JOURNAL OF ALGEBRAIC STATISTICS

Volume 13, No. 2, 2022, p. 1592 - 1600

https://publishoa.com
ISSN: 1309-3452

Comparative Study on Robust Ranking Technique and Magnitude

Ranking Method for Fuzzy Linear Programming Problem
*1 E. Kuppusamy, 2* V.E. Sasikala
*1
M.Phil Scholar, 2* Corresponding Author, Research Supervisor,
Department of Mathematics, Vels Institute of Science, Technology & Advanced Studies (VISTAS), Chennai, India.

ABSTRACT
In this paper, we first consider the fuzzy linear problem, Here coefficients are fuzzy numbers, we convert the problem
into crisp form using the robust technique after using the simplex method we obtain the optimal solution for the given
problem, secondary we apply the Magnitude method and got the solution comparison study illustrated.

Keywords: Fuzzy Linear programming problem, Robust Ranking Method, Triangular Fuzzy Number, Magnitude
Ranking Method, Simplex Method.

1. Introduction:
In fuzzy environment the concept of decision making introduced Bellman and Zadeh. Many researchers
used this concept. However, fuzzy linear programming not all parts of the problem to be fuzzy, for example only the
right hand side or the objective function coefficients were fuzzy. The fuzzy linear programming problems in which all
the parameters as well as the variables are represented by fuzzy numbers is known as FFLP problems. FFLP problems
can be divided into two categories:
(1) FFLP problems with equality constraints
(2) FFLP problems with inequality constraints.
Some authors have different methods for solving FFLP problems with inequality constraints. In all methods
first the FFLP problem is converted into crisp linear programming problem and then the obtained crisp linear
programming problem is solved to find the fuzzy optimal solution of the FFLP problems.
Fuzzy set introduced by Zadeh in 1965.A. Nagoor Gani et al created the new operations for the triangular
fuzzy numbers for solving fuzzy linear programming. Because there is a some issues for the usual arithmetic operations.
Arun Pratap Singh compared the Robust method and Centroid technique for the Fuzzy assignment
problems. K. Kalaiarasi et al used the RRT for the fuzzy optimization in Assignment. Monalisha Pattnaik presented the
two phase optimization using the RRT.
Stephen Dinagar et al applied the RRT for Fully Fuzzy Integer Linear
Programming Problems. H. R. Maleki et al approached fuzzy linear problem for the parabolic method. D. Selvi et al used
the Magnitude ranking for the fuzzy assignment to get the optimum solution. K. Ganesan et al tested the fuzzy linear
programming problem for the trapezoidal fuzzy numbers. That fuzzy problem author solved without converting the
problem into the crisp model, the solution obtain the best solution in this research. Amit Kumar et al suggested the new
method for solving the fuzzy linear problems.
In this paper we have to solve the fuzzy linear problems using the RRT and MRM, finally we have comparison
between optimal solution made the table. The further section as follows; section two introduces the concept of fuzzy sets
and fuzzy numbers etc., Third section presents the General form of Fuzzy Linear programming Problems. Four
investigates a numerical example and five contains the comparison table, six finished the conclusions based on our
discussion.

2. Preliminaries:
Definition: 2.1
“ A fuzzy set 𝐴̃ is called fuzzy number if it satisfies the following criteria
1. 𝐴̃ must be normal
2. 𝐴̃ is convex fuzzy set

1592
JOURNAL OF ALGEBRAIC STATISTICS
Volume 13, No. 2, 2022, p. 1592 - 1600
https://publishoa.com
ISSN: 1309-3452

3. 𝐴̃(x) is piece wise continuous.”

4.
Definition : 2.2
“A fuzzy number 𝐴̃ on R is said to be a triangular fuzzy number if its membership
function 𝐴̃: R → [0, 1] has the following characteristics:
𝑥− 𝑎1
, 𝑎1 ≤ 𝑥 ≤ 𝑎2
𝑎2 − 𝑎1
𝐴̃(x) = { 𝑎3 − 𝑥 , 𝑎 < 𝑥 ≤ 𝑎
2 3
𝑎3 − 𝑎2
0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
We denote this triangular fuzzy number 𝐴̃ = (𝑎1 , 𝑎2 , 𝑎3 ).”
We use F(R) to denote the set of all triangular fuzzy numbers.”

Definition : 2.3
“ If ã is a fuzzy number, the Robust ranking index is defined by
1
R(𝑎̃) = 0.5 ∫0 (𝑎𝛼𝐿 , 𝑎𝛼𝑈 ) dα
Where (𝑎𝛼𝐿 , 𝑎𝛼𝑈 ) = {(b-a)α + a, d-(d-c)α } is the α - cut of the fuzzy number”𝑎̃.”

Definition : 2.4
“The magnitude of the triangular fuzzy number ã is defined by,
1 1
Mag ã = ∫0 (𝑎3 + 3𝑎1 − 𝑎2 )f(r)dr
2
Where the function f(r) is a non-negative and increasing function on (0,1). In the real life applications f(r) can be
chosen by the decision maker according to the situation.

Definition : 2.5
“Define a ranking function R : F(R) → R which maps each fuzzy number into the real line, where a natural order
exists and defining order on F(R) by
1. 𝑎̃ ≥ b̃ iff R(𝑎̃) ≥ R(b̃)
2. 𝑎̃ ≤ b̃ iff R(𝑎̃) ≤ R(b̃)
3. 𝑎̃ = b̃ iff R(𝑎̃) = R(b̃)”

3. Fuzzy Linear Programming Problems:

The general form fuzzy linear programming problem defined as
Maximize 𝑊 ̃ (X) = ∑𝑛𝑖=1 𝑐̃𝑖 𝑎̃𝑖
Subject to 𝑔𝑗 ̃)
̃(𝑐 ̃
𝑗 ≤ or ≈ or ≥ 𝑏𝑖 , i = 1,2,3,…,m

& 𝑐̃𝑗 ≥ 0̃, for j = 1,2,3,…,n

4. Numerical Examples:

Consider a fuzzy linear programming problem

Max 𝑊 ̃ = (5,7,9) 𝑐̃1 + (6,8,10) 𝑐̃2
Subject to
(1,2,3) 𝑐̃1 + (2,3,4) 𝑐̃2 ≤ (4,6,8)
(4,5,6) 𝑐̃1 + (3,4,5) 𝑐̃2 ≤ (8,10,12) & 𝑐̃1 , 𝑐̃2 ≥ 0
Now convert the given fuzzy problem into crisp form using the Robust
Ranking method,
Max 𝑊 ̃ = R(5,7,9) 𝑐̃1 + R(6,8,10) 𝑐̃2
Subject to
R (1,2,3) 𝑐̃1 +R (2,3,4) 𝑐̃2 ≤ R(4,6,8)
R(4,5,6) 𝑐̃1 + R(3,4,5) 𝑐̃2 ≤ R(8,10,12) &𝑐̃1 , 𝑐̃2 ≥ 0

1593
JOURNAL OF ALGEBRAIC STATISTICS
Volume 13, No. 2, 2022, p. 1592 - 1600
https://publishoa.com
ISSN: 1309-3452

The membership function of the triangular fuzzy number (5,7,9) is

𝑥− 5
,5 ≤ 𝑥 ≤ 7
2
̃
𝐴(x) = { 9 − 𝑥
,7 < 𝑥 ≤ 9
2
0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

1
R(𝑎̃) = 0.5 ∫0 (𝑎𝛼𝐿 , 𝑎𝛼𝑈 ) dα

Where (𝑎𝛼𝐿 , 𝑎𝛼𝑈 ) = {(b-a)α + a, d-(d-c)α }

The α - cut of the fuzzy number (5,7,9) is (𝑎𝛼𝐿 , 𝑎𝛼𝑈 ) = (2α+5,9-2α).
1
R(5,7,9)= 0.5 ∫0 {(5 − 7) α + 5+9-(9-7)α } dα
1 1
= 0.5 ∫0 2 α + 14-2α } dα = 0.5 ∫0 (14 ) dα =7.
The membership function of the triangular fuzzy number (6,8,10) is
𝑥− 6
,6 ≤ 𝑥 ≤ 8
2
̃
𝐴(x) = {10 − 𝑥
, 8 < 𝑥 ≤ 10
2
0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
The α - cut of the fuzzy number (6,8,10) is (𝑎𝛼𝐿 , 𝑎𝛼𝑈 ) = (2α+6,10-2α).
1
R(6,8,10)= 0.5 ∫0 {(8 − 6) α + 6+10-(10-8)α } dα
1 1
= 0.5 ∫0 2 α + 16-2α } dα = 0.5 ∫0 (16 ) dα =8.

Similarly we get,

R(1,2,3)=2, R(2,3,4)=3, R(4,6,8)=6, R(4,5,6)=5, R(3,4,5)=4, R(8,10,12)=10.

we obtain crisp linear programming problem is
Max 𝑊 ̃ = 7 𝑐̃1 + 8 𝑐̃2
Subject to
2 𝑐̃1 + 3 𝑐̃2 ≤ 6
5 𝑐̃1 + 4 𝑐̃2 ≤ 10 &𝑐̃1 , 𝑐̃2 ≥ 0

Now we solve the problem using the simplex method,

The problem is converted to canonical form by adding slack variables. As the constraints are '≤' type we should add
slack variables s̃1 and s̃2 .
Max 𝑊 ̃ = 7 𝑐̃1 + 8 𝑐̃2 + 0 𝑠̃1 + 0 𝑠̃2
Subject to
2 𝑐̃1 + 3 𝑐̃2 + 𝑠̃1 = 6
5 𝑐̃1 + 4 𝑐̃2 + 𝑠̃2 = 10 &𝑐̃1 , 𝑐̃2 , 𝑠̃1 , 𝑠̃2 ≥ 0
Therefore we get the Iteration 1 table as follows:
Table 4.1
Iteration-1 Cj 7 8 0 0

MinRatio
B CB XB 𝒄̃𝟏 𝒄̃𝟐 𝒔̃𝟏 𝒔̃𝟐 𝑿𝑩
𝒄̃𝟐
6
𝑠̃1 0 6 2 (3) 1 0 =2→
3

10
𝑠̃2 0 10 5 4 0 1 4
=2.5

1594
JOURNAL OF ALGEBRAIC STATISTICS
Volume 13, No. 2, 2022, p. 1592 - 1600
https://publishoa.com
ISSN: 1309-3452

Z=0 Zj 0 0 0 0

0
Zj-Cj -7 -8↑ 0

Negative minimum Zj-Cj is -8 and its column index is 2. So, the entering variable is 𝒄̃𝟐 .
Minimum ratio is 2 and its row index is 1. So, the leaving basis variable is 𝒔̃𝟏 .
∴ The pivot element is 3.

Table 4.2
Iteration-2 Cj 7 8 0 0

MinRatio
B CB XB 𝒄̃𝟏 𝒄̃𝟐 𝒔̃𝟏 𝒔̃𝟐 𝑿𝑩
𝒄̃𝟏
2
𝒄̃𝟐 8 2 0.6667 1 0.3333 0 0.6667
=3

2
𝑠̃2 0 2 (2.3333) 0 -1.3333 1 2.3333
=0.8571→

Z=16 Zj 5.3333 8 2.6667 0

Zj-Cj -1.6667↑ 0 2.6667 0

Negative minimum Zj-Cj is -1.6667. So, the entering variable is 𝑐̃1 .

Minimum ratio is 0.8571. So, the leaving basis variable is 𝑠̃2 .
∴ The pivot element is 2.3333.

Table 4.3

Iteration-3 Cj 7 8 0 0

B CB XB 𝒄̃𝟏 𝒄̃𝟐 𝒔̃𝟏 𝒔̃𝟐 MinRatio

𝒄̃𝟐 8 1.4286 0 1 0.7143 -0.2857

𝒄̃𝟏 7 0.8571 1 0 -0.5714 0.4286

Z=17.4286 Zj 7 8 1.7143 0.7143

Zj-Cj 0 0 1.7143 0.7143

Since all Zj-Cj ≥ 0

Hence, optimal solution is
𝒄̃𝟏 =0.8571,𝒄̃𝟐 =1.4286
Max 𝑾 ̃=17.4286

Magnitude Ranking Method:

Max 𝑊 ̃ = (5,7,9) 𝑐̃1 + (6,8,10) 𝑐̃2

Subject to

1595
JOURNAL OF ALGEBRAIC STATISTICS
Volume 13, No. 2, 2022, p. 1592 - 1600
https://publishoa.com
ISSN: 1309-3452

(1,2,3) 𝑐̃1 + (2,3,4) 𝑐̃2 ≤ (4,6,8)

(4,5,6) 𝑐̃1 + (3,4,5) 𝑐̃2 ≤ (8,10,12)
1 1
Applying Mag ã = ∫0 (𝑎3 + 3𝑎1 − 𝑎2 )f(r)dr
2
1 1 1 1
→Mag(5,7,9) = ∫ (9 + 3(5) − 7)r dr= 2 ∫0 (17)r dr = 4.25
2 0
1 1 1 1
Mag (6,8,10) = ∫ (10 + 3(6) − 8)r dr = 2 ∫0 (20)r dr = 5
2 0
Therefore we get in the similar manner,
Mag(1,2,3)=1,
Mag(2,3,4)=1.5,
Mag(4,6,8)=3.5,
Mag(4,5,6)=2.6,
Mag(3,4,5)=3.25,
Mag(8,10,12)=6.5
Max 𝑊 ̃ = 4.25 𝑐̃1 + 5 𝑐̃2 + 0 𝑠̃1 + 0 𝑠̃2
Subject to
1 𝑐̃1 + 1.5 𝑐̃2 + 𝑠̃1 = 3.5
2.6 𝑐̃1 + 3.25 𝑐̃2 + 𝑠̃2 = 6.5 &𝑐̃1 , 𝑐̃2 , 𝑠̃1 , 𝑠̃2 ≥ 0
The iteration table given below as follows:
Table 4.4
Iteration-1 Cj 4.25 5 0 0

MinRatio
B CB XB 𝒄̃𝟏 𝒄̃𝟐 𝒔̃𝟏 𝒔̃𝟐 𝑿𝑩
𝒄̃𝟐

3.5
𝑠̃1 0 3.5 1 (1.75) 1 0 1.75
=2→

6.5
𝑠̃2 0 6.5 2.6 3.25 0 1 3.25
=2

Z=0 Zj 0 0 0 0

Zj-Cj -4.25 -5↑ 0 0

Table 4.5
Iteration-2 Cj 4.25 5 0 0

MinRatio
B CB 𝑿𝑩 𝒄̃𝟏 𝒄̃𝟐 𝒔̃𝟏 𝒔̃𝟐 𝑿𝑩
𝒄̃𝟏
2
𝒄̃𝟐 5 2 0.5714 1 0.5714 0 0.5714
=3.5

0
𝑠̃2 0 0 (0.7429) 0 -1.8571 1 0.7429
=0→

Z=10 Zj 2.8571 5 2.8571 0

Zj-Cj -1.3929↑ 0 2.8571 0

1596
JOURNAL OF ALGEBRAIC STATISTICS
Volume 13, No. 2, 2022, p. 1592 - 1600
https://publishoa.com
ISSN: 1309-3452

Table 4.6

Iteration-3 Cj 4.25 5 0 0

MinRatio
B CB XB 𝒄̃𝟏 𝒄̃𝟐 𝒔̃𝟏 𝒔̃𝟐 𝑿𝑩
𝒔̃𝟏
2
𝒄̃𝟐 5 2 0 1 (2) -0.7692 2
=1→

𝒄̃𝟏 4.25 0 1 0 -2.5 1.3462 ---

Z=10 Zj 4.25 5 -0.625 1.875

Zj-Cj 0 0 -0.625↑ 1.875

Table 4.7

Iteration-4 Cj 4.25 5 0 0

B CB XB 𝒄̃𝟏 𝒄̃𝟐 𝒔̃𝟏 𝒔̃𝟐 MinRatio

𝒔̃𝟏 0 1 0 0.5 1 -0.3846

𝒄̃𝟏 4.25 2.5 1 1.25 0 0.3846

Z=10.625 Zj 4.25 5.3125 0 1.6346

Zj-Cj 0 0.3125 0 1.6346

Hence, optimal solution is

𝒄̃𝟏 = 4.25 ,𝒄̃𝟐 = 0

Max 𝑾 ̃=10.625

Example 2:
Min z = (1, 1, 1) 𝑥1 + (2, 1, 2) 𝑥2 Subject to (4, 1, 0)𝑥1 +(−3, 2, 1)x2 ≥ (2, 1, 2)
(−3, 1, 2) 𝑥1 + (2, 1, 1) 𝑥2 ≥ (1, 0, 1) & 𝑥1, 𝑥2 ≥ 0
Converting to crisp problem using RRT, we have
Min z = 𝑥1 +1.5𝑥2
0.75𝑥1 + 0.5𝑥2 ≥1.5
0.5𝑥1 + 1.25𝑥2 ≥0.5

Therefore we could solve the problem now

Min z = 𝑥1 +1.5𝑥2 +0𝑠1 + 0𝑠2 + 𝑀𝐴1 +𝑀𝐴2
Sub to
0.75𝑥1 + 0.5𝑥2 − 𝑠1 +𝐴1 =1.5
0.5𝑥1 + 1.25𝑥2 − 𝑠2 +𝐴2 =0.5
and x1, x2, S1, S2, A1, A2 ≥0

1597
JOURNAL OF ALGEBRAIC STATISTICS
Volume 13, No. 2, 2022, p. 1592 - 1600
https://publishoa.com
ISSN: 1309-3452

Table 4.8

Iteration-1 Cj 1 1.5 0 0 M M

MinRatio
B CB XB x1 x2 S1 S2 A1 A2 𝑿𝑩
𝒙𝟐

A1 M 1.5 0.75 0.5 -1 0 1 0 1.5/0.5=3

A2 M 0.5 0.5 (1.25) 0 -1 0 1 0.5/1.25=0.4→

Z=2M Zj 1.25M 1.75M -M -M M M

Zj-Cj 1.25M-1 1.75M-1.5↑ -M -M 0 0

Table 4.9
Iteration-2 Cj 1 1.5 0 0 M

MinRatio
B CB XB x1 x2 S1 S2 A1 𝑿𝑩
𝒙𝟏

A1 M 1.3 0.55 0 -1 0.4 1 1.3/0.55=2.3636

x2 1.5 0.4 (0.4) 1 0 -0.8 0 0.4/0.4=1→

Z=1.3M+0.6 Zj 0.55M+0.6 1.5 -M 0.4M-1.2 M

Zj-Cj 0.55M-0.4↑ 0 -M 0.4M-1.2 0

Table 4.10
Iteration-3 Cj 1 1.5 0 0 M

MinRatio
B CB XB x1 x2 x2 S2 A1 𝑿𝑩
s𝟐

A1 M 0.75 0 -1.375 -1 (1.5) 1 0.75/1.5=0.5→

x1 1 1 1 2.5 0 -2 0 ---

Z=0.75M+1 Zj 1 -1.375M+2.5 -M 1.5M-2 M

Zj-Cj 0 -1.375M+1 -M 1.5M-2↑ 0

Table 4.11
Iteration-4 Cj 1 1.5 0 0

B CB XB x1 x2 S1 S2 MinRatio

S2 0 0.5 0 -0.9167 -0.6667 1

1598
JOURNAL OF ALGEBRAIC STATISTICS
Volume 13, No. 2, 2022, p. 1592 - 1600
https://publishoa.com
ISSN: 1309-3452

x1 1 2 1 0.6667 -1.3333 0

Z=2 Zj 1 0.6667 -1.3333 0

Zj-Cj 0 -0.8333 -1.3333 0

Hence the solution is

x1=2, 𝑥2 =0→ Min Z=2

Method :2

Applying Magnitude method, Min z=0.75𝑥1 +1.75𝑥2

2.75𝑥1 -2.5𝑥2 ≥1.75
-2.5𝑥1 + 1.5𝑥2 ≥1

Min z=0.75𝑥1 +1.75𝑥2 +0𝑠1 + 0𝑠2 + 𝑀𝐴1+𝑀𝐴2

2.75𝑥1 -2.5𝑥2 − 𝑠1 +𝐴1 ≥1.75
-2.5𝑥1 + 1.5𝑥2 − 𝑠2 + 𝐴2 ≥1

Table 4.12
Iteration 1 Cj 0.75 1.75 0 0 M M

MinRatio
B CB XB x1 x2 S1 S2 A1 A2 𝑿𝑩
𝒙𝟐

A1 M 1.75 (2.75) -2.5 -1 0 1 0 1.75/2.75=0.6364→

A2 M 1 -2.5 1.5 0 -1 0 1 ---

Z=2.75M Zj 0.25M -M -M -M M M

0.25M-
Zj-Cj -M-1.75 -M -M 0 0
0.75↑
Table 4.13
Iteration-2 Cj 0.75 1.75 0 0 M

B CB XB x1 x2 S1 S2 A2 MinRatio

x1 0.75 0.6364 1 -0.9091 -0.3636 0 0

A2 M 2.5909 0 -0.7727 -0.9091 -1 1

Z=2.5909M+0.4773 Zj 0.75 -0.7727M-0.6818 -0.9091M-0.2727 -M M

Zj-Cj 0 -0.7727M-2.4318 -0.9091M-0.2727 -M 0

Hence, optimal solution is arrived as :

x1=0.6364, x2=0

Min Z=0.4773

1599
JOURNAL OF ALGEBRAIC STATISTICS
Volume 13, No. 2, 2022, p. 1592 - 1600
https://publishoa.com
ISSN: 1309-3452

5. Comparison:

Table 5.1 comparison table

Robust ranking technique Magnitude ranking method

𝒄̃𝟏 0.8571 4.25

𝒄̃𝟐 1.4286 0

̃
Max 𝑾 17.4286 10.625

Robust ranking technique Magnitude ranking method

x1 2 0.6364

x2 0 0

Min Z 2 0.4773

Conclusion

In the article, we compare the two ranking methods for the fuzzy linear programming maximization problem we obtain
the best solution for the Robust ranking technique and minimization problem we obtain the best solution for the
Magnitude method. We confirmed the accuracy of the problem using numerical examples.

References

1. Zadeh, L.A. Fuzzy Sets, Information and Control, 8 (3), p. 338-353, (1965)
2. Zimmermann, H.J, Fuzzy Set Theory and Its Applications, (4th Ed.).
3. Bellman, R. E., Zadeh, L. A., Decision-marking in a fuzzy environment, Management Science, 17(4), (1970), B141-
B164.
4. A. Nagoor Gani,A New Operation on Triangular Fuzzy Number for Solving Fuzzy Linear Programming
Problem,Applied Mathematical Sciences, Vol. 6, 2012, no. 11, 525-532.
5. Arun Pratap Singh, A Comparative Study of Centroid Ranking Method and Robust Ranking Technique in Fuzzy
Assignment Problem, Global Journal of Technology and Optimization, Volume 12:3 , (2021).
6. Dr. K. Kalaiarasi , S.Sindhu, Dr. M. Arunadev, Optimization of fuzzy assignment model with triangular fuzzy
numbers using Robust Ranking technique, International Journal of Innovative Science, Engineering and Technology, Vol.
1 Issue 3, May (2014).
7. Monalisha Pattnaik, Applied Robust ranking method in two phase fuzzy Optimization linear programming problems,
Scientific Journal of Logistics, 10 (4), 399-408, (201,).
8. D. Stephen Dinagar1 and M. Mohamed Jeyavuthin1,Fully Fuzzy Integer Linear Programming Problems Under
Robust Ranking Techniques, International Journal of Mathematics And its Applications, 6(3)(2018), 19–25.
9. H. R. Maleki and m. Mashinchi, fuzzy number linear programming: a probabilistic approach, j. Appl. Math. &
computing vol. 15(2004), no. 1 - 2, pp. 333 - 341.
10. D. Selvi, r. Queen mary & g. Velammal, method for solving fuzzy assignment problem using magnitude ranking
technique, international Journal of Applied and Advanced Scientific Research, February - 2017.
11. K. Ganesan , P. Veeramani, Fuzzy linear programs with trapezoidal fuzzy numbers, Ann Oper Res (2006) 143: 305–
315.

1600