An efficient exact procedure for the analysis of the state-dependent polling model

Polling models are used to model and analyze systems in which a single server attends to multiple queues at which requests arrive for service. State-Dependent (SD) polling models form a class of polling models in which the server performs a setup at a polled queue only if that queue is non-empty. SD polling models facilitate the analysis of many real-world systems. For example, recent developments in the field of radio frequency data communication, with its application to warehouse management and material handling systems, will provide a significant impetus to research on SD polling models. Exact analysis of the SD polling model is, however, inherently complex. The lack of an efficient exact procedure to analyze the SD polling model has hitherto limited its application, with real-world systems typically being modeled by the simpler state-independent polling model. Naturally, that results in a loss of modeling power and provides questionable values for desired performance measures. In this paper, we present an efficient exact analytic technique to compute the desired performance measures for the SD polling model.

Acknowledgments

Contributed by the Engineering, Statistics and Applied Probability Department