Importance Sampling for Stochastic Timed Automata

We present an importance sampling framework that combines symbolic analysis and simulation to estimate the probability of rare reachability properties in stochastic timed automata. By means of symbolic exploration, our framework first identifies states that cannot reach the goal. A state-wise change of measure is then applied on-thefly during simulations, ensuring that dead ends are never reached. The change of measure is guaranteed by construction to reduce the variance of the estimator with respect to crude Monte Carlo, while experimental results demonstrate that we can achieve substantial computational gains.