The Analysis of Performance Measures of Generalized Trapezoidal Fuzzy Queuing Model with an Unreliable Server

V.R. Bindu Kumari. Int. Journal of Engineering Research and Application www.ijera.com
ISSN : 2248-9622, Vol. 7, Issue 5, ( Part -3) May 2017, pp.136-139
www.ijera.com DOI: 10.9790/9622-070503136139 136 | P a g e
The Analysis of Performance Measures of Generalized
Trapezoidal Fuzzy Queuing Model with an Unreliable Server
V.R. Bindu Kumari 1,
Dr. R. Govindarajan2
1
Assistant Professor, Department of Mathematics, St. Thomas College of Arts & Science , Koyambedu,
Chennai-600 107, Tamilnadu, India.
2
Rtd. Head & Professor, PG & Research Department of Mathematics, D.G. Vaishnav College, Chennai-600
106, Tamilnadu, India.
ABSTRACT
The purpose of this research paper was to propose a method which can be utilized to determine the different
types of performance measures on the basis of the crisp values for the fuzzy queuing model which has an
unreliable server and where the rate of arrival, the rate of service, the rate of breakdown and the rate of repair
are all expressed as the fuzzy numbers. In this case the inter arrival time, the time of service, the rates of
breakdown and the rates of repair are all triangular functions and are also expressed as the Trapezoidal fuzzy
numbers. The main intent is to transform the fuzzy inter arrival time, the time of service, the rates of breakdown
and the rates of repair into the crisp values by using the Ranking function method. Then the crisp values are
applied in the classical formulas for the performance measure. In the fuzzy environment the ranking fuzzy
numbers are very helpful in making the decisions. The ranking function method is one of the most reliable
method, is simpler to apply in comparison to other methods and can be utilized to solve the different types of
queuing problems. In this research paper a numerical example is also provided for both the triangular and the
trapezoidal fuzzy number so that a practical insight into the problem can be provided.
Key words: Fuzzy sets, Fuzzy queues, Fuzzy ranking, Membership Functions, Unreliable server
I. INTRODUCTION
It is not always possible to keep the server in
operational mode at all times in most of the queuing
models which can lead to the interruption in the
services. The main reason behind the interruption of
the services is that the server breaks down. Also
during the weekends and the holidays there can be
the schedules service breakdown of the server. The
first ordinary M/G/I queuing system which has
interrupted services and priorities was first proposed
by Gaver [1]. Then Sengupta [2] extended the system
proposed by Gaver [1] to the GI/G/I case. The
behavior of the of the unreliable server and the
impact of the unreliable server on the server
breakdowns and repairs of the M/G/I queuing models
was first established by the Li [3] and Wang [4].
They also analyzed the controllable M/HK/I queuing
system with the unreliable server. The complex two –
unit system whereby the facility for repair is lead to
the random breakdown is analyzed by Gurukajan and
Srinivasan [5]. With the help of the extension
principle conceptualized by Zadeh [6], the fuzzy
queue with an unreliable server is transformed in to a
family of crisp queues with an unreliable server.
Taking into consideration the notational convenience,
the researcher in this research paper has denoted the
model by FM/FM (FM, FM)/I. In the given model the
first FM represents the fuzzified exponential arrival,
the second FM represents the service, the third FM
depicts the breakdown and the final FM represents
the rate of repair. On the contrary in case of the M/M
(M, M)/I model first to the fourth represents the
exponential inter-arrival (Poisson), service ,
breakdown and the rates of repair respectively.
Jain [7] first proposed the method of
ranking. For the purpose of ordering of the fuzzy
quantities in [0, 1], Yager [8] conceptualized the four
indices. A method for the ranking of the fuzzy
numbers is given by the Kaufman and Gupta [9]. A
subjective method for the ranking of the fuzzy
numbers is given by the Campos and Gonzalez [10].
Further Liou and Wang [11] proposed a ranking
method which was based on the integral value index.
Another ranking method was presented by Cheng
[12] which was based on the distance method.
Interestingly, the overall possibility distributions of
the fuzzy numbers were considered in their
evaluation by Kwang and Lee [13], on the basis of
which they proposed a ranking method. A ranking
method was proposed by Modarres and Nezhad [14]
which was dependent upon the preference function
that evaluates the fuzzy numbers on a point by point
basis and at every point the most preferred number is
determined. A centroid index method for the ranking
of the fuzzy numbers was presented by the Deng and
Liu [15]. The concept of centriod was utilized by
Liang et al. [16] and Wang and Lee [17], in order to
develop the ranking index. The method for the
ranking of the generalized trapezoidal fuzzy numbers
was presented by Chen and Chen [18]. An interesting
RESEARCH ARTICLE OPEN ACCESS

method for the ranking of the trapezoidal fuzzy
numbers was introduced by Abbasbandy and Hajjari
[19], on the basis of the left and the right spreads at
certain α-levels of trapezoidal fuzzy numbers.
Different other researchers have also analyzed the
ranking methods.
In the main intent behind presenting this
particular research paper is to develop a method
which will enable to provide performance measures
in terms of the crisp numbers for the fuzzy queuing
model utilizing an unreliable server. This method will
be based on four fuzzy variables which will be
fuzzified exponential arrival, rate of service, rate of
breakdown and rate of repair. In order to reach the
crisp values the ranking function method has been
utilized.
Definition
A membership function mapping the
elements of the domain space or the universe of
discourse Z to the unit interval [0,1] is utilized for the
determination of the fuzzy set.
i.e Ã = {(z,μ (z)); z Є Z}.
In the above equation μ : Z → [0,1] is a
mapping which is known as the degree of the
membership function of the fuzzy set Ã and μA (z) is
said to the value of membership of z Є Z in the fuzzy
set Ã The membership grades are normally
represented using the real numbers between the range
[0,1].
Also, in the universe of discourse Z a fuzzy set of Ã
is known as the normal fuzzy set is there is atleast
one z Є Z exist so that μ (z) = 1.
i) The fuzzy set can be explained as a membership
function mapping element for a space in a domain or
in a universe of discourse X for the unit of interval
[0,1]. For example A= {x, μA(x)); x ϵ X}, in that case
μA : X [0,1], can be said to be mapping which can
be described as the degree of membership function of
the fuzzy set A. Here μA(x) can be said to the
membership value that is related to x ϵ X of the fuzzy
set A. These grades of membership are generally
represented by real numbers within the range of [0,
1].
ii) The normal fuzzy can be explained as the fuzzy
set A that belongs to the universe of discourse X and
points out that there is atleast one x ϵ X in such a
manner that μA(x) = 1.
iii) The fuzzy set A can be said to be convex fuzzy
set only if for any x1, x2 ϵ X, the inequality μA
{λx1+(1-λ) x2} ≥ min { μA (x1), μA (x2)}. 0≤λ≤ 1, has
to be satisfied by the membership function A.
iv) A trapezoidal fuzzy number can be represented as
A (a,b,c,d;1) which has the membership function μ(x)
is basically expressed as :
μ(x) = (x-a)/(b-a), a≤ x≤b
1, b≤ x≤c
(d-x)/(d-c), c≤ x≤d
0, otherwise
v) The α-cut of a fuzzy number A(x) is generally
defined as:
(α) = [ x: μ(x) ≥ α, α ϵ [0,1]}
It is possible to add two trapezoidal fuzzy numbers in
the following manner:
(a1,b1,c1,d1) +(a2,b2,c2,d2) = (a1+a2,b1+b2,
c1+c2,d1+d2).
vi) Generalized fuzzy number: A generalized fuzzy
number may be defined as the fuzzy set Ã, which can
be measured on the universal set of the real numbers
R, if the membership function has the below given
characteristics:
a)  A
: R → [0,w] is continuous;
b)  A
(z) = 0 for all z ∈(−∞,a1
] [ a4
,∞);
c)  A
(z) strictly increasing on [a1
,a2
] and strictly
decreasing on [a3
,a4
];
d) A
~ (z) = w, for all z ∈[a2
,a3
], where 0 < w ≤ 1.
ix) Gneralized trapezoidal fuzzy numbers : The
generalised trapezoidal fuzzy numbers Ã (z) is generally
expressed as Ã ( a1 , a2 , a3, a4; 1), where the mebership
function is illustrated as:
[w(z-a1)/ (a2-a1)], a1  z  a2
μ (z) = w , a2  z  a3
[w(z-a4)/ (a3-a4)] , a3  z  a4
0,otherwise
Fuzzy Queuing Model With An Unreliable Server
There are two major groups in which the
queuing models can be divided into. One of the group
can be said to be those group which describes the
actual events or real events and the other group
provides an optimum view point about the what
should be aspired. The descriptive models basically
describes the average values and probabilities that are
related to the performance measures which can be
utilized to describe the system after the completion of
the setting up of the patterns of arrivals and the
service, the server numbers, the capacity of the
system and the discipline of the queue. On the other
hand the second category of models which are also
termed as the queuing decision models is basically
utilized to estimate the different parameters which
can be used to optimize the models. In general the
basic system of queuing generally consists of the
queues that are organized and the units are generally
dealt according to their order of arrival. In most of
the queuing models the waiting discipline can be
seen, but there are certain privileged classes of units
which hold certain degree of priority in the work
mode of the system are defined as per their efficiency
and hierarchy. For example the status of the message

that has to be transmitted in a system of
telecommunication. The priority discipline model is
generally followed more closely by the real queuing
system than any other model for queuing that are
available in the present times. In the priority model
the urgent tasks are performed first than the other
tasks and the main clients are given special
preferences.
There have been several methods which
have been proposed so as to resolve the design and
the control of the queuing models in case when the
coefficients of the costs and the arrival or the patterns
of the service are known exactly. In certain
conditions these parameters can be said to be not
precise, thus in that case the unit cost of waiting
overtime for every client can vary overtime. The
main problem that has been found in such type of
models with that of the fuzzy cost efficient is that it
can be very difficult to make a proper estimate of the
waiting unit cost, most importantly when the nature
of the human affects the working of the system. Due
to the wide range of situations the cost of service and
the costs related to waiting and inactivity can be
highly uncertain, which are beyond the control of the
humans. In case of the conventional queuing decision
models, the quality of the decision making can be
undermined by the initial fuzzy information. This
point out the need to study in greater depth the
problems related to decision making which has the
fuzzy queuing models.
Let us take into consideration a fuzzy
queuing system which has an unreliable server and
the there are two different types of breakdown of the
system. In the type 1 of breakdown, there can be
breakdown of the server where there are no
customers present in the system and in the type 2 of
breakdown, there is a breakdown of the server when
there is at least one customer present at the system.
Let us assume that the arrival of the customers takes
place in a single-server facility like that of a Poisson
process and have the fuzzy rate of λ. Furthur let us
assume that the service times are in the form of the
exponential distribution which has the fuzzy rate μ, and
following the Poisson process the serverr might have
the breakdown which has the fuzzy rate of α and
finnaly exponential distribution which is followed by
the repair has the fuzzy rate as β. Let these functions
are known approximately as the convex fuzzy sets. let
us take into consideration that the memebership
functions of the λ,μ,α and β are expressed by ϕλ (x), ϕμ
(y), ϕα (s) and ϕβ (t). Thus we get the following fuzzy
sets:
λ = {(x, ϕλ (x)) │ x∈ X }
μ = {(y, ϕμ (y)) │ y∈ Y }
α = {(s, ϕα (s)) │ s∈ S }
β = {(t, ϕβ (t)) │ t∈ T }
in the above equation X,Y, S and T can be defined as
the crisp universal sets of the arrival rates, rates of
service, rates of breakdown and the rates of repair
respectively. Let us assume that f(x,y,s,t) represents
the characterists of the system which is of interest. As
x,y,s,t are denoted as the fuzzy numbers thus
f(x,y,s,t) are also fuzzy numbers. Let us take into
assumption that the membership function of the
expected time for which the system will remain idle
in the breakdowns of type 1 and type 2, be expressed
by A and B.
then,
A = f(x,y,s,t) = ty x(s t)
y(s t)
and, B = f(x,y,s,t) = ty  x(s t)
ty
In case of the steady state, it is necessary that 0 <
ty x(s t) < 1 and
y(s t)
0 < ty x(s t) < 1.
ty
Ranking Function Method – Algorithm
In order to solve the problems that has been
provided we have to first defuzzify the fuzzy
numbers into the crisp numbers, by utilizing the
fuzzy numbers Ranking function R: F(R)→ R where
F(R) is a set of fuzzy numbers which are generally
expressed on the set of real numbers, which is used
for the mapping of each fuzzy numbers on the real
line, where a natural order exists.
Let the convex trapezoidal fuzzy number
Ã(z) = Ã ( a1 , a2 , a3 , a4 : w) ,
=
Numerical Example
Let us take into consideration the railway
computer reservation system where the passengers are
utilising a single channel reservation arrival as per the
poisson process. There may be interruption in the time
taken for the reservation process (due to the problems
in the computer, server or the problems in the printers)
according to the Poisson process. The time of recovery
of the problems caused due to interruption is followed
by the exponential distribution. When the interruption
is over the reservation process resumes again. The
railway wishes to evaluate the total number of hours the
system has been in idle condition so as to analyse the
single-channel reservation system. It is known that this
system will follow the FM/FM(FM,FM)/1/∞ and the
expected time for which the system is going to remain
idle can be calculated from the procedure that has been
proposed below.

Solution
Now let us take into consideration that
arrival rate, rate of service, rate of breakdown and the
rates of repair are the trapezoidal fuzzy numbers
which are generally expressed as :
 = [2,4,5,7;1], = [12,14,16,20;1],
 = [0.05, 0.1, 0.2, 0.5;1],
β = [2,3,5,6;1] per hour.
As per equation we can write that,The ranking index
of λ is given by,
R(λ) = R(2,4,5,7;1) = (24 5+7) = 4.5
4
Progressing furthur on the similar basis we get,
R(  ) = R(12,14,16,20;1) =
(12 14+16  20) = 15.5
4
R( ) = R(0.05, 0.1,0.2, 0.5;1) =
(0.050.2 0.5) = 0.2125
4
R(β ) = R(2,3,5,6;1) = (2 ) =4
4
As per the equations we can write that,
The expected time for which the sysytem will remain
idle can be expressed as:
Type 1: ty x(s t)
y(s t)
= 4(15.5) 4.25(0.21254) = 0.659
15.5(0.21254)
Type 2: ty x(s t)
ty
= 415.5) 4.5(0.2125 4) = 0.694
4(15.5)
II. CONCLUSION
In this particular research paper, the Fuzzy
set theory has been applied on a queuing model
which has an unreliable server. For the evaluation of
the system performance the fuzzy queuing model
with the unreliable server has put in use for the
operations and the mechanisms for service. With the
help of the ranking indices the fuzzy problem has
been transformed into the crisp problem. As the
performance measures like that of the expected time
for which the system will remain idle for two types
are the crisp values and thus it will be helpful for the
manager to take best and the optimum decisions.
Finally it is concluded that the Ranking function
method is a very effective way which can be utilized
to find out the problems related to the fuzzy queues
with an unreliable server. The approach that has been
shown in this research paper gives a practical insight
for the operations and systems managers and the
practitioners.
REFERENCES
[1]. W. Li, D. Shi, X. Chao, Reliability analysis
of M/G/I queuing systems with server
breakdowns and vacations,Journal of
Applied probability 34 (1997) 546-555.
[2]. K.H. Wang, H.J. Kao, G. Chen, Optimal
management of aremarkable and non-
reliable server in an infinite and a finite
M/Hk/I queuing systems ,Quality
Technology and quantitative Management
1(2) (2004) 325-339.
[3]. Gurukajan, M. and Srinivasan, B. (1995). A
complex two unit system with random
breakdown of repair facility.
Microelectronics and Reliability, 35(2): 299-
322.
[4]. L.A. Zadeh, Fuzzy sets as a basis for a
theory of possibility, fuzzy sets and systems
1 (1978) 3-28.
[5]. Jain R (1976) Decision-making in the
presence of fuzzy variables. IEEE
Transactions on Systems,Fuzzy Inf. Eng.
(2010) 1: 37-47 47
[6]. Yager R R (1981) A procedure for ordering
fuzzy subsets of the unit interval.
Information Sciences 24(2): 143
[7]. Kaufmann A, Gupta M M (1988) Fuzzy
mathematical models in engineering and
management science. Elsevier Science
Publishers, Amsterdam, Netherlands
[8]. Campos L, Gonzalez A (1989) A subjective
approach for ranking fuzzy numbers. Fuzzy
Sets and Systems 29(2): 145
[9]. Liou T S, Wang M J (1992) Ranking fuzzy
numbers with integral value. Fuzzy Sets and
Systems 50(3): 247
[10]. Cheng C H (1998) A new approach for
ranking fuzzy numbers by distance method.
Fuzzy Sets and Systems 95(3): 307
[11]. Kwang H C, Lee J H (1999) A method for
ranking fuzzy numbers and its application to
decision making. IEEE Transaction on
Fuzzy Systems 7(6): 677
[12]. Modarres M, Sadi Nezhad S (2001) Ranking
fuzzy numbers by preference ratio. Fuzzy
Sets and Systems 118(3): 429

The Analysis of Performance Measures of Generalized Trapezoidal Fuzzy Queuing Model with an Unreliable Server

More Related Content

What's hot (18)

Similar to The Analysis of Performance Measures of Generalized Trapezoidal Fuzzy Queuing Model with an Unreliable Server (20)

Recently uploaded (20)

The Analysis of Performance Measures of Generalized Trapezoidal Fuzzy Queuing Model with an Unreliable Server