See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/369969914

Analysis Of Fuzzy Queueing Models To Optimize The Average System Size In

Packet-Switching Network

Article · April 2023

CITATIONS READS

0 157

4 authors, including:

Ebenesar Anna Bagyam Suganthi Poomalai

Karpagam Academy of Higher Education RMK Engineering College
49 PUBLICATIONS 72 CITATIONS 21 PUBLICATIONS 55 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Ebenesar Anna Bagyam on 12 April 2023.

The user has requested enhancement of the downloaded file.

Journal of Survey in Fisheries Sciences 10(1) 234-240 2023

Analysis Of Fuzzy Queueing Models To Optimize The Average

System Size In Packet-Switching Network
Ebenesar Anna Bagyam J1*, Suganthi P2, Sudha R3, Manoj Kumar Chande4
1
*Department Of Mathematics, Avinashilingam Institute For Home Science And Higher
Education For Women Coimbatore, India
2
Department Of Mathematics, R.M.K. Engineering College, Chennai, India
3
Department Of Mathematics, Dr. N.G.P. Institute Of Technology, Coimbatore, India
4
Department Of Applied Mathematics, Shri Shankaracharya Institute Of Professional
Management And Technology, Raipur, Chhattisgarh, India

*Corresponding author: - Ebenesar Anna Bagyam J

*Department Of Mathematics, Avinashilingam Institute for Home Science And Higher
Education For Women Coimbatore, India, E-mail: [email protected]

Abstract
Single server batch arrival retry queue with varying modes of breakdowns and two stages of restoration is
considered under a fuzzy situation. Assuming the arrival, working, retry, breakdown, and repair rates are fuzzy;
a mathematical programming method is proposed to construct the membership function for the average system
size of the prescribed model. To convert a fuzzy repairable retrial queue into a family of crisp queues, the
alpha-cut technique and Zadeh's extension concept are utilised. Trapezoidal fuzzy numbers are used to
illustrate the strength of the proposed approach. Ranking fuzzy numbers play a huge role in decision-making
under fuzzy conditions. This ranking method is the most reliable, simple to apply, and used to find the
defuzzification of the syatem measures. A particular application in packet-switching network is given for a
better understanding of the model.

Key words: Fuzzy queue, Server Breakdown, packet switching network, Zadeh’s extension principle, alpha
cut approach, Trapezoidal fuzzy number

1 Introduction and Relevant Literature applications, statistical data can be acquired

The remarkable extension of queues to the individually; for example, the arrival rate and
fuzzy world has practical consequences for working rate are better expressed by language
decision analysis, operations research, phrases such as rapid, modest, or sluggish than
computer technology, and abstract theory. by a probability distribution based on
Because of its practical applicability in real statistical theory. This type of indefinite
life, the fuzzy retrial queuing model has evidence will correctly identify the structure
recently gained much interest. Fuzzy retrial performance measurement.
queues have a variety of real-world
applications, including communication Zadeh (1965, 1973, 1978) established the idea
networks, production models, financial of fuzziness to deal with imperfect
sectors, tollbooths, and service stations, information. The notion of fuzzy set theory is
among others. well-known for modeling imprecision or
uncertainty from mental processes. Yager
When the characteristics of the queuing (1986) discussed the extension concept and the
system, such as the arrival rate and working ranging approach for defuzzification.
rate, are identified precisely, effective methods Unfortunately, their method only supplied
for examining has been devised. However, simple remedies. In other words, the
these parameters may not be given accurately performance measures' membership functions
in some situations due to uncontrolled are not fully explained. Li and Lee (1989)
circumstances. In various practical suggested an all-encompassing method for
234
Journal of Survey in Fisheries Sciences 10(1) 234-240 2023

queuing systems in a fuzzy environment. managed by the membership functions. By

Analytical data for the M/F/1 and FM/FM/1 extending the fuzzy environment, the
systems are offered to demonstrate the proposed approach may more accurately
methodology. Fuzzy queues are far more represent fuzzy queues with an unstable
realistic than the often utilised crisp queues in server, and the analytical findings associated
many real-world scenarios. Kao et al. (1999) with this queuing model will be informative
employed computational programming to and relevant for system developers and users.
construct the membership functions of four
fundamental fuzzy queues with one or two The system features of a queueing model with
fuzzy variables: M/F/1, F/M/1, F/F/1, and fuzzy customer arrival, retry, and operative
FM/FM/1, where F represents fuzzy time and rates were built by Ke et al. (2007). In this
FM signifies fuzzified exponential time. case, a fuzzy retrial queue is converted using
Zimmermann (2001) highlighted various real- the -cut technique into a family of traditional
world applications of fuzzy set theory. crisp retrial queues. A collection of parametric
non-linear programmes has been built to
In a recent study on fuzzy queueing theory characterise the family of crisp retrial queues
based on possibility theory, Buckley et al . using the membership functions of the system
(2001) explore the fundamental findings. characteristics. More data is offered for
Then, he use this to solve two application management's usage since the membership
problems: the first is a machine servicing issue functions express and regulate the system
to finding the ideal amount of repair teams, features. The fuzzy retrial-queue is more
and the second is a queueing decision issue to correctly represented by extending this model
finding the ideal number of servers). to the fuzzy nature, and analytical findings are
improved.
Chen (2005) demonstrated how to build
membership functions for performance In order to handle fuzzy threshold-based space
indicators in bulk service queuing systems priority buffers, Wang et al. (2009) created a
with fuzzy arrival and working rates. Two discrete-time queueing model and tested its
fuzzy queues that are frequently seen in effectiveness in real-world scenarios. The
transportation management are used as analysis of the pertinent performance metrics,
examples to show the viability of the such as the packet loss probability of high-
suggested technique. Machine interference is a priority traffic and of low-priority traffic, is
prevalent problem in manufacturing and done using a matrix-analytic technique.
production activities. Because of According to intuition, the fuzzy threshold
uncontrollable reasons, the parameters of the produces a reduced packet loss probability for
machine interference problem may be low-priority packets by adapting effectively to
ambiguous. Chen (2006) suggested a various input traffic flow conditions and the
mathematical programming technique for packet loss rate criteria of high-priority
constructing the membership function of the packets. Kalyanaraman et al. (2010)
machine interference system's measure, with investigated a retry queueing system with
the machine breakdown rate and working rate interruptions in a fuzzy environment.
being fuzzy numbers.
Viswanathan et al. (2015) developed the non-
Ke and Lin (2006) defined the membership linear parametric programs to give a
functions of the key variables of a queuing description about crisp retrial family queues
model with an unstable server, in which the with Coxian 2 vacation. Comparative analysis
customer arrival rate and operation rate, as of both crisp and fuzzy retrial two phase
well as the server failure rate and restoration queueing model with Bernoulli vacation and
rate, are all vague values. An efficient restricted admission was given by Ebenesar
approach is offered to locate the best answers Anna Bagyam and Udaya Chandrika (2019).
at each level of possibility. More information
is available for management to utilise as a Fuzzy Markovian queues, in which all of the
result of the system attributes being stated and system constraints are fuzzy numbers, were

235
 Analysis Of Fuzzy Queueing Models To Optimize The Average System Size In Packet-Switching Network

studied by Chen et al. (2020) for the optimum it is susceptible to one of the M modes of
and equilibrium techniques. Using approaches breakdowns. The failing server requires
from fuzzy logic and queuing theory, this work necessary repair based on the kind of
investigates the membership functions of the breakdown and optional common repair.
optimal and equilibrium procedures in both When a server breaks, it stops serving clients
tangible and intangible circumstances. In order and waits for the necessary repair to begin. The
to characterise the family of crisp techniques, time spent waiting for the server to be repaired
we build two parametric nonlinear is referred to as setup time. As quickly as the
programmes using Zadeh's extension concept necessary restoration is accomplished, the
and the alpha - cut methodology. Then, in customer may choose common repair with a
single and multi-server models, the particular probability p or deny with the
membership functions of the tactics are complimentary probability. Before the service
determined. To estimate the stability approach provider went down, the client who was now
in the fuzzy intelligence, the grated mean receiving service either stayed in the service
integration approach is also used. position with probbaility r or left the amenity
area with complementary probability and kept
Kannadasan and Padmavathi (2021) discussed coming back with certain rate. Once repaired,
about the fuzzy techniques with use of the server either waits for the same client or
hexagonal fuzzy numbers and the authors have resumes serving the client who was stopped.
acquired the fuzzy environment with the This period of waiting is known as reserved
presence of numerical results. Revathi and time. Until the interrupted consumer exits the
Selvakumari (2021) given the priority system, the server cannot accept new clients.
disciplines of fuzzy retrial models. Fuzzy If the server is busy or unavailable, it is
retrial queue model is applied to most of the referred to as blocked. (Ebenesar Anna
telecommunication systems. In this paper, we Bagyam and Udayachandrika, 2018)
can find the suitable application of the
prescribed model in a packet-switching 2.1 Notations Used
network. As a result, a single server batch  - Arrival rate  - service rate  - retrial
arrival retry queue with varying modes of rate
breakdowns and two stages of restoration  - failure rate 1, 2 - necessary and common
would be appropriate for the above-stated reapir rates
problem. The major goal of this article is to  - setup rate m1 and m2 - First two moments
analyze the mean system size under fuzzy of the batch arrival
conditions.
2.2 Solution Methodology
The remaining part of the paper is organised as Assume that the arrival rate (), retrial rate (),
follows:: The construction of a fuzzy service rate (), setup rate (), essential repair
repairable queue is given in Section 2. By rate (1) and common repair rate (2) are
considering arrival, service, and repair rates as ~ ~,  , ~
a trapezoidal fuzzy numbers, numerical results represented by fuzzy numbers  , ~ ,  1
~
are given in Section 4. Section 5 presents and 2 respectively. Let  ~ ( x),  ~ ( y),
results and discussion based on the numerical
~ (s) ,  ~ (v1 ) ,  ~ (v 2 ) and  ~ ( v 3 ) denote
illustration. Finally, in the last section, the 1 2

conclusion has been drawn by highlighting the the membership functions of the
novelty of the investigation. corresponding parameters. Then we have
~
 = {(x,  ~ ( x )) / x  X}
2 Construction of Fuzzy Repairable Retrial ~
Queue  = {(y,  ~ ( y )) / y  Y}

2.1 Overview of the Model ~ = {(s, ~ ( s)) / s  S}
 
Single server batch arrival queue with M ~ = {(v1, ~ ( v )) / v1  V1}
 1
modes of breakdown and two stages of repair
~
with retrial is considered under a fuzzy 1 = {(v2,  ~ ( v 2 )) / v2  V2}
1
environment. Although the server is running,

236
Journal of Survey in Fisheries Sciences 10(1) 234-240 2023

~
and 2 = {(v3, ~ ( v 3 )) / v3  V3}
M
x m1
2 T1 = 1 – [1 +   i T4] + m1 (1 – A
s i =1

where X, Y, S, V1, V2 and V3 are the crisp sets (x))

of arrival, retry, service, setup, essential repair x m12 M
m
~ T2 = [ (1 +   i T4) + 2 ] (1 – A
and optional repair rates respectively. Ls(  , ~  s i =1 2
,~,  , ~ ~
1 , 2 ) is fuzzy numbers since  , ~
~
, T
(x)) + 3
~,  , ~ ~ 2
 1 and 2 are all fuzzy numbers. Based M
2 2 M
on Zadeh’s extension principle (Zadeh, 1978), T3 = x2 m12 [ 2 (1 +   i T4)2 +   i [
~ ~ s i =1 s i =1
the membership function Ls (  , ~ ,  ~,  , 
1,
~ r 1 1 1 p
2 ) is defined as  L ( ~ , ~ , ~ , ~ , ~ , ~ ) (z) = ( + + + )
S 1 2   v1 v 2 v3
 L~ (z) Hence,  L~ (z) = 1 1 1 p p
S S + + 2+ 2
+ 2+ + ]] +
sup min { ~ ( x ),  ~ ( y),  ~ (s),   ( v1 ),  ~ ( v 2 ), v1 v2 v1 v2 v3 v1 v 3 v pv 2 3
1
 M
x m2
 ~ (v 2 ),  ~ (v 3 )} (1) [1 +   i T4]
1 2 s i =1
where  = {x  X, y  Y, s  S, v1  V1, v2 1 1 p r
 V2, v3  V3 / T4 = + + +
v1 v 2 v3 
M
x m1
[1 +   i T4] + (1 – A  (x)) m1 < 1} y
s and A (x) =
i =1
y+x
(2)
T x m1 M
Ls(x, y, s, v1, v2, v3) = 2 [1 +   i ~
The crisp intervals for the -cuts of  , ~ ~,
, 
T1 s i =1
~ ~ ~
T4 ] (3)  ,  and  and upper and lower bound of
1 2
the invervals are tabulated as below:

Table 1 Crisp Intervals, Upper and Lower Bounds

Crisp Intervals Upper Bound Lower Bound
() [ min {x /  ~ (x)  }, max {x /  ~ (x)  }] −1
max  ~ () = x U min  −~1 () = x L
xX xX 
() [ min {y /  ~ ( y)  }, max {y /  ~ ( y)  }] max  −~1 () = y U min  −~1 () = y L
yY yY
() [ min {s / ~ (s)  }, max {s / ~ (s)  }] max  −~1 () = s U min  −~1 () = s L
s S s S
() [ min {v1/  ~ ( v1 ) }, max {v1/  ~ (v1 )  }] max  −1 () = v 1 U , min  −~ 1 () = v1 L
v1  V1 v1  V1
1() [ min {v2/  ~ (v 2 ) }, max {v2/  ~ (v 2 )  }] max  −~1 () = v 2 U min  −~1 () = v 2 L
v 2  V2 1 v 2  V2 1 1 1

2() [ min {v3/  ~ ( v 3 )  }, max {v3/  ~ ( v 3 )  }] −1

max  ~ () = v 3 
U −1
min  ~ () = v 3 
L
v 3  V3 2 v 3  V3 2 2 2

To build the membership function  L~ (z) - to x L  x  x U , y L  y  y U , s L  s  s U ,

s ,
L U
cuts approach is used. ( L s )  and (L s )  are v1 L  v1  v 1 U , v 2 L  v2  v 2 U and v 3 L 
the lower bound and the upper bound of the v3  v 3 U . The membership function  L~ (z)
s
-cuts of  L~ (z) respectively. Thus, ( L s ) L = can be defined as follows:
s

min Ls (x, y, s, v1, v2, v3) and (L s ) U  L~ (z) = L(z), if (L s ) L =0  z  (L s ) L =1

s

= max Ls (x, y, s, v1, v2, v3), subject  L~ (z) = 1, if (L s ) L =1  z  (L s ) U=1

s and
 L~ (z) = R (z), if (L s ) U=1  z  (L s ) U=0
s

237
 Analysis Of Fuzzy Queueing Models To Optimize The Average System Size In Packet-Switching Network

where L(z) = inverse of ( L s ) L = Left shape the membership function  L~ (z) for different
s

L q ) U values of  is shown in Fig. 1. Moreover,

function and R(z) = inverse of = Right  L~ (z) appears a trapezoidal like structure as
shape function respectively. s

the arrival rate, retry rate, working rate, setup

2.3 Defuzzification - Ranking Method rate and repair rate (essential and optional) are
For defuzzify the fuzzy value of mean system trapezoidal fuzzy members. Figure 2 reports
size into a crisp one Yager ranking index the -cuts of expected system size (LL- Lower
method is used. bound and LU - Upper bound) for different -
The formula for Yager ranking index is 1 values.
2
1
 [(L s )  + (L s )  ] d ,
L U

0
~
Where [ ( L s ) L , (L s ) U ] is the -cuts of L s .

3 Numerical Illustration
Assume that arrival rate, retry rate, working
rate, setup rate, essential repair rate and
optional repair rates are trapezoidal fuzzy Fig. 1 Mean System Size’s Membership
~
number such that  = [0.5, 1, 1.5, 2], ~  = [25, Function
~ ~
26, 27, 28],  = [28, 29, 30, 31],  = [5, 6, 7,
~ ~
8], 1 = [15, 16, 17, 18] and 2 = [17, 18, 19,
20] with the fixed values M = 4, i = 2.5, m1 = Mean System Size
2, m2 = 6, pi = 0.6, ri = 0.5 and i = 5 (i = 1, 2,
3, 4).
The values of ( L s ) L and (L s ) U are obtained
as
(L s ) L = [( + 1) (− 2 8 + 241 7 − 9153 6
+ 26506 5 + 7631928 4 Fig. 2 -cut values of Mean System Size
− 239135100 3 + 3300827064 2 −
22186570720  4 Discussion on the Results
+ 59110464000)] / [10 ( − 8) ( − 18) ( − The range of the system length for possibility
20) (5 5 – 515 4 level alpha = 1 is about [1.4479, 3.9387],
+ 19061 3 – 308961 2 + 2091714  − suggesting that it is certainly feasible that the
4244640)] expected number of consumers in the system
and falls between 1.4479, 3.9387, however this is
(L s ) U = [− ( − 4) (− 2 8 − 193 7 − 4596 uncertain. The range of the system length for
possibility level alpha = 0 is around [0.4835,
6 + 95723 5 + 7010268 4 18.6644] at the other extreme. This range
+ 149453025 3 + 1557960000 2 + suggests that the mean system size will never
8015423125  fall below 0.4835 or increase above 18.6644.
+ 16420016250)] / [10 ( + 5) ( + 15) ( + 4.6877 is the predicted system size when
17) (5 5 + 440 4 utilising the Yager ranking index approach.
+ 13331 3 + 163872 2 + 699000  + The information mentioned above will be very
276000)] beneficial when developing a queueing
system.
Because of complexity in finding the closed
form expression of devloping we analyze 5 Practical Application of the model
 L~ (z) numerically. The numerical result of Our retrial queue has potential uses in a
s
packet-switching network, where messages

238
Journal of Survey in Fisheries Sciences 10(1) 234-240 2023

are split into IP packets before being

transmitted, in addition to its theoretical 6 Conclusion and Future Work
appeal. For instance, packet-switching The well-established classical retrial queuing
technologies constitute the foundation of the systems, which are difficult but usually make
majority of contemporary WAN protocols, assumptions that are too far from reality. But
such as TCP/IP, X.25, and Frame Relay. A by using fuzzy retrial queue it describe
router is a connected device in a packet- realistic scenarios. Fuzzy retrial queues are
switching network that is used to transfer more precise and perfect than conventional
packets from a host server to a remote host. To lineups. The model's output from this study
transfer a package from a source host to a may be used to analyze the variables impacting
destination, the base station must first transmit packet-switching networks. In the subsequent
the package to the router to which it is linked research, we will deal with generalizing
before sending the package to the remote host. alternative ways to assist decision-makers in
deciding the path for modification in
Suppose that packages arrive at the source host repairable retrial queues.
through a Poisson mechanism. All packages
received by the host are immediately sent to its REFERENCES
router. If the router is available, the package is 1. Buckley, J.J., Feuring, T. and Hayashi, Y.,
accepted and sent right away, with the 2001. Fuzzy queueing theory revisited.
assumption that the transmission time will be International Journal of Uncertainty,
evenly split. If not, the router rejects the Fuzziness and Knowledge.-Based Sys-
package owing to current malfunctions or tems, 9(5):527-537.
MTU (Maximum Transmission Unit) 2. Chen G, Liu Z and Zhang J, 2020.
restrictions on the TCP/IP network path. Analysis of strategic customer behavior in
According to FCFS, in this scenario, the fuzzy queueing systems. Journal of
blocked package must be retransmitted at a Industrial and Management Optimization,
later time and is kept in the source host's buffer 2020, 16 (1): 371-386.
(referred to as the orbit). Furthermore, the 3. Chen S P, 2005. Parametric Nonlinear
router may malfunction whether it is idle or Programming Approach to Fuzzy Queues
when sending packets, due to outside assaults with Bulk Service. European Journal of
or other technical issues. Operations Research, 163(1): 434-444.
4. Chen S P 2006. A Mathematics Program-
We assume that until a packet arrives at the ming Approach to the Machine
router, the network administrator in charge of Interference Problem with Fuzzy Para-
fault management always conducts certain meters. Applied Mathematics and
auxiliary activities while the router is idle and Computation 174(1): 374-387.
is always on service when the router is active. 5. Ebenesar Anna Bagyam J. and Udaya
If a router fails while transmitting a packet, the Chandrika K., 2019. Fuzzy Analysis of
network administrator can immediately repair bulk arrival two phase retrial queue with
it, and the router will resume broadcasting the vacation and admission control. The
paused packet as fast as the repair is Journal of Analysis, 27(1): 209 – 232.
completed. If the router fails while it is 6. Ebenesar Anna Bagyam, J. and Udaya
inactive, the restoration may not begin until the Chandrika, K., 2018. MX/G/1 Retrial
next packet from an outside source comes or Queue with different modes of failure,
until the orbit when the network manager Journal of Physics: Conference series,
emerges and immediately begins the router 1742 - 6588, 1139 012031.
repair operation. The delayed period refers to 7. Kannadasan, G. and Padmavathi, V. 2021.
the span between the epoch of the passive Classical Fuzzy Retrial Queue with
failure and the epoch of the next packet Working Vacation using Hexagonal
arriving. Once the repair for the passive failure Fuzzy Numbers. Journal of Physics:
is complete, the packet that came even during Conference Series. 2070.
delayed interval can sometimes be transmitted 8. Kao, C., Li, C.C., and Chen, S.P. 1999.
immediately. Parametric programming to the analysis

239
 Analysis Of Fuzzy Queueing Models To Optimize The Average System Size In Packet-Switching Network

of fuzzy queues, Fuzzy sets and Systems,

107(1): 93-100.
9. Kalyanaraman.R, Thillaigovidan.N and
Kannadasan.G, 2010. A single server
fuzzy queue with unreliable server,
International Journal of Computational
Cognition, 8(1): 1 - 4.
10. Ke J.C. and Lin C.H., 2006. Fuzzy
analysis of queueing systems with an
unreliable server: A nonlinear program-
ming approach, Applied Mathematics and
Computation, 175(1): 330-346.
11. Ke J.C., Huang H.I. and Lin C.H., 2007.
On retrial queueing model with fuzzy
parameters, Physica A: Statistical
Mechanics and its Applications, 374 (1):
272-280.
12. Li, R. J., and Lee, E.S., 1989. Analysis of
fuzzy queues, Computers and Mathema-
tics with Applications, 17(7): 1143-1147.
13. Revathi S and Selvakumari K, 2021. Cost
Analysis of Fuzzy Queueing Models,
Advances and Applications in Mathe-
matical Sciences, 21(2): 743-747.
14. Viswanathan A, Udaya Chandrika K and
Ebenesar Anna Bagyam J, 2015. Fuzzy
Retrial Queueing System with Coxian - 2
Vacation. International Journal of Applied
Engineering Research. 10(8): 21315-
21322.
15. Wang Y. C., Wang J. S. and Tsai F. H.,
2009. Analysis of discrete-time space
priority queue with fuzzy threshold,
Journal of Industrial and Management
Optimization, 5 (2), 467-479.
16. Yager R .R. 1986. A Characterization of
the extension principle, Fuzzy sets and
system, 18(3): 205-217.
17. Zadeh L .A. 1978. Fuzzy sets as a basis
for theory of possibility, Fuzzy sets and
system, 1(1):3-28.
18. Zadeh L. A., 1973. Outline of a New
Approach to the Analysis of Complex
Systems and Decision Processes, IEEE
Transactions on Systems, Man and
Cybernetics, 3(1): 28-44.
19. Zadeh L.A., 1965. Fuzzy Sets,
Information and Control 8: 338-353.
20. Zimmermann, H.J., 2001. Fuzzy set
Theory and Its Applications, Kluwer
Academic, Boston.

240

View publication stats