Large Deviations Analysis of Queueing Networks

Kavita Ramanan Department of Mathematical Sciences Carnegie Mellon University United States of America h
Tuesday, 2 Jun 2009 (all day)
A-212 (STCS Seminar Room)
Many stochastic systems are governed by events that, though they have a small probability of occurrence, are crucial to performance.
Large deviations is an asymptotic theory that allows
for the estimation of the probabilities of such rare events.
Motivated by concrete examples arising in wireless communications
and computer systems that cannot be analyzed using classical
large deviations theory, we describe several new tools that we
developed for large deviations analysis.