World Academy of Science, Engineering and Technology

International Journal of Mathematical, Computational Science and Engineering Vol:8 No:3, 2014

Some Application of Random Fuzzy Queueing System

Based On Fuzzy Simulation
Behrouz Fathi-Vajargah, Sara Ghasemalipour

Abstract—This paper studies a random fuzzy queueing system Let Θ be a nonempty set, and P the power set of Θ (i.e., the
that the interarrival times of customers arriving at the server and largest σ- algebra over Θ). Each element in P is called an
the service times are independent and identically distributed random event. In order to present an axiomatic definition of credibility,
fuzzy variables. We match the random fuzzy queueing system with
the random fuzzy alternating renewal process and we do not use from it is necessary to assign to each event A a number Cr{A}
α-pessimistic and α-optimistic values to estimate the average chance which indicates the credibility that A will occur. In order
of the event ”random fuzzy queueing system is busy at time t”, we to ensure that the number Cr{A} has certain mathematical
employ the fuzzy simulation method in practical applications. Some properties which we intuitively expect a credibility to have,
theorem is proved and finally we solve a numerical example with we accept the following four axioms:
International Science Index Vol:8, No:3, 2014 waset.org/Publication/9997864

fuzzy simulation method.

Axiom 1. (Normality) Cr{Θ} = 1.
Keywords—Random fuzzy variables, Fuzzy simulation, Queueing Axiom 2. (Monotonicity)Cr{A} ≤ Cr{B} for A ⊂ B.
system, Interarrival times. Axiom 3. (Self-Duality) Cr{A} + Cr{Ac } = 1 for any event
A.
I. I NTRODUCTION Axiom 4. (Maximality) C{∪i Ai } = supi Cr{Ai } for any
events {Ai } with supi Cr{Ai } < 0.5.
UEUEING system is the most applications of alternating
Q renewal process and has been widely applied to
many practical problems. In classic queueing systems, the
Definition 1 (Liu and Liu [4]) The set function Cr is
called a credibility measure if it satisfies the normality,
interarrival times of customers and service times of servers
monotonicity, self-duality, and maximality axioms.
are characterized as random variables. In such cases, Ning,
Tang, and Zhao[1] proposed the concept of random fuzzy
Definition 2 Let Θ be a nonempty set, P the power
queueing systems with finite capacity, where the interarrival
set of Θ, and Cr a credibility measure. Then the triplet
times and service times were characterized as random fuzzy
(Θ, P, Cr) is called a credibility space.
variables. We proposes a theorem on the limit value of the
average chance of the random fuzzy event ”the queueing
Product credibility measure may be defined in multiple
system is busy at time t”. In this paper we use fuzzy simulation
ways. We accepts the following axiom.
to estimate he average chance of the random fuzzy event
Axiom 5. (Product Credibility Axiom) Let Θk be nonempty
”the queueing system is busy at time t” instead of the
sets on which Crk are credibility measures, k = 1, 2, ..., n,
α0 -pessimistic value and the α0 -optimistic value method.
respectively, and Θ = Θ1 × Θ2 × ... × Θn . Then
In Section II, we introduce some basic definition about
credibility measure and random fuzzy variables. In Section III, Cr{(θ1 , θ2 , ..., θn )} = Cr1 {θ1 } ∧ Cr2 {θ2 } ∧ ... ∧ Crn {θn }
we consider the random fuzzy queueing system and obtain the (1)
Ch {random fuzzy queueing system is busy at time t} and in for each (θ1 , θ2 , ..., θn ) ∈ Θ.
Section IV we illustrate the fuzzy simulation and in Section V, Let (θk , Pk , Crk ), k = 1, 2, ..., n be credibility spaces,
an example is solved by the simulation method. Θ = Θ1 × Θ2 × ... × Θn and Cr1 ∧ Cr2 ∧ ... ∧ Crn .
Then (Θ, P, Cr) is called the product credibility space of
(θk , Pk , Crk ), k = 1, 2, ..., n.
II. D EFINITIONS AND PRELIMINARIES
Definition 3 A fuzzy variable is a measurable function
Credibility theory, founded by Liu [2] in 2004 and refined
from a credibility space (Θ, P, Cr) to the set of real numbers.
by Liu [3] in 2007, is a branch of mathematics for studying
the behavior of fuzzy phenomena. The emphasis in this
Example 1. Take (Θ, P, Cr) to be {θ1 , θ2 } with
section is mainly on credibility measure, credibility space,
Cr{θ1 } = Cr{θ2 } = 0.5. Then the function
fuzzy variable, membership function, credibility distribution,
0, θ = θ1
expected value, random fuzzy variable and its expected value, ξ(θ) = is a fuzzy variable.
1, θ = θ2
independence, identical distribution,
Definition 4 A fuzzy variable ξ is said to be
B. Fathi-Vajargah is with the Department of Statistics, Faculty of Mathe- (a) Nonnegative if Cr{ξ < 0} = 0;
matical Sciences, University of Guilan, Iran (e-mail: [email protected]).
S. Ghasemalipour is with Anonymous University, (e-mail: (b) Positive if Cr{ξ ≤ 0} = 0;
[email protected]). (c) Continuous if Cr{ξ = x} is a continuous function of x.

36
 World Academy of Science, Engineering and Technology
International Journal of Mathematical, Computational Science and Engineering Vol:8 No:3, 2014

The same result can be obtained when r0 > 0.•

Definition 5 Let ξ be a fuzzy variable defined on the
credibility space (Θ, P, Cr). Then its membership function is Definition 9 A random fuzzy variable is a function from
derived from the credibility measure by the credibility space (Θ, P, Cr) to the set of random variables.
μ(x) = (2Cr{ξ = x}) ∧ 1, x ∈ . (2)
Definition 10 The expected value of a random fuzzy
Membership function represents the degree that the variable ξ is defined by
fuzzy variable ξ takes some prescribed value. There are  ∞
several methods reported in the past literature. Anyway, the E[ξ] = Cr{θ ∈ Θ|E[ξ(θ)] ≥ r}dr
0
membership degree μ(x) = 0 if x is an impossible point, and  0
μ(x) = 1 if x is the most possible point that ξ takes. − Cr{ξ ≤ r}Cr{θ ∈ Θ|E[ξ(θ)] ≤ r}dr, (8)
−∞
Definition 6 Let ξ be a fuzzy variable defined on the Proposition 2. Let ξ be a random fuzzy variable defined on
credibility space (Θ, P (Θ), P os), and α ∈ (0, 1]. Then (Θ, P, Cr). Then, for any θ ∈ Θ, E[ξ(θ)] is a fuzzy variable
ξα = inf{r|μξ (r) ≥ α}, ξα = sup{r|μξ (r) ≥ α}, (3) provided that E[ξ(θ)] is finite for fixed θ ∈ Θ.

are called α-pessimistic value and α-optimistic value of ξ, Definition 11 The random fuzzy variables ξ and η are
International Science Index Vol:8, No:3, 2014 waset.org/Publication/9997864

respectively. said to be identically distributed if

Definition 7(Liu [5]) The credibility distribution sup inf {P r{ξ(θ) ∈ B}} sup = inf {P r{η(θ) ∈ B}}
Cr{A}≥α θ∈A Cr{A}≥α θ∈A
Φ :  → [0, 1] of a fuzzy variable ξ is defined by (9)
Φ(x) = Cr{θ ∈ Θ|ξ(θ) ≤ x}. (4) for any α ∈ (0, 1] and Borel set B of real numbers.

That is, F (x) is the credibility that the fuzzy variable ξ Definition 12 The random fuzzy variables ξi , i = 1, ..., n are
takes a value less than or equal to x. Generally speaking, said to be independent if
the credibility distribution F is neither left-continuous nor (1) ξi (θ), i = 1, ..., n are independent random variables for
right-continuous. each θ ∈ Θ.
(2) E[ξi (.)], i = 1, ..., n are independent fuzzy variables.
There are many ways to define an expected value operator
for fuzzy variables. The most general definition of expected Definition 13 Let ξ be a random fuzzy variable on the
value operator of fuzzy variable was given by Liu and Liu possibility space (Θ, P (Θ), Cr). Then the average chance of
[5]. This definition is applicable to not only continuous fuzzy random fuzzy event ξ ≤ 0 is defined as
variables but also discrete ones.  1
Ch{ξ ≤ 0} = Cr{θ ∈ Θ|P r{ξ(θ) ≤ 0} ≥ p}dp. (10)
Definition 8 Let ξ be a fuzzy variable. Then the expected 0
value of ξ is defined by
III. R ANDOM FUZZY QUEUEING SYSTEM
 ∞  0
E[ξ] = Cr{ξ ≥ r}dr − Cr{ξ ≤ r}dr, (5) Consider a random fuzzy queueing system with single
0 −∞ server, that the interarrival times of customers arriving at the
provided that at least one of the two integrals is finite. server are independent and identically distributed random
In particular, if the fuzzy variable ξ is positive (i.e. Cr{ξ ≤ fuzzy variables, ξi ∼ EXP (λi ), where λi are fuzzy variables
0} = 0), then defined on the credibility space (Θi , P (Θi ), Cri ), i = 1, 2, ...,
 ∞ and the service times are independent and identically
E[ξ] = Cr{ξ ≥ r}dr. (6) distributed random fuzzy variables, ηi ∼ EXP (μi ),
0 where μi are fuzzy variables defined on credibility space
Proposition 1. Let ξ be a fuzzy variable defined on the (Γi , P (Γi ), Cri ), i = 1, 2, .... ξi and ηi are independent.
credibility space (Θ, P (Θ), P os). Then we have We denote this random fuzzy queueing system with
 RF/RF/1/F CF S/K/∞, where RF denotes the interarrival
1 1 
E[ξ] = [ξ + ξα ]dα. (7) times and service times are random fuzzy variables and the
2 0 α queue discipline is first come, first served(FCFS).
Proof. Let ξ is normalized, i.e., there exists a real number r0 In classic stochastic queueing systems, the event
such that μξ (r0 ) = 1 and if r0 > 0, then the equation (5) can {system is busy at time t} is one of the most
be rewritten as important applications of queueing systems (see [6,7]).
 +∞  r0 It is clear that the busy period and idle period in the
1
E[ξ] = [r0 + Cr(ξ ≥ r)dr + r0 − Cr(ξ ≤ r)dr] random fuzzy queueing systems follow a random fuzzy
2 r0 −∞
alternating renewal process. In random fuzzy alternating

1 1  renewal processes, the event {system is on at time t},
= (ξ + ξα )dα,
2 0 α is a random fuzzy event. Therefore, the event

37
 World Academy of Science, Engineering and Technology
International Journal of Mathematical, Computational Science and Engineering Vol:8 No:3, 2014

λ λ
{random f uzzy queueing system is busy at time t} is Since, E[ μλ ] is finite, than 2+ μα + μα is an integrable function
α α
a random fuzzy event. In what follows, we will concentrate of α. It follows from Fatou’s lemma that
on the average chance of the event random fuzzy queueing  1
system is busy at time t. By definition 8, we have lim inf (Pα (t) + Pα (t))dα
t→∞ 0
 1
Ch{random f uzzy queueing system is busy at time t} ≥ lim inf (Pα (t) + Pα (t))dα,
0 t→∞
 1
= Cr{θ ∈ Θ|P r{queueing system is busy at time t} ≥ p}dp
and  1
0 lim sup (Pα (t) + Pα (t))dα
t→∞ 0
 1
Set P(t)=Pr{queueing system is busy at time t}, then P(t)
≤ lim sup (Pα (t) + Pα (t))dα.
is a fuzzy variable and if Pα 0 and Pα0 are the α0 -pessimistic 0 t→∞
values and the α0 -optimistic values P (t) and E[ μλ ] < 1. Since λα , λα , μα , μα are almost surely continuous at point α,
then
Theorem 1. Assume that, in a random fuzzy queueing
system RF/RF/1/F CF S/K/∞, the fuzzy variables λ has lim Ch{the f uzzy queueing system is busy at time t}
t→∞
the same α0 -pessimistic values and the α0 -optimistic values  1 
1 1 1
= lim (Pα (t)+Pα (t))dp = lim (P  (t)+Pα (t))dp
International Science Index Vol:8, No:3, 2014 waset.org/Publication/9997864

λi and the fuzzy variables μ has the same α0 -pessimistic

2 t→∞ 0 2 0 t→∞ α
values and the α0 -optimistic values μi and are continuous at 
the point α0 , α0 ∈ [0, 1], then we have 1 1 λα λ
= (  + α )dα
2 0 μα μα
λ
lim Ch{the f uzzy queueing system is busy at time t} = E[ ].
t→∞ μ
λ The proof is completed.•
= E[ ]. (11)
μ
Remark 1. It is obvious that
Proof. From definition 13 and proposition 1, it follows that
lim Ch{the f uzzy queueing system is idle at time t}
t→∞
Ch{the f uzzy queueing system is busy at time t} λ
= 1 − E[ ]. (12)
 1 μ
= Cr{θ ∈ Θ|P (t)(θ) ≥ p}dp IV. F UZZY SIMULATION METHOD
0
 In order to evaluate the expected value of a fuzzy variable,
∞
Liu and Liu [2] designed a fuzzy simulation for both discrete
= Cr{θ ∈ Θ|P (t)(θ) ≥ p}dp
0 and continuous cases.
 (a) Discrete fuzzy vector: assume that f is a function, and
1
1 ξ= (ξ1 , ..., ξm ) is discrete fuzzy vector whose joint possibility
E[P (t)] = (Pα (t) + Pα (t))dp
2 0 distribution function is defined by
⎧
The following results have been proved in [8] ⎪
⎪ μ1 , u = u1
⎨
μ2 , u = u2
λα μξ (u) = (13)
lim Pα (t) = , ⎪
⎪ ...
t→∞ μα ⎩
μn , u = un
and where μu = min1≤i≤m μ(i) (ui ) and u = (u1 , ..., um ) ∈ m
λ
lim Pα (t) = α . and μ(i) are the possibility distribution function of ξi for i =
t→∞ μα 1, 2, ..., m.
It follows from the definition of the limit that there exist two Let ai = f (ui ). Without loss of generality ,we assume that
real numbers t1 and t2 with t1 ≥ 0 and t2 ≥ 0 such that for a1 ≤ a2 ≤ ... ≤ an ,then the expected value is given by
any t ≥ t1 n

λ E[f (ξ)] = ai pi , (14)

0 ≤ Pα (t) ≤ α , i=1
μα
where
λα 1 n 1 i
0 ≤ Pα (t) ≤ . pi =(∨ μj − ∨n+1 i−1
j=i+1 μj ) + (∨j=1 μj − ∨j=0 μj ) (15)
μα 2 j=i 2
where (μ0 = μn+1 = 0) for i = 1, 2, ..., n.
Therefor, we have for any t ≥ max(t1 , t2 )
(b) Continuous fuzzy vector: assume that ξ is a continuous
λα λα fuzzy vector with a possibility distribution function μ. In this
0 ≤ Pα (t) + Pα (t) ≤ 2 + + .
μα μα case, we can estimate the expected value by the formula (16).

38
 World Academy of Science, Engineering and Technology
International Journal of Mathematical, Computational Science and Engineering Vol:8 No:3, 2014

TABLE I
T HE AVERAGE CHANCE OF RANDOM FUZZY QUEUEING SYSTEM IS BUSY VI. C ONCLUSION
AT TIME t WITH FUZZY SIMULATION METHOD
In this paper, we considered a fuzzy queueing system that
N umber of iteratins T he average chance
the interarrival times of customers and service times of servers
500 0.6355
are fuzzy random variables and we simulate the average
chance of the random fuzzy event ”the queueing system is
1000 0.6370
busy at time t” by fuzzy simulation. In this method we simulate
5000 0.6393
the expected value of E[ μλ ] (or when it is difficult to compute
10000 0.6422
the expected value) and obtain the real value. If we increase
20000 0.6423
the number of iterations or improve the random numbers in
50000 0.6432
simulation algorithm, we approach to real solution.
70000 0.6432

R EFERENCES
[1] Y. Ning, W. Tang, R. Zhao, Analysis on random fuzzy queueing systems
with finite capacity, Technical Report, 2006.
[2] B. Liu, Y. Liu, Expected value of fuzzy variable and fuzzy expected value
model, IEEE Transactions on Fuzzy Systems 10 (2002) 445-450.
[3] B. Liu, A survey of credibility theory, Fuzzy Optimization and Decision
Making, Vol 5, No. 4, pp. 387-408, 2006.
International Science Index Vol:8, No:3, 2014 waset.org/Publication/9997864

[4] Y.K. Liu, B.D. Liu, Expected value operator of random fuzzy variable and
random fuzzy expected value models, International Journal of Uncertainty,
Fuzziness and Knowledge-Based Systems 11 (2003) 195-215.
[5] B. Liu, A survey of credibility theory, Fuzzy Optimization and Decision
Making, Vol 5, No. 4, pp. 387-408, 2006.
[6] S. Li. Some properties of fuzzy alternating renewal processes, Mathemat-
ical and Computer Modelling 54 (2011) 1886-1896.
[7] D. Gross, C. M. Harris, Fundamentals of Queueing Theory, Wiley, New
York, 1998.
[8] Y. F. Ning, Busy period of random fuzzy queueing systems, International
conference, Hong Kong, 19-22, 2007.

Fig. 1. The convergence of the fuzzy simulation method

V. A PPLICATION E XAMPLE
Now, for illustrating the method, we consider an example.
We present an application practical of this model how using
fuzzy simulation method to estimate the average chance.

Example 1. Let us consider a random fuzzy queueing system.

λ and μ are trapezoidal fuzzy variables, λ = (1/2/3/4) and
μ = (3/4/5/6). We want to calculate the average chance
of ”random fuzzy queueing system is busy at time t”. From
Theorem 1, for estimating the E[ μλ ], we use the simulation
method in Section IV. The simulation results are shown in
Table I and Fig 1.

Table I and Fig 1 show that the average chance of random

fuzzy queueing system is busy at time t after 70000 times is
equal 0.6432, in fact it remains at 0.6432, level. Also, from
Remark 1,

lim Ch{the f uzzy queueing system is idle at time t}

t→∞

λ
= 1 − E[ ] = 1 − 0.6432 = 0.3568
μ

39