FUZZY

MATHEMATICAL Boudhayan Kundu

IMH/10070/19

PROGRAMMING
INTRODUCTION
Fuzzy Linear Programming (FLP) programs will appear when a kind of ambiguity is occurred in
the parameters of the model and/or in the lack of deterministic information in the classical linear
programming (LP) programs. In this case, the mentioned model will be established based on a
type of ambiguity instead of the crisp data such as fuzzy inputs and so on. In the literature of
fuzzy optimization, there are a lot of researches which are focused on fuzzy linear programming.
An interesting approach is using parametric programming methods in the fuzzy linear
programming. Afterward, various types of LP problems along with their solving methods have
been presented by different authors. Also, this is important to distinguish between flexibility and
uncertainty which appears in the parameters of models including the aim coefficients, the
coefficients matrix, and its constraints and also in the right-hand-side data. Flexibility concept is
modelled using fuzziness and reflects the concept that feasibility of the solutions for every
constraint will be valid based on its satisfaction degree. Furthermore, every constraint has an
accepted tolerance which is predetermined by the decision maker. Moreover, there is uncertainty
that concerns an objective variability in the desired model parameters (random uncertainty) or the
lack of some information of the parameter values.
VERDE GAY'S APPROACH
Verde gay (1982) showed that an LP problem with the crisp target and some fuzzy
constraints is equal to a common parametric LP model and thus, we are able to use
parametric approaches to solve these FLP problems. Cadenas and Verdegay (2000)
provided a multi-objective LP problem, where the objective coefficients in the
objective function appeared as fuzzy numbers, and also applied fuzzy ordering
approach to solve these models. Nasseri and Bavandi (2017) considered a Stochastic
Interval-Valued Linear Fractional Programming (SIVLFP) problem, where in their
model, the coefficients and scalars of the objective function are fractional intervals,
and technological coefficients and the quantities of the constraints in the mentioned
model were random variables with the specific distribution.
PROBLEM STATEMENT
Consider the following LPP problem that we are going to solve using fuzzy
mathematical programming using Verde gay’s Approach
Given objective function
Max :
Subject to the constraints:
FORMULATION OF
MEMBERSHIP FUNCTION
It is to be noted that all the constraints are fuzzy in nature and the tolerance
parameters are given as

We can formulate the membership functions for each constraint using the general
formula :
FOR EACH CONSTRAINT
𝜇 𝑖
Graph
𝜇1
of the membership function(s)

1 1

𝐴𝑥𝑖 𝐴𝑥 1
𝟎
0𝑏𝑖𝑏𝑖+ 𝑝 𝑖 16 21
𝜇 2 𝜇 3

1 1

𝐴𝑥2 𝐴𝑥3
0 70 110 0 90 120
The problem now takes the form : Substituting :

Where, Where,
SOLVING USING TORA
When ,
USING THE SIMPLEX TABLE
WE GET
Hence, we have :

Varying the value of in the range of we obtain the following table

0 0 12.75 0 3.25 102.75
0.1 0 13.11 0 3.39 105.64
0.2 0 13.47 0 3.53 108.53
0.3 0 13.83 0 3.67 111.42
0.4 0 14.19 0 3.81 114.31
0.5 0 14.55 0 3.95 117.20
0.6 0 14.91 0 4.09 120.09
0.7 0 15.27 0 4.23 122.98
0.8 0 15.63 0 4.37 125.87
0.9 0 15.99 0 4.51 128.76
1 0 16.35 0 4.56 131.65
THANK YOU