Dynamic load balancing in parallel queueing systems: Stability and optimal control

We consider a system of parallel queues with dedicated arrival streams. At each decision epoch a decision-maker can move customers from one queue to another. The cost for moving customers consists of a fixed cost and a linear, variable cost dependent on the number of customers moved. There are also linear holding costs that may depend on the queue in which customers are stored. Under very mild assumptions, we develop stability (and instability) conditions for this system via a fluid model. Under the assumption of stability, we consider minimizing the long-run average cost. In the case of two-servers the optimal control policy is shown to prefer to store customers in the lowest cost queue. When the inter-arrival and service times are assumed to be exponential, we use a Markov decision process formulation to show that for a fixed number of customers in the system, there exists a level S such that whenever customers are moved from the high cost queue to the low cost queue, the number of customers moved brings the number of customers in the low cost queue to S. These results lead to the development of a heuristic for the model with more than two servers.

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

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

[3]  J. H. van Schuppen,et al.  Distributed routing for load balancing , 1989, Proc. IEEE.

[4]  J. Dai.  On Positive Harris Recurrence of Multiclass Queueing Networks: A Unified Approach Via Fluid Limit Models , 1995 .

[5]  Sean P. Meyn.  The policy iteration algorithm for average reward Markov decision processes with general state space , 1997, IEEE Trans. Autom. Control..

[6]  Sean P. Meyn.  The Policy Improvement Algorithm for Markov Decision Processes , 1997 .

[7]  Zhen Liu,et al.  Optimal Load Balancing on Distributed Homogeneous Unreliable Processors , 1998, Oper. Res..

[8]  Ali R. Hurson,et al.  Scheduling and Load Balancing in Parallel and Distributed Systems , 1995 .

[9]  Christos G. Cassandras,et al.  Optimal routing and buffer allocation for a class of finite capacity queueing systems , 1992 .

[10]  Ger Koole,et al.  On the Pathwise Optimal Bernoulli Routing Policy for Homogeneous Parallel Servers , 1996, Math. Oper. Res..

[11]  L. Sennott.  Stochastic Dynamic Programming and the Control of Queueing Systems , 1998 .

[12]  Mark H. A. Davis.  Piecewise‐Deterministic Markov Processes: A General Class of Non‐Diffusion Stochastic Models , 1984 .

[13]  David Finkel,et al.  Load balancing in a multi-server queuing system , 1990, Comput. Oper. Res..

[14]  Yung-Terng Wang,et al.  Load Sharing in Distributed Systems , 1985, IEEE Transactions on Computers.

[15]  Mark H. A. Davis.  Piecewise-deterministic Markov processes , 1993 .

[16]  Nicholas Bambos,et al.  Dynamic on-line task scheduling on parallel processors , 2001, Perform. Evaluation.

[17]  Debasish Ghose,et al.  Scheduling Divisible Loads in Parallel and Distributed Systems , 1996 .

[18]  M. Lewis.  Average optimal policies in a controlled queueing system with dual admission control , 2001, Journal of Applied Probability.

[19]  Edward D. Lazowska,et al.  Adaptive load sharing in homogeneous distributed systems , 1986, IEEE Transactions on Software Engineering.

[20]  Marcel F. Neuts,et al.  Two M/M/1 Queues with Transfers of Customers , 2002, Queueing Syst. Theory Appl..

[21]  Sean P. Meyn,et al.  Stability and convergence of moments for multiclass queueing networks via fluid limit models , 1995, IEEE Trans. Autom. Control..