Probabilistic Model Checking has had successful applications in
various areas, including the verification of embedded protocol
software, real-time systems and even biological systems. However,
many interesting models have a statespace which is too large to fit
in memory of todays computers or are even infinite. Predicate
abstraction is a technique with which many nondeterministic models
containing infinite statespace where successfully handled. There has
also been research on predicate abstraction for probabilistic
systems. Former results had the restriction that for predicate
abstraction to be used, the whole statespace of the models had to be
generated, which made this technique was still infeasible in many
cases. In the talk I am going to hold, I introduce PASS, which is a
software tool allowing to do predicate abstraction of systems
containing probabilities without the need of generating the
statespace of the original model beforehand.