On the Analysis of Exponential Queuing Systems with Randomly Changing Arrival Rates: Stability Conditions and Finite Buffer Scheme with a Resume Level

Abstract This paper considers a single server exponential queue with random fluctuations in the intensity of the arrival process. The motivation being the modelling of random changes in traffic patterns. This random intensity model does not obey the independence assumption made in queuing theory. Necessary and sufficient conditions for the stability or ergodicity of the queuing process are obtained via analytic techniques using Jury's stability criteria, often used to study discrete time control systems. The effect of such fluctuations is then studied for a finite resume level queue which is often used in flow control. Exact performance measures are computed and are compared with existing results.