Markovian Arrival Process Research Articles

Optimal Control of State-Dependent Service Rates in a MAP/M/1 Queue

In this paper, we study the optimal control of service rates in a queueing system with a Markovian arrival process (MAP) and exponential service times. The service rate is allowed to be state dependent, i.e., we can adjust the service rate according to the queue length and the phase of the MAP. The cost function consists of holding cost and operating cost. The goal is to find the optimal service rates that minimize the long-run average total cost. To achieve that, we use the matrix-analytic methods (MAM) together with the sensitivity-based optimization (SBO) theory. A performance difference formula is derived, which can quantify the difference of the long-run average total cost under any two different settings of service rates. Based on the difference formula, we show that the long-run average total cost is monotone in the service rate and the optimal control is a bang–bang control. We also show that, under some mild conditions, the optimal control policy of service rates is of a quasi-threshold-type. By utilizing the MAM theory, we propose a recursive algorithm to compute the value function related quantities. An iterative algorithm to efficiently find the optimal policy, which is similar to policy iteration, is proposed based on the SBO theory. We further study some special cases of the problem, such as the optimality of the threshold-type policy for the M/M/1 queue. Finally, a number of numerical examples are presented to demonstrate the main results and explore the impact of the phase of the MAP on the optimization in the MAP/M/1 queue.

TTL approximations of the cache replacement algorithms LRU(m) and h-LRU

Computer system and network performance can be significantly improved by caching frequently used information. When the cache size is limited, the cache replacement algorithm has an important impact on the effectiveness of caching. In this paper we introduce time-to-live (TTL) approximations to determine the cache hit probability of two classes of cache replacement algorithms: h-LRU and LRU(m). These approximations only require the requests to be generated according to a general Markovian arrival process (MAP). This includes phase-type renewal processes and the IRM model as special cases. We provide both numerical and theoretical support for the claim that the proposed TTL approximations are asymptotically exact. In particular, we show that the transient hit probability converges to the solution of a set of ODEs (under the IRM model), where the fixed point of the set of ODEs corresponds to the TTL approximation.We use this approximation and trace-based simulation to compare the performance of h-LRU and LRU(m). First, we show that they perform alike, while the latter requires less work when a hit/miss occurs. Second, we show that as opposed to LRU, h-LRU and LRU(m) are sensitive to the correlation between consecutive inter-request times. Last, we study cache partitioning. In all tested cases, the hit probability improved by partitioning the cache into different parts—each being dedicated to a particular content provider. However, the gain is limited and the optimal partition sizes are very sensitive to the problem’s parameters.

Queueing systems with correlated arrival flows and their applications to modeling telecommunication networks

We give a brief survey of literature devoted to studying queueing systems with Markovian and batch Markovian arrival processes and their application to modeling telecommunication networks.

A stochastic inventory system with two modes of service and retrial of customers

This paper analyses an inventory model with positive service time and retrial of customers. The customers arrive according to a Markovian Arrival Process with representation ( $$D_{0} ,D_{1}$$ ). The service times are assumed to be of Phase type distribution with representation $$\left( {\eta ,U} \right)$$ . When the inventory level depletes to the reorder point s, the service is given at a reduced rate and the service time distribution has the representation $$\left( {\eta ,\alpha U} \right)$$ , where $$0 < \alpha < 1$$ . An arriving customer, who finds the inventory level zero or the server busy, enters into an orbit of infinite capacity and will retry for service from there. The lead time follows an exponential distribution with rate β. We analyze the system using Matrix Analytic Method. Some important performance measures in the steady state are obtained. A suitable cost function for the expected total cost is constructed and analyzed numerically and graphically.

Analysis of unreliable BMAP/PH/N type queue with Markovian flow of breakdowns

Unreliability of components is the inherent feature of many real world systems and its account is vital for correct prediction of performance measures of the system. Multi-server queueing model considered in this paper allows to evaluate characteristics of the systems under much more general assumptions about the probabilistic distributions describing behavior of the system than models known in existing literature. We analyse the multi-server queue with infinite buffer and the Batch Markovian Arrival Process (BMAP) of customers. The servers are identical and independent of each other. Service time of a customer has phase-type (PH) distribution. Servers are subject to breakdowns and repairs. Breakdowns occurrence moments are defined by the Markovian Arrival Process (MAP). The breakdown causes a failure of any server, which is not under repair. When a server fails the repair period starts immediately. The duration of this period has PH distribution. A customer whose service is interrupted occupies an idle server, if any, and continues his/her service. If he/she does not see an idle server, the customer goes to the buffer with some probability and permanently leaves the system with the complementary probability. We derive the constructive ergodicity condition and calculate the stationary distribution and the main performance characteristics of the system. Illustrative numerical examples are presented.

Canonical representation of order 2 transient Markov and rational arrival processes

Similar to other processes that are modulated by background Markov chains the matrix representation of a transient Markov arrival process is not unique and the use of a convenient unique canonical form is essential for practical computations.The paper presents a set of 5 Markovian forms which provide a unique and minimal representation for all members of the TMAP(2) class. In the course of the derivation we also show the identity of the TMAP(2) and the TRAP(2) classes.

A Queueing System with Heterogeneous Impatient Customers and Consumable Additional Items

AbstractA single-server queueing system with a marked Markovian arrival process of heterogeneous customers is considered. Type-1 customers have limited preemptive priority over type-2 customers. There is an infinite buffer for type-2 customers and no buffer for type-1 customers. There is also a finite buffer (stock) for consumable additional items (semi-products, half-stocks, etc.) which arrive according to the Markovian arrival process. Service of a customer requires a fixed number of consumable additional items depending on the type of the customer. The service time has a phase-type distribution depending on the type of the customer. Customers in the buffer are impatient and may leave the system without service after an exponentially distributed amount of waiting time. Aiming to minimize the loss probability of type-1 customers and maximize throughput of the system, a threshold strategy of admission to service of type-2 customers is offered. Service of type-2 customer can start only if the server is idle and the number of consumable additional items in the stock exceeds the fixed threshold. Stationary distributions of the system states and the waiting time are computed. In the numerical example, we show some interesting effects and illustrate a possibility of application of the presented results for solution of optimization problems.

Dynamic Power Management With Optimal Time-Out Policies

In the dynamic power management (DPM), it is quite important to switch a high-power consuming state to a low-power consuming one at a suitable timing. This problem has been formulated as Markov decision processes (MDPs) with state-dependent control in the past literature. However, this approach may not be often feasible in many practical situations, because the state-dependent policy requires that all the states on a request arrival process must be observed through an online monitoring. To overcome this problem, we develop a simple time-out policy in the DPM, which can be regarded as the optimal timing to take a GO-SLEEP action during an idle period of the transaction system. We derive the optimal time-out policy minimizing the expected power consumption per unit time in the steady state analytically under the assumption that the request arrival process is given by a Markovian arrival process with an arbitrary number of phases. In numerical experiments with real read/write data for a hard disk unit and CPU utilization data, we estimate the expected power consumption per unit time under two DPM policies, namely, the time-out policy and the MDP-based policy, and quantitatively compare their effectiveness on the power reduction.

On preemptive-repeat LIFO queues

In this paper, we study the basic properties of last-in first-out (LIFO) preemptive-repeat single-server queues in which the server needs to start service from scratch whenever a preempted customer reaches the server. In particular, we study the question of when such queues are stable (in the sense that the equilibrium time-in-system is finite-valued with probability one) and show how moments of the equilibrium customer sojourn time can be computed when the system is stable. A complete analysis of stability is provided in the setting of Poisson arrivals and in that of the Markovian arrival process. The stability region depends upon the detailed structure of the interarrival and service time distributions and cannot be expressed purely in terms of expected values. This is connected to the fact that such preemptive-repeat queues are not work conserving.

Stochastic gene expression conditioned on large deviations

The intrinsic stochasticity of gene expression can give rise to large fluctuations and rare events that drive phenotypic variation in a population of genetically identical cells. Characterizing the fluctuations that give rise to such rare events motivates the analysis of large deviations in stochastic models of gene expression. Recent developments in non-equilibrium statistical mechanics have led to a framework for analyzing Markovian processes conditioned on rare events and for representing such processes by conditioning-free driven Markovian processes. We use this framework, in combination with approaches based on queueing theory, to analyze a general class of stochastic models of gene expression. Modeling gene expression as a Batch Markovian Arrival Process (BMAP), we derive exact analytical results quantifying large deviations of time-integrated random variables such as promoter activity fluctuations. We find that the conditioning-free driven process can also be represented by a BMAP that has the same form as the original process, but with renormalized parameters. The results obtained can be used to quantify the likelihood of large deviations, to characterize system fluctuations conditional on rare events and to identify combinations of model parameters that can give rise to dynamical phase transitions in system dynamics.

Performance modeling of QoS differentiation in optical packet switching via FDL access limitation

We present an exact analytical model for single-wavelength quality of service (QoS) differentiation in a two-class optical packet switch. In this system, QoS differentiation is achieved by limiting the set of fiber delay lines (FDLs) to the low-priority class, whereas the high-priority class is allowed to access the entire FDL bank. The analytical model is based on multi-regime Markov fluid queues and is extensible to multi-class systems with more than two classes. Markovian arrival process packet arrivals and phase-type distributed packet sizes are considered for the purpose of generality. The proposed analytical model is validated through simulations. The numerical results provide insight into determining appropriate subsets of FDLs allowed for the access of the low-priority class. The results also show that it is possible to direct almost all the packet losses to the low-priority class under moderate loads.

Novel Queueing Model for Multimedia Over Downlink in 3.5G Wireless Network

In this paper, a model for multimedia transmission over downlink shared channel in 3.5G wireless network is presented. The multimedia stream consists of multiplesubstreams that are aggregated into one real-time and onenonreal-time flows. Correlation with each flow and between flows is assumed. Additionally, we propose a combined time-space priority buffer management scheme to optimise quality of service requirements for each flow. The problem is formulated in terms of a queue with two priority classes, one of which has time priority while the another has space priority. The input is described by the Batch Marked Markovian Arrival Process (BMMAP). Service time distributions are of PH (phase) type dependent on the class of a customer. The buffer is finite, but the customers of a class having higher priority for taking into the service from a buffer (time priority) can occupy only a part of this buffer. Queueing system's behavior is described in terms ofmulti-dimensional continuous time skip-free to the left Markov chain. It allows to exploit an effective algorithm for calculation of the stationary distribution of the queueing system. Loss probability for customers of both classes is calculated. Waiting time distribution for priority customers is calculated.

On the Queue Length Distribution in BMAP Systems

Batch Markovian Arrival Process – BMAP – is a teletraffic model which combines high ability to imitate complexstatistical behaviour of network traces with relative simplicity in analysis and simulation. It is also a generalization of a wide class of Markovian processes, a class which in particular include the Poisson process, the compound Poisson process, the Markovmodulated Poisson process, the phase-type renewal process and others. In this paper we study the main queueing performance characteristic of a finite-buffer queue fed by the BMAP, namely the queue length distribution. In particular, we show a formula for the Laplace transform of the queue length distribution. The main benefit of this formula is that it may be used to obtain both transient and stationary characteristics. To demonstrate this, several numerical results are presented.

Analysis of an infinite-buffer batch-size-dependent service queue with Markovian arrival process

This paper considers an infinite-buffer batch-service queue with Markovian arrival process, generally distributed and batch-size-dependent service time. We obtain a bivariate vector generating function of queue length and server content distribution at departure epoch of a batch. The complete joint distribution of queue length, server content and phase of the arrival process at departure epoch is extracted in terms of roots of the associated characteristic equation. By employing these probability vectors we also perceive the joint distribution at arbitrary and pre-arrival epochs. Our analytic procedure and results are demonstrated using some numerical examples for phase-type as well as deterministic service time distributions with high and low values of model parameters. The occurrence of multiple roots are also investigated in case of phase-type service time distribution. Finally, we also investigate the influence of correlation of the arrival process on the behavior of the system performance.

Analysis of an MAP/PH/1 Queue with Flexible Group Service

AbstractA novel customer batch service discipline for a single server queue is introduced and analyzed. Service to customers is offered in batches of a certain size. If the number of customers in the system at the service completion moment is less than this size, the server does not start the next service until the number of customers in the system reaches this size or a random limitation of the idle time of the server expires, whichever occurs first. Customers arrive according to a Markovian arrival process. An individual customer’s service time has a phase-type distribution. The service time of a batch is defined as the maximum of the individual service times of the customers which form the batch. The dynamics of such a system are described by a multi-dimensional Markov chain. An ergodicity condition for this Markov chain is derived, a stationary probability distribution of the states is computed, and formulas for the main performance measures of the system are provided. The Laplace–Stieltjes transform of the waiting time is obtained. Results are numerically illustrated.

Gerber–Shiu analysis with two-sided acceptable levels

In this paper, insurer’s surplus process moved within upper and lower levels is analyzed. To this end, a truncated type of Gerber–Shiu function is proposed by further incorporating the minimum and the maximum surplus before ruin into the existing ones (e.g. Gerber and Shiu (1998), Cheung et al. (2010a)). A key component in our analysis of this proposed Gerber–Shiu function is the so-called transition kernel. Explicit expressions of the transition function under two different risk models are obtained. These two models are both generalizations of the classical Poisson risk model: (i) the first model provides flexibility in the net premium rate which is dependent on the surplus (such as linear or step function); and (ii) the second model assumes that claims arrive according to a Markovian arrival process (MAP). Finally, we discuss some applications of the truncated Gerber–Shiu function with numerical examples under various scenarios.

일량분석에 의한 이산시간 MAP/G/1 대기행렬시스템의 통합적 분석

본 논문에서는 이산시간 마코비안 도착과정(discrete-time Markovian arrival process, 이하 이산시간 MAP)을 갖는 대기행렬시스템의 주요성능척도들을 분석하기 위한 통합적 접근방법을 제시한다. 기존의 이산시간 MAP/G/1 대기행렬시스템의 연구들을 보면 동일한 시스템에 대하여 시스템 내 고객수와 대기시간 등을 분석할 때 서로 다른 방법으로 접근하였기에 이 둘을 동시에 살펴보고자 할 때는 추가적인 시간과 노력이 뒤따랐다. 따라서 하나의 시스템을 여러 방면에서 포괄적으로 분석할 수 있는 통합적인 접근방법은 시스템을 설계하고 관리하는 입장에서 볼 때 중요한 분석의 틀이 된다. 본 논문에서는 이산시간 MAP/G/1 시스템의 안정상태 일량 분포를 유도하고 이를 이용하여 임의고객의 대기시간, 체재시간 분포를 유도한다. 체재시간 분포로부터 이탈시점 고객수 분포를 구하고 이탈시점 고객수와 임의시점 고객수와의 관계로부터 고객수 분포를 유도한다. In this paper, We suggest a unified approach for the analysis of discrete-time MAP/G/1 queueing system. Many researches on the D-MAP/G/1 queue have been used different approach to analyze system queue length and waiting time for the same system. Therefore, a unified framework for analyzing a system is necessary from a viewpoint of system design and management. We first derived steady-state workload distribution, and then waiting time and sojourn time are derived by the result of workload analysis. Finally, system queue length distribution is derived with generating function from the sojourn time distribution.

Joint Insolvency Analysis of a Shared MAP Risk Process: A Capital Allocation Application

In recent years, multivariate insurance risk processes have received increasing attention in risk theory. First-passage-time problems in the context of these insurance risk processes are of primary interest for risk management purposes. In this article we study joint-ruin problems of two risk undertakers in a proportionally shared Markovian claim arrival process. Building on the existing work in the literature, joint-ruin–related quantities are thoroughly analyzed by capitalizing on existing results in certain univariate insurance surplus processes. Finally, an application is considered where the finite-time and infinite-time joint-ruin probabilities are used as risk measures to allocate risk capital among different business lines. The proposed joint-ruin allocation principle enables us to not only capture the risk dynamics over a given time horizon, but also overcome the “cross-subsidizing” effect of many existing allocation principles.

On the MAP/PH/c retrial queueing system with two types of nonpersistent calls

In this paper, we establish a performance model of a cell in a cellular mobile wireless network with two types of calls arrivals: originating calls and handoff calls. Using the Markovian Arrival Processes to describe the calls arrival processes and PH type distributions to model the calls service times and the cell-sojourn times of any ongoing calls, we investigate the impact of retrial and nonpersistent phenomenon on the performance measures of the system. The algorithms for calculating the stationary state probabilities of the model and determining the optimal number of guard channels and total channels are elaborated. Main performance measures are obtained. Finally, we provide numerical results to illustrate the optimization problems and the impact of retrial and nonpersistent phenomena on performance measures. We find that the performance measures are mainly affected by the two types of calls’ arrivals and service patterns, but the retrial rate plays a little crucial role.

Efficient Computational Analysis of Stationary Probabilities for the Queueing System BMAP/G/1/N With or Without Vacation(s)

We consider a finite-buffer single-server queue with batch Markovian arrival process. In the case of finite-buffer batch arrival queue, there are different customer rejection/acceptance strategies such as partial batch rejection, total batch rejection, and total batch acceptance policy. We consider partial batch rejection strategy throughout our paper. We obtain queue length distributions at various epochs such as pre-arrival, arbitrary, and post-departure as well as some important performance measures, like probability of loss for the first, an arbitrary, and the last customer of a batch, mean queue lengths, and mean waiting times. The corresponding queueing model under single and multiple vacation policy has also been investigated. Some numerical results have been presented in the form of tables by considering phase-type and Pareto service time distributions. The proposed analysis is based on the successive substitutions in the Markov chain equations of the queue-length distribution at an embedded post-departure epoch of a customer. We also establish relationships among the queue-length distributions at post-departure, arbitrary, and pre-arrival epochs using the classical argument based on Markov renewal theory and semi-Markov processes. Such queueing systems find applications in the performance evaluation of teletraffic part in 4G A11-IP networks.