Keynote Speaker: Patricia Bouyer-Decitre
Title: On the (Approximate) Analysis of Stochastic Real-Time Systems
Abstract
Analyzing stochastic systems with real-time features is notably difficult, and often calls for approximation technics. In this talk I will present a crucial property of stochastic systems, which allows for provably sound approximated model-checking technics. This property, called decisiveness, was first proposed by Abdulla et al in 2007 for infinite-state discrete Markov chains. This was later extended to general stochastic transition systems, and general approximation schemes were proven correct for large classes of properties. Finally it turns out that several classes of real-time stochastic systems that have been considered in the literature do actually satisfy the property, which implies that they can safely be analysed. Finally, I plan to discuss ongoing work on statistical simulation. This talk is based on joint works with Nathalie Bertrand, Thomas Brihaye, Pierre Carlier, Mickael Randour, Cedric Riviere and Pierre Vandenhove.