Multi-objective Linear

Programming Problems involving

Fuzzy Parameters

Multi-objective optimization problem occurs in various real life

applications such as Aerodynamic design, Medical decision making,
Industrial neural network design etc. In many models the data can rarely
be determined exactly with certainty and precision. We may consider the
intervals of real numbers and be sure that the data fluctuates in these
intervals. It is clear that precision in decision making is very important
and any error may give rise to high expenses in application. That is why
in such problems, the data is considered in the form of fuzzy numbers.

Bellman and Zadeh proposed the concept of decision making in fuzzy

environment. The first formulation of fuzzy linear programming was
presented by Zimmerman. A multi objective approach for fuzzy linear
programming problems was proposed by Tanaka et al. Pandian used sum
of objectives for solving multi-objective programming problems.

In this article, we propose a method to solve multi-objective linear

problems with all the involved parameters, cost coefficients in the
objective functions, technological coefficients in the constraints and the
resources, being represented by Triangular fuzzy numbers. The problem
is first converted into equivalent crisp form giving rise to multi-objective
linear problem which is solved then using Pareto’s optimality technique.

1
1.2 PRELIMINARIES

1.2.1 Definition: Let X be a classical set of objects which should be

evaluated with regards to a fuzzy statement. Then the set of ordered
pairs = ,
where :X is a fuzzy set in X. The evaluation function
is called the membership function or the grade of
membership of x in .

1.2.2 Definition: A Fuzzy set is called normalised fuzzy set if ,

=1

1.2.3 Definition: Let be a fuzzy set in X and [0,1] a real

number. Then the classical set = is called
-level set or -cut of .
= {x X │ (x)> α} is called strong -cut of .

1.2.4 Definition: A fuzzy set in a convex set X is called convex

if ,
( +(1- ) ) ≥ Min ( ), , [0,1] .

1.2.5 Definition: A convex normalised fuzzy set

on the real line R such that ,
(i) there exist exactly one with the membership degree
=1 and
(ii) is piecewise continuous in R, is called a fuzzy
number.

1.2.6 Definition: A convex normalised fuzzy set =

on the real line R is called a fuzzy interval if,

2
i) there exists more than one real number x with a
membership degree =1.
ii) is piecewise continuous in R .

1.2.7 Definition: A fuzzy number is a triangular fuzzy number

denoted by (c, l, r) where c, l, r are real numbers and c is the peak
value of and membership function is given by,

, for c-l ≤ x ≤ c

, for c ≤ x ≤ c +r

=0 , elsewhere

Consider two triangular fuzzy number and where,

= (𝑠1 , 𝑙1 ,𝑟1 ) and
= (𝑠2 , 𝑙2 ,𝑟2 )
+ = (𝑠1+ 𝑠2 , 𝑙1 + 𝑙2 , 𝑟1 + 𝑟2 )
= ( k𝑠1 , k𝑙1 , k𝑟1 ) , for k ≥ 0

1.2.8 Definition: The partial ordering ≤ is defined by

≤ iff. ,
s1 ≤ s2, s1-l1 ≤ s2-l2 , s1 + r1 ≤ s2+r2 .

1.2.9 Theorem: (Pareto’s Optimality) For a problem with k

objective functions, The point x* X is a Pareto Optimal solution
if there does not exist x X such that if,
≥ for all i and for at least one j.

3
2.1 Linear Programming Problems with Fuzzy
Parameters

The Linear Programming problem with fuzzy parameters are given as ,

Maximize 𝑍̃ = ∑𝑛𝑗=1 𝑐̃ j xj

Such that ,

∑𝑛𝑗=1 𝑎̃ij xj ≤ 𝑏̃i , For i=1, 2,3, ..... m,

xj ≥ 0, j =1,2,3....n …….. (I)

Where 𝑐̃ j ,𝑎̃ ij , 𝑏̃ i are fuzzy numbers and xj are fuzzy variables whose
states are fuzzy numbers.

Considering the fuzzy parameters as triangular fuzzy numbers the

problem (I) can be given as ,

Maximizing 𝑍̃ = ∑𝑛𝑗=1(𝑐𝑠, 𝑐𝑙, 𝑐𝑟)j xj

Such that,
∑𝑛𝑗=1(𝑎𝑠, 𝑎𝑙, 𝑎𝑟)ij xj ≤ (bs,bl,br)i , i =1, 2,3....m

xj ≥ 0 , j=1, 2,3... n …….. (II)

where (cs, cl, cr)j s the jth fuzzy coefficient in the objective function,
(𝑎𝑠, 𝑎𝑙, 𝑎𝑟)𝑖𝑗 is the fuzzy coefficient of jth variable in the ith constraints,
(bs,bl,br)i is the ith fuzzy resource .

4
These problem can be solved by converting (II) into equivalent crisp
multi –objective linear problem as given below,

Maximize (Z1, Z2, Z3)

Where,

𝑍1 =∑𝑛𝑗=1 𝑐𝑠𝑗 𝑥𝑗 , 𝑍2 =∑𝑛𝑗=1 𝑐𝑙𝑗 𝑥𝑗 , 𝑍3 =∑𝑛𝑗=1 𝑐𝑟𝑗 𝑥𝑗

Such that,
∑𝑛𝑗=1 𝑎𝑠ij xj ≤ bsi , i= 1, 2,3...m

∑𝑛𝑗=1(𝑎𝑠ij – alij)xj ≤ bsi – bli , i= 1, 2, 3...m

∑𝑛𝑗=1(𝑎𝑠ij + arij)xj ≤ bsi + bri ,i= 1, 2, 3...m

xj ≥ 0, j=1, 2, 3... n.

5
2.2 Multi-Objective Fuzzy linear Programming
Problem

The general form of Multi-Objective Linear Programming Problem is

given as ,

Maximize 𝑍̃1, 𝑍̃ 2, 𝑍̃3, ….,𝑍̃k

Where, 𝑍̃k = ∑𝑛𝑗=1 𝑐̃𝑗𝑘 xj , k= 1,2,….,k

Such that ,

∑𝑛𝑗=1 𝑎̃ij xj ≤ 𝑏̃i , For i=1,2,3, .....m

xj ≥ 0, j =1,2,3....n. ……..(III)

Where 𝑐̃𝑗𝑘 , 𝑎̃𝑖𝑗 , 𝑏̃i are fuzzy numbers and xj are fuzzy variables whose
states are fuzzy numbers.

Considering the objective function coefficient, technological coefficient

and resource coefficient as triangular fuzzy numbers (s, l, r) , the problem
(III) can be rewritten as,

Maximize 𝑍̃1, 𝑍̃ 2, 𝑍̃3, ….,𝑍̃k

Where, 𝑍̃k =∑𝑛𝑗=1(𝑐𝑠, 𝑐𝑙, 𝑐𝑟)𝑘𝑗 xj , For k= 1,2,….,k

Such that ,
∑𝑛𝑗=1(𝑎𝑠, 𝑎𝑙, 𝑎𝑟)𝑖𝑗 xj ≤ (𝑏𝑠, 𝑏𝑙, 𝑏𝑟)𝑖 ,For i=1,2,3, .....m

xj ≥ 0 , j =1,2,3....n. ……..(IV)

6
(𝑐𝑠, 𝑐𝑙, 𝑐𝑟)𝑘𝑗 is the jth fuzzy coefficient in the kth objective function
(𝑎𝑠, 𝑎𝑙, 𝑎𝑟)𝑖𝑗 is the fuzzy coefficient of the jth variable in the ith
constraint, (𝑏𝑠, 𝑏𝑙, 𝑏𝑟)𝑖 is the ith fuzzy resource.

7
2.3 Solution for Multi-Objective Fuzzy Linear
Programming Problem

Here there are k fuzzy objective function 𝑍̃1, 𝑍̃ 2, 𝑍̃ 3, ….,𝑍̃k and m fuzzy
constraints 𝐺̃ 1 , 𝐺̃ 2 , 𝐺̃ 3 ,…., 𝐺̃ m where ,

𝐺̃ i = ∑𝑛𝑗=1 𝑎̃ij xj ≤ 𝑏̃i , for i=1,2,3, .....m.

The resulting fuzzy decision is given as,

𝑍̃1 ∩ 𝑍̃2 ∩ 𝑍̃3 ∩….∩ 𝑍̃k ∩ 𝐺̃ 1 ∩ 𝐺̃ 2 ∩…∩ 𝐺̃ m.

In terms of corresponding membership values for the fuzzy objectives and

fuzzy constraints the decision is ,

= ,

An Optimal solution X* is one at which the membership function of the

resultant decision is maximum.

i.e. =max ( , )

In our considered problem (IV) the coefficient of objective function is

represented by triangular fuzzy numbers. Therefore each objective function
gives rise to 3 crisp objective functions. Hence the converted problem will
involve 3k objective function to be optimized as below,

Maximize (𝑍11 , 𝑍21 , 𝑍31 , 𝑍12 , 𝑍22 , 𝑍32 ,….., 𝑍2𝑘 , 𝑍3𝑘 )

Where, 𝑍1𝑘 = ∑𝑛𝑗=1(𝑐𝑠)𝑘𝑗 xj ;

𝑍2𝑘 = ∑𝑛𝑗=1(𝑐𝑙)𝑘𝑗 xj ;

𝑍3𝑘 = ∑𝑛𝑗=1(𝑐𝑟)𝑘𝑗 xj ;

Such that , ∑𝑛𝑗=1 𝑎𝑠𝑖𝑗 xj ≤ 𝑏𝑠𝑖 , i= 1,2,3...m

∑𝑛𝑗=1( 𝑎𝑠𝑖𝑗 − 𝑎𝑙𝑖𝑗 ) xj ≤ 𝑏𝑠𝑖 − 𝑏𝑙𝑖 , i= 1,2,3...m

8
 ∑𝑛𝑗=1( 𝑎𝑠𝑖𝑗 + 𝑎𝑟𝑖𝑗 ) xj ≤ 𝑏𝑠𝑖 + 𝑏𝑟𝑖 , i= 1,2,3...m

xj ≥ 0 j=1, 2, 3... n

The weighted objective function of the above problem using pareto’s

method can be represented as,

Max 𝑤11 𝑍11 +𝑤12 𝑍21 +𝑤13 𝑍31 +𝑤21 𝑍12 +𝑤22 𝑍22 +…..+𝑤𝑘3 𝑍3𝑘

Where, 𝑍1𝑘 = ∑𝑛𝑗=1(𝑐𝑠)𝑘𝑗 xj ;

𝑍2𝑘 = ∑𝑛𝑗=1(𝑐𝑙)𝑘𝑗 xj ;

𝑍3𝑘 = ∑𝑛𝑗=1(𝑐𝑟)𝑘𝑗 xj ;

Such that , ∑𝑛𝑗=1 𝑎𝑠𝑖𝑗 xj ≤ 𝑏𝑠𝑖 , i= 1,2,3...m

∑𝑛𝑗=1( 𝑎𝑠𝑖𝑗 − 𝑎𝑙𝑖𝑗 ) xj ≤ 𝑏𝑠𝑖 − 𝑏𝑙𝑖 , i= 1,2,3...m

∑𝑛𝑗=1( 𝑎𝑠𝑖𝑗 + 𝑎𝑟𝑖𝑗 ) xj ≤ 𝑏𝑠𝑖 + 𝑏𝑟𝑖 , i= 1,2,3...m

xj ≥ 0 , j= 1,2,3,….,n ……. (V)

Solution can be obtained by solving the problem with different weights.

9
2.4 Numerical Example

Consider the Multi-objective Fuzzy Linear Problem ,

Maximize 𝑍̃1 , 𝑍̃ 2

Where, 𝑍̃1 = 𝑐̃1 x1 + 𝑐̃2 x2

𝑍̃ 2 = 𝑐̃3 x1 + 𝑐̃4 x2

Subject to the constraints

(3,2,1) x1+(6,4,1) x2 ≤ (13,5,2)

(4,1,2) x1+(6,5,4) x2 ≤ (7,4,2)

Where 𝑐̃1 =(7,10,14) ,𝑐̃2 = (20,25,35),𝑐̃3 =(10,14,25),𝑐̃4 =(25,35,40) ;

Where the membership function of the 𝑐̃1 , 𝑐̃2 , 𝑐̃3 , 𝑐̃4 are,
𝑥−7
µ𝑐̃1 (x) = , for 7 < x ≤ 10
3

14−𝑥
= , for 10 < x ≤ 14
4

= 0 , otherwise
𝑥−20
µ𝑐̃2 (x) = , for 20 < x ≤ 25
5

35−𝑥
= , for 25 < x ≤ 35
10

= 0 , otherwise
𝑥−10
µ𝑐̃3 (x) = , for 10 < x ≤ 14
4

25−𝑥
= , for 14 < x ≤ 25
9

= 0 , otherwise

10
 𝑥−25
µ𝑐̃4 (x) = , for 25 < x ≤ 35
10

40−𝑥
= , for 35 < x ≤ 40
5

= 0 , otherwise

It is equivalent to solving the Multi objective linear programming

problem,
7𝑥1 + 20𝑥2 , 10𝑥1 + 25𝑥2 ,
Maximize ( )
14𝑥1 + 35𝑥2 , 25𝑥1 + 40𝑥2
Subject to the constraints ,

3x1+6x2≤ 13

x1+2x2≤ 8

4x1+7x2≤ 15

4x1+6x2≤ 7

3x1+ x2 ≤ 3

6x1+10x2≤ 9

x1 , x2 ≥ 0

This is same as solving ,

Maximize w1(7𝑥1 + 20𝑥2 ) + w2(10𝑥1 + 25𝑥2 ) + w3(14𝑥1 + 35𝑥2 ) +

w4(25𝑥1 + 40𝑥2 )

Subject to the constraints,

3x1+6x2≤ 13

x1+2x2≤ 8

11
 4x1+7x2≤ 15

4x1+6x2≤ 7

3x1+ x2 ≤ 3

6x1+10x2≤ 9

x1 , x2 ≥ 0

Such that ∑ wi = S , i = 1,2,3,4.

Standard Optimization technique is used to solve the problem by

using different weights in such that their sum S is 2 and solving the resultant
single objective Linear programming problem.

For example, by taking w1 = w4 =0 and w2 = w3=1 we get,

Multi objective linear programming problem,

Maximize (w) = 24x1+60x2

Subject to the constraints,

3x1+6x2≤ 13

x1+2x2≤ 8

4x1+7x2≤ 15

4x1+6x2≤ 7

3x1+ x2 ≤ 3

6x1+10x2≤ 9

x1 , x2 ≥ 0

Solving we get, (𝑥1∗ , 𝑥2∗ ) =(0,0.9).

12
 Optimum solution is obtained as ,

f(𝑥1∗ , 𝑥2∗ ) = 0.9(𝑐̃2 , 𝑐̃4 )

𝑥−40.5
µ𝑓(𝑥1∗ ,𝑥2∗ ) (x) = , for 40.5 < x ≤ 54
13.5

67.5−𝑥
= , for 54 < x ≤ 67.5
13.5

=0 , otherwise

Following table lists the solution for above Multi objective linear
programming problem, for various weights and it also shows that the
solutions are independent of weights (wi ,i =1,2,3,4) ,

Sr. No (𝒙∗𝟏 , 𝒙∗𝟐 )

w1 w2 w3 w4
1 0 1 1 0 (0, 0.9)

2 0 1 0.5 0 (0, 0.9)

3 0.2 0.4 0.5 0.2 (0, 0.9)
4 0.1 0.2 0.3 0.4 (0, 0.9)
5 0 0.3 0 0.4 (0, 0.9)
6 0.2 0.4 0.6 0.8 (0, 0.9)

7 0.5 0 0.5 0 (0, 0.9)

8 0 1 1 0 (0, 0.9)

9 0 0 0 0.5 (0, 0.9)

10 0.3 0.1 1 1 (0, 0.9)
11 0.5 0.5 0.5 0.5 (0, 0.9)
12 0 0 0.5 0.5 (0, 0.9)
13 0.2 0.5 0.5 0.5 (0, 0.9)
14 0.1 0.2 0.3 0.4 (0, 0.9)

15 0 0.2 0 0.2 (0, 0.9)

13
2.5 CONCLUSION :

In this project, solution procedure of multi objective fuzzy linear

programming and its application were presented. Here the coefficient of
objective and constraint functions was represented as Triangular fuzzy
number. The problem is then converted to crisp linear programming
problem based on the concept of fuzzy decision and the fuzzy model under
fuzzy environments proposed by Bellman and Zadeh (1970).

14
 REFERENCES

[1] Buckley, J.J., Fuzzy Probabilities: New Approach and Application. Springer
(2005).
[2] Mangasarian, O.L., “Non-Linear Programming” Mc-Graw Hill Book Co. New
York. (1969).
[3] Zadeh, L.A. , Fuzzy sets as a basic for theory of possibility, FSSI, 3-28,(1978) .

[4] Zimmermann, H.J., Using Fuzzy sets in operational research, EJOR 13, 201-
216,(1983).
[5] Bellman R. E. , Zadeh L.A., Decision Making in A Fuzzy Environment,
Management Science, vol. 17, 1970, pp. 141-164.
[6] Zimmerman H. J., Fuzzy Programming and Linear Programming with Several
Objective Functions, Fuzzy Sets and Systems, vol. 1, 1978, pp. 45-55.
[7] Tanaka H, Ichihashi H, Asai K, Formulation of fuzzy linear programming
problem based on comparison of fuzzy numbers, Control Cybernetics 3 (3): 185-
194. (1991).
[8] Pandian P., Multi-objective Programming Approach for Fuzzy Linear
Programming Problems, Applied Mathematical Sciences, Vol. 7, no. 37, 1811 -
1817, 2013.

15