. Montazer G. (2006).

DESIGN AND IMPLEMENTATION OF A FUZZY EXPERT DECISION SUPPORT SYSTEM FOR VENDOR SELECTION - Case Study in OIEC IRAN(Oil

Industerial Engineering and Construction).

In Proceedings of the Eighth International Conference on Enterprise Information Systems - AIDSS, pages 243-248

DOI: 10.5220/0002498202430248

 SciTePress

of different factors. For example cost factors are

price , logistics costs (transportation, inventory,

administration, customs, risk and damage, handling

and packaging), operating costs, after sales service

costs. (Bhutta, 2001) reviews the status of

methodology literature in supplier selection, a total

of 154 papers from 68 refereed journals are

reviewed and classified into various categories such

as Mathematical Models, Criteria, Case Study,

Literature review, Conceptual. (Kumara, 2004) has

formulated a vendor selection problem as a fuzzy

mixed integer goal programming vendor selection

problem that includes three primary goals:

minimizing the net cost, minimizing the net

rejections, and minimizing the net late deliveries.

There are some restrictive assumptions in the

aforementioned formulating; For example, only one

item is supposed to be purchased from one vendor.

Also, (Kumar a, 2005) formulated Vendor selection

problem as a fuzzy Multi-objective Integer

Programming incorporating three important goals:

cost-minimization, quality-maximization and

maximization of on-time-delivery-with the realistic

constraints such as meeting the buyers’ demand,

vendors’ capacity, vendors’ quota flexibility, etc. In

the proposed model, various input parameters have

been treated as vague data with a linear membership

function of fuzzy type with the same restriction

pointed above. However, each company selects its

own special criteria and a unique approach for

vendor selection. In here some applicable common

approaches in Iran will be described.

2.1 Common Vendor Selection

Approaches in Iran

Sealed bid evaluation is most common approach for

vendor selection in Iran. The common procedure is

that first technical scoring will be done based on the

technical or quality evaluation. In the quality

evaluation Step vendor’s capacity for performing the

projects is estimated base on such factors as work

experience, management staff, technical staff,

manufacturing abilities, financial abilities, and good

background in other projects, creativity and

innovation, among others. Technical evaluation is

based on such criteria as exact consideration of

buyer or client technical request, complete vendor

documents, consideration of international standards,

quality of installation and supervision and other

technical factors.

The technical and commercial committee

estimates the technical score of each vendor based

on the abovementioned criteria. Vendors obtaining

higher technical score than a specific threshold are

approved technically and their commercial quotation

will be unsealed. In this Step, all quotations will be

apple to apple based on special declared conditions.

One of the common approaches for sake of making

the quotations apple to apple is that the offered price

will be divided by the technical score. Another

approach is to consider a ratio for technical and

commercial, for example 30 for commercial and 70

for technical score. Obviously, the ratio can be

different depending on the conditions of each

project.

The above-mentioned approaches are popular

methods in the governmental companies. In many

private ones which do not allow this status, such

other methods are used that in many cases, technical

evaluation is done by accept or reject and no scoring

methods are done. In this way, the lowest price is the

winner although the difference in price may be much

less valuable than the difference in quality. Thus,

decision making for selecting the right vendor is

complicated and time consuming job which needs a

committee of technical and commercial experts.

Decision making in these committees are based on

linguistic criteria. As an illustration, the price of a

proposal is “high” and the other is “very high”.

3 FUZZY EXPERT SYSTEM

An expert system is a computing system capable of

representing and reasoning about some knowledge-

rich domain with a view to solving problems and

giving advice(Jackson, 1990).

Fuzzy set theory provides a framework for handling

the uncertainties. (Zadeh, 1965) initiated the fuzzy

set theory.(Bellman, 1970) presented some

applications of fuzzy theories to the various

decision-making processes in a fuzzy environment.

In fuzzy sets every object is to some extent member

of a set and to some extent it is member of another

set. Thus, unlike the crisp sets membership is a

continuous concept in fuzzy sets. Fuzzy is used in

cases which the variables are linguistic and there is

uncertainness in the problem. Fuzzy expert decision

support system is an expert system that uses fuzzy

logic instead of Boolean logic. It can be seen as

special rule-based systems that use fuzzy logic in

their knowledge base and derive conclusions from

user inputs and fuzzy inference process (Kandel A,

1992) while fuzzy rules and the membership

functions make up the knowledge base of the

system. In other words a “fuzzy if-then” rule is a “if-

then” rule which some of the terms are given with

continuous functions.(Li-Xin,Wang 1994)

ICEIS 2006 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

244

Figure 1: Architecture of Fuzzy Expert Decision Support System for Vendor Selection.

Most common fuzzy systems are: pure fuzzy

systems, Takagi-Sugeno-Kang (TSK) and fuzzy

system with fuzzifying and defuzzifying parts(Li-

Xin Wang, 1994) .Since in the system developed in

this paper the input and output are real numbers, the

last kind is used. This system has a fuzzifier module

in the input that changes the real numbers to fuzzy

sets and a defuzzifier module in the output that

changes the fuzzy sets to real numbers. The

architecture of the system is composed of three main

blocks as shown in figure 1.

A-Fuzzy inference engine: A program which

analyzes the rules and knowledge aggregated in the

database and finds the logical result. There are

different selection for the fuzzy inference engine

depending on the aggregation, implication and

operators used for s-norm and t-norms (Li-Xin

Wang, 1994)

B- User Interface: Users of this system are

organizational decision makers that enter the real

number of all variables via user interface. Also, user

interface shows the result scoring. C-Fuzzy rule

base: Experts’ experience is used to build up the

fuzzy rules. These rules are conditional statements

and can be represented as “IF x is Xi and y is Yi and

… THEN o is Oi” Where x and y are linguistic input

variables. Xi and Yi are possible linguistic values for

x and y; respectively.

4 DESIGN OF THE FUZZY

EXPERT SYSTEM

The goal of a fuzzy expert DSS is to take in

subjective, partially true facts randomly distributed

over a sample space, and build a knowledge-based

expert system to produce useful decisions(Vadiee N,

1994). The overview of the framework is shown in

Fig 2.There are 7 fundamental Steps in the

development of a fuzzy expert DSS. Details of these

Steps are as follows:

4.1 Identification and Analysis of the

Problem

As mentioned above, in many bids in different

organizations the winner is selected just by the price

factor and other important factors such as ‘quality’

and ‘delivery time’ are not considered.

4.2 Identification of Critical Factors

and Membership Functions

This Step involves the compilation of a list of

critical factors based on a literature review and in

depth interviews with expert people who are

involved in the procurement and bid evaluation

process. This survey shows that there are three

importance factors for vendor selection which are of

great customer consideration. They are price, quality

and delivery time.

DESIGN AND IMPLEMENTATION OF A FUZZY EXPERT DECISION SUPPORT SYSTEM FOR VENDOR

SELECTION - Case Study in OIEC IRAN(Oil Industerial Engineering and Construction)

245

4.3 Fuzzy Rules Construction

Fuzzy expert DSS makes decisions and generate

output values based on knowledge provided by the

designer in the form of IF _condition_ THEN

_action_ rules. The rule base specifies qualitatively

how the output parameter “overall rating” of the

vendor proposal is determined for various instances

of the input parameters of “price”, “quality” and

“delivery time”.There will be nine rules out of our

depth interviews as below:

Rule 1: IF “Wanted Price” is cheap AND

“Vendor Price” is cheap THEN “Price Matching” is

high.

Rule 2: IF “Wanted Price” is cheap AND

“Vendor Price” is moderate THEN “Price matching”

is medium.

Rule 3: IF “Wanted Price” is cheap AND

“Vendor Price” is high THEN “Price matching” is

low.

Rule 4: IF “Wanted Price” is moderate AND

“Vendor Price” is cheap THEN “Price matching” is

medium.

Rule 5: IF “Wanted Price” is moderate AND

“Vendor Price” is moderate THEN “Price matching”

is high.

Rule 6: IF “Wanted Price” is moderate AND

“Vendor Price” is expensive THEN “Price

matching” is medium.

Rule 7: IF “Wanted Price” is expensive AND

“Vendor Price” is low THEN “Price matching” is

low.

Rule 8: IF “Wanted Price” is expensive AND

“Vendor Price” is moderate THEN “Price matching”

is medium.

Rule 9: IF “Wanted Price” is expensive AND

“Vendor Price” is expensive THEN “Price

Matching” is high.

We will have the same inference for “quality” and

“delivery time”.

4.4 Fuzzification

Fuzzification refers to the process of taking a crisp

input value and transforming it into the degree

required by the terms(E.W.T. Ngai., 2003). The

“fuzzified” values are determined by intersecting the

input value to the fuzzy membership function. In the

present study triangular membership functions have

been used to define the fuzzy sets for the linguistic

values of “price”, “quality” and “delivery time”. The

same triangular membership functions have been

defined for “wanted Price”, “wanted quality” and

“wanted delivery time”. The membership function of

“price matching” indicates the degree of matching in

price between “price” and the customer’s “wanted

Price”. It takes “low”, “medium” and “expensive” as

its linguistic terms. The same approach is used to

define “quality matching” and “delivery time

matching”. Due to the fact that all input values are

normalized, fuzzification input will be between 0

and 1. For instance, an input value of “price” .4

results in a degree of membership in the set labeled

“cheap” of 0.25 and a degree of membership in the

set labeled “medium” of 0.75 (see Fig. 3)

4.5 Fuzzy Inference Generation

Fuzzy inference is guided by the fuzzy rules. The

standard max–min inference algorithm was used in

the fuzzy inference process, as it is a commonly

used fuzzy inference strategy (E.W.T.Ngai.,

2003).Mamdani inference is used as equation 1:

Step 1: Identification and analysis of

the problem

Step 2: Identification of critical

factors and membership functions

definition

Step 3: Fuzzy rules construction

Step 4: fuzzification

Step 5: fuzzy inference module

generation

Step 6: defuzzification

Step7: comparison of the overall

rating for all vendors

Figure 2: Fuzzy Inference Process for vendor selection.

Figure 3: Fuzzy Membership Function for Price.

0.25

0.5

0.75

0 0.15 0.3 0.5 0.7 0.85 1

Normalized Price

MemberShip Functio

Chea

Mediu

ensive

(1)

ICEIS 2006 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

246

In the max–min composition fuzzy inference

method, the min operator is used for the AND

conjunction (set intersection) and the max operator

is used for the OR disjunction (set union) in order to

evaluate the grade of membership of the antecedent

clause in each rule.

4.6 Defuzzification

When the inference process is complete, the

resulting data for each output of the fuzzy

classification system are a collection of fuzzy sets or

a single, aggregate fuzzy set. The process of

computing a single number that best represents the

outcome of the fuzzy set evaluation is called

defuzzification (E.W.T. Ngai., 2003). There are

several existing methods that can be used for

defuzzification. These include the methods of

maximum or the average heights methods, and

others. These methods tend to jump erratically on

widely non-contiguous and non-monotonic input

values(Diego, 1999). We chose the centroid method,

also referred to as the “center-of-gravity (COG)”

method, as it is frequently used and appears to

provide a consistent and well-balanced approach.

(Klir ,G.j.T.A ,1998).

For each output using this defuzzification method,

the resultant fuzzy sets are merged into a final

aggregate shape and the centroid of the aggregate

shape computed.

4.7 Comparison of the Overall

Rating for All Vendors

The overall ratings for all vendors are obtained by

passing measures of their initial factors and

weightings through the proposed fuzzy logic model.

The final score is calculated via defuzzification. The

system finally ranks all vendors according to their

final scores and displays them in descending order.

5 IMPLEMENTATION AND

EVALUATION OF THE FUZZY

EXPERT DSS FOR VENDOR

SELECTION

Considering the fact that the system should be

capable of evaluating all bids whether big or small,

the ‘highest price’, ’highest quality’ and ‘longest

delivery time’ are used for normalizing inputs. Thus,

the designed fuzzy system will have inputs between

0 and 1. For using the same system for all bids in all

projects we define a weigh for each of the factors

price, quality and delivery time and the final score

that is a number between 0 and 1 will be the average

of the system outputs considering the factors.

A prototype system is designed by use of Matlab

fuzzy toolbox. Once the prototype system is built,

testing and evaluation of the prototype system can

be performed. The designed system was tested for

10 bids in OIEC for different bids. An example is

shown in table 1 for a bid evaluation. In this

example vendor3 has the most score and will be the

winner of the bid. In this example as it is clear

although vendor2 has less offered price, considering

the whole factors together vendor3 will be the best

selection.

Comparison of outcome with the decision of

transaction committee shows that the system works

properly and can be used instead of the transaction

committee. Evaluation is achieved through

interviews with the experts and users. We

particularly asked the potential users about the

effectiveness and the usability of the prototype

system. Also we asked them to tell us what they

considered to be the strengths and weaknesses of the

prototype system, and how it should be improved.

From fifteen interviews, 12 agreed that the proposed

expert system is seen to be a promising system for

supporting the selection of right vendor based on the

positive results of its evaluation and three of them

needed more time for more careful evaluations.

Table 1: An Example of the bid evaluation in the fuzzy expert decision support system.

Bid Evaluation

Price(

q =7)

Delivery

(

q =4)

Quality

(

q =5)

Price

Match

Delivery

Match

Quality

Match

Overall

Score

Offered 145,555,050 14 weeks 6 Vendor1

Normalize .7661 .56 .8571

.71 .564 .837 .7131

Offered 131,211,253 25 weeks 5 Vendor2

Normalize .6906 1 .7143

.91 .439 .595 .6938

Offered 141,022,800 7 weeks 5.5 Vendor3

Normalize .7422 .28 .7857

.76 .781 .686 .799

Offered 190,000,000 10 weeks 7 Vendor4

Normalize 1 .4 1

.51 .623 1 .6913

Wanted 130,000,000 8 weeks 7 Wanted

Normalize

.6842 .32 1

--- --- --- ---

DESIGN AND IMPLEMENTATION OF A FUZZY EXPERT DECISION SUPPORT SYSTEM FOR VENDOR

SELECTION - Case Study in OIEC IRAN(Oil Industerial Engineering and Construction)

247

6 CONCLUSIONS

This paper has described a new method for design of

a fuzzy expert DSS, which is used to assist

companies by facilitating vendor selection. Three

critical factors for vendor selection which are above

all considerable for the customer are price, quality

and delivery time. The designed system evaluates

the degree of match between vendor offer and

customer need and considering the importance of

each factor in each bid the final score of each vendor

is determined. Comparing this system with the

conventional one, the new system is much less time

consuming and the need to have different transaction

committee meetings for decision making is omitted.

Performance evaluation is done in a case study in a

lead oil company (OIEC) in Iran using expert

validation and prototype testing. The results of the

prototype evaluation are satisfactory and support the

view that the system has performed its functions as

expected.

REFERENCES

Bellman, R. E., & Zadeh, L. A. (1970). "Decision making

in a fuzzy environment." Management Sciences 17:

B141–B164.

Bhutta, M. K. S. (2001). "SUPPLIER SELECTION

PROBLEM: METHODOLOGY LITERATURE

REVIEW."

Darlington (2000). "The essence of expert system."

England, Prentice-hall.

E.W.T. Ngai., F. K. T. W. (2003). "Design and

development of a fuzzy expert system for hotel

selection." Omega (The international journal og

Mangment sience) 31: 275 – 286.

Klir ,G.j.,Folger, T.A. (1998). "Fuzzy set ,Uncertainty ,and

information." Prentice-Hall.

Garfamy, R. M. (2004). "Supplier Selection and

Business,Process Improvement:An exploratory

multiple-case study." Doctoral Program in Business

Creation, Strategy and Management,Faculty of

Economics and Business studies Department of

Business

Economics, University of Autnoma de barcelona.

Jackson P (1990). "Introduction to Expert Systems,2nd

edition, Addison-Wesley, England,."

Johnson, F. (2002). "Purchasing & Supply Management."

McGraw-Hill Twelfth Edition: 2-3.

Kandel A (1992). "Fuzzy expert systems." Boca Raton,

FL: CRC Press.

Kumara, P. V., R. Shankar (2004). "A fuzzy goal

programming approach for vendor selection problem

in a supply chain." Computer & Industerial

engineering 46: 69-85.

Li-Xin Wang (1994,). "A course in fuzzy systems and

control." Prentice Hall

Manoj Kumar a, P. V., Ravi Shankar (2005). "A fuzzy

programming approach for vendor selection problem

in a supply chain." Internatinal journal of Production

economics.

Matthews, C. (2003). "A formal specification for a fuzzy

expert system." Information and Software Technology

45: 419–429.

Metaxiotis, k. s., Psarras,J.E., Askounis,D.T. (2002).

"GENESYS: an expert system for production

scheduling." Industrial Management & Data Systems,:

309-317.

Nunamaker JF, C. M., Purdin DM. (1990). "Systems

development in information systems research." Journal

of Management Information Systems 7: 89–106.

Vadiee N, J. M. T. (1994). "he promising future of fuzzy

logic." IEEE Expert 9(4): 36–38.

Weber, C. A., Current, J. and Desai, A. (2000a). "An

optimization approach to determining the number of

vendors to employ,." Supply Chain Management: An

International Journal 2(2): 90-98.

Weber, C. A., Current, J. R. and Desai, A., (2000). " 2000,

VENDOR: A structured approach to vendor selection

and negotiation." Journal of Business Logistics, 21(1):

135-167.

Weber and John R. Current (2004), " A multiobjective

approach to vendor selection" in proceedings of

Second World Conference on POM and 15th Annual

POM Conference, Cancun, Mexico, April 30.

Zadeh, L. A. (1965). "Fuzzy sets." Information and

Control 8: 338–353.

ICEIS 2006 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

248