Article ID: | iaor1992133 |
Country: | United States |
Volume: | 38 |
Issue: | 6 |
Start Page Number: | 1052 |
End Page Number: | 1064 |
Publication Date: | Nov 1990 |
Journal: | Operations Research |
Authors: | Harrison J. Michael, Wein Lawrence M. |
Keywords: | queues: theory |
The authors consider a multiclass closed queueing network with two single-server stations. Each class requires service at a particular station, and customers change class after service according to specified probabilities. There is a general service time distribution for each class. The problem is to schedule the two servers to maximize the long-run average throughput of the network. By assuming a large customer population and nearly balanced loading of the two stations, the scheduling problem can be approximated by a dynamic control problem involving Brownian motion. A reformulation of this control problem is solved exactly and the solution is interpreted in terms of the queueing network to obtain a scheduling rule. (The authors conjecture, quite naturally, that the resulting scheduling rule is asymptotically optimal under