You are on page 1of 10

Development of Fuzzy Multi-Criteria Group

Decision Support System Software

• Muhammad ‘Ammar bin Awang


• Dr. Sundar Raju Raviraja
What is multi-criteria decision support
system (MCDSS)?

• A system that helps with making decisions under


multiple, but conflicting criteria.

• Provides a systematic procedure to help decision


makers choose the most desirable and satisfactory
alternative under uncertain situations.

• Example: Making decision to choose a new car.


Why we need MCDSS?
• A decision can be made according to the
decision maker’s preference. But if the
problem becomes more complicated,
computer aid is necessary.

• Relying on factual knowledge alone is not


always sufficient in making a decision.

• Value judgments, depending on the role and


the goals of each decision maker, are among
the critical issues requiring attention.
What is wrong with MCDSS?

• Most MCDSS only explored on numeric values but


ignore qualitative inputs or natural language which
are often used by decision makers to express
thinking.

• The decision maker needs this (qualitative inputs) to


assert the preferences.

• This might limit the capability of the system.


Can we improve the traditional
MCDSS?
• To maximize the usability of MCDSS, we must
include the qualitative inputs or natural
language to the system.

• But when using natural language as a label


for a set, the boundaries within the objects
which belong to the set become fuzzy.

• Therefore, fuzzy sets theory and natural


language is used to deal with this problem
Have fuzzy MCDSS been developed before?

• In 2009, Jasleen Kaur Sethi developed Fuzzy Logic based Decision Maker
System for choosing cars using MATLAB.
• The idea of this system is user can input the scale that is how good or bad
a car is according to different parameters. The scale can range for 0 to 5
with 0 for poor, 1 for average and so on.
• Next a fuzzy set containing the parameters and their scale of importance is
made. A score specifying the scale of importance is given to each
parameter for all cars in that category. Then implication process is done.
• In, implication, union of the complemented fuzzy set and the scale of
importance entered by the user is performed that is maximum value is
taken.
• Finally to get the output, the intersection operation for all the cars in that
category is performed that is a min function is applied across all relevant
members. The car with the maximum score becomes the best appropriate
car for the user.
• But Sethi’s system only involved one decision maker
only.

• In real world, sometimes when it comes to make a


decision, it involve more than one decision maker.

• In that situation, Sethi’s algorithm will not work.


• Also in 2009, during 22nd International Conference On
Industrial, Engineering and Other Applications of Applied
Intelligent Systems in Taiwan, Yeh et al. present a paper about
multi-criteria group decision support for information systems
project selection.

• This system involved multiple decision makers and each


decision maker have ‘power’ to determine the relative
importance of each evaluation criteria.

• This is perfect!
• But Yeh et al. system only focused on evaluating and
selecting IS projects for development needs.

• We cannot use their system to help us making


decision under different situation.
What we are going to do?
• We want to design and create a fuzzy MCDSS software
that can be useful within broad scope of problem-
solving scenarios, towards “universal DSS”.

• Using Yeh et al. algorithm, we will create a fuzzy


MCDSS software that can be used for multi-purpose
decision making processes.

• Users can determine the main and sub evaluation


criteria, their weights, and evaluate the performance
according to the number of decision makers and
evaluation weights of criteria.

You might also like