Near-Optimal Switching Strategies for a Tandem Queue

Motivated by various applications in logistics, road traffic and production management, we investigate two versions of a tandem queueing model in which the service rate of the first queue can be controlled. The objective is to keep the mean number of jobs in the second queue as low as possible, without compromising the total system delay (i.e. avoiding starvation of the second queue). The balance between these objectives is governed by a linear cost function of the queue lengths. In the first model, the server in the first queue can be either switched on or off, depending on the queue lengths of both queues. This model has been studied extensively in the literature. Obtaining the optimal control is known to be computationally intensive and time consuming. We are particularly interested in the scenario that the first queue can operate at larger service speed than the second queue. This scenario has received less attention in literature. We propose an approximation using an efficient mathematical analysis of a near-optimal threshold policy based on a matrix-geometric solution of the stationary probabilities that enables us to compute the relevant stationary measures more efficiently and determine an optimal choice for the threshold value.

[1]  Marcel F. Neuts,et al.  Efficient Algorithmic Solutions to Exponential Tandem Queues with Blocking , 1980, SIAM J. Algebraic Discret. Methods.

[2]  Howard J. Weiss.  The Computation of Optimal Control Limits for a Queue with Batch Services , 1979 .

[3]  Vaidyanathan Ramaswami,et al.  A logarithmic reduction algorithm for quasi-birth-death processes , 1993, Journal of Applied Probability.

[4]  Sean R Eddy,et al.  What is dynamic programming? , 2004, Nature Biotechnology.

[5]  Martin L. Puterman,et al.  Markov Decision Processes: Discrete Stochastic Dynamic Programming , 1994 .

[6]  Marcel F. Neuts,et al.  Matrix-geometric solutions in stochastic models - an algorithmic approach , 1982 .

[7]  M. Neuts.  A General Class of Bulk Queues with Poisson Input , 1967 .

[8]  Steven A. Lippman,et al.  Applying a New Device in the Optimization of Exponential Queuing Systems , 1975, Oper. Res..

[9]  Richard M. Feldman,et al.  An M/M/1 queue with a general bulk service rule , 1985 .

[10]  Ad Ridder.  Large-deviations analysis of the fluid approximation for a controllable tandem queue , 2003 .

[11]  Zvi Rosberg,et al.  Optimal control of service in tandem queues , 1982 .

[12]  F Avram.  Optimal Control of Fluid Limits of Queuing Networks and Stochasticity Corrections, Lectures in Applied Mathematics , 1997 .

[13]  M. Kwiatkowska,et al.  Solving Infinite Stochastic Process Algebra Models Through Matrix-Geometric Methods , 1999 .

[14]  Sean P. Meyn.  Control Techniques for Complex Networks: Workload , 2007 .

[15]  Ger Koole.  Convexity in tandem queues , 1999 .

[16]  R. Weber,et al.  Optimal control of service rates in networks of queues , 1987, Advances in Applied Probability.