Analysis of a finite-source customer assignment model with no state information

In this paper we analyse a closed queueing network in which customers have to be assigned to parallel queues. The routing decision may not depend on the numbers of customers in the queues. We present an algorithm and we show that it computes an average optimal policy in case of exponential service times. The algorithm also works for non-exponential service times, in which case periodic policies are found.

[1]  Arie Hordijk,et al.  Undiscounted Markov decision chains with partial information; an algorithm for computing a locally optimal periodic policy , 1994, Math. Methods Oper. Res..

[2]  P. Moran,et al.  Reversibility and Stochastic Networks , 1980 .

[3]  A. Hordijk,et al.  Optimal static customer routing in a closed queuing network , 2000 .

[4]  Arie Hordijk,et al.  Insensitivity for Stochastic Networks , 1983, Computer Performance and Reliability.

[5]  Richard F. Serfozo,et al.  Markovian network processes: Congestion-dependent routing and processing , 1989, Queueing Syst. Theory Appl..

[6]  Ger Koole,et al.  Analysis of a Customer Assignment Model with No State Information , 1994, Probability in the Engineering and Informational Sciences.

[7]  Nico M. van Dijk Queueing networks and product forms - a systems approach , 1993, Wiley-Interscience series in systems and optimization.