You are on page 1of 15

Applied Mathematical Modelling 36 (2012) 255–269

Contents lists available at ScienceDirect

Applied Mathematical Modelling
journal homepage: www.elsevier.com/locate/apm

A batch arrival retrial queue with general retrial times under
Bernoulli vacation schedule for unreliable server and delaying repair
Gautam Choudhury a,⇑, Jau-Chuan Ke b
a
Mathematical Sciences Division, Institute of Advanced Study in Science and Technology, Paschim Boragaon, Guwahati 781035, Assam, India
b
Department of Applied Statistics, National Taichung Institute of Technology, No. 129, Section 3, Sanmin Road, Taichung 404, Taiwan, ROC

a r t i c l e i n f o a b s t r a c t

Article history: This paper deals with the steady state behaviour of an Mx/G/1 queue with general retrial
Received 4 June 2009 time and Bernoulli vacation schedule for an unreliable server, which consists of a break-
Received in revised form 18 May 2011 down period and delay period. Here we assume that customers arrive according to com-
Accepted 22 May 2011
pound Poisson processes. While the server is working with primary customers, it may
Available online 16 July 2011
breakdown at any instant and server will be down for short interval of time. Further con-
cept of the delay time is also introduced. The primary customer finding the server busy,
Keywords:
down or vacation are queued in the orbit in accordance with FCFS (first come first served)
Stationary system size distribution
Random breakdown
retrial policy. After the completion of a service, the server either goes for a vacation of ran-
Delay time dom length with probability p or may continue to serve for the next customer, if any with
Repair time probability (1  p). We carry out an extensive analysis of this model. Finally, we obtain
Retrial time some important performance measures and reliability indices of this model.
Reliability indices Ó 2011 Elsevier Inc. All rights reserved.

1. Introduction

Retrial queues (or queues with repeated attempts) are characterized by the feature that a customer who finds the server
busy upon arrival is obliged to leave the service area and repeat his demand after some time called ‘‘retrial time’’. Between
trials, the blocked customer joins a pool of unsatisfied customers called ‘‘orbit’’. Queues in which customers are allowed to
conduct retrials have been widely used to model many practical problems in telephone switching systems, telecommunica-
tion networks and computers competing to gain service from a central processing unit. Moreover, retrial queues are also
used as mathematical models for several computer systems: Packet switching networks, shared bus local area networks
operating under the carrier-sense multiple access protocol and collision avoidance star local area networks, etc. For a review
of main results and methods, the reader is referred to the survey papers by Yang and Templeton [1], Falin [2], Kulkarni and
Liang [3] and the book by Falin and Templeton [4]. For more recent references see the bibliographical overviews in [5–7].
Further, a comprehensive comparison between retrial queues and their standard counterpart with classical waiting line
can be found in Artalejo and Falin [8].
Many of the queueing systems with repeated attempts operate under the classical retrial policy, where each block of cus-
tomers generate a stream of repeated attempts independently of the rest of the customers in the orbit. However, there is a
second kind of policy, called constant retrial policy, which arises naturally in problems where the server is required to search
for customers (e.g., see [9]) and in communication protocols of type carrier sense multiple access (CSMA). The latter disci-
pline was introduced by Fayolle [10], who investigated an M/M/1 retrial queue in which the repeat customers from a queue
and only the head customers of the orbit queue can request a service after an exponentially distributed retrial time with

⇑ Corresponding author.
E-mail addresses: gc.iasst@gmail.com (G. Choudhury), jauchuan@ntit.edu.tw (J.-C. Ke).

0307-904X/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.apm.2011.05.047

Most of the analysis for retrial queues concerns the exhaustive service schedule [21]. who assumed that each customer in the orbit generates a stream of repeated attempts that are independent of the customer in the orbit and state of the server. even in the most recent studies.17–20]. Section 3 deals with the derivations of the stability criteria for existence of stationary regime and the steady state joint distribution of the server state and the number of customers in the system..25–27] have recently appeared in the queueing literature in which concept of general retrial times have been introduced along with Bernoulli vacation schedule under the FCFS orbit retrial policy. repeated attempts with general retrial time. If the queue is not empty then service begins with specific probability q or a vacation period begins with probability p = 1  q. then whole batch joins the retrial group. Gomez-Corral [51] dis- cussed extensively an M/G/1 retrial queue with FCFS discipline and general retrial times. the server waits for the first customer to arrive. Moreover. Takin and Sengupta [34] and Tang [35]. i. The first investigation with general retrial time done by Kapyrin [13].26] discussed more complicated queueing situations with retrials and batch arrivals. Ke / Applied Mathematical Modelling 36 (2012) 255–269 some parameter e (say). [38] studied a repairable M/G/1 retrial queueing model from the viewpoint of reliability for the first time. the service channel undergoes repair instantaneously. gated service policy [22] and recently modified vacation policy [23]. retrial rate is (1  d0. Other wise. where the number of individual primary customers arrives to the system according to a compound Poisson process with arrival rate k.e. Sengupta [33]. On the other hand the batch arrival retrial queue was first considered by Falin [40]. its applicability is connected with the per- formance evaluation of Local Area Networks operating under transmission protocols like the CSMA/CD (carrier sense multiple access with collision detection). X2. no work have been found that combine these features together for unreliable server queueing systems with batch arrivals. where . details of which may be found in [9. [12]. then one packet is selected to be transmitted automatically and the rest is stored in a buffer (i.[15] have developed an approximation method to obtain the steady state performance for the model of Kapyrin. Thirurengadan [30] and Mitrany and Avi-ltzhak [31] for some fundamental works. At the end of a vacation period service begins if a customer is present in the queue. as soon as it fails. In this connection. Section 5 deals with the system size distribution at a departure epoch. Choudhury. then they are divided into a number of packets in order to be transmitted to the destination station. Later. several retrial mod- els have been analyzed with general retrial times. The mathematical model We consider an MX/G/1 queueing system. production and quality control problems. . Farahmand [11] called this discipline a retrial queue with FCFS orbit retrial policy.-C. . in this article an attempt has been made to study an MX/G/ 1 retrial queue with general retrial times and Bernoulli vacation schedule for an unreliable server.e. who assumed the following operating rule: ‘‘If the server is busy at the arrival epoch. which consists of a server’s breakdown period and delayed period. A number of papers [13. Recently. a bus in the engineering terminology) is idle. Yang et al. etc. Although some aspects have been discussed separately on queueing systems with service interruptions. we give a brief description of the mathematical model. Some recent papers [41–43.. see Choi et al. Such type of queueing models occur in many real life situations where the server may be used for other secondary jobs. To this end.256 G. In Section 2. [12] generalized this retrial policy by considering an M/M/1 retrial queue with general retrial times. i. The size of successive arriving batches is X1. On the other hand. [32]. Retrial queueing systems with general service times and non-exponential retrial times have also received considerable attention during last decade. Some important particular cases of this model are discussed briefly in Section 6. Among some early papers on service interruptions.e. The cost effectiveness maximization of this model is discussed with an application to Wire- less Network System in Section 8. While Li et al. another important characteristic for considering the retrial model with general retrial times is that we always obtain analytical solu- tion in term of closed form expression. On the other hand retrial queues that take into account servers failures and repairs were introduced by Aissani [36] and Kulkarni and Choi [37].j denotes Kronecker’s delta). customized manufacturing. In such a context. If the transmission median (i. Existence of the Stochastic decomposition property is also dem- onstrated in Section 4. . then one of the arriving units starts its service and the rest join the retrial group’’. However. the retrial group). 2. where concept of repeated attempts with Bernoulli vacation schedule is also introduced under different retrial policy. J. maintenance activities. Choi et al. The rest of the paper is organized as follows.. whereas if the server is free. we refer the readers to see those by Gaver [28]. Hence to fill up to this gap. Keilson and Servi [24] introduced Bernoulli vacation schedule: if the queue is empty after service com- pletion then the server becomes inactive. Avi-ltzhak and Naor [29]. Applications arise naturally in call centers with multi task employees. messages of variable length arrive at the stations and. among others have studied some queueing systems with interruptions wherein one of the underlying assumption is that.. if the bus is busy then the entire packet must be stored in the buffer and the station will retry the transmission later on. begins a vacation period. Bernoulli sche- dule vacation. A wide class of retrial policies for governing the vacation mechanism has also been discussed in the literature. A more complete description of this mechanism can be found in Yang and Temple- ton [1]. The study of queueing models with service interruptions goes back to the 1950s. Choudhury and Deka [39] investigate such a repairable M/G/1 retrial queueing model with two phases of service. Subsequently. Allowing server to take vacations makes the queueing model more realistic and flexible in studying real word queueing sit- uations. this methodology was found to be incorrect by Falin [14]. telecommunica- tion and computer networks. Some important performance measures and reliability indices of this model are derived in Section 7.n)e (where di. the methodology will be based on inclusion of supplementary variables.e. when number of units in the orbit is ‘n’. In recent years. for instance to serve customers in other systems. Wang et al. however.

uj ¼ ka½1 b1 f1 þ aðd1 þ g 1 Þg þ kpa½1 c1 þ a½1 ð1  C  ðkÞÞ  1. . server’s delay times.i. n P 1} of the customers are i. Theorem 3. a[1](1  C⁄(k)) + ka[1]{b1(1 + a(g1 + d1)) + pc1} = a[1](1  C⁄(k)) + qH < 1 is a sufficient condition for ergodicity The same inequality is also necessary for ergodicity. LST G⁄(h) = E[ehG] and finite kth moment gk.i.d random variable with d.f. elapsed delay time and elapsed repair time of the server as supplementary .d random variables with d.e.i. server’s repair times.e. we have reasonable conclusion that the term k a[1]{1 + a(g1 + d1)} + pka[1]c1 represents a batch arrival during service time. for k P 1. server is fixed).This type of queueing model is known as MX/G/1 Bernoulli vacation queue with unreliable server during which a breakdown period and a delay period occur. pp. . s 2 Z+ and e > 0 such that the mean drift us = E[f(Xn+1)  f(Xn)jXn = s] is finite for all s 2 Z+ and us 6 e for all s 2 Z+ except perhaps a finite number. in a orbit in order to seek service again and again until it finds the server free. namely uj < 1 for all j P 0 and there exists j0 2 Z+ such that uj P 0 for j P j0. J. G.d. Similar interpretation can be provided for j = 0. Choudhury. C(x) and LST C⁄(h) = E[ehR].j = 0 for j < i  k and i > 0. Ke / Applied Mathematical Modelling 36 (2012) 255–269 257 X1.. The term a[1] (1  C⁄(k))  1 refers to the contribution to the orbit size due to a batch arrival during the retrial time excluding the arbitrary customer of the arriving batch whose service commences so that he/she no longer belongs to the orbit.e. where qH = ka[1]{b1(1 + a(g1 + d1)) + pc1}. V(y). server’s life times.i. X2. n 2 Z+} is an irreducible and a periodic Markov chain. The condition uj < 0 ensure that the orbit size does not grow indefinitely in course of time. the server is ready to start its remaining service to the primary cus- tomers and in this case the service times are cumulative. Obviously. random variables with probability distribution function (d. of the customer and elapsed vacation time. . It is easy to see that {Xn.f.f.) random variables with probability mass function (p. for k P 1. [32]. Proof. . on vacation or down. The time between successive repeated attempts.where P = (pi.-C. where if a primary customer who finds the server busy. it may breakdown at any time and the server will be down for a short interval of time.) B(x).g. While the server is serving the primary customers. for k P 1.. X(t)} as t ? 1 exist and positive under the same condition {Xn. together with a batch arrival when the server is on vacation both leaving for the orbit.i. we take f(s) = s to obtain ( ka½1 b1 f1 þ aðg 1 þ d1 Þg þ pka½1 c1  1.i.n P 1. V( ). Kaplin’s condition is satisfied because there is a suchk such that for pi.d.e. service times and retrial time random variables are mutually independent of each others. Next. attempts have been made to obtain the PGF of the joint distribution of the state of the server and number in the system (by system size here we mean including one is being served. if any) by treating elapsed retrial time.d random variable with d. i. 187–188) that the steady state probabilities of our bivariate Markov process Z(t) = {N(t).The service times {Bn. j ¼ 1. j ¼ 0. The model of this nature without a delay period and single unit arrival case was first investigated by Li et al. which states that an irreducible and aperiodic Markov chain is positive recurrent if there exists a non-negative function f(s).e. n 2 Z+}.m. if any.1.e. are identically and independently distributed (i. . which we may referred to as generalized service times. The service interruption. Notice that in our case. it can be shown from Burke’s theorem (e. Hence if a[1](1  C⁄(k)) + qH < 1 and G( ).f. then he/she leaves the service area and joins a group of unsatisfied customers. i. n P 1} of the server are i. LST c⁄(h) = E[ehV] and kth finite moment ck. We define this waiting time as delay time. n P 1g of the server are also assumed to be i. C( ) and B( ) satisfy the regularity conditions. As soon as breakdown occurs it is sent for repair during which the server stops providing service to the primary customers and waits for repair to start. Further.Now further develop- ment for such a type of model we may further introduce the concept of repeated attempts with FCFS orbit retrial policy.. . D(y). server’s life times are generated by exogenous Poisson process with rates a. 2. To prove the positive recurrent we may use the Foster’s criterion. n 2 Z+}. then the system is stable.. i. i. where p + q = 1.. From the mean draft uj = ka[1]{1 + a(g1 + d1)} + pka[1]c1 + a[1](1  C⁄(k))  1 for j P 1. retrial times {Rn. after each service completion the server may go for a vacation of random length V with probability p or with probability q he may serve the next unit. [44]..e. which we may refer to as waiting period of the server. The inequality a[1](1  C⁄(k)) + qH < 1 is a necessary and sufficient condition for the system to be stable. n P 1} of the customers are assumed to be a sequence of i. elapsed service time.f) an = Pr{X = n}. we can guarantee nonerodicity if the Markov chain {Xn. n 2 Z+} satisfies Kaplin’s condition. The vacation times fV n .j) is the one step transition matrix of {Xn. service channel is interrupted for a short interval of times. Laplace Stielt- jes Transform (LST) b⁄(h) = E[ehB] and finite kth moment bk. if and only if a[1](1  C⁄(k)) + qH < 1. i. After each service completion the server takes Bernoulli vacation. i.f. see Cooper [45]. probability generating function (PGF) a(z) = E[zX] and first two factorial moments a[1] and a[2]. which we may call some short of disaster during serving the customers. i. Joint distribution of number in the orbit and state of the server We first investigate the necessary and sufficient condition for the system to be stable of our model. Then a[1](1  C⁄(k)) + qH P 1 implies the nonergodicity of the Markov chain. we assume that input process. The delay times {Dn. n P 1g of the server are assumed to be i. D( ).... 3. we are considering preemptive resume for service time. In our case.e. As noted in Sennott et al.d random vari- ables with d. h Since the arrival process is a Poisson process. LST D⁄(h) = E[ehD] and kth finite moment dk. G(y). The repair times fGn . Immediately after the broken server is repaired (i.

XðtÞ ¼ H0 ðtÞ.. if the ser v er is idle with zero customer in the system at time t.1. > > > > > 4. ðx. x < R0 ðtÞ 6 x þ dxg. yÞ þ ½k þ nðyÞK n ðx. x > 0. t!1 Q n ðyÞdy ¼ lim Pr fNðtÞ ¼ n. yÞ ¼ k ak K nk ðx. t!1 wn ðxÞdx ¼ lim Pr fNðtÞ ¼ n. let D0(t). y < H0 ðtÞ 6 y þ dyjB0 ðtÞ ¼ xg. D(y) and G(y) are continuous at y = 0 respectively. Choudhury.0 Þ ak Q nk ðyÞ. n P 1. YðtÞ ¼ > > > 3. if the ser v er is busy at time t. ð3:5Þ dy k¼1 . D( ) and G( ). XðtÞ ¼ D0 ðtÞ. respectively at timet. X(t)}. B0(t). V( ). J. Assuming the system is in steady state conditions. XðtÞ ¼ R0 ðtÞ. R0(t). nP0 t!1 and for fixed values of x and n P 1 Rn ðx. V(0) = 0. The steady state equations The Kolmogorov forward equations to govern the system under steady state conditions (e. ðx. B(x) is continuous at x = 0 and V(y). yÞ ¼ k ak Rnk ðx. y > 0. : 5. yÞ > 0: t!1 Further. respectively elapsed repair time of the server during which breakdown occurs in the system and elapsed vacation time of the server at time t. yÞ. Now we define following limiting probabilities: P0.258 G. H0(t) and V0(t). X(t) = D0(t) if Y(t) = 4. t!1 Pn ðxÞdx ¼ lim Pr fNðtÞ ¼ n. x > 0. be the elapsed delay time. X(t) = V0(t) if Y(t) = 3. number of customers in the retrial group) at time t.e. XðtÞ ¼ V 0 ðtÞ.0 ¼ lim Pr fNðtÞ ¼ 0. n P 1. XðtÞ ¼ 0g. B(1) = 1. n P 1. Further.g. V0(t). G(0) = 0. In addition. it is assumed that C(0) = 0. G(1) = 1 f and that C(x). where X(t) = 0 if Y(t) = 0.-C. if the ser v er is waiting for repair at time t. we introduced the following random variable: 8 > 0. ð3:2Þ dx k¼1 0 d Xn Q n ðyÞ þ ½k þ cðyÞQ n ðyÞ ¼ kð1  dn. see Cox [46]) can be written as follows: d w ðxÞ þ ½k þ gðxÞwn ðxÞ ¼ 0. Ke / Applied Mathematical Modelling 36 (2012) 255–269 variables. X(t) = R0(t) if Y(t) = 1. x < B0 ðtÞ 6 x þ dxg. > > > < 2. lðxÞdx ¼ . n P 0. yÞdy ¼ lim Pr fNðtÞ ¼ n. so that dCðxÞ dBðxÞ dVðyÞ dDðyÞ dGðyÞ gðxÞdx ¼ . t!1 K n ðx. C(1) = 1. n P 1. fðyÞdy ¼ and nðyÞdy ¼ 1  CðxÞ 1  BðxÞ 1  VðyÞ 1  DðyÞ 1  GðyÞ are the first order differential (hazard rate) functions of C( ). yÞdy. yÞ þ ½k þ fðyÞRn ðx. if the ser v er is on v acation at time t. B( ). n P 1. ð3:1Þ dx n Xn Z 1 d Pn ðxÞ þ ½k þ a þ lðxÞP n ðxÞ ¼ k ak Pnk ðxÞ þ fðyÞRn ðx. yÞ > 0. D0(t) and H0(t) are introduced in order to obtain a bivariate Markov process Z(t) = {N(t). Let N(t) be the orbit size (i. 3. respectively. and X(t) = H0(t) if Y(t) = 5. B0(t) be the elapsed retrial time and service time of the customers.. y < D0 ðtÞ 6 y þ dyjB0 ðtÞ ¼ xg. yÞ. XðtÞ ¼ B0 ðtÞ. if the ser v er is idle during retrial time at time t. D(0) = 0. > > > > 1. yÞdy ¼ lim Pr fNðtÞ ¼ n. y < V 0 ðtÞ 6 y þ dyg. X(t) = B0(t) if Y(t) = 2. D(1) = 1. if the ser v er is under repair at time t: So that the supplementary variables R0(t). ð3:4Þ dy k¼1 d Xn K n ðx. n P 1. cðyÞdy ¼ . ð3:3Þ dy k¼1 d Xn Rn ðx. V(1) = 1. B(0) = 0.

y. The model solution To solve the system Eqs. x > 0. (3. 0Þ. n P 1. 0. x > 0. n P 1 ð3:8Þ i¼1 0 0 and at y = 0: Z 1 Q n ð0Þ ¼ p lðxÞPnþ1 ðxÞdx. yÞdy. ð3:10Þ Z 1 K n ðx. zÞ ¼ zn K n ðx.-C. y.0 ¼ cðyÞQ 0 ðyÞdy þ q lðxÞP1 ðxÞdx. Choudhury. zÞ½1  CðxÞ expfkxg. (3. n¼1 n¼1 X 1 X 1 X 1 wðx. zÞ ¼ zn wn ð0Þ and aðzÞ ¼ zn a n : n¼1 n¼1 n¼1 Let b(z) = k(1  a(z)). n P 1. y) occurring in Eqs. wð0. 0Þ ¼ aPn ðxÞ. ð3:13:1Þ Pðx. ð3:11Þ 0 R1 where 0 w0 ðxÞdx ¼ 0 occurring in Eq. yÞdx dy ¼ 1: n¼0 0 n¼1 0 0 0 0 0 0 ð3:12Þ 3. zÞ ¼ Kðx.11) let us define following PGFs for jzj < 1: X 1 X 1 Rðx. we get a set of differential equations of Lagrangian type whose solutions are given by: wðx. 0.2. Also. zÞ½1  VðyÞ expfbðzÞyg.n ðxÞdx þ wn ðxÞdx þ Rn ðx. then proceeding in usual manner with Eqs. zÞ ¼ zn P n ðxÞ.0 þ gðxÞwn ðxÞdx. (3. zÞ ¼ zn Q n ð0Þ. 0Þ ¼ fðyÞRn ðx. 0. ð3:13:2Þ Q ðy.4) and (3.7) by zn and then taking summation over all possible values of n P 1. zÞ ¼ Qð0. Pð0. zÞ ¼ zn Rn ðx. Again multiplying both sides of Eq. we get . ð3:13:3Þ Rðx. yÞ. zÞ½1  GðyÞ expfbðzÞyg. 0Þ. These set of equations are to be solved under the boundary conditions at x = 0: Z 1 Z 1 wn ð0Þ ¼ cðyÞQ n ðyÞdy þ q lðxÞPnþ1 ðxÞdx. ð3:9Þ 0 also at y = 0 and fixed values of x Rn ðx. n P 0. ð3:13:4Þ and Kðx. yÞ. y. (3.2). 0. y. n¼0 n¼0 X 1 X 1 Pðx. n¼1 n¼1 X 1 X 1 Kðx. ð3:6Þ 0 0 where P0(x) = 0. zÞ ¼ zn Pn ð0Þ. Rðx. zÞ½1  DðyÞ expfbðzÞyg. zÞ ¼ zn K n ðx. (3. G. zÞ ¼ zn wn ðxÞ. ð3:7Þ 0 0 X n Z 1 Z 1 Pn ð0Þ ¼ k ai wni ðxÞdx þ kan P0. zÞ ¼ Rðx. J. R0(x. y) = 0 = K0(x. zÞ ¼ zn Rn ðx. zÞ½1  BðxÞ expfAðzÞxg. zÞ ¼ zn Q n ðyÞ.1)–(3. Ke / Applied Mathematical Modelling 36 (2012) 255–269 259 Z 1 Z 1 kP0. (3. n P 1.1)–(3.5) respectively. the normalizing condition is given by 1 Z X 1 1 Z X 1 Z 1 Z 1 Z 1 Z 1 Z 1  P0. yÞdx dy þ K n ðx. Kðx.5).0 þ Q n ðyÞdy þ Pi.8). ð3:13:5Þ ⁄ ⁄ where A(z) = b(z) + a(1  D (b(z))G (b(z))). n¼1 n¼1 X 1 X 1 Q ðy. zÞ ¼ Pð0. zÞ ¼ wð0. Q ð0.

0. zÞ½1  BðxÞ expfAðzÞxg ð3:17Þ and Kðx. 0. Ke / Applied Mathematical Modelling 36 (2012) 255–269 wð0.11). Theorem 3. zÞ ¼ ð3:20Þ BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z P0. we get zPð0. ð3:24Þ ½1  a½1 ð1  C  ðkÞÞ  qH  ka½1 C  ðkÞP0.2. ð3:29Þ BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z p½1  a½1 ð1  C  ðkÞ  qH ÞbðzÞ½1  VðyÞ expfbðzÞyg Qðy. zÞ½aðzÞð1  C  ðkÞÞ þ C  ðkÞ ð3:15Þ and Q ð0.0 þ wð0. zÞ ¼ kaðzÞP0. zÞ ¼ aPð0. (3. zÞD ðbðzÞÞ: ð3:18Þ Now from Eqs. (3. zÞb ðAðzÞÞ: ð3:16Þ Again from Eqs. zÞ ¼ Q ð0. Under the stability condition a[1](1  C⁄(k)) + qH < 1. zÞ ¼ . (3.27) represents steady state probability that the server is idle but available in the system.2. (3.12).0 : ð3:14Þ Similarly from Eqs. zÞ ¼ . ð3:19Þ BðzÞfC ðkÞ þ aðzÞð1  C  ðkÞÞg  z  P 0. Let z ? 1 in (3. 1Þ ¼ : ð3:26Þ ½1  a½ ð1  C  ðkÞÞ  qH  Hence from normalizing condition (3.16)–(3.0 ½1  BðxÞ½1  DðyÞ Rðx.0 ½1  BðxÞ½1  GðyÞ and Kðx.10) and (3. ð3:30Þ BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z .0 Pð0. zÞb ðAðzÞÞ  kP0.0 pbðzÞb ðAðzÞÞC  ðkÞ and Q ð0. ð3:28Þ C  ðkÞ½BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z z½1  a½1 ð1  C  ðkÞ  qH ÞbðzÞ½1  BðxÞ expfAðzÞxg Pðx. zÞ ¼ z1 pPð0. we have Rðx. we get P0. zÞ ¼ . zÞ ¼ . we get ½1  a½1 ð1  C  ðkÞÞ  qH  P0. ð3:23Þ ½1  a½1 ð1  C  ðkÞÞ  qH  pka½1 C  ðkÞP 0. Choudhury.260 G.9). 1Þ ¼ . zÞ ¼ Rðx. zÞ ¼ . Thus we sum- marize our results in the following Theorem 3. ð3:21Þ BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z where B(z) = {q + pc⁄(b(z))}b⁄(A(z)). the joint distribution of the state of the server and the size of the orbit has the following partial PGFs as k½1  a½1 ð1  C  ðzÞÞ  qH ½z  aðzÞBðzÞ½1  CðxÞ expfkxg wðx. 1Þ ¼ . J.19). 1Þ ¼ ð3:25Þ ½1  a½1 ð1  C  ðkÞÞ  qH  ka½1 C  ðkÞP0.0 zbðzÞC  ðkÞ Pð0.8) and (3.0 ½1  BðxÞ Pðx.0 ¼ : ð3:27Þ C  ðkÞ Note that the Eq.-C. we obtain the following equation by the L’ Hospital’s rule ka½1 C  ðkÞP0. y.18). zÞc ðbðzÞÞ þ qz1 Pð0. y. 0. 1Þ ¼ : ð3:22Þ ½1  a½1 ð1  C  ðkÞÞ  qH  This gives ka½1 C  ðkÞP0.0 ½1  VðyÞ Qðy.0 k½z  aðzÞBðzÞ wð0.

the marginal PGFs of the server’s state orbit size distribution are given by ½1  a½1 ð1  C  ðkÞÞ  qH ½z  aðzÞBðzÞð1  C  ðkÞÞ wðzÞ ¼ .32) with respect to y.39) and (3. Let Uj and Xj be the stationary distribution of the number of customers in the system and orbit respectively under P P1 j the steady state condition. zÞð1  D ðbðzÞÞÞ Rðx. Ke / Applied Mathematical Modelling 36 (2012) 255–269 261 az½1  a½1 ð1  C  ðkÞÞ  qH bðzÞ½1  BðxÞ expfAðzÞxg  ½1  DðyÞ expfbðzÞyg Rðx. respectively. zÞD ðbðzÞÞð1  G ðbðzÞÞÞ Kðx.1. Again. zÞ ¼ Rðx. Under the stability condition a[1](1  C⁄(k)) + qH < 1. we may write . integrating Eq. h Corollary 3.0. ð3:42Þ fq þ pc z respectively. ð3:39Þ 0 bðzÞ Z 1 aPðx. ð3:33Þ C  ðkÞ½BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z ½1  a½1 ð1  C  ðkÞÞ  qH zbðzÞ½1  b ðAðzÞÞ PðzÞ ¼ . we get formulae (3. y. P(z).40) with respect to x. Integrating (3.37). (3.35). Further utilizing the relationship between orbit size distribution and system size distribution for an MX/G/1 type retrial queue.29) with respect to x and using the well known result of renewal theory Z 1 ½1  f  ðsÞ esx ð1  FðxÞÞdx ¼ . zÞ ¼ . G.36) and (3.2.30) with respect to y. ð3:35Þ ½BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z az½1  a½1 ð1  C  ðkÞÞ  qH ½1  b ðAðzÞÞ½1  D ðbðzÞÞ RðzÞ ¼ ð3:36Þ AðzÞ½BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z az½1  a½1 ð1  C  ðkÞÞ  qH ½1  b ðAðzÞÞD ðbðzÞÞ½1  G ðbðzÞÞ and RðzÞ ¼ : ð3:37Þ AðzÞ½BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z Proof. then its corresponding PGFs UðzÞ ¼ 1 j j¼0 z Uj and XðzÞ ¼ j¼0 z Xj are given by ½1  a½1 ð1  C  ðkÞÞ  qH ð1  zÞ½q þ pc ðbðzÞÞb ðAðzÞÞ UðzÞ ¼ ð3:41Þ fq þ pc ðbðzÞÞgfC  ðkÞ þ aðzÞð1  C  ðkÞÞgb ðAðzÞÞ  z ½1  a½1 ð1  C  ðkÞÞ  H ð1  zÞ q and XðzÞ ¼  ðbðzÞÞgfC  ðkÞ þ aðzÞð1  C  ðkÞÞgb ðAðzÞÞ . zÞdy ¼ .34). With the help of PGFs w(z).0 þ wðzÞ þ PðzÞ þ zQðzÞ þ RðzÞ þ KðzÞ: By direct calculation we can obtain (3.-C. ð3:32Þ BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z where b(z) = k(1  a(z)). ð3:34Þ AðzÞ½BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z p½1  a½1 ð1  C  ðkÞÞ  qH b ðAðzÞÞ½1  c ðbðzÞÞ Q ðzÞ ¼ . A(z) = b(z) + a(1  D⁄(b(z))G⁄(b(z))) and B(z) = b⁄(A(z))  {q + pc⁄(b(z))}. y.41). we get Z 1 aPðx. we claim in formula (3. ð3:31Þ BðzÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z az½1  a½1 ð1  C  ðkÞÞ  qH D ðbðzÞÞbðzÞ½1  BðxÞ expfAðzÞxg  ½1  GðyÞ expfbðzÞyg Kðx.31) and (3. respec- tively. ð3:38Þ 0 s where F( ) is distribution function of a random variable and f⁄( ) represents LST of it. Choudhury. y. Theorem 3.28) and (3.3. Proof. K(z) and P0. J. zÞ ¼ . we get formula (3. we get the dis- tribution of the number of customers in the system at a random epoch having PGF UðzÞ ¼ P0. The result follows directly from Theorem 3. integrating (3. zÞdy ¼ : ð3:40Þ 0 bðzÞ Further integrating (3. h Next we are interested to investigate marginal orbit size distribution due to system state of the server. zÞ ¼ Kðx. Q(z).33) and (3. Similarly. R(z). y.

Then the sequence Nn = N(sn+) form a Markov chain which is the embedded Markov renewal process of the continuous time Markov process Z(t) as t ? 1.2) of Krishna Kumar et al.42) is consistent with the expression (4. [26]. mean of additional system size due to retrials is found to be qH ð1  C  ðkÞÞa½1 ½ka2 p½c2 þ 2c1 b1 f1 þ aðd1 þ g 1 Þg½1  C  ðkÞa½1 Lv ¼ v0 ð1Þ ¼ þ ½1  a½1 ð1  c& ðkÞÞ  qH  2ð1  qH Þ½1  a½1 ð1  c& ðkÞÞ  qH  ½ka½1 2 ½b2 f1 þ aðd1 þ g 1 Þg2 þ ab1 fd2 þ 2d1 g 1 þ g 2 g½1  C  ðkÞa½1 þ :  2ð1  qH Þ½1  a½1 ð1  c& ðkÞÞ  qH  Remark 4. we obtain (3. 1.g.. Distribution of number of customers in the system at a departure epoch In this section first we derive the PGF of the stationary system size distribution a departure epoch. [15] and Yang and Tampleton [1]. we consider the epochs at which the required service time requested by an unit expires.e.-C.0 þwð1Þ which is equal to ½1  a½1 ð1  C  ðkÞÞ  qH ½fq þ pc ðbðzÞÞgb ðAðzÞÞ  z vðzÞ ¼ : ð1  qH Þ½fq þ pc ðbðzÞÞgb ðAðzÞÞfC  ðkÞ þ aðzÞð1  C  ðkÞÞg  z More specifically. i.43). The expected number of customers in the orbit during the retrial time.40) by putting C⁄(k) = 1 in the formula (3.1)] and 2.. i. Stochastic decomposition for the retrial models has also been found in Yang et al. The additional system size distribution due to retrials [represented by the second term of right hand side of expression (4. Under the steady state condition a[1](1  C⁄(k)) + qH < 1.1.1). h 5. see Doshi [47].41) in (3. we may call it additional system size distribution due to retrial time. the PGF of the steady state system size distribution at a departure epoch for this type of model is given by ½1  a½1 ð1  C  ðkÞÞ  qH ½1  aðzÞfq þ pc ðbðzÞÞgb ðAðzÞÞ pðzÞ ¼ : ð5:1Þ a½1 ½fq þ pc ðbðzÞÞgb ðAðzÞÞ  z Proof. e. Thus we have ð1  qH Þð1  zÞ½q þ pc ðbðzÞÞb ðAðzÞÞ U0 ðzÞ ¼ : ð4:2Þ fq þ pc ðbðzÞÞgb ðAðzÞÞ  z Note that for a = 0 (i.1. Stochastic decomposition In this section we study the stochastic decomposition property of the system size distribution of our model.42). we observe that the system size distribution at a departure epoch of the unreliable MX/ G/1 retrial queue with delaying repair and generalized repeated attempts under Bernoulli vacation schedule can be decomposed into distributions of two independent random variables viz. and Fuhrmann and Cooper [48]. Let sn be the time of nth total service completion epoch.e. The literature on vacation models recognizes this property one of the most interesting features in this mater.e. The key result of this section is now stated in Theorem 5.262 G. J. The existence of the stochastic decomposition property for our model can be demonstrated easily by showing that UðzÞ ¼ U0 ðzÞvðzÞ. Ke / Applied Mathematical Modelling 36 (2012) 255–269 XðzÞBðzÞ ¼ UðzÞ: ð3:43Þ Now utilizing (3. which can obtained from formula (3..1. The system size distribution of an MX/G/1 queue with unreliable server and delaying repair under Bernoulli vacation sche- dule [represented by first term of right hand side of expression (4. ð4:1Þ X where U0(z) is the PGF of the system size distribution of an M /G/1 queue with unreliable server and delaying repair under Bernoulli vacation schedule.. Choudhury.0 þwðzÞ idle ¼ ½P0. Theorem 5.40). for reliable server) the above result is consistent with the result obtained by Choudhury and Madan [49] and v(z) is the PGF of the conditional distribution of the number of customers in the orbit given that the system is ½P 0.1)]. h 4. It should be noted here that for a = 0 above expression (3. From expression (4. Next we define the following limiting probabilities .

J. i. ð5:2Þ 0 0 where B0 is the normalizing constant. we get on simplification B0 ½1  a½1 ð1  C  ðkÞÞ  qH bðzÞfq þ pc ðbðzÞÞgb ðAðzÞÞ pðzÞ ¼ : ð5:3Þ fq þ pc ðbðzÞÞgb ðAðzÞÞ  z Utilizing normalizing condition p(1) = 1..g. then we have ½1  aðzÞ pðzÞ ¼ /ðzÞ ¼ HðzÞ/ðzÞ. we may write . If we compare expression (3. Let W(z) be the PGF of the number of customers in the system at immediately after those time points at which a customer departs or an idle period is terminated of an Mx/G/1 retrial queue with Bernoulli vacation schedule and unreliable server. the PGF of a batch of customers who arrived during our generalized service time. ð5:8Þ ½BðzÞ  z where C0 is a constant to be evaluated. j P 0: Now since the arrival process is compound Poisson process.30). Choudhury. This is due to randomness property of the arriving size of the batches.2) by both sides of z and then taking summation over j P 0 and utilizing Eqs. Section 4) WðzÞ ¼ W0 ðzÞvðzÞ. pp. Now utilizing B(z) in (5. then multiplying Eq. Cooper [45]. Then following argument of PASTA we note that a departing customer will see ‘j’ units in the system just after a departure if and only if there were (j + 1) units in the system just before the departure and therefore we may write Z 1 Z 1 pj ¼ B0 q lðxÞPjþ1 ðxÞdx þ B0 cðxÞQ j ðxÞdx. 237]: C 0 ½BðzÞ  zaðzÞ W0 ðzÞ ¼ . j P 0}. ð5:5Þ a½1 ð1  zÞ where HðzÞ ¼ a=½1aðzÞ ð1Þð1zÞ is the PGF of the number of customers placed before an arbitrary test customer (tagged customer) in an admission batch in which the tagged customer arrives.5) it is observed that the PGF of system size distribution at a departure epoch of this type of model is the convolution of distributions of three independent random variables. in the well- known result of Gross and Harris [16] which is of the form [see Section 5. Ke / Applied Mathematical Modelling 36 (2012) 255–269 263 pj ¼ n!1 lim P r fNn ¼ jg. as expected.1. Now form above expression (5. p. then under steady state condition a[1](1  C⁄(k)) + qH < 1 we have ½1  a½1 ð1  C  ðkÞÞ  qH ½fq þ pc ðbðzÞÞgb ðAðzÞÞ  zaðzÞ WðzÞ ¼ : ð5:6Þ ð1 þ a½1  qH Þ½fq þ pc ðbðzÞÞgfC  ðkÞ þ aðzÞð1  C  ðkÞÞgb ðAðzÞÞ  z Proof. we get required expression (5.2. h Remark 5.1).8). ð5:7Þ where W0(z) is the PGF of the stationary queue size distribution at which a customer departs or at termination epoch of an idle period for an Mx/G/1 type of retrial queue with Bernoulli vacation schedule and unreliable server.. h Next.-C. 187– 188) that the departure point probabilities {pj} exist and positive under the same condition a[1](1  C⁄(k)) + qH < 1. (5. it follows from Burke’s theorem (e. Theorem 5.1. The interpretation of the other two random variables is provided in Remark 4. we get B0 ¼ ½ka½1 1 : ð5:4Þ Now inserting (5. P Let pðzÞ ¼ 1 j j j¼0 z pj be the PGF of {pj. This number is given as backward recurrence time in the discrete time renewal process where renewal points are generated by the arrival size random variable and a[1] = a0 (1).9. (3. G.. we are interested about the distribution for the number of customers in the system immediately after those time points at which a customer departs or an idle period is ended.1). j P 0.3). This is the relation ship between the system size distributions at a random point of time and at departure point of time for MX/G/1 type of retrial queue with unreliable server.4) in (5.e. The result follows directly by utilizing stochastic decomposition property for an Mx/G/1 retrial queue with Bernoulli vacation schedule and unreliable server viz (e.g. one of which is the number of customers placed before a tagged customer in a batch in which the tagged customer arrives.1.29) and (3.40) with expression (5. This can be obtained easily by replacing the original service time distribution by our generalized service time distribution or equivalently B(z) = {q + pc⁄(b(z))}b⁄(A(z)).

for single unit arrival case)..e. there is no breakdown in the system) and a(z) = z (i. then we have C⁄(k) = m(k + m)1 and therefore from Eqs. there is no Bernoulli vacation in the system) and a = 0 (i.. ð6:7Þ fC  ðkÞ þ aðzÞð1  C  ðkÞÞgb ðAðzÞÞ  z ½1  a½1 ð1  C  ðkÞÞ  qH ð1  zÞ XðzÞ ¼ ð6:8Þ fC ðkÞ þ aðzÞð1  C  ðkÞÞgb ðAðzÞÞ  z  ½1  a½1 ð1  C  ðkÞÞ  qH ð1  aðzÞÞb ðAðzÞÞ and pðzÞ ¼ . [50]. which are consistent with existing literature.13) of Choi et al.e. .264 G..e. [26]. then we get on simplification ½q  a½1 ð1  C  ðkÞÞ  qH ð1  zÞðq þ pzÞb ðAðzÞÞ UðzÞ ¼ . respectively: ð6:6Þ a½1 ½fq þ pzgfC  ðkÞ þ aðzÞð1  C  ðkÞÞgb ðAðzÞÞ  z These are the PGFs of the number of customers in the system at a service completion epoch.. Further. h 6.1) is consistent with expression (2.40). the above expressions are consistent with the result obtained by Gomez-Corral [51]. (5. It should be noted here this type of model was considered by Krishna kumar et al. orbit size distribution and the system size distribution at a departure epoch of the unreliable MX/G/1 retrial queue with delaying repair and constant repeated attempts under Bernoulli vacation schedule. Ke / Applied Mathematical Modelling 36 (2012) 255–269 C 0 ½fq þ pc ðbðzÞÞgb ðAðzÞÞ  zaðzÞ W0 ðzÞ ¼ : ð5:9Þ ½fq þ pc ðbðzÞÞgb ðAðzÞÞ  z Since W0(1) = 1.e. Choudhury. we get ð1  qH Þ C0 ¼ : ð5:10Þ ð1 þ a½1  qH Þ Hence formula (5.41) and (5.-C.1) yields ½ðk þ mÞð1  qH Þ  a½1 kð1  zÞfq þ pc ðbðzÞÞgb ðAðzÞÞ UðzÞ ¼ . the number of customers in the orbit and number of customers in the system at a departure epoch respectively service completion epoch for the unreliable MX/G/1 retrial queue with Bernoulli feedback mechanism and generalized repeated attempts. (3. J. (3. ð6:4Þ fq þ pzgfC  ðkÞ þ aðzÞð1  C  ðkÞÞgb ðAðzÞÞ  z ½q  a½1 ð1  C  ðkÞÞ  qH ð1  zÞ XðzÞ ¼ ð6:5Þ fq þ pzgfC  ðkÞ þ aðzÞð1  C  ðkÞÞgb ðAðzÞÞ  z ½q  a½1 ð1  C  ðkÞÞ  qH ð1  aðzÞÞðq þ pzÞb ðAðzÞÞ and pðzÞ ¼ . (3.40).1). Some particular case In this section we discuss briefly some important particular cases of this type of model..4)–(6. respectively: ð6:9Þ a½1 ½fC  ðkÞ þ aðzÞð1  C  ðkÞÞgb ðAðzÞÞ  z These are the corresponding PGFs of the system size distributions and orbit size distribution of the unreliable MX/G/1 retrial queue with delaying repair and generalized repeated attempts.e.9). for single unit arrival case) the above expression (6.6) yields ½1  a½1 ð1  C  ðkÞÞ  qH ð1  zÞb ðAðzÞÞ UðzÞ ¼ .e. for reliable server) and a(z) = z (i.6) follows by inserting (5.7).10) and (4. Suppose that retrial time distribution is exponential with parameter m > 0. Again if we take z = c⁄(b(z)) in the Eqs. if we take p = 0 (i.41) and (5. then qH = ka[1]b1{1 + a(g1 + d1)} < 1 and expres- sion (6.1) in (5.. there is no Bernoulli feedback in the system). Note that a = 0 (i. ð6:1Þ fm þ kaðzÞgfq þ pc ðbðzÞÞgb ðAðzÞÞ  zðk þ mÞ ½ðk þ mÞð1  qH Þ  a½1 kð1  zÞ XðzÞ ¼ ð6:2Þ fm þ kaðzÞgfq þ pc ðbðzÞÞgb ðAðzÞÞ  zðk þ mÞ ½ðk þ mÞð1  qH Þ  a½1 kð1  aðzÞÞfq þ pc ðbðzÞÞgb ðAðzÞÞ and pðzÞ ¼ : ð6:3Þ a½1 ½fm þ kaðzÞgfq þ pc ðbðzÞÞgb ðAðzÞÞ  zðk þ mÞ These are the PGFs of the system size distribution at a random point of time. (3. Note that for p = 0 (i.

ð7:3Þ a½2 where a½R ¼ 2a½1 is the mean residual batch size. J. If the system is in steady state conditions. Note that PNE ¼ limz!1 wðzÞ. Ke / Applied Mathematical Modelling 36 (2012) 255–269 265 7. ½1a½1 ð1C  ðkÞÞqH  (ii) the probability that the server is idle but system is nonempty is P NE ¼ C  ðkÞ qH (iii) the probability that the server is busy is PB = ka[1]b1 (iv) the probability that the server is waiting for repair is PW = aka[1]b1d1 (v) the probability that the server is under repair is PR = aka[1]b1g1 (vi) the probability that the server is on vacation is PV = pka[1]c1 Proof.2.5). then we have ½1a ð1C  ðkÞÞq  ½1 H (i) the probability that the server is idle and system is empty is PE ¼ C  ðkÞ . Choudhury. G. First of all we derive the system state probabilities and results are summaries in the following theorem.26) in (7. Theorem 7. h . (3. P W ¼ limz!1 RðzÞ. Let Tb and Tc be the length of a busy period and length of a busy cycle. E(NS) and E(ND) be the expected number of units in the orbit and in the system at stationary point of time and at a departure epoch respectively. h Secondly. the stated formulae follow by direct calculation. PB ¼ limz!1 PðzÞ. Theorem 7. and P V ¼ limz!1 Q ðzÞ and P E ¼ 1  fPNE þ P B þP V þ PW þ P R g.1. P R ¼ limz!1 KðzÞ. Once we have obtained the expected number of units in the orbit. then under the steady state conditions. we have ð1  C  ðkÞÞða½1  1Þ b1 f1 þ aðd1 þ g 1 Þg pc1 EðT b Þ ¼ þ þ ð7:4Þ ka½1 ½1  a½1 ð1  C  ðkÞÞ  qH  ½1  a½1 ð1  C  ðkÞÞ  qH  ½1  a½1 ð1  C  ðkÞÞ  qH  C  ðkÞ and EðT c Þ ¼ : ð7:5Þ ka½1 ½1  a½1 ð1  C  ðkÞÞ  qH  Proof.7).0 1  1Þ. ð7:6Þ ka½1 and EðT c Þ ¼ ½ka½1 P0. Proof.6) and (7. Theorem 7. we derive the mean orbit size and the mean system size of this model under stability conditions.-C.40).3. respectively. ð7:1Þ 2½1  a½1 ð1  C ðkÞÞ  qH  ½1  a½1 ð1  C  ðkÞÞ  qH  EðNS Þ ¼ EðNO Þ þ qH ð7:2Þ and EðND Þ ¼ EðNS Þ þ a½R .41) and (5. h Next we derive the mean busy period and expected length of a busy cycle under the steady state condition.4) and (7. h Remark 7.0 1 : ð7:7Þ Inserting (3. System performance measures Our next object is to provide explicit expression for the system state probabilities and some important performance mea- sures of the system. then under the steady state condition we have qH ð1  C  ðkÞÞa½1 ½ka½1 ½b2 f1 þ aðd1 þ g 1 Þg2 þ ab1 fd2 þ 2d1 g 1 þ g 2 g EðNO Þ ¼  þ ½1  a½1 ð1  C ðkÞÞ  qH  2½1  a½1 ð1  C  ðkÞÞ  qH  p½ka½1 2 ½c2 þ 2c1 b1 f1 þ aðd1 þ g 1 Þg ½a½1 ð1  C  ðkÞÞ þ qH  þ  þ a½R . The results follow directly by applying the argument of alternating renewal process which lead to well known result 1 EðT b Þ ¼ ð½P0.1. The mean waiting time of W in the steady state can be easily obtained with the help of Little’s formula viz – L0 = ka[1]W. we get (7. Let E(NO). (3.1) with respect to z and then taking limit z ? 1 by using the L’ Hospital’s rule. respectively. The results follow directly by differentiating Eqs. it is trivial to obtain the other related per- formance measure viz – mean waiting time.

22) and (3.  Exponential delayed time with a mean d1 = 1 for access point. h 8. Typically. it will wait another amount of time and then try again. Further. then Cs Cr ¼ . so the wireless devices can access the resources of the wired network. such that the steady state availability of the server will be AV = limt?1AV(t). In wireless network.  Geometric batch size with mean a[1] = 1. Proof. The availability of the server and failure frequency of the server under the steady state condition are given by ½1  a½1 fð1  C  ðkÞÞ þ kb1 ðC  ðkÞ þ 1 þ aðg 1 þ d1 ÞÞ þ pkc1 g AV ¼ ð7:8Þ C  ðkÞ and Mf ¼ aka½1 b1 : ð7:9Þ . the probability that the server is either working for a customer or in an idle period. It serves as an interface point and bridges between wireless devices and wired network. a (retrial rate of requests in the buffer) andp(the probability of access point performing maintenance) on the cost rate and cost effectiveness listed above. respectively.266 G. Wireless access point provides services to a WLAN. Because the system performance may be heavily affected by access point breakdown.9). one request is selected to be served and the rest of the requests will join the buffer and retry again after a random period. in the queueing terminology. In addition. Let Cr be the expected cost rate per busy cycle. It fur- ther provides wireless linkage between wireless devices that may be out of range of each other. the cost effectiveness is defined as AV Ce ¼ : EðT c Þ We now present some numerical examples to study the effect of various parameters k (request arrival rate). buffer in the access point. When requests arrive at the access point.  Exponential service time with a mean b1 = 0. (3. in practice the access point may fail and can be repaired. J.25 for requests. 1Þdx. and maintenance activities correspond to the server. access point. it is well worth to investigate such system from the queueing theory viewpoint. Theorem 7. wireless device’s connection requests arrive at the access point following the Poisson stream. Each time it tries but fails. the orbit.26).4. the retrial discipline. For convenience. therefore from Eqs. access point is used to connect between wireless and wired networks.8) and (7. At the completion of a service.0 þ Pðx. access point may perform one of maintenance activities such as virus scan to keep the access point functioning well or serve the next connection request. Choudhury. In this scenario. 1Þdx: 0 R1 R1 Now since 0 ½1  BðxÞdx ¼ 0 xdBðxÞ ¼ b1 . as well as reliability point of view. we get (7. For access point. The repair time depends on the degree of failure of access point. that is. Ke / Applied Mathematical Modelling 36 (2012) 255–269 Finally. A real world application with the cost effectiveness maximization model In this section. In the buffer.0 for the request arrival. we consider two reliability indices of the system viz – the system availability and failure frequency under the steady state conditions. and the vacation policy. The result follows directly by considering the following equations Z 1 AV ¼ P0. EðT c Þ where Cs is out of pocket cost per cycle. each message waits a certain amount of time and requires the service again. we develop the cost effectiveness defined as (availability)/(the expected out of pocket cost rate) to be an alternative cost criterion. . which reflects the efficiency per dollar outlay. Let AV(t) be the system availability at time ‘t’. This criterion is useful for the effective use of avail- able money. retransmission policy. respectively. There is a daemon program implemented at access point to manage the connection requests from buffer. 0 Z 1 Mf ¼ a Pðx. This criterion is helpful in the situation that the benefits obtained from the investment are not reducible to mon- etary terms.-C. the settings of system’s parameters are given by:  Cs = 100. we present a possible application and some numerical examples in some situations to explain that our model can represent the possible application reasonably well.

 Exponential vacation time with a mean c1 = 0. Ke / Applied Mathematical Modelling 36 (2012) 255–269 267 40 1 Cost Effectness Cost Rate 20 CrExp 0.04 30 Cost Effectness Cost Rate CrExp 20 CeExp CrErl CeErl CrHyp 10 CeHyp 0.6 0.5 CeExp CrErl CeErl CrHyp CeHyp 0 0 0 0. Cost rate and cost effectiveness for various values of k. 40 0.6 0.3 0. Choudhury. J.2 0.-C.4 0. 2.9 1 λ Fig.7 0.035 0 0.8 0.5 0.5 0.5 for access point. .9 1 α Fig.7 0. G.3 0.1 0.  Exponential repair time with a mean g1 = 1 for access point.2 0.4 0. Cost rate and cost effectiveness for various values of a.8 0.1 0. 1.

which consists of a breakdown period and delay period. Moreover. 2 and 3 reveal that the effect of various retrial rates and vacation probabilities on the cost rate and the cost effectiveness. p = 0. in Figs.0 for different retrial time distribution.0 for different retrial time distribution.1 0. while the cost effectiveness decreases first and then in- creases. Future investigation can try to incorporate a more complex system (such as non-Markovian input) with different failure criteria.0. (I) We choose a = 0. more connection requests can be treated using the same computing power no matter the connection requests being served or waiting in the buffer.0 to 1. Along with more and more requests being served. Choudhury.0. Figs. for example. 1.05.-C.5. new arriving requests with the same destination can be served quickly. and vary the values of k from 0.0 for different retrial time distribution. p = 0. This research presents an extension of the vacation model theory and the analysis of the model will provide a useful performance evaluation tool for more general situations arising in practical applications.03 CrExp CeExp CrErl CeErl CrHyp CeHyp 0 0.0 to 1. . Acknowledgement We thank Editor. Cost rate and cost effectiveness for various values of p.0 to 1.02 0 0. we can see that the cost rate decreases as retrial rate or vacation probability increases from 0. Professor Cross.05. This is because that the increasing of retrial rate or vacation probability repre- sents more connection requests are waiting in the buffer. In addition. (III) We choose k = 0.8 0. Some extensive numerical computations were performed to study the effect of the system parameters on the system characteristics. Conclusions In this paper we analyzed an MX/G/1 queue with general retrial time and Bernoulli vacation schedule for an unreliable server. we also observe that the cost effectiveness increases as retrial rate or vacation probability increases from 0. As shown in Fig. For illustrative purpose. Exponential (denoted by Exp). J.4 0. The reason is that access point has to take more computing power to deal with the arriving requests in the initial period.3 0. Ke / Applied Mathematical Modelling 36 (2012) 255–269 40 0.04 Cost Effectness Cost Rate 20 0.0 to 1. respectively. 3. A cost effectiveness maximization model was also developed to demonstrate the efficiency per dollar outlay.5. respectively. and vary the values of p from 0. and referees for constructive suggestions to improve its presentation.6 0.9 1 p Fig. access point is used to connect between wireless and wired networks in wireless network (presented in Section 8). From the viewpoint of access point.6. From the figures.7 0. the cost rate increases first and then decreases. and vary the values of a from 0. with retrial rate is 2. The distributions of some important system characteristics for such a system were derived. 2-stage Erlang (Erl) and 2-stage Hyper-exponential (Hyp) are con- sidered as retrial time. when the arrival rate for requests increases from 0.0 to 1.6. (II) We choose k = 0.2 0.268 G. a = 0. the effect of balking/reneging on a service system or more complex system may also be a problem of interest. 9.5 0. 1–3 for the following three cases.0. The results of cost rate and effectiveness are shown.0 to 1.

G. 7 (1990) 127–168. J. J. P. 24 (1997) 493–504. [46] D. [4] G. Appl. Comput. Oscillating random walk models for GI/G/1 vacation systems with Bernoulli schedule. Res. [5] J. Res. K. [21] J. Thirurengadan. Appl. Appl. K. . A. Farahmand. Optimization 17 (1986) 649–677. H. A survey of retrial queues. in: O. Ahn. Model. Log. Ukrain Math. Probab. [33] B. Acta Math. Queueing systems with vacations-A survey. Sengupta.C. 11 (1963) 62–71.). Res. Manage. [13] V. [38] J. Q.R.C. Elsevier. 43 (2002) 15–30.G. K. J.G.B. Falin.M. Falin. Aissani. [45] R. 1997. 6 (1988) 93–123. Brill. F. Oper. 33 (1985) 1117–1129. A single server retrial queue with general retrial time and Bernoulli schedule. An M/G/1 retrial queueing system with two phases of service subject to the server breakdown and repair. Retrial Queues. Cox. An M/G/1 retrial queue with an additional phase of second service and general retrial time. Comput. Math. Appl. Choudhury. 1997. J. Humblet. X. Retrial queues with collision arising from unslottted CDMA/CD protocol.-M. Technol. J. Keilson. Krishna Kumar.I. Moren. 7 (1993) 29–46. L. 26 (1998) 285–300. Falin. Proc. Tang. K. 51 (1955) 433– 441.W. Choudhury. On the M/G/1/1 queueing system with repeated orders and unreliable. [8] J. Math.A.. A two phase batch arrival retrial queueing system with Bernoulli vacation schedule. Mean drifts and the non-ergodicity of Markov chain. Res. Comput. 33 (2009) 3186–3196. Sci. [50] B. Shi. Rev. Falin. 19–34. Y. Modified vacation policy for M/G/1 retrial queue with balking and feedback.G. John Willey and Sons. [15] T. [32] W. 141 (2006) 163–191. Res.T. Kapyrim. Artalejo. Shin. Chao. [2] G. [10] G. Servi. [39] G. Sengupta. D.J. A simple telephone exchange with delayed feedback. Artalejo. Falin. Naor. D. [26] B.C. B. 28 (1976) 337–340. Krishna Kumar. Appl.B. Sci.R. Templeton. 30 (1999) 171–187. Oper. A classical bibliography of research on retrial queues: Progress in 1990–1999. Manage.M. Soc. Tijms (Eds. Cambridge Philos. Oper. Single line queue with repeated attempts. J. Eng. Choi. Appl. 16 (1968) 628–638. Choudhury. 161 (2005) 353–364. Wang. [12] B. Queueing Syst. [43] J. [28] D.. 38 (2001) 363–380. Deka. Stochastic analysis of a single server retrial queue with general retrial time. Yang. H. M. 2 (1987) 201–233.R. Sennot. [11] K. Harris. Ke. in: J. M.A. 1 (2005) 30–51. 1 (1986) 29–66. Queueing with breakdowns. Arivudainambi. Posner. Li. 38 (1990) 308–318. Comput. [25] B.R. J. K. Model. 31 (1983) 783–789. Takine. Wenhui. Perform. [30] K. Queueing Syst. Accessible bibliography on retrial queues. P. 19 (2002) 177–194. Madan. H. G. Cohen. Eval. A batch arrival retrial queue with two phases of service and Bernoulli vacation schedule. J. Ke. London. Artalejo. F. Dshalalow (Ed. Rhee. Comput.L. D. pp.W. Res.C. Avi-Itzhak. A waiting line with interrupted service including priorities. J. 15 (2002) 101–129. Elsevier. Queueing Syst. We.-B 24 (1962) 73–90. A queue with service interruptions in an alternating random environment. C. The analysis of non Markovian stochastic process by the inclusion of supplementary variables. Retrial queues revisited.G. Oper. Indus. [36] A. J. Oper.-C. Templeton. Arivudainambi. Gomez –Corral. A. Aggregate arrival of customers in one line systems with repeated calls. 11 (1992) 335–356. Appl. [6] J. J. [41] G. 245–253. Prob.M.D. Li. 2nd ed. Boxma.P. B. Reliability analysis of M/G/1 queueing system with server breakdowns and vacations. forthcoming. Probab. Moreno. [35] Y.-C. New York. A bibliographical guide to analysis of retrial queues through matrix analytic techniques. A survey on retrial queue. 11 (1963) 303–320. CRC. 37 (1997) 315–321. A single server queue with service interruptions. Choudhury. Vijay Kumar. Ann. An approximation method for the M/G/1 retrial queue with general retrial times. Queueing Syst. [44] L. J. Stat. Math. Res. 57 (2009) 433–443. D. Comput. [49] G. Reliab.I.G2)/1 retrial queue under Bernoulli vacation schedules with general repeated attempts and starting failures. Yang. [18] G. [47] B. J. J. Inform.C. Park. Tweedie. [31] I. A many server queue with service interruptions.C. Res. An M/G/1 retrial queue with balking and retrials during service. Choudhury.H. Queueing Syst. 30 (1999) 1–6. 34 (1997) 546–555. Microelectron. Cao.I. Choi. [17] I. Retrial queues with server subject to breakdowns and repairs. Oper. Int. M[x]/(G1. Sci. Stochastic decomposition in M/G/1queue with generalized vacations. Model. Gross. The M/G/1 retrial queue with Bernoulli schedule and general retrial time. Fundamentals of Queueing Theory. Math. [23] J. Gomez-Corral. [20] X. Oper. Templeton. R. Gaver. Chapman and Hall. 159 (2004) 651–666. Fuhrmann. Mitrany. On the MX/G/1retrial queue with Bernoulli schedule and general retrial times.-M.G.D. Doshi. Ke / Applied Mathematical Modelling 36 (2012) 255–269 269 References [1] T. The M/G/1 retrial queue with retrial rate control policy. Cooper. Arivudainambi. 7 (1990) 191–208. Chang. 188 (2007) 1455–1466. Single line repeated orders queueing systems. Math. [48] S. Inform. Queueing Syst. Res. New York.). Comput.K. Amsterdam.I. Deka.J. 12 (2) (2006) 1– 18. [27] Z. Res. [19] B. Oper. Soc. Amsterdam. [37] V. 1986. Math. Res. Comput. 46 (1999) 561–581. pp. 6 (1990) 223–228. A single server M/G/1 queueing system subject to breakdowns – Some reliability and queueing problems. teletraffic analysis and computer performance evaluation. Appl. Some queueing problems with the service station subject to breakdowns. Chang.W. Krishna Kumar. Li. Analysis of an M/G/1 queue with constant repeated attempts and server vacations. Choi. Cooper. Sin. An M/G/1 retrial queueing system with two phase service and preemptive resume. B. [51] A. Eng. [14] G. Int. Hlyanka.D. 1981. Naval Res. [22] G. Wang. Fayolle. 65 (2008) 714–724.R. 23 (1986) 790–802. J. Asia-Pacific J. [34] T. Comput. P. Int. Langaries. Gated pooling models with customers in orbit. 113 (2002) 61–79.-C. Oper. [9] P. 1985. Queueing Syst. Math. Choudhury. Frontiers in Queueing. Eur. [3] V. Ser. Reliability analysis of the retrial queue with server breakdowns and repairs. Oper. Introduction to Queueing Theory. Ann. Appl. [24] J. Comput. Analysis of a single server retrial queue with FCFS orbit and Bernoulli vacation. J. Liang.D. A batch arrival queue with Bernoulli vacation schedule under multiple vacation policy.I. Avi-ltzhak. TOP 7 (1999) 187–211.H. [16] D. Artalejo. 162 (2005) 855–880. An M/G/1 retrial queue with recurrent customers and general retrial times. A study of stationary distributions of a queueing system with recurring demands. Math. Standard and retrial queueing systems: A comparative analysis. Atencia. Kulkarni. [42] G. Math. 76 (1994) 552–562. [29] B.L. W. Oper. R. Cybernatics 13 (1977) 548–590. P. [40] G.. Vijayakumar. R. Oper. [7] A. Queueing Syst. Res. Kulkarni. 20 (2009) 1–14.