A Discrete-Time Geo/G/1 Retrial Queue with General Retrial Times

Size: px
Start display at page:

Download "A Discrete-Time Geo/G/1 Retrial Queue with General Retrial Times"

Transcription

1 Queueing Systems 48, 5 21, Kluwer Academic Publishers. Manufactured in The Netherlands. A Discrete-Time Geo/G/1 Retrial Queue with General Retrial Times IVAN ATENCIA iatencia@ctima.uma.es Departamento de Matemática Aplicada, E.T.S.I. Telecomunicación, Universidad de Málaga, Campus de Teatinos, Málaga, Spain PILAR MORENO mpmornav@dee.upo.es Departamento de Economía y Empresa, Facultad de Ciencias Empresariales, Universidad Pablo de Olavide, Ctra. de Utrera, km. 1, Sevilla, Spain Received 21 November 2002; Revised 25 November 2003 Abstract. We consider a discrete-time Geo/G/1 retrial queue in which the retrial time has a general distribution and the server, after each service completion, begins a process of search in order to find the following customer to be served. We study the Markov chain underlying the considered queueing system and its ergodicity condition. We find the generating function of the number of customers in the orbit and in the system. We derive the stochastic decomposition law and as an application we give bounds for the proximity between the steady-state distributions for our queueing system and its corresponding standard system. Also, we develop recursive formulae for calculating the steady-state distribution of the orbit and system sies. Besides, we prove that the M/G/1 retrial queue with general retrial times can be approximated by our corresponding discrete-time system. Finally, we give numerical examples to illustrate the effect of the parameters on several performance characteristics. Keywords: discrete-time model, general retrial times, Markov chain, recursive formulae, stochastic decomposition AMS subject classification: 60K25 1. Introduction There is a great potential for using the discrete-time queues in the performance analyses of computer and communication networks. The discrete-time queueing system has been found to be more appropriate in modeling computer and telecommunication systems because the basic time unit in these systems is a binary code. Moreover, the discretetime system can be used to approximate the continuous system in practice. The earliest work on discrete-time queue was presented by Meisling [13]. Recently, due to the fast progress of computer and telecommunication network technologies, the discrete-time queueing models have received more attention from queueing researchers. In last years, a number of queueing models have been analysed in discrete-time, details of which may be found in recent books [5,10,15,17].

2 6 ATENCIA AND MORENO Queueing systems with repeated attempts are characteried by the fact that a customer finding the server busy upon arrival must leave the service area and repeat his request for service after some random time. Between trials, the blocked customer joins a group of unsatisfied customers called orbit. Retrial queues have been widely used to model many practical problems in telephone switching systems, telecommunication networks and computers competing to gain service from a central processing unit. For a detailed review of the main results and the literature on this topic the reader is referred to [1,2,7]. In the past, the study of the retrial queues has been focused on the continuous case, but recently Yang and Li [18] have extended the study to discrete-time systems. Nevertheless, there are many papers on continuous-time retrial queues, but very little is known about discrete-time retrial queues. In all models of discrete-time retrial queues (see [4,6,11,12,16,18]), the time between retrials for any customer is assumed to be geometrically distributed. Our objective here is to extend the general retrial policy to the discrete-time retrial queues. The inherent difficulty with non-geometrical retrial times stems from the fact that the model must, in some way, keep track of the elapsed retrial time for each of possibly a very large number of customers. We avoid this problem by considering a Geo/G/1 retrial queue with general retrial times where only one customer may attempt retrials from orbit. Thus, the retrial discipline does not depend on the orbit sie. An important characteristic of the general retrial times policy is that we always obtain analytical solutions in terms of closed-form expressions. The general retrial times policy arises naturally in problems where the server is required to search for customers (see [14] in continuous-time), that is, this policy is related to many service systems where, after each service completion, the processor will spend a random amount of time in order to find the next item to be processed. The rest of the paper is organied as follows. In the next section, we give the model description of the considered queueing system. In section 3, we study the Markov chain and the stability condition of the system. Also, the distribution of the number of customers in the orbit and in the system is obtained. Besides, we obtain several performance measures of the system. In section 4, we find the stochastic decomposition property on the steady-state system sie and as an application we provide upper and lower estimates for the distance between the steady-state distributions for our queueing system and its corresponding standard system. In section 5 we derive a recursive algorithm for computing the steady-state probabilities of the numbers of customers in the orbit and in the system. In section 6, we analyse the relation between the continuous-time system and the discrete-time system. Finally, we discuss numerical results in section Description of the queueing system Let us consider a discrete-time single server retrial queue where the time axis is segmented into a sequence of equal time intervals (called slots). Further, let the time axis be marked by 0, 1,..., m,... It is assumed that all queueing activities (arrivals, departures and retrials) take place around the slot boundaries. For mathematical clarity,

3 DISCRETE-TIME Geo/G/1 RETRIAL QUEUE 7 we assume that a potential departure occurs in the interval (m,m), and a potential arrival or retrial occurs in the interval (m, m + ); that is, arrivals and retrials occur at the moment immediately after the slot boundaries and the departures occur at the moment immediately before the slot boundaries. Customers arrive according to a geometrical arrival process with rate p, i.e., p is the probability that an arrival occurs in a slot. If an arriving customer finds the server free, he immediately begins his service. Otherwise, the customer leaves the service area and enters a group of blocked customers called orbit in accordance with an FCFS discipline. We will assume that only the customer at the head of the orbit is allowed for access to the server. It is always supposed that retrials and services can be started only at slot boundaries and their durations are integral multiples of a slot duration. Successive interretrial times of any customer are governed by an arbitrary distribution {a i } i=0 with generating function A(x) = i=0 a i x i. Service times are independent and identically distributed with general distribution {s i } i=1, generating function S(x) = i=1 s ix i and nth factorial moments β n. After service completion, the served customer leaves the system forever and will have no further effect on the system. In order to avoid trivial cases, we assume 0 <p<1. The traffic intensity is given by ρ = pβ The Markov chain At time m + (the instant immediately after time slot m), the system can be described by the process Y m = (C m,ξ 0,m,ξ 1,m,N m ), where C m represents the server state (0 or 1 according to the server is free or busy, respectively) and N m is the number of customers in the retrial group. If C m = 0and N m > 0, ξ 0,m denotes the remaining retrial time. If C m = 1, ξ 1,m corresponds to the remaining service time of the customer being served. It can be shown that {Y m,m N} is the Markov chain of our queueing system, whose states space is { } (0, 0); (0,i,k): i 1, k 1; (1,i,k): i 1, k 0. Our objective is to find the stationary distribution π 0,0 = lim P [C m = 0, N m = 0], m π 0,i,k = lim P [C m = 0, ξ 0,m = i, N m = k]; i 1, k 1, m π 1,i,k = lim P [C m = 1, ξ 1,m = i, N m = k]; i 1, k 0, m of the Markov chain {Y m,m N}. The one-step transition probabilities p yy = P [Y m+1 = y Y m = y] are given by the formulae p (0,0)(0,0) = p, p (1,1,0)(0,0) = p,

4 8 ATENCIA AND MORENO if i 1, k 1 p (0,i+1,k)(0,i,k) = p, p (1,1,k)(0,i,k) = pa i, if i 1, k 0 p (0,0)(1,i,k) = ps i, k = 0, p (0,1,k+1)(1,i,k) = ps i, p (0,j,k)(1,i,k) = ps i, j 1, k 1, p (1,1,k)(1,i,k) = ps i, p (1,1,k+1)(1,i,k) = pa 0 s i, p (1,i+1,k 1)(1,i,k) = p, k 1, p (1,i+1,k)(1,i,k) = p, where p = 1 p. The Kolmogorov equations for the stationary distribution are π 0,0 = pπ 0,0 + pπ 1,1,0, (1) π 0,i,k = pπ 0,i+1,k + pa i π 1,1,k ; i 1, k 1, (2) π 1,i,k = δ 0k ps i π 0,0 + ps i π 0,1,k+1 + (1 δ 0k )ps i π 0,j,k + ps i π 1,1,k + pa 0 s i π 1,1,k+1 + (1 δ 0k )pπ 1,i+1,k 1 + pπ 1,i+1,k ; i 1, k 0, (3) and the normaliation condition is π 0,0 + π 0,i,k + π 1,i,k = 1. i=1 k=1 i=1 k=0 In order to solve (1) (3) we introduce the generating functions ϕ 0 (x, ) = π 0,i,k x i k, ϕ 1 (x, ) = i=1 k=1 i=1 k=0 j=1 π 1,i,k x i k and the auxiliary generating functions ϕ 0,i () = π 0,i,k k, i 1, ϕ 1,i () = k=1 π 1,i,k k, i 1. k=0

5 DISCRETE-TIME Geo/G/1 RETRIAL QUEUE 9 Multiplying (2), (3) by k and summing over k, these equations become ϕ 0,i () = pϕ 0,i+1 () + pa i ϕ 1,1 () pa i π 1,1,0, i 1, (4) ϕ 1,i () = ( p + p)ϕ 1,i+1 () + ps i ϕ 0 (1,)+ pa 0 + p s i ϕ 1,1 () + p s iϕ 0,1 () pa 0 s iπ 1,1,0 + ps i π 0,0, i 1. (5) By substituting equation (1) into equations (4), (5), we get ϕ 0,i () = pϕ 0,i+1 () + pa i ϕ 1,1 () pa i π 0,0, i 1, (6) ϕ 1,i () = ( p + p)ϕ 1,i+1 () + ps i ϕ 0 (1,)+ pa 0 + p s i ϕ 1,1 () + p s iϕ 0,1 () + p( a 0) s i π 0,0, i 1. (7) Then, multiplying (6), (7) by x i and summing over i leads to x p ϕ 0 (x, ) = p [ ] A(x) a 0 ϕ1,1 () pϕ 0,1 () p [ ] A(x) a 0 π0,0, (8) x [ ] x ( p + p) pa0 + p ϕ 1 (x, ) = S(x) ( p + p) ϕ 1,1 () x + p S(x)ϕ 0,1() + ps(x)ϕ 0 (1,)+ p( a 0) S(x)π 0,0. (9) Choosing x = 1in(8) yields pϕ 0 (1,)= p(1 a 0 )ϕ 1,1 () pϕ 0,1 () p(1 a 0 )π 0,0 and substituting the above equation into (9), weget [ ] x ( p + p) + pa0 (1 ) ϕ 1 (x, ) = S(x) ( p + p) ϕ 1,1 () x p(1 ) + S(x)ϕ 0,1 () pa 0(1 ) S(x)π 0,0. (10) Setting x = p in (8) and x = p + p in (10), we obtain p [ ] A( p) a 0 π0,0 = p [ ] A( p) a 0 ϕ1,1 () pϕ 0,1 (), [ ] pa 0 (1 ) + pa0 (1 ) S( p + p)π 0,0 = S( p + p) ( p + p) ϕ 1,1 () p(1 ) + S( p + p)ϕ 0,1 (). And from this system of equations, we find the generating functions ϕ 1,1 () and ϕ 0,1 (): ϕ 1,1 () = pa( p)(1 )S( p + p) pa( p)(1 )S( p + p) [( p + p) S( p + p)] π 0,0, (11)

6 10 ATENCIA AND MORENO ϕ 0,1 () = p[a( p) a 0 ][( p + p) S( p + p)] π 0,0 pa( p)(1 )S( p + p) [( p + p) S( p + p)] p. (12) Lemma 1. (1) pa( p)(1 )S( p + p) [( p + p) S( p + p)] > 0 holds for 0 <1if and only if ρ<p+ pa( p). (2) The following limits exist if and only if ρ<p+ pa( p): lim 1 lim 1 1 pa( p)(1 )S( p + p) [( p + p) S( p + p)] = 1 ( p + p) S( p + p) pa( p)(1 )S( p + p) [( p + p) S( p + p)] = p + pa( p) ρ, ρ p p + pa( p) ρ. Using lemma 1, the auxiliary generating functions ϕ 1,1 () and ϕ 0,1 () are defined for [0, 1) and in = 1 are extended by continuity if and only if ρ<p+ pa( p). Substituting (11), (12) into (8), (10), we have the generating functions A(x) A( p) px[( p + p) S( p + p)]π 0,0 ϕ 0 (x, ) = x p pa( p)(1 )S( p + p) [( p + p) S( p + p)], S(x) S( p + p) pxa( p)(1 )( p + p)π 0,0 ϕ 1 (x, ) = x ( p + p) pa( p)(1 )S( p + p) [( p + p) S( p + p)]. Using the normaliation condition π 0,0 + ϕ 0 (1, 1) + ϕ 1 (1, 1) = 1, we can find the unknown constant π 0,0 : p + pa( p) ρ π 0,0 =. (13) A( p) Obviously from (13), asπ 0,0 > 0, we obtain that ρ<p+ pa( p) is a necessary condition for the ergodicity of the Markov chain. Let us observe that this condition can be written as p(β 1 1) < pa( p) where the first member is the expected number of external customers who arrive per service interval and the second one represents the expected number of repeated customers who enter service at the epoch at which a service starts. Therefore, this condition indicates that external customers must arrive per service interval more slowly than repeated customers can enter service at the epoch at which a service starts (on the average). Accordingly, if the condition ρ<p+ pa( p) is fulfilled, then the system is stable and consequently this condition is also sufficient. We summarie the above results in the following theorem. Theorem 1. The Markov chain {Y m,m N} is ergodic if and only if ρ<p+ pa( p).

7 DISCRETE-TIME Geo/G/1 RETRIAL QUEUE 11 The generating functions of the stationary distribution of the chain are given by A(x) A( p) px[( p + p) S( p + p)]π 0,0 ϕ 0 (x, ) = x p pa( p)(1 )S( p + p) [( p + p) S( p + p)], S(x) S( p + p) pxa( p)(1 )( p + p)π 0,0 ϕ 1 (x, ) = x ( p + p) pa( p)(1 )S( p + p) [( p + p) S( p + p)], where π 0,0 = p + pa( p) ρ. A( p) Corollary 1. (1) The marginal generating function of the number of customers in the retrial group when the server is idle is given by A( p)( p + p)[s( p + p) ]π 0,0 π 0,0 + ϕ 0 (1,)= pa( p)(1 )S( p + p) [( p + p) S( p + p)]. (2) The marginal generating function of the number of customers in the retrial group when the server is busy is given by A( p)( p + p)[1 S( p + p)]π 0,0 ϕ 1 (1,)= pa( p)(1 )S( p + p) [( p + p) S( p + p)]. (3) The probability generating function of the number of customers in the retrial group (i.e., of the variable N)isgivenby ()= π 0,0 + ϕ 0 (1,)+ ϕ 1 (1,) A( p)(1 )( p + p)π 0,0 = pa( p)(1 )S( p + p) [( p + p) S( p + p)]. (4) The probability generating function of the number of customers in the system (i.e., of the variable L)isgivenby () = π 0,0 + ϕ 0 (1,)+ ϕ 1 (1,) A( p)(1 )( p + p)s( p + p)π 0,0 = pa( p)(1 )S( p + p) [( p + p) S( p + p)]. In the next corollary we present some performance measures for the system at the stationary regime. Corollary 2. (1) The system is free with probability π 0,0 = p + pa( p) ρ. A( p)

8 12 ATENCIA AND MORENO (2) The system is occupied with probability ϕ 0 (1, 1) + ϕ 1 (1, 1) = (3) The server is idle with probability ρ p[1 A( p)]. A( p) (4) The server is busy with probability π 0,0 + ϕ 0 (1, 1) = 1 ρ. ϕ 1 (1, 1) = ρ. (5) The mean number of customers in the retrial group is E[N] = (1) = 2 p(ρ p)[1 A( p)]+p2 β 2. 2[p + pa( p) ρ] (6) The mean number of customers in the system is E[L] = (1) = ρ + 2 p(ρ p)[1 A( p)]+p2 β 2. 2[p + pa( p) ρ] (7) The mean time a customer spends in the system (including the service time) is given by W = E[L] p = β p(β 1 1)[1 A( p)]+pβ 2. 2[p + pa( p) ρ] The proof of both corollaries is trivial and thus omitted. Remark 1. The stationary distribution of the server state π 0,0 + ϕ 0 (1, 1) = 1 ρ, ϕ 1 (1, 1) = ρ, depends on the service time distribution only through its mean β 1 and does not depend on the interretrial time distribution. Remark 2. We observe a relation between the following generating functions () = ()S( p + p), and as a consequence we find the formula n ( ) n (n (1) = p m β m (n m (1), n 1, m m=0 where (n (1) and (n (1) are the nth factorial moments for the distribution of the random variables L and N, respectively.

9 DISCRETE-TIME Geo/G/1 RETRIAL QUEUE 13 Remark 3 (Special case). When a 0 = 1, () reduces to (1 ρ)(1 )S( p + p) () =, S( p + p) which is the probability generating function for the number of customers (including the one in service, if any) in the standard Geo/G/1/ queueing system (see [10]). This result is not surprising because when a 0 = 1, the customer at the head of the orbit immediately commences his service whenever the server is idle. This system is equivalent to the Geo/G/1/ queue with random service order whose queue sie distribution is the same as that of the standard Geo/G/1/ queue since it is not affected by the service discipline adopted. 4. Stochastic decomposition This section deals with the analysis of the stochastic decomposition property of the system sie distribution. The first result on stochastic decomposition was given by Fuhrmann and Cooper [8]. Later, this property has been discussed by several authors. Generally speaking, the stochastic decomposition relates one performance characteristic for the system with vacations to the corresponding one for the same model without vacations. The stochastic decomposition property states that the system sie distribution decomposes into two random variables, one of which corresponds to the system sie of the queueing model without vacations. The second random variable is the system sie given that the server is on vacation. We now note that the probability generating function of the number of customers in the system can be expressed as () = (1 ρ)(1 )S( p + p) S( p + p) π 0,0 + ϕ 0 (1,) π 0,0 + ϕ 0 (1, 1), where the first fraction is the probability generating function of the number of customers in the Geo/G/1/ queueing system and the second fraction is the probability generating function of the number of customers in the orbit given that the server is idle. In fact, this is the stochastic decomposition law for our queueing system; i.e., the total number of customers in our queueing system can be represented by the sum of two independent random variables: one is the total number of customers in the corresponding regular system and the other is the number of customers in the orbit given that the server is idle. This result can be summaried in the following theorem. Theorem 2. The total number of customers in the system under study (L) can be represented as the sum of two independent random variables, one of which is the total number of customers in the Geo/G/1/ queueing system (L 0 ) and the other is the number of repeated customers given that the server is free (M). That is, L = L 0 + M.

10 14 ATENCIA AND MORENO It is not surprising that the system under study satisfies this property, since our system can be considered as a standard queue with server vacation. In this vacation model, the server begins vacation when a service concludes and there is neither arrival nor retrial. The duration of the vacation depends on the arrival process and the interretrial times. Vacation finish whenever an external customer arrives or the server chooses the customer at the head of the orbit. Under these considerations, the stochastic decomposition property observed previously for our system is consistent with the one reported by Fuhrmann and Cooper [8] for M/G/1 vacation models. As an application of the stochastic decomposition property we now study a measure of the proximity between the steady-state distributions for the standard Geo/G/1/ queueing system and our queueing system. The importance of the following bounds is to provide upper and lower estimates for the distance between both distributions. Theorem 3. The following inequalities hold (ρ p)[1 A( p)] 2 P [L = j] P [L0 = j] (ρ p)[1 A( p)] 2. A( p) (1 ρ)a( p) j=0 The proof of the preceding theorem follows the steps given in the paper [3] and accordingly it is omitted. Finally, let us observe that the distance j=0 P [L = j] P [L 0 = j] between the distributions of the random variables L and L 0 decreases as A( p) approaches Calculation of steady-state probabilities This section is devoted to develop some recursive formulae for calculating the more characteristic stationary distributions associated with our system. Theorem 4. The steady-state distribution of the orbit sie is given by the following recursive formulae p + pa( p) ρ ψ 0 = P [N = 0]=, (14) A( p)s( p) where ψ k = P [N = k]= k 1 n=0 [b k n pa( p)c k n ]ψ n, k 1, (15) A( p)s( p) ( ) i 1 b n = B i p i n p n, n 1, n 1 i=n ( ) i c n = s i+1 p i n p n, n 1, n i=n and B i = j=i+1 s j is the probability that a service lasts more than i slots.

11 DISCRETE-TIME Geo/G/1 RETRIAL QUEUE 15 Proof. Let () be the function S( p + p) () = pa( p) p + p [ 1 1 ] S( p + p) = p + p ω n n. In order to obtain the sequence {ω n } n=0 we will use the properties of the generating functions and Newton s binomial (see [18, theorem 3]) getting [ ] () = A( p)s( p) bn pa( p)c n n. n=1 Since () () = p + pa( p) ρ, comparing the coefficients of k on both sides of this equation leads to k ψ 0 ω 0 = p + pa( p) ρ, ψ n ω k n = 0, k 1. But ω 0 = A( p)s( p) and ω n = pa( p)c n b n for n 1, and as consequence we find equations (14), (15). Theorem 5. The steady-state distribution of the system sie is given by the following recursive formulae p + pa( p) ρ φ 0 = P [L = 0]=, (16) A( p) k 1 n=0 φ k = P [L = k]= [b k n pa( p)c k n ]φ n +[p + pa( p) ρ]d k (17) A( p)s( p) for k 1, where ( ) i d n = s i p i n p n, n 1. n Proof. i=n Let us observe the equation n=0 () () = [ p + pa( p) ρ ] S( p + p), (18) where () and its expression in power series are given in the proof of the previous theorem. From the relation S( p + p) = S( p) + d n n and after comparing the coefficients of k on both sides in equation (18), we get n=1 φ 0 ω 0 = [ p + pa( p) ρ ] S( p), k φ n ω k n = [ p + pa( p) ρ ] d k, k 1. n=0 n=0

12 16 ATENCIA AND MORENO we obtain equa- Then, taking into account the expression of the sequence {ω n } n=0 tions (16), (17). In what continues, we will denote by π 0,,k = i=1 π 0,i,k the stationary probability that there are k 1 repeated customers and the server is idle, and π 1,,k = i=1 π 1,i,k the stationary probability that there are k 0 repeated customers and the server is busy. Theorem 6. The steady-state distribution of the orbit sie when the server is idle is given by the following recursive formulae where p + pa( p) ρ π 0,0 =, A( p) π 0,,k = [b k pa( p)c k ]π 0,0 + (1 δ 1k ) k 1 n=1 [b k n pa( p)c k n ]π 0,,n A( p)s( p) p + pa( p) ρ e k, k 1, A( p)s( p) e n = i=n ( ) i B i p i n p n+1, n 1. n Proof. Firstly, we observe the equation [ π0,0 + ϕ 0 (1,) ] () = [ p + pa( p) ρ ] S( p + p), (19) 1 where the function () and its development in power series are given in the proof of the theorem 4. From the equality S( p + p) 1 = S( p) e n n and after equaliing the coefficients of k on both sides of equation (19), we have π 0,0 ω k + n=1 π 0,0 ω 0 = [ p + pa( p) ρ ] S( p), k π 0,,k ω k n = [ p + pa( p) ρ ] e k, k 1. n=1 Finally, considering the expression of the sequence {ω n } n=0, the proof of this theorem is completed. Theorem 7. The steady-state distribution of the orbit sie when the server is busy is given by the following recursive formulae

13 DISCRETE-TIME Geo/G/1 RETRIAL QUEUE 17 p + pa( p) ρ π 1,,0 = A( p) π 1,,k = k 1 1 S( p), S( p) n=0 [b k n pa( p)c k n ]π 1,,n +[p + pa( p) ρ]e k, k 1. A( p)s( p) Proof. We now observe the following equation ϕ 1 (1, ) () = [ p + pa( p) ρ ] 1 S( p + p), (20) 1 where the definition of () and its expression in power series are given in the proof of the theorem 4. Since 1 S( p + p) 1 = 1 S( p) + e n n, n=1 equalling the coefficients of k on both sides of equation (20) yields π 1,,0 ω 0 = [ p + pa( p) ρ ][ 1 S( p) ], k π 1,,n ω k n = [ p + pa( p) ρ ] e k, k 1. n=0 To conclude this proof, it will be enough to keep in mind the expression of the sequence {ω n } n=0. Remark 4. If the service times are geometrically distributed with generating function S(x) = (1 s)x/(1 sx), then the coefficients b n, c n, d n and e n have the following closed-form expressions for n 1. b n = pn s n (1 ps), c n n = pn s n (1 s) (1 ps), n+1 d n = pn s n 1 (1 s) (1 ps) n+1, e n = pn+1 s n (1 ps) n+1 6. Relation with the continuous-time system This section concerns the analysis of the relation between the continuous-time system and the discrete-time system. We will show that the continuous-time M/G/1 retrial queue with general retrial times can be approximated by the corresponding discrete-

14 18 ATENCIA AND MORENO time system; for this purpose, time is slotted into small intervals of equal length, so the approximation approaches the exact value when the length of the interval tends to ero. We consider the continuous-time M/G/1 retrial queue with general retrial times (see [9] for details) where customers arrive according to a Poisson process with rate λ. Upon arrival, the customer who finds the server busy leaves the service area and joins the retrial group in accordance with an FCFS discipline (that is, only the customer at the head of the orbit is allowed for access to the server). Successive interretrial times of any customer are governed by an arbitrary probability distribution function Ɣ(x) with corresponding Laplace Stieltjes transform γ(s). Customer service times are identically and independently distributed random variables with a common distribution function B(x), Laplace Stieltjes transform β(s) and a finite mean µ 1. Interarrival times, retrial times and service times are assumed to be mutually independent. If we suppose that time is divided into intervals of equal length, the continuoustime system can be approximated by a discrete-time system for which p = λ, a i = (i+1) i dɣ(x), i 0 and s i = i (i 1) db(x), i 1, where must be chosen sufficiently small so that p is probability. Our immediate objective is to prove that lim 0 () is the probability generating function of the number of customers in the M/G/1 retrial queueing system with general retrial times (obtained by Góme-Corral [9]). Firstly, it is not difficult to prove the following equalities using the definition of Lebesgue integration: lim 0 lim ρ = 0 λµ 1, A( p) = γ(λ), lim S( p + p) = β( λ(1 ) ). 0 The proof of these relationships is omitted here since the technique used can be found in [18, theorem 5]. Taking into account the above results, we get the next relation lim () = lim 0 0 (1 )( p + p)s( p + p)[p + pa( p) ρ] pa( p)(1 )S( p + p) [( p + p) S( p + p)] (1 )(1 λ(1 ) )S( p + p)[λ + (1 λ )A( p) ρ] = lim 0 (1 λ )A( p)(1 )S( p + p) [1 λ(1 ) S( p + p)] [ = γ(λ) λ ] (1 )β(λ(1 )) µ γ(λ)(1 )β(λ(1 )) [1 β(λ(1 ))], which coincides with the probability generating function of the number of customers in the M/G/1 retrial queue with general retrial times (see [9, equation (16)]).

15 DISCRETE-TIME Geo/G/1 RETRIAL QUEUE Numerical results In this section, we present some numerical results to illustrate the effect of varying parameters on the main performance measures of our system. In figures 1(a) and (b), we consider that the service time distribution is geometrical with mean β 1 = 2 and the retrial times are governed by a geometrical distribution with generating function A(x) = (1 r)/(1 rx). In figure 1(a) the probability that the system is busy is plotted versus the retrial rate r. We have presented three curves which correspond to p = 0.2, 0.3, 0.4, respectively. As we expect, the utiliation factor increases with increasing retrial rate r and increasing p. The same discussion holds for figure 1(b), which illustrates the behaviour of E[L] as function of r. Moreover, we observe that, as r approaches the ergodicity condition, the mean system sie tends to infinite (due to the system becomes unstable) and, as a consequence, the probability that the system is occupied converges to 1. In figures 1(c) and (d), we assume that the parameter p is equal to 0.2, the service times follow a geometrical distribution with mean β 1 = 2 and the retrial times have a binomial distribution with generating function A(x) = (rx + 1 r) n. (a) (b) (c) (d) Figure 1. (a) The utiliation factor versus r. (b) The mean system sie versus r. (c) The utiliation factor versus r. (d) The mean system sie versus r.

16 20 ATENCIA AND MORENO The influence of the parameter r on the utiliation factor is shown in figure 1(c). The highest curve in figure 1(c) corresponds to the highest value of n. As intuition tells us, the probability that the system is busy increases with increasing values of n. From this graphic, we see that the utiliation factor increases with increasing values of r,which also agrees with the intuitive expectations. Substantially, the same effects are shown in figure 1(d), which represents the behaviour of E[L] as function of r. We would like to remark that both the utiliation factor and the mean system sie are consistent (when r 0) with the characteristics of the standard Geo/G/1/ queue given by ρ = 0.4 ande[l] = , respectively, that is, we have the limiting results [ lim ϕ0 (1, 1) + ϕ 1 (1, 1) ] = ρ and lim E[L] =ρ + p2 β 2 r 0 r 0 2(1 ρ). Besides, we note that the utiliation factor and the mean system sie in the classi- Table 1 The steady-state distribution of the orbit sie. r = 0 r = 0.3 r = 0.6 r = 0.9 ψ ψ ψ ψ ψ ψ ψ ψ ψ ψ ψ Table 2 The steady-state distribution of the system sie. r = 0 r = 0.3 r = 0.6 r = 0.9 φ φ φ φ φ φ φ φ φ φ φ

17 DISCRETE-TIME Geo/G/1 RETRIAL QUEUE 21 cal Geo/G/1/ queue are a lower bound for the corresponding characteristics in our queueing system. An important feature of this work is in the recursion scheme provided by theorems 4 and 5. The formulae (14) (17) have been implemented in Matlab. Tables 1 and 2 summarie our numerical experiments for p = 0.2, service times geometrically distributed with mean β 1 = 2 and retrial times governed by a geometrical distribution with generating function A(x) = (1 r)/(1 rx). Acknowledgement This research is supported by the DGINV through the project BFM References [1] J.R. Artalejo, A classified bibliography of research on retrial queues: Progress in , Top 7(2) (1999) [2] J.R. Artalejo, Accessible bibliography on retrial queues, Math. Comput. Modelling 30 (1999) 1 6. [3] J.R. Artalejo and G.I. Falin, Stochastic decomposition for retrial queues, Top 2 (1994) [4] I. Atencia and P. Moreno, Discrete-time Geo [X] /G H /1 retrial queue with Bernoulli feedback, To appear in Comput. Math. Appl. [5] H. Bruneel and B.G. Kim, Discrete-Time Models for Communication Systems Including ATM (Kluwer Academic, Boston, 1993). [6] B.D. Choi and J.W. Kim, Discrete-time Geo 1, Geo 2 /G/1 retrial queueing system with two types of calls, Comput. Math. Appl. 33(10) (1997) [7] G.I. Falin and J.G.C. Templeton, Retrial Queues (Chapman & Hall, London, 1997). [8] S.W. Fuhrmann and R.B. Cooper, Stochastic decompositions in the M/G/1 queue with generalied vacations, Oper. Res. 33(5) (1985) [9] A. Góme-Corral, Stochastic analysis of a single server retrial queue with general retrial times, Naval Res. Logistics 46 (1999) [10] J.J. Hunter, Mathematical Techniques of Applied Probability, Vol. 2, Discrete-Time Models: Techniques and Applications (Academic Press, New York, 1983). [11] H. Li and T. Yang, Geo/G/1 discrete time retrial queue with Bernoulli schedule, European J. Oper. Res. 111(3) (1998) [12] H. Li and T. Yang, Steady-state queue sie distribution of discrete-time PH/Geo/1 retrial queues, Math. Comput. Modelling 30 (1999) [13] T. Meisling, Discrete time queueing theory, Oper. Res. 6 (1958) [14] M.F. Neuts and M.F. Ramalhoto, A service model in which the server is required to search for customers, J. Appl. Probab. 21 (1984) [15] H. Takagi, Queueing Analysis: A Foundation of Performance Evaluation, Discrete-Time Systems, Vol. 3 (North-Holland, Amsterdam, 1993). [16] M. Takahashi, H. Osawa and T. Fujisawa, Geo [X] /G/1 retrial queue with non-preemptive priority, Asia-Pacific J. Oper. Res. 16(2) (1999) [17] M.E. Woodward, Communication and Computer Networks: Modelling with Discrete-Time Queues (IEEE Computer Soc. Press, Los Alamitos, CA, 1994). [18] T. Yang and H. Li, On the steady-state queue sie distribution of the discrete-time Geo/G/1 queue with repeated customers, Queueing Systems 21 (1995)

18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

A discrete-time Geo/G/1 retrial queue with starting failures and second optional service

A discrete-time Geo/G/1 retrial queue with starting failures and second optional service Computers and Mathematics with Applications 53 (2007) 115 127 www.elsevier.com/locate/camwa A discrete-time Geo/G/1 retrial queue with starting failures and second optional service Jinting Wang, Qing Zhao

More information

A Two Phase Service M/G/1 Vacation Queue With General Retrial Times and Non-persistent Customers

A Two Phase Service M/G/1 Vacation Queue With General Retrial Times and Non-persistent Customers Int. J. Open Problems Compt. Math., Vol. 3, No. 2, June 21 ISSN 1998-6262; Copyright c ICSRS Publication, 21 www.i-csrs.org A Two Phase Service M/G/1 Vacation Queue With General Retrial Times and Non-persistent

More information

Discrete-time Retrial Queue with Bernoulli Vacation, Preemptive Resume and Feedback Customers

Discrete-time Retrial Queue with Bernoulli Vacation, Preemptive Resume and Feedback Customers Journal of Industrial Engineering and Management JIEM, 2015 8(4): 1236-1250 Online ISSN: 2013-0953 Print ISSN: 2013-8423 http://dx.doi.org/10.3926/jiem.1487 Discrete-time Retrial Queue with Bernoulli Vacation,

More information

Non-Persistent Retrial Queueing System with Two Types of Heterogeneous Service

Non-Persistent Retrial Queueing System with Two Types of Heterogeneous Service Global Journal of Theoretical and Applied Mathematics Sciences. ISSN 2248-9916 Volume 1, Number 2 (211), pp. 157-164 Research India Publications http://www.ripublication.com Non-Persistent Retrial Queueing

More information

ON THE NON-EXISTENCE OF PRODUCT-FORM SOLUTIONS FOR QUEUEING NETWORKS WITH RETRIALS

ON THE NON-EXISTENCE OF PRODUCT-FORM SOLUTIONS FOR QUEUEING NETWORKS WITH RETRIALS ON THE NON-EXISTENCE OF PRODUCT-FORM SOLUTIONS FOR QUEUEING NETWORKS WITH RETRIALS J.R. ARTALEJO, Department of Statistics and Operations Research, Faculty of Mathematics, Complutense University of Madrid,

More information

Inventory Ordering Control for a Retrial Service Facility System Semi- MDP

Inventory Ordering Control for a Retrial Service Facility System Semi- MDP International Journal of Engineering Science Invention (IJESI) ISS (Online): 239 6734, ISS (Print): 239 6726 Volume 7 Issue 6 Ver I June 208 PP 4-20 Inventory Ordering Control for a Retrial Service Facility

More information

An M/G/1 Retrial Queue with Non-Persistent Customers, a Second Optional Service and Different Vacation Policies

An M/G/1 Retrial Queue with Non-Persistent Customers, a Second Optional Service and Different Vacation Policies Applied Mathematical Sciences, Vol. 4, 21, no. 4, 1967-1974 An M/G/1 Retrial Queue with Non-Persistent Customers, a Second Optional Service and Different Vacation Policies Kasturi Ramanath and K. Kalidass

More information

M/M/1 Retrial Queueing System with Negative. Arrival under Erlang-K Service by Matrix. Geometric Method

M/M/1 Retrial Queueing System with Negative. Arrival under Erlang-K Service by Matrix. Geometric Method Applied Mathematical Sciences, Vol. 4, 21, no. 48, 2355-2367 M/M/1 Retrial Queueing System with Negative Arrival under Erlang-K Service by Matrix Geometric Method G. Ayyappan Pondicherry Engineering College,

More information

M/M/1 Retrial Queueing System with N-Policy. Multiple Vacation under Non-Pre-Emtive Priority. Service by Matrix Geometric Method

M/M/1 Retrial Queueing System with N-Policy. Multiple Vacation under Non-Pre-Emtive Priority. Service by Matrix Geometric Method Applied Mathematical Sciences, Vol. 4, 2010, no. 23, 1141 1154 M/M/1 Retrial Queueing System with N-Policy Multiple Vacation under Non-Pre-Emtive Priority Service by Matrix Geometric Method G. AYYAPPAN

More information

Retrial queue for cloud systems with separated processing and storage units

Retrial queue for cloud systems with separated processing and storage units Retrial queue for cloud systems with separated processing and storage units Tuan Phung-Duc Department of Mathematical and Computing Sciences Tokyo Institute of Technology Ookayama, Meguro-ku, Tokyo, Japan

More information

A Batch Arrival Retrial Queue with Two Phases of Service, Feedback and K Optional Vacations

A Batch Arrival Retrial Queue with Two Phases of Service, Feedback and K Optional Vacations Applied Mathematical Sciences, Vol. 6, 212, no. 22, 171-187 A Batch Arrival Retrial Queue with Two Phases of Service, Feedback and K Optional Vacations D. Arivudainambi and P. Godhandaraman Department

More information

M/M/3/3 AND M/M/4/4 RETRIAL QUEUES. Tuan Phung-Duc, Hiroyuki Masuyama, Shoji Kasahara and Yutaka Takahashi

M/M/3/3 AND M/M/4/4 RETRIAL QUEUES. Tuan Phung-Duc, Hiroyuki Masuyama, Shoji Kasahara and Yutaka Takahashi JOURNAL OF INDUSTRIAL AND doi:10.3934/imo.2009.5.431 MANAGEMENT OPTIMIZATION Volume 5, Number 3, August 2009 pp. 431 451 M/M/3/3 AND M/M/4/4 RETRIAL QUEUES Tuan Phung-Duc, Hiroyuki Masuyama, Shoi Kasahara

More information

An M/M/1 Retrial Queue with Unreliable Server 1

An M/M/1 Retrial Queue with Unreliable Server 1 An M/M/1 Retrial Queue with Unreliable Server 1 Nathan P. Sherman 2 and Jeffrey P. Kharoufeh 3 Department of Operational Sciences Air Force Institute of Technology Abstract We analyze an unreliable M/M/1

More information

Preemptive Resume Priority Retrial Queue with. Two Classes of MAP Arrivals

Preemptive Resume Priority Retrial Queue with. Two Classes of MAP Arrivals Applied Mathematical Sciences, Vol. 7, 2013, no. 52, 2569-2589 HIKARI Ltd, www.m-hikari.com Preemptive Resume Priority Retrial Queue with Two Classes of MAP Arrivals M. Senthil Kumar 1, S. R. Chakravarthy

More information

Geo (λ)/ Geo (µ) +G/2 Queues with Heterogeneous Servers Operating under FCFS Queue Discipline

Geo (λ)/ Geo (µ) +G/2 Queues with Heterogeneous Servers Operating under FCFS Queue Discipline American Journal of Applied Mathematics and Statistics, 5, Vol. 3, No., 54-58 Available online at http://pubs.sciepub.com/aams/3// Science and Education Publishing DOI:.69/aams-3-- Geo ()/ Geo () +G/ Queues

More information

On Tandem Blocking Queues with a Common Retrial Queue

On Tandem Blocking Queues with a Common Retrial Queue On Tandem Blocking Queues with a Common Retrial Queue K. Avrachenkov U. Yechiali Abstract We consider systems of tandem blocking queues having a common retrial queue, for which explicit analytic results

More information

The discrete-time Geom/G/1 queue with multiple adaptive vacations and. setup/closedown times

The discrete-time Geom/G/1 queue with multiple adaptive vacations and. setup/closedown times ISSN 1750-9653, England, UK International Journal of Management Science and Engineering Management Vol. 2 (2007) No. 4, pp. 289-296 The discrete-time Geom/G/1 queue with multiple adaptive vacations and

More information

System with a Server Subject to Breakdowns

System with a Server Subject to Breakdowns Applied Mathematical Sciences Vol. 7 213 no. 11 539 55 On Two Modifications of E 2 /E 2 /1/m Queueing System with a Server Subject to Breakdowns Michal Dorda VSB - Technical University of Ostrava Faculty

More information

A Class of Geom/Geom/1 Discrete-time Queueing System with Negative Customers

A Class of Geom/Geom/1 Discrete-time Queueing System with Negative Customers ISSN 749-3889 (print), 749-3897 (online) International Journal of Nonlinear Science Vol.5(2008) No.3,pp.275-280 A Class of Geom/Geom/ Discrete-time Queueing System with Negative Customers Li Ma Deptartment

More information

Submitted Manuscript. A discrete-time single-server queue with balking: economic applications

Submitted Manuscript. A discrete-time single-server queue with balking: economic applications Submitted Manuscript A discrete-time single-server queue with balking: economic applications Journal: Manuscript ID: Journal Selection: Applied Economics APE-0-0.R Applied Economics JEL Code: C0 - General

More information

On Tandem Blocking Queues with a Common Retrial Queue

On Tandem Blocking Queues with a Common Retrial Queue On Tandem Blocking Queues with a Common Retrial Queue K. Avrachenkov U. Yechiali Abstract We consider systems of tandem blocking queues having a common retrial queue. The model represents dynamics of short

More information

A discrete-time priority queue with train arrivals

A discrete-time priority queue with train arrivals A discrete-time priority queue with train arrivals Joris Walraevens, Sabine Wittevrongel and Herwig Bruneel SMACS Research Group Department of Telecommunications and Information Processing (IR07) Ghent

More information

Asymptotic study of a busy period in a retrial queue

Asymptotic study of a busy period in a retrial queue PUB. IRMA, LILLE 2011 Vol. 71, N o VI Asymptotic study of a busy period in a retrial queue Y. Taleb a, F. Achemine a, D. Hamadouche a, A. Aissani b djhamad@mail.ummto.dz, taleb.youcef@yahoo.fr, achemine_f2001@yahoo.fr,

More information

Citation Operational Research (2012), 12(2):

Citation Operational Research (2012), 12(2): TitleAn explicit solution for a tandem q Author(s) Phung-Duc, Tuan Citation Operational Research (2012), 12(2): Issue Date 2012-08 URL http://hdl.handle.net/2433/158216 RightThe final publication is available

More information

1 IEOR 4701: Continuous-Time Markov Chains

1 IEOR 4701: Continuous-Time Markov Chains Copyright c 2006 by Karl Sigman 1 IEOR 4701: Continuous-Time Markov Chains A Markov chain in discrete time, {X n : n 0}, remains in any state for exactly one unit of time before making a transition (change

More information

Lecture 20: Reversible Processes and Queues

Lecture 20: Reversible Processes and Queues Lecture 20: Reversible Processes and Queues 1 Examples of reversible processes 11 Birth-death processes We define two non-negative sequences birth and death rates denoted by {λ n : n N 0 } and {µ n : n

More information

M/M/1 Queueing System with Delayed Controlled Vacation

M/M/1 Queueing System with Delayed Controlled Vacation M/M/1 Queueing System with Delayed Controlled Vacation Yonglu Deng, Zhongshan University W. John Braun, University of Winnipeg Yiqiang Q. Zhao, University of Winnipeg Abstract An M/M/1 queue with delayed

More information

A FAST MATRIX-ANALYTIC APPROXIMATION FOR THE TWO CLASS GI/G/1 NON-PREEMPTIVE PRIORITY QUEUE

A FAST MATRIX-ANALYTIC APPROXIMATION FOR THE TWO CLASS GI/G/1 NON-PREEMPTIVE PRIORITY QUEUE A FAST MATRIX-ANAYTIC APPROXIMATION FOR TE TWO CASS GI/G/ NON-PREEMPTIVE PRIORITY QUEUE Gábor orváth Department of Telecommunication Budapest University of Technology and Economics. Budapest Pf. 9., ungary

More information

Contents Preface The Exponential Distribution and the Poisson Process Introduction to Renewal Theory

Contents Preface The Exponential Distribution and the Poisson Process Introduction to Renewal Theory Contents Preface... v 1 The Exponential Distribution and the Poisson Process... 1 1.1 Introduction... 1 1.2 The Density, the Distribution, the Tail, and the Hazard Functions... 2 1.2.1 The Hazard Function

More information

Two Heterogeneous Servers Queueing-Inventory System with Sharing Finite Buffer and a Flexible Server

Two Heterogeneous Servers Queueing-Inventory System with Sharing Finite Buffer and a Flexible Server Two Heterogeneous Servers Queueing-Inventory System with Sharing Finite Buffer and a Flexible Server S. Jehoashan Kingsly 1, S. Padmasekaran and K. Jeganathan 3 1 Department of Mathematics, Adhiyamaan

More information

THIELE CENTRE. The M/M/1 queue with inventory, lost sale and general lead times. Mohammad Saffari, Søren Asmussen and Rasoul Haji

THIELE CENTRE. The M/M/1 queue with inventory, lost sale and general lead times. Mohammad Saffari, Søren Asmussen and Rasoul Haji THIELE CENTRE for applied mathematics in natural science The M/M/1 queue with inventory, lost sale and general lead times Mohammad Saffari, Søren Asmussen and Rasoul Haji Research Report No. 11 September

More information

Stability Condition of a Retrial Queueing System with Abandoned and Feedback Customers

Stability Condition of a Retrial Queueing System with Abandoned and Feedback Customers Available at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 1932-9466 Vol. 10, Issue 2 December 2015), pp. 667 677 Applications and Applied Mathematics: An International Journal AAM) Stability Condition

More information

ON MAIN CHARACTERISTICS OF THE M/M/1/N QUEUE WITH SINGLE AND BATCH ARRIVALS AND THE QUEUE SIZE CONTROLLED BY AQM ALGORITHMS

ON MAIN CHARACTERISTICS OF THE M/M/1/N QUEUE WITH SINGLE AND BATCH ARRIVALS AND THE QUEUE SIZE CONTROLLED BY AQM ALGORITHMS K Y B E R N E T I K A V O L U M E 4 7 ( 2 0 1 1 ), N U M B E R 6, P A G E S 9 3 0 9 4 3 ON MAIN CHARACTERISTICS OF THE M/M/1/N QUEUE WITH SINGLE AND BATCH ARRIVALS AND THE QUEUE SIZE CONTROLLED BY AQM

More information

Performance Evaluation of Queuing Systems

Performance Evaluation of Queuing Systems Performance Evaluation of Queuing Systems Introduction to Queuing Systems System Performance Measures & Little s Law Equilibrium Solution of Birth-Death Processes Analysis of Single-Station Queuing Systems

More information

A Heterogeneous Two-Server Queueing System with Balking and Server Breakdowns

A Heterogeneous Two-Server Queueing System with Balking and Server Breakdowns The Eighth International Symposium on Operations Research and Its Applications (ISORA 09) Zhangjiajie, China, September 20 22, 2009 Copyright 2009 ORSC & APORC, pp. 230 244 A Heterogeneous Two-Server Queueing

More information

Queueing systems. Renato Lo Cigno. Simulation and Performance Evaluation Queueing systems - Renato Lo Cigno 1

Queueing systems. Renato Lo Cigno. Simulation and Performance Evaluation Queueing systems - Renato Lo Cigno 1 Queueing systems Renato Lo Cigno Simulation and Performance Evaluation 2014-15 Queueing systems - Renato Lo Cigno 1 Queues A Birth-Death process is well modeled by a queue Indeed queues can be used to

More information

J. MEDHI STOCHASTIC MODELS IN QUEUEING THEORY

J. MEDHI STOCHASTIC MODELS IN QUEUEING THEORY J. MEDHI STOCHASTIC MODELS IN QUEUEING THEORY SECOND EDITION ACADEMIC PRESS An imprint of Elsevier Science Amsterdam Boston London New York Oxford Paris San Diego San Francisco Singapore Sydney Tokyo Contents

More information

Figure 10.1: Recording when the event E occurs

Figure 10.1: Recording when the event E occurs 10 Poisson Processes Let T R be an interval. A family of random variables {X(t) ; t T} is called a continuous time stochastic process. We often consider T = [0, 1] and T = [0, ). As X(t) is a random variable

More information

Derivation of Formulas by Queueing Theory

Derivation of Formulas by Queueing Theory Appendices Spectrum Requirement Planning in Wireless Communications: Model and Methodology for IMT-Advanced E dite d by H. Takagi and B. H. Walke 2008 J ohn Wiley & Sons, L td. ISBN: 978-0-470-98647-9

More information

Chapter 1. Introduction. 1.1 Stochastic process

Chapter 1. Introduction. 1.1 Stochastic process Chapter 1 Introduction Process is a phenomenon that takes place in time. In many practical situations, the result of a process at any time may not be certain. Such a process is called a stochastic process.

More information

A Simple Solution for the M/D/c Waiting Time Distribution

A Simple Solution for the M/D/c Waiting Time Distribution A Simple Solution for the M/D/c Waiting Time Distribution G.J.Franx, Universiteit van Amsterdam November 6, 998 Abstract A surprisingly simple and explicit expression for the waiting time distribution

More information

c~msshset ea::-institut OFs TECH-~-~~ENOLGY~ wor :: :a:.-p i ': ----.: _- ; ; - L i: : ; - i. ; i.. i.. r - - i-; i. -.?. _'.: :: : '.

c~msshset ea::-institut OFs TECH-~-~~ENOLGY~ wor :: :a:.-p i ': ----.: _- ; ; - L i: : ; - i. ; i.. i.. r - - i-; i. -.?. _'.: :: : '. wor :: :a:.-p,..,!. i. - -., : -!'.,:.:....- '. ' -: : _'.: :: : ~~~~~~- : 00 :0; 0 :0orkig: : ; : 0V:3 a 0-:0; r-x '-:'::-,,,.'-::, -,!- :':- ': ~, - : :: ', i ': ----.: _- ; i. ; -- 1. ; - ; - ' : L

More information

Batch Arrival Queueing System. with Two Stages of Service

Batch Arrival Queueing System. with Two Stages of Service Int. Journal of Math. Analysis, Vol. 8, 2014, no. 6, 247-258 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/10.12988/ijma.2014.411 Batch Arrival Queueing System with Two Stages of Service S. Maragathasundari

More information

TCOM 501: Networking Theory & Fundamentals. Lecture 6 February 19, 2003 Prof. Yannis A. Korilis

TCOM 501: Networking Theory & Fundamentals. Lecture 6 February 19, 2003 Prof. Yannis A. Korilis TCOM 50: Networking Theory & Fundamentals Lecture 6 February 9, 003 Prof. Yannis A. Korilis 6- Topics Time-Reversal of Markov Chains Reversibility Truncating a Reversible Markov Chain Burke s Theorem Queues

More information

STEADY-STATE BEHAVIOR OF AN M/M/1 QUEUE IN RANDOM ENVIRONMENT SUBJECT TO SYSTEM FAILURES AND REPAIRS. S. Sophia 1, B. Praba 2

STEADY-STATE BEHAVIOR OF AN M/M/1 QUEUE IN RANDOM ENVIRONMENT SUBJECT TO SYSTEM FAILURES AND REPAIRS. S. Sophia 1, B. Praba 2 International Journal of Pure and Applied Mathematics Volume 101 No. 2 2015, 267-279 ISSN: 1311-8080 (printed version); ISSN: 1314-3395 (on-line version) url: http://www.ijpam.eu doi: http://dx.doi.org/10.12732/ijpam.v101i2.11

More information

Queues and Queueing Networks

Queues and Queueing Networks Queues and Queueing Networks Sanjay K. Bose Dept. of EEE, IITG Copyright 2015, Sanjay K. Bose 1 Introduction to Queueing Models and Queueing Analysis Copyright 2015, Sanjay K. Bose 2 Model of a Queue Arrivals

More information

Burst Arrival Queues with Server Vacations and Random Timers

Burst Arrival Queues with Server Vacations and Random Timers Burst Arrival Queues with Server acations and Random Timers Merav Shomrony and Uri Yechiali Department of Statistics and Operations Research School of Mathematical Sciences Raymond and Beverly Sackler

More information

Introduction to Markov Chains, Queuing Theory, and Network Performance

Introduction to Markov Chains, Queuing Theory, and Network Performance Introduction to Markov Chains, Queuing Theory, and Network Performance Marceau Coupechoux Telecom ParisTech, departement Informatique et Réseaux marceau.coupechoux@telecom-paristech.fr IT.2403 Modélisation

More information

A Direct Approach to Transient Queue-Size Distribution in a Finite-Buffer Queue with AQM

A Direct Approach to Transient Queue-Size Distribution in a Finite-Buffer Queue with AQM Appl. Math. Inf. Sci. 7, No. 3, 99-915 (213) 99 Applied Mathematics & Information Sciences An International Journal A Direct Approach to Transient Queue-Size Distribution in a Finite-Buffer Queue with

More information

Part I Stochastic variables and Markov chains

Part I Stochastic variables and Markov chains Part I Stochastic variables and Markov chains Random variables describe the behaviour of a phenomenon independent of any specific sample space Distribution function (cdf, cumulative distribution function)

More information

Link Models for Circuit Switching

Link Models for Circuit Switching Link Models for Circuit Switching The basis of traffic engineering for telecommunication networks is the Erlang loss function. It basically allows us to determine the amount of telephone traffic that can

More information

NEW FRONTIERS IN APPLIED PROBABILITY

NEW FRONTIERS IN APPLIED PROBABILITY J. Appl. Prob. Spec. Vol. 48A, 209 213 (2011) Applied Probability Trust 2011 NEW FRONTIERS IN APPLIED PROBABILITY A Festschrift for SØREN ASMUSSEN Edited by P. GLYNN, T. MIKOSCH and T. ROLSKI Part 4. Simulation

More information

A preemptive repeat priority queue with resampling: Performance analysis

A preemptive repeat priority queue with resampling: Performance analysis Ann Oper Res (2006) 46:89 202 DOI 0.007/s0479-006-0053-4 A preemptive repeat priority queue with resampling: Performance analysis Joris Walraevens Bart Steyaert Herwig Bruneel Published online: 6 July

More information

Stationary Analysis of a Multiserver queue with multiple working vacation and impatient customers

Stationary Analysis of a Multiserver queue with multiple working vacation and impatient customers Available at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 932-9466 Vol. 2, Issue 2 (December 207), pp. 658 670 Applications and Applied Mathematics: An International Journal (AAM) Stationary Analysis of

More information

LIMITS FOR QUEUES AS THE WAITING ROOM GROWS. Bell Communications Research AT&T Bell Laboratories Red Bank, NJ Murray Hill, NJ 07974

LIMITS FOR QUEUES AS THE WAITING ROOM GROWS. Bell Communications Research AT&T Bell Laboratories Red Bank, NJ Murray Hill, NJ 07974 LIMITS FOR QUEUES AS THE WAITING ROOM GROWS by Daniel P. Heyman Ward Whitt Bell Communications Research AT&T Bell Laboratories Red Bank, NJ 07701 Murray Hill, NJ 07974 May 11, 1988 ABSTRACT We study the

More information

A Study on M x /G/1 Queuing System with Essential, Optional Service, Modified Vacation and Setup time

A Study on M x /G/1 Queuing System with Essential, Optional Service, Modified Vacation and Setup time A Study on M x /G/1 Queuing System with Essential, Optional Service, Modified Vacation and Setup time E. Ramesh Kumar 1, L. Poornima 2 1 Associate Professor, Department of Mathematics, CMS College of Science

More information

Overload Analysis of the PH/PH/1/K Queue and the Queue of M/G/1/K Type with Very Large K

Overload Analysis of the PH/PH/1/K Queue and the Queue of M/G/1/K Type with Very Large K Overload Analysis of the PH/PH/1/K Queue and the Queue of M/G/1/K Type with Very Large K Attahiru Sule Alfa Department of Mechanical and Industrial Engineering University of Manitoba Winnipeg, Manitoba

More information

Data analysis and stochastic modeling

Data analysis and stochastic modeling Data analysis and stochastic modeling Lecture 7 An introduction to queueing theory Guillaume Gravier guillaume.gravier@irisa.fr with a lot of help from Paul Jensen s course http://www.me.utexas.edu/ jensen/ormm/instruction/powerpoint/or_models_09/14_queuing.ppt

More information

Priority Queueing System with a Single Server Serving Two Queues M [X 1], M [X 2] /G 1, G 2 /1 with Balking and Optional Server Vacation

Priority Queueing System with a Single Server Serving Two Queues M [X 1], M [X 2] /G 1, G 2 /1 with Balking and Optional Server Vacation Available at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 1932-9466 Vol. 11, Issue 1 (June 216), pp. 61 82 Applications and Applied Mathematics: An International Journal (AAM) Priority Queueing System

More information

GI/M/1 and GI/M/m queuing systems

GI/M/1 and GI/M/m queuing systems GI/M/1 and GI/M/m queuing systems Dmitri A. Moltchanov moltchan@cs.tut.fi http://www.cs.tut.fi/kurssit/tlt-2716/ OUTLINE: GI/M/1 queuing system; Methods of analysis; Imbedded Markov chain approach; Waiting

More information

An M/M/1 Queue in Random Environment with Disasters

An M/M/1 Queue in Random Environment with Disasters An M/M/1 Queue in Random Environment with Disasters Noam Paz 1 and Uri Yechiali 1,2 1 Department of Statistics and Operations Research School of Mathematical Sciences Tel Aviv University, Tel Aviv 69978,

More information

Multiserver Queueing Model subject to Single Exponential Vacation

Multiserver Queueing Model subject to Single Exponential Vacation Journal of Physics: Conference Series PAPER OPEN ACCESS Multiserver Queueing Model subject to Single Exponential Vacation To cite this article: K V Vijayashree B Janani 2018 J. Phys.: Conf. Ser. 1000 012129

More information

WORKING PAPER SLOAN SCHOOL OF MANAGEMENT MASSACHUSETTS ALFRED P. CAMBRIDGE, MASSACHUSETTS INSTITUTE OF TECHNOLOGY

WORKING PAPER SLOAN SCHOOL OF MANAGEMENT MASSACHUSETTS ALFRED P. CAMBRIDGE, MASSACHUSETTS INSTITUTE OF TECHNOLOGY :ib S1988 f ;*: A.^.T/ ALFRED P. WORKING PAPER SLOAN SCHOOL OF MANAGEMENT A DISTRIBUTIONAL FORM OF LITTLES LAW J. Keilson* GTE Laboratories Incorporated and Mass. Inst, of Technology and L. D. Servi* GTE

More information

2905 Queueing Theory and Simulation PART III: HIGHER DIMENSIONAL AND NON-MARKOVIAN QUEUES

2905 Queueing Theory and Simulation PART III: HIGHER DIMENSIONAL AND NON-MARKOVIAN QUEUES 295 Queueing Theory and Simulation PART III: HIGHER DIMENSIONAL AND NON-MARKOVIAN QUEUES 16 Queueing Systems with Two Types of Customers In this section, we discuss queueing systems with two types of customers.

More information

Markov Chains CK eqns Classes Hitting times Rec./trans. Strong Markov Stat. distr. Reversibility * Markov Chains

Markov Chains CK eqns Classes Hitting times Rec./trans. Strong Markov Stat. distr. Reversibility * Markov Chains Markov Chains A random process X is a family {X t : t T } of random variables indexed by some set T. When T = {0, 1, 2,... } one speaks about a discrete-time process, for T = R or T = [0, ) one has a continuous-time

More information

Relating Polling Models with Zero and Nonzero Switchover Times

Relating Polling Models with Zero and Nonzero Switchover Times Relating Polling Models with Zero and Nonzero Switchover Times Mandyam M. Srinivasan Management Science Program College of Business Administration The University of Tennessee Knoxville, TN 37996-0562 Shun-Chen

More information

CPSC 531: System Modeling and Simulation. Carey Williamson Department of Computer Science University of Calgary Fall 2017

CPSC 531: System Modeling and Simulation. Carey Williamson Department of Computer Science University of Calgary Fall 2017 CPSC 531: System Modeling and Simulation Carey Williamson Department of Computer Science University of Calgary Fall 2017 Motivating Quote for Queueing Models Good things come to those who wait - poet/writer

More information

A Vacation Queue with Additional Optional Service in Batches

A Vacation Queue with Additional Optional Service in Batches Applied Mathematical Sciences, Vol. 3, 2009, no. 24, 1203-1208 A Vacation Queue with Additional Optional Service in Batches S. Pazhani Bala Murugan Department of Mathematics, Annamalai University Annamalainagar-608

More information

Stability of the two queue system

Stability of the two queue system Stability of the two queue system Iain M. MacPhee and Lisa J. Müller University of Durham Department of Mathematical Science Durham, DH1 3LE, UK (e-mail: i.m.macphee@durham.ac.uk, l.j.muller@durham.ac.uk)

More information

Stationary Probabilities of Markov Chains with Upper Hessenberg Transition Matrices

Stationary Probabilities of Markov Chains with Upper Hessenberg Transition Matrices Stationary Probabilities of Marov Chains with Upper Hessenberg Transition Matrices Y. Quennel ZHAO Department of Mathematics and Statistics University of Winnipeg Winnipeg, Manitoba CANADA R3B 2E9 Susan

More information

Exact Simulation of the Stationary Distribution of M/G/c Queues

Exact Simulation of the Stationary Distribution of M/G/c Queues 1/36 Exact Simulation of the Stationary Distribution of M/G/c Queues Professor Karl Sigman Columbia University New York City USA Conference in Honor of Søren Asmussen Monday, August 1, 2011 Sandbjerg Estate

More information

Contents LIST OF TABLES... LIST OF FIGURES... xvii. LIST OF LISTINGS... xxi PREFACE. ...xxiii

Contents LIST OF TABLES... LIST OF FIGURES... xvii. LIST OF LISTINGS... xxi PREFACE. ...xxiii LIST OF TABLES... xv LIST OF FIGURES... xvii LIST OF LISTINGS... xxi PREFACE...xxiii CHAPTER 1. PERFORMANCE EVALUATION... 1 1.1. Performance evaluation... 1 1.2. Performance versus resources provisioning...

More information

Queueing Systems with Customer Abandonments and Retrials. Song Deng

Queueing Systems with Customer Abandonments and Retrials. Song Deng Queueing Systems with Customer Abandonments and Retrials by Song Deng A dissertation submitted in partial satisfaction of the requirements for the degree of Doctor of Philosophy in Engineering - Industrial

More information

On the Pathwise Optimal Bernoulli Routing Policy for Homogeneous Parallel Servers

On the Pathwise Optimal Bernoulli Routing Policy for Homogeneous Parallel Servers On the Pathwise Optimal Bernoulli Routing Policy for Homogeneous Parallel Servers Ger Koole INRIA Sophia Antipolis B.P. 93, 06902 Sophia Antipolis Cedex France Mathematics of Operations Research 21:469

More information

Non Markovian Queues (contd.)

Non Markovian Queues (contd.) MODULE 7: RENEWAL PROCESSES 29 Lecture 5 Non Markovian Queues (contd) For the case where the service time is constant, V ar(b) = 0, then the P-K formula for M/D/ queue reduces to L s = ρ + ρ 2 2( ρ) where

More information

The Erlang Model with non-poisson Call Arrivals

The Erlang Model with non-poisson Call Arrivals The Erlang Model with non-poisson Call Arrivals Thomas Bonald To cite this version: Thomas Bonald. The Erlang Model with non-poisson Call Arrivals. Sigmetrics / Performance 2006 - Joint International Conference

More information

Delay characteristics in discrete-time GI-G-1 queues with non-preemptive priority queueing discipline

Delay characteristics in discrete-time GI-G-1 queues with non-preemptive priority queueing discipline Delay characteristics in discrete-time GI-G-1 queues with non-preemptive priority queueing discipline Joris Walraevens, Bart Steyaert, Herwig Bruneel SMACS Research Group, University of Ghent, Vakgroep

More information

System occupancy of a two-class batch-service queue with class-dependent variable server capacity

System occupancy of a two-class batch-service queue with class-dependent variable server capacity System occupancy of a two-class batch-service queue with class-dependent variable server capacity Jens Baetens 1, Bart Steyaert 1, Dieter Claeys 1,2, and Herwig Bruneel 1 1 SMACS Research Group, Dept.

More information

A tandem queueing model with coupled processors

A tandem queueing model with coupled processors A tandem queueing model with coupled processors Jacques Resing Department of Mathematics and Computer Science Eindhoven University of Technology P.O. Box 513 5600 MB Eindhoven The Netherlands Lerzan Örmeci

More information

Exercises Stochastic Performance Modelling. Hamilton Institute, Summer 2010

Exercises Stochastic Performance Modelling. Hamilton Institute, Summer 2010 Exercises Stochastic Performance Modelling Hamilton Institute, Summer Instruction Exercise Let X be a non-negative random variable with E[X ]

More information

The M/G/1 FIFO queue with several customer classes

The M/G/1 FIFO queue with several customer classes The M/G/1 FIFO queue with several customer classes Boxma, O.J.; Takine, T. Published: 01/01/2003 Document Version Publisher s PDF, also known as Version of Record (includes final page, issue and volume

More information

EQUILIBRIUM STRATEGIES IN AN M/M/1 QUEUE WITH SETUP TIMES AND A SINGLE VACATION POLICY

EQUILIBRIUM STRATEGIES IN AN M/M/1 QUEUE WITH SETUP TIMES AND A SINGLE VACATION POLICY EQUILIBRIUM STRATEGIES IN AN M/M/1 QUEUE WITH SETUP TIMES AND A SINGLE VACATION POLICY Dequan Yue 1, Ruiling Tian 1, Wuyi Yue 2, Yaling Qin 3 1 College of Sciences, Yanshan University, Qinhuangdao 066004,

More information

Queueing Systems: Lecture 3. Amedeo R. Odoni October 18, Announcements

Queueing Systems: Lecture 3. Amedeo R. Odoni October 18, Announcements Queueing Systems: Lecture 3 Amedeo R. Odoni October 18, 006 Announcements PS #3 due tomorrow by 3 PM Office hours Odoni: Wed, 10/18, :30-4:30; next week: Tue, 10/4 Quiz #1: October 5, open book, in class;

More information

6 Solving Queueing Models

6 Solving Queueing Models 6 Solving Queueing Models 6.1 Introduction In this note we look at the solution of systems of queues, starting with simple isolated queues. The benefits of using predefined, easily classified queues will

More information

M/G/1 and Priority Queueing

M/G/1 and Priority Queueing M/G/1 and Priority Queueing Richard T. B. Ma School of Computing National University of Singapore CS 5229: Advanced Compute Networks Outline PASTA M/G/1 Workload and FIFO Delay Pollaczek Khinchine Formula

More information

Queueing Theory I Summary! Little s Law! Queueing System Notation! Stationary Analysis of Elementary Queueing Systems " M/M/1 " M/M/m " M/M/1/K "

Queueing Theory I Summary! Little s Law! Queueing System Notation! Stationary Analysis of Elementary Queueing Systems  M/M/1  M/M/m  M/M/1/K Queueing Theory I Summary Little s Law Queueing System Notation Stationary Analysis of Elementary Queueing Systems " M/M/1 " M/M/m " M/M/1/K " Little s Law a(t): the process that counts the number of arrivals

More information

Intro Refresher Reversibility Open networks Closed networks Multiclass networks Other networks. Queuing Networks. Florence Perronnin

Intro Refresher Reversibility Open networks Closed networks Multiclass networks Other networks. Queuing Networks. Florence Perronnin Queuing Networks Florence Perronnin Polytech Grenoble - UGA March 23, 27 F. Perronnin (UGA) Queuing Networks March 23, 27 / 46 Outline Introduction to Queuing Networks 2 Refresher: M/M/ queue 3 Reversibility

More information

M/G/1 and M/G/1/K systems

M/G/1 and M/G/1/K systems M/G/1 and M/G/1/K systems Dmitri A. Moltchanov dmitri.moltchanov@tut.fi http://www.cs.tut.fi/kurssit/elt-53606/ OUTLINE: Description of M/G/1 system; Methods of analysis; Residual life approach; Imbedded

More information

Sensitivity Analysis for Discrete-Time Randomized Service Priority Queues

Sensitivity Analysis for Discrete-Time Randomized Service Priority Queues Sensitivity Analysis for Discrete-Time Randomized Service Priority Queues George Kesidis 1, Takis Konstantopoulos 2, Michael Zazanis 3 1. Elec. & Comp. Eng. Dept, University of Waterloo, Waterloo, ON,

More information

Research Article Performance of an M/M/1 Retrial Queue with Working Vacation Interruption and Classical Retrial Policy

Research Article Performance of an M/M/1 Retrial Queue with Working Vacation Interruption and Classical Retrial Policy Advances in Operations Research Volume 216, Article ID 453831, 9 pages http://dx.doi.org/1.1155/216/453831 Research Article Performance of an M/M/1 Retrial Queue with Working Vacation Interruption and

More information

Stochastic Models in Computer Science A Tutorial

Stochastic Models in Computer Science A Tutorial Stochastic Models in Computer Science A Tutorial Dr. Snehanshu Saha Department of Computer Science PESIT BSC, Bengaluru WCI 2015 - August 10 to August 13 1 Introduction 2 Random Variable 3 Introduction

More information

Optimal and Equilibrium Retrial Rates in Single-Server Multi-orbit Retrial Systems

Optimal and Equilibrium Retrial Rates in Single-Server Multi-orbit Retrial Systems Optimal and Equilibrium Retrial Rates in Single-Server Multi-orbit Retrial Systems Konstantin Avrachenkov, Evsey Morozov, Ruslana Nekrasova To cite this version: Konstantin Avrachenkov, Evsey Morozov,

More information

L. Lakatos, S. V. Serebriakova

L. Lakatos, S. V. Serebriakova L. Lakatos S. V. Serebriakova Eötvös Lorand University Budapest Hungary V. M. Glushkov Institute of Cybernetics of NAS of Ukraine Kyiv Ukraine e-mail: lakatos@inf.elte.hu svitlana.pustova @ gmail.com Abstract

More information

Finite source retrial queues with two phase service. Jinting Wang* and Fang Wang

Finite source retrial queues with two phase service. Jinting Wang* and Fang Wang Int. J. Operational Research, Vol. 3, No. 4, 27 42 Finite source retrial queues with two phase service Jinting Wang* and Fang Wang Department of Mathematics, Beijing Jiaotong University, Beijing, 44, China

More information

Statistics 150: Spring 2007

Statistics 150: Spring 2007 Statistics 150: Spring 2007 April 23, 2008 0-1 1 Limiting Probabilities If the discrete-time Markov chain with transition probabilities p ij is irreducible and positive recurrent; then the limiting probabilities

More information

Name of the Student:

Name of the Student: SUBJECT NAME : Probability & Queueing Theory SUBJECT CODE : MA 6453 MATERIAL NAME : Part A questions REGULATION : R2013 UPDATED ON : November 2017 (Upto N/D 2017 QP) (Scan the above QR code for the direct

More information

Stochastic inventory system with two types of services

Stochastic inventory system with two types of services Int. J. Adv. Appl. Math. and Mech. 2() (204) 20-27 ISSN: 2347-2529 Available online at www.ijaamm.com International Journal of Advances in Applied Mathematics and Mechanics Stochastic inventory system

More information

PBW 654 Applied Statistics - I Urban Operations Research

PBW 654 Applied Statistics - I Urban Operations Research PBW 654 Applied Statistics - I Urban Operations Research Lecture 2.I Queuing Systems An Introduction Operations Research Models Deterministic Models Linear Programming Integer Programming Network Optimization

More information

A Queueing System with Queue Length Dependent Service Times, with Applications to Cell Discarding in ATM Networks

A Queueing System with Queue Length Dependent Service Times, with Applications to Cell Discarding in ATM Networks A Queueing System with Queue Length Dependent Service Times, with Applications to Cell Discarding in ATM Networks by Doo Il Choi, Charles Knessl and Charles Tier University of Illinois at Chicago 85 South

More information