Performance evaluation and analysis of a discrete queue system with multiple working vacations and non-preemptive priority

Abstract

In this paper, we introduce a discrete time Geo/Geo/1 queue sys-tem with non-preemptive priority and multiple working vacations. We assume that there are two types of customers in this queue system named "Customers of type-I" and "Customers of type-II". Customer of type-II has a higher prior-ity with non-preemption than Customer of type-I. By building a discrete time four-dimensional Markov Chain which includes the numbers of customers with different priorities in the system, the state of the server and the service state, we obtain the state transition probability matrix. Using a birth-and-death chain and matrix-geometric method, we deduce the average queue length, the average waiting time of the two types of customers, and the average busy pe-riod of the system. Then, we provide some numerical results to evaluate the effect of the parameters on the system performance. Finally, we develop some benefit functions to analyse both the personal and social benefit, and obtain some optimization results within a certain range.