Importance sampling simulation of queues with time-varying rates

Abstract: 

In this talk we consider rare events in Markovian queues with time-varying arrival rates (nonhomogeneous Poisson arrivals) and time-varying service rates. These rare events comprise level crossings at fixed time epochs, and level crossings of the maximum in a busy cycle. We apply importance sampling simulation for estimating efficiently the rare-event probabilities. We discuss some rare-event analysis, and the complexity analysis of the proposed importance sampling algorithms. The analysis is illustrated by numerical results.

The seminar is followed by wine and finger food in the staff room RC-3082. All attendees are welcome!