World Applied Sciences Journal 8 (5): 543-549, 2010

ISSN 1818-4952
IDOSI Publications, 2010

Corresponding Author: Dr. Nuran Gzel, Department of Mathematics, Faculty of Art and Sciences, Yildiz Technical
University, Davutpasa, Istanbul, Turkey
543
Fuzzy Transportation Problem with the Fuzzy Amounts and the Fuzzy Costs

Nuran Gzel

Department of Mathematics, Faculty of Art and Sciences,
Yildiz Technical University, Davutpasa, Istanbul, Turkey

Abstract: In this paper, we investigated a fuzzy transportation problem with fuzzy quantities, in which are
bounded fuzzy triangular numbers and fuzzy transportation costs per unit that are bounded upper fuzzy
numbers. The problem is solved at two stages. In first stage is calculated maximum satisfactory level
satisfying balance between fuzzy supplies and fuzzy demands. The given transportation problem is
rearranged according to the maximum satisfactory level. In second stage, by considering the unit
transportation costs, the arranged problem has investigated all of optimal solution connected with
satisfactory level for quantities on all intervals that constituted from breaking points of the unit
transportation costs, from zero to maximum satisfactory level. We derive two different satisfactory levels to
the problem: one is breaking points
p
, for p = 1,2,, of transportation costs C = C () and the other is
s

values for s = 1,2,, that are violated positive condition for optimal solution X
*
= X
*
() in intervals [
p-1
,

p
], for p = 1,2,. Optimal solutions could be different by depending on the constructed [
s-1
,
s
] intervals
for s = 1,2, on the constructed [
p-1
,
p
] intervals for p = 1,2,. We proposed a new technique in which
find all different optimal solutions to the fuzzy transportation problem with respect to in the [
s-1
,
s
]
intervals for s = 1,2, on the [
p-1
,
p
] intervals for p = 1,2,.

Key words: Fuzzy transportation problem breaking points membership functions

INTRODUCTION

The transportation is called the transportation
problem that transports a homogenous product from m
sources to n different destinations to minimize total
transportation cost. There are numerous solution
methods for transportation problem when prices and
quantities are given as crisp numbers [1, 2]. Several
variations of transportation methods have been using
with the table methods such as the northwest-
corner method, the shortcut method and Russels
approximation method [1, 2]. Some others [1-3] have
been using special techniques for linear programming
(LP) problem because the classic single objective
transportation problem is a special case of the linear
programming problem. However, recently fuzzy
programming approach started to use the optimal
solutions of multi objective or single objective
transportation problem [4-8]. For instance, Wahed [4]
presented a fuzzy programming approach to determine
the compromise solution of multi objective
transportation problem (MOTP). Kikuchi [6] proposed
a simple adjustment method that finds the most
appropriate set of crisp numbers. Wahed et al. [8]
presented an interactive fuzzy goal programming
approach to determine the preferred compromise
solution for MOTP.
In reality, it is not possible to determine both
quantities and transportation unit prices, but the fuzzy
numbers gives best approximation of them. A model
solving the transportation problem is given in [5] when
quantities are fuzzy and prices are crisp. The model
uses the table method for solution. Again, in [6] is
given a method determining quantities that is satisfied
the higher satisfactory level while quantities is only
fuzzy. This method uses LP model in solution.
OhEigeartaigh [1, 9] considered the case where the
membership functions of the fuzzy demands are
triangular forms for transportation problems and
solved it using table method. Chanas and Kulej [1, 10]
provided an approach to solve a fuzzy linear
programming problem with triangular membership
functions of fuzzy resources. Geetha and Nair [11]
formulated a stochastic version of the time minimizing
transportation problem and developed an algorithm
based on parametric programming to solve it when
transportation time is considered to be independent,
positive normal random variables. Chanas et al. [7] are
analyzed the transportation problem with fuzzy supply
values of deliverers and with fuzzy demand values of
World Appl. Sci. J., 8 (5): 543-549, 2010
544

the receivers. Liu and Kao [12] developed a procedure
to derive the fuzzy objective value of the fuzzy
transportation problem, in that the cost coefficients and
the supply and demand quantities are fuzzy numbers
basing on extension principle. Ahlatcioglu, Sivri and
Gzel [13] proposed a solution algorithm finding all
fuzzy optimal solution of the transportation problem
that the cost coefficients and the supply and demand
quantities are fuzzy numbers. Chanas and Kuchta [14]
proposed an algorithm that solves the transportation
problem with fuzzy supply and demand values and
integrality condition imposed on the solution. Ammar
and Youness [15] investigated the efficient solutions
and stability of fuzzy multiobjective transportation
problem and proposed an algorithm for the
determination of the stability set. Gen et al. [16]
describes an implementation of genetic algorithm to
solve Bicriteria Solid Transportation Problem (BSTP).
Li et al. [17] proposed neural network approach for
multicriteria solid transportation problem.
In this paper, we investigated a fuzzy
transportation problem with fuzzy quantities, in which
are bounded fuzzy triangular numbers and fuzzy
transportation costs per unit that are bounded upper
fuzzy numbers. The problem is solved at two stages. In
first stage is calculated maximum satisfactory level
satisfying balance between fuzzy supplies and fuzzy
demands. The given transportation problem is
rearranged according to the maximum satisfactory
level. In second stage, considering the unit
transportation costs, the arranged problem has
investigated all of optimal solution connected with
satisfactory level for quantities on all intervals that
constituted from breaking points of the unit
transportation costs, from zero to maximum satisfactory
level. We derive two different satisfactory levels to the
problem: one is breaking points
p
, for p = 1,2,, of
transportation costs C = C() and the other is
s
values
for s = 1, 2,, that are violated positive condition for
optimal solution X
*
= X
*
() in intervals [
p-1
,
p
], for
p = 1, 2,. Depending on the constructed [
s-1
,
s
]
intervals for s = 1, 2, on the constructed [
p-1
,
p
]
intervals for p = 1, 2,. optimal solutions could be
different on these intervals. We proposed a technique
in which find all different optimal solutions with
respect to in the [
s-1
,
s
] intervals for s = 1, 2, on
the [
p-1
,
p
] intervals for p = 1, 2, for the fuzzy
transportation problem.
We have shown that there is an optimal solution
each interval constituting by the iterative breaking
points like to stochastic order after the breaking points
for the satisfactory level that are changed the values of
transportation costs, respectively are determined.
However, It is transferred a new optimal solution by

using duality while optimal solutions lose to feasibly,
because satisfactory level depending on quantities
increases. The breaking points of satisfactory levels
depending on quantities and changing optimal solutions
are obtained successively. Because of the special
structure of the problem, all of these operations are
realized using the table and the optimal solutions are
obtained to all fuzzy cases that may be encountered in
future to problem.

THE MODEL

It is called the transportation problem that
transports a homogeneous product from m sources to n
destinations with the cheapest cost in the desired level.
If the data is given as fuzzy numbers, then this
problem is called the fuzzy transportation problem.

Data
m: Number of sources,
n: Number of the demand centers,
A
i
: Fuzzy quantity of i th source,
B
j
: Fuzzy quantity of j th demand,
C
ij
: Fuzzy unit transportation cost from i th source to
jth demand center.

Demand: x
ij
: The fuzzy quantities which is transported
from ith source to jth demand center. This
transportation problem may be stated mathematically in
general form:

Objective: min
m n
ij ij
i 1 j 1
C X

Constraints: The source constraints:


ij i
j
X A (i 1,...m)

(2.1)

The destination constraints:

ij j
i
X B ( j 1,...,n)

The balance conditions:
i j
i j
A B


Non-negative constraints:
ij
X 0(for i,j )
A
i
and B
j
are fuzzy quantities and shown as
1 2 3
i i i i
A (A , A , A ) and
1 2 3
j j j j
B ( B , B , B ) fuzzy triangle
numbers [5]. The unit transportation cost is given as
fuzzy number
2 3
ij ij ij
C ( , C , C ) . Let (A
i
), (B
j
) and
(C
ij
) denote their membership functions of A
i
, B
j
, C
ij
.
We have
World Appl. Sci. J., 8 (5): 543-549, 2010
545


( ) { }
( ) { }
( ) { }
i i i i i
j j j j j
ij ij ij ij ij
A a , ( A ) , / a A(A ) ,
B b , ( B) , / b A( B) ,
C c , (C ) , / c A(C ) ,



(2.2)

where A (A
i
), A (B
j
) and A (C
ij
) are the supports of A
i
,
B
j
and C
ij
. Denote the -cuts of A
i
and B
j
as:

{ }
{ }
i
i i i
A L U
i i i
i i i
min a A(A ) / ( A) ,
( A) ( A ) , ( A )
max a A(A ) / (A )

1
1 1
]
1

]


1 2 1 3 3 2
i i i i i i
A (A A) , A (A A ) 1 +
]

{ }
{ }
j
j j j
B L U
j j j
j j j
min b A( B) / (B ) ,
( B) (B ) , ( B)
max b A( B) / (B )

1

1
1
]
1

1
]
(2.3)

1 2 1 3 3 2
j j j j j j
B (B B) , B (B B ) 1 +
]


These intervals indicate where supply and demand
lie at possibility level [12].
Denote the -cuts of C
ij
as:

( ) { }
ij
U
ij ij ij ij ij
C
(C ) , C , max c A(C ) / (C )

1
1

1 1
]
]
(2.4)

3 3 2
ij ij ij
, C (C C ) 1
]
.

These intervals indicate where the unit shipping
cost lie at possibility level .
In the first step, based on the extension principle,
using the membership function of supply and demand
can be defined as:

n
1 2 1 3 3 2
i i i ij i i i
j 1
m
1 2 1 3 3 2
j j j ij j j j
i 1
Maks
A (A A ) x A (A A ), i 1,...,m,
B (B B ) x B (B B ), j 1,...,n,

+
+

(2.5)

From Optimal solution of problem (2.5) is obtained
maks . For supplies, if some j:

n
2
ij i
j 1
x A

then the related constraints are arranged as:

n
1 2 1
ij i i i
j 1
x A (A A) f or i 1,...,m

Otherwise the related constraints are arranged as:

n
3 3 2
ij i i i
j 1
x A (A A ) , f or i 1,...,m.

For demands, if some i:

n
2
ij i
j 1
x B

then the related constraints are arranged as:

m
1 2 1
ij j j j
i 1
x B (B B )

for J = 1,,n. Otherwise the related constraints are
arranged as:

m
3 3 2
ij j j j
i 1
x B (B B ), j 1,...,m

In the first step, supplies and demands have
arranged according to . In second step, from twice
intersection of the unit shipping costs in each cells with
related to is obtained breaking points of costs.
Membership functions of each of fuzzy numbers A
i
, B
j

and C
ij
for i = 1,,m; j = 1,,n can be written as:


1
i i
1
1 2 i i
i i i 2 1
i i
i i
3
2 3 i i
i i i 2 3
i i
3
i i
0 , A A
A A
, A A A
A A
( A)
A A
, A A A
A A
0 , A A

'

(2.6)


1
j j
1
j j 1 2
j j j 2 1
j j
j j
3
j j 2 3
j j j 2 3
j j
3
j j
0 , B B
B B
, B B B
B B
(B )
B B
, B B B
B B
0 , B B

'

(2.7)


2
ij ij
3
ij ij 2 3
ij ij ij ij ij 2 3
ij ij
3
ij ij
1 , C C
C C
(C ) , C C C
C C
0 , C C

'

(2.8)

There are one
1
i i
A ( )

and
1
j j
B ( )

in each
[ ] 0, , because the membership functions are
World Appl. Sci. J., 8 (5): 543-549, 2010
546



Fig. (a):



Fig. (b):

monotone for 0<<1. If these numbers are explained
with -parameter, then are obtained

1 1 1 2 1
i i i i i
A ( ) A (A A )

+
and
1 3 3 2
i i i i i
A ( ) A (A A )
+


from
i
(A
i
) = and

1 1 2 1
j j j j j
B ( ) B (B B )

+
and
1 3 3 2
j j i i i
B ( ) B (B B )
+


from
j
(B
j
) = .
The membership functions of these numbers are
shown in Fig. (a).
In the satisfactory level always is satisfied equality
between the total demand and the total supply
(
i j
i j
A B

). Since must satisfy the equation

i j
i j
= max min { A , B )}
_ _


,
,



is in [0,1] interval. In the corresponding ,1 1
]
,
the problem becomes balanced by adding to an artificial
source and is solved. Because of maximum satisfactory
level, , ,1 1
]
interval doesnt show directly the
satisfactory level in transportation (Fig. (b)).

The problem is balanced problem at . In solving
of this problem the satisfactory level can increase as
much as = 1. In solutions involving quantities out of
the interval value becomes <1.

DEFINING OF BREAKING POINTS IN
PRICES AND FINDING OF CRISP PRICES

Because the membership function of C
ij
is
monotone decreasing in (C
ij
) (0,1) interval,

1
ij ij
(C ) C ( )

is obtained. In this state, the transportation costs are
defined in terms of satisfactory level (Fig. c).
That is,

3 3 2
ij ij ij ij
C C (C C )

By means of the twice intersection of

3 3 2
ij ij ij ij
C C (C C )

lines, the N = m.n piece values of (01) for each k
are obtained and then these values change the order of
c
ij
. These values of are called the breaking points [3].
The order of C
ij
s in each subinterval that are formed
between the iterative breaking points dont change.
Therefore, the representative point can be selected in
this interval. This selected point can be any point of the
interval, but the optimal solution of transportation
problem doesnt change. However, when the interval
changes, the ordering among costs change. Therefore,
the optimal solution of transportation problem changes
also. Because of this, for each interval that is formed
the iterative breaking points of are searched the
optimal solutions one by one. Thus, all optimal
solutions are investigated.



Fig. (c):
World Appl. Sci. J., 8 (5): 543-549, 2010
547

Table 1:
Supply (Ai)

11
C =(- ,3,5)
12
C =(- ,6,7)
13
C =(- ,6,11) (2, 9, 11)

21
C =(- ,7,9)
22
C =(- ,9,15)
23
C =(- ,10,18) (3, 8, 12)

31
C =(- ,9,13)
32
C =(- ,9,16)
33
C =(- ,9,10) (4, 9, 14)
Demand Bj (2, 5, 6) (14, 15, 17) (5, 10, 13)

Let this interval [
p-1
,
p
] (p = 1,2,) be subinterval
by forming iterative breaking points. By selecting
p p 1 p
[ , ]

, we solve the transportation problem for

crisp costs, by using

3 3 2
ij ij ij ij
C C (C C ) .

As you know, for the transported quantity X
ij
from
ith source to jth demand center the fuzzy transportation
problem is stated as follows,

3 3 2
ij i i i
i
X A (A A )

or
1 2 1
ij i i i
i
X A (A A ) +

for some j,

1 2 1
ij j j j
j
X B (B B ) +

and
3 3 2
ij j j j
j
X B (B B )

for some i. objective: min
ij ij
i j
C X

.

DEFINING OF INTERVALS

Depending on that changes in [0,1] interval the
quantities of source and demand centers and optimal
solutions will change. Thus, optimal solutions
* *
ij ij
x x ( ) are linear. But, while variations of change
the transported quantities continuously, non-negative
constraint is violated. which violated non-negative
constraint of optimal solution is called breaking points
of . After value in this breaking point at least one
component of optimal solution is negative. Therefore,
the feasible condition is violated. The feasible condition
is gained again using the dual problem. If the variation
in the positive direction of again violates the feasible
condition then same operations are continued.
Therefore, the different optimal solution can be
obtained as the number of intervals as. When the costs
and quantities are given fuzzy form simultaneously, the
numbers of optimal solutions are p.r, where p is the
number of interval that is formed by in [0,1], r is the
Table 2:
11
C =5-2
12
C =7-
13
C =11-5
14
C =0
1
a =11-2
21
C =9-2
22
C =15-6
23
C =18-8
24
C =0
2
a =12-4
31
C =13-4
32
C =16-7
33
C =10-
34
C =0
3
a =14-5
1
b=2+3
2
b =14+
3
b=5+5
4
b=16-20

number of interval that is formed by in [0, ] . But,
some optimal solutions are same.

Example: We consider the following fuzzy
transportation problem. In the example, the information
on the fuzzy quantity supplied and demanded as well as
the fuzzy transportation costs per unit can be arranged
in the following Table 1. In the table, the fuzzy
transportation costs per unit are indicated by the figures
in the small rectangles. The fuzzy quantities supplied
and demanded appear at the rights and along the bottom
of the table. The fuzzy quantities to be transported from
the sources to the different destinations are represented
by the decision variables (x
11
, x
12
,,x
21
, x
22
,) in the
cells.
Now, we solve the given transportation problem
with fuzzy cost and fuzzy quantity.


1 2 3
1 2 3
4 5 6
4 5 6
7 8 9
7 8 9
1 4 7
1 4 7
2 5 8
2 5 8
3 6 9
3 6 9
1 2 3 4 5 6 7 8 9
Max
x x x 2 7
x x x 11 2
x x x 3 5
x x x 12 4
x x x 4 5
x x x 14 5
x x x 2 3
x x x 6
x x x 14
x x x 17 2
x x x 5 5
x x x 13 3
x , x , x , x , x , x , x , x , x 0

+ + +
+ +
+ + +
+ +
+ + +
+ +
+ + +
+ +
+ + +
+ +
+ + +
+ +

(5.1)

Solution of problem (5.1) is = 0.8 and x
11
= 3.4,
x
12
= 6, x
13
= 0, x
21
= 0, x
22
= 8.8, x
23
= 0, x
31
= 1, x
32
=
0, x
33
= 9. From this solution are obtained
1
= 9.4,

2
= 8.8,
3
= 10 and b
1
= 4.4, b
2
= 14.8, b
3
= 9 where
World Appl. Sci. J., 8 (5): 543-549, 2010
548

Table 3:
11-2
1
a =11-2
11
C =5-2
12
C =7-
13
C =11-5
14
C =0
2 3 + 3 3 + t 7 10 +t
2
a =12-4
21
C =9-2
22
C =15-6
23
C =18-8
24
C =0
+t 5+5 9 10 t
3
a =14-5
31
C =13-4
32
C =16-7
33
C =10-
34
C =0
1
b=2+3
2
b =14+
3
b=5+5
4
b=16-20

9
1
11, 8
2
12, 9
3
14 and 2b
1
5, 14b
2
15,
5b
3
10. The given fuzzy transportation problem
can be arranged as the fuzzy transportation problem
in Table 2.
where quantities in artificial demand center are

3
i
i 1
a 37-11

and
3
j
j 1
b 21+9 .

=0.8 is found from 37-11 = 21+9. The satisfactory
level a will be in [ 0, ]=[ 0,0.8] interval. Therefore,
quantity in the artificial demand center is

4 i j
i j
b = a b=16-20



so in this problem, the total available quantity in the
sources is exactly equal to the total quantity required by
the destinations, that is, the problem is balanced
transportation problem.
The breaking points of s are intersection points of
lines C
ij
= C
ij
() Breaking values of in [0,1] interval
are 0, 1/4, 2/3 and 1. Intervals among these iterative
values are [0, 1/4], [1/4, 2/3], [2/3, 1] Now, for interval
[0, 1/4], let us solve the transportation problem with
fuzzy quantity. Let us = 0 as representative of the
first interval for . Then, the obtained optimal solution
is given in Table 3. For the values in intervals of

1
0
4
and
7
0
10
;

The values of in the supply and demand
centers are distributed homogenous to cell that are
occurred optimal transportation. The first value that
optimal solution violates non-negative condition is
x
24
= 7-10 = 0. Thus, the present optimal solution is
valid for interval [0, 7/10].
Table 4:
11-2
1
a =11-2
11
C =5-2
12
C =7-
13
C =11-5
14
C =0
2 3 + 10 7
2
a =12-4
21
C =9-2
22
C =15-6
23
C =18-8
24
C =0
7 10 + 5+5 16-20
3
a =14-5
31
C =13-4
32
C =16-7
33
C =10-
34
C =0
1
b=2+3
2
b =14+
3
b=5+5
4
b=16-20

Table 5:

-------------------------------------------------------------------
a [0,1/4] [1/4,2/3] [2/3,1]
7
0,
10
1
1
]

12
11 2 x
12
11 2 x
11
2 3 + x

21
2 3 + x
21
2 3 + x
12
9 5 x

22
3 3 + x
22
3 3 + x
22
5 6 + x

24
7 10 x
24
7 10 x
24
7 10 x

33
5 5 + x
33
5 5 + x
33
5 5 + x

34
9 10 x
34
9 10 x
34
9 10 x
7 8
,
10 10
1
1
]

12
11 2 x
12
11 2 x
11
2 3 + x

21
2 3 + x
21
2 3 + x
12
9 5 x

22
10 7 x
22
10 7 x
22
12 4 x

32
7 10 + x
32
7 10 + x
32
7 10 + x

33
5 5 + x
33
5 5 + x
33
5 5 + x

34
16 20 x
34
16 20 x
34
16 20 x

For the values in intervals of

1
0
4
and
7 8
;
10 10


In Table 4, the obtained solutions for interval
[7/10, 8/10] are optimal and feasible. The obtained
optimal solutions for intervals of [1/4, 2/3] and
[2/3,1] are given in Table 5.
World Appl. Sci. J., 8 (5): 543-549, 2010
549

CONCLUSIONS

We consider the fuzzy transportation problem that
needs to be delivered with fuzzy quantities and fuzzy
transportation cost per unit at the time.
In the proposed study is calculated all different
optimal solutions connected with on all of the
constituted intervals ordering all of satisfactory levels

s
s, (s = 1,2,) obtained from positive condition of
decision variables for quantities and all of satisfactory
levels
p
s, (p = 1,2,) calculated breaking points for
transportation costs per unit. Thus, all of the optimal
solutions are obtained to all the fuzzy cases that may be
encountered in future to problem.

REFERENCE

1. Kwak, N.K., 1973. Mathematical Programming
with Business Applications. Mc Graw-Hill Book
Company, New York, pp: 88-149.
2. Greig, D.M., 1980. Optimisation, Lonman Group
Limited, London, pp: 100-111.
3. Ringuest, J.L. and D.B. Rink, 1987. Interactive
solutions for the linear multiobjective
transportation problem. European Journal
Research, North-Holland, 32: 96-106.
4. Wahed, W.F.A.E., 2001. A multi objective
transportation problem under fuzziness. Fuzzy
Sets and Systems , 117: 27-33.
5. Lai, Y.J. and C.L. Hwang, 1992. Fuzzy
Mathematical Programmig, Springer-Verlag,
Berlin.
6. Kikuchi, S., 2000. A method to defuzzify the fuzzy
number: Transportation problem application.
Fuzzy Sets and Systems , 116: 3-9.
7. Chanas, S., W. Kolodziejczyk and A. Machaj,
1984. A fuzzy Approach to the Transportation
Problem. Fuzzy Sets and Systems , 13: 211-221.

8. Wahed, W.F.A.E. and S.L. Lee, 2006. Interactive
fuzzy goal programming for multi objective
transportation problems, Omega, 34: 158-166.
9. OhEigeartaigh, M., 1982. A fuzzy transportation
algorithm. Fuzzy Sets and Systems , 8: 235-243.
10. Chanas, S. and M. Kulej, 1984. A fuzzy linear
programming problem with equality constrains.
Control and Cybernetics, 13: 195-201.
11. Geetha, S. and K.P. Nair, 1994. A Stochastic
Bottleneck Transportation Problem. J. Opl. Res.
Soc., 45 (5): 583-588.
12. Liu, S.-T. and C. Kao, 2004. Solving Fuzzy
Transportation Problems based on extension
princible. European Journal of Operational
Research, 153: 661-674.
13. Ahlatcioglu, M., M. Sivri and N. Gzel, 2002/1.
Transportation of the fuzzy amounts using the
fuzzy cost. Journal of Marmara for Pure and
Applied Sciences, Istanbul, Turkey, Vol: 18.
14. Chanas, S. and D. Kuchta, 1998. Fuzzy integer
transportation problem. Fuzzy Sets and Systems ,
98: 291-298.
15. Ammar, E.E. and E. Youness, 2005. A, Study on
multiobjective transportation problem with fuzzy
numbers. Applied Mathematics and Computation,
16 (2): 241-253.
16. Gen, M., K. Ida and Y. Li, 1994. Solving Bicriteria
Solid Transportation Problem, Humans,
information and Technology. IEEE International
Conference, 2: 1200-1207.
17. Li, Y., K. Ida, M. Gen and R. Kobuchi, 1997.
Neural Network Approach for Multicriteria Solid
Transportation Problem. Computers Industrial
Engineering, 33 (3-4): 465-468.