Approximations for Non-Stationary Queues

Professor Peter Glynn

ABSTRACT

Traditional queueing theory typically assumes that the exogenous traffic flows into the system are stationary stochastic processes. However, most real-world queueing environments exhibit non-stationarities due to time-of-day effects, day-of-week effects, seasonalities, or marco-economic effects. In this talk, we will discuss both modeling issues that arise in trying to capture non stationarities and associated approximations for non-stationary queues.