Strongly asymptotically optimal design and control of production and service systems

In this paper we consider production and service systems that can be modeled as single or multiple stage queueing networks. We provide a formal definition of strong asymptotic optimality in the context of design and control of such queueing systems. We describe a simple approach to obtain strongly asymptotically optimal design and control policies for these systems. We illustrate our approach through some examples. In particular we obtain a strongly asymptotically optimal workload allocation for a multiple center service system modeled by the open Jackson network. The objective here is to minimize the expected total holding cost. A simple counter example shows that the much celebrated balanced workload allocation is not even asymptotically optimal for minimizing the expected number of jobs in the system. We show that an index policy is strongly asymptotically optimal for the scheduling control problem in a single stage (GI/G/1) service system. We also obtain a strongly asymptotically optimal allocation of classes of customers to a multiple center service system. This allocation agrees with the traditional wisdom of forming service stations to process similar tasks that reduce fluctuations in processing times. Suggestions for further work on this topic are summarized as well.