A Progress Measure for Explicit-State Probabilistic Model-Checkers

Verification of the source code of a probabilistic system by means of an explicit-state model-checker is challenging. In most cases, the model-checker will either run out of memory or will simply not terminate within any reasonable amount of time. We introduce the notion of a progress measure for such a model-checker. The progress measure returns a number in the interval [0, 1]. This number provides us a quantitative measure of the amount of progress the model-checker has made verifying a particular linear time property. The larger the number, the more progress the model-checker has made. We also show how to compute the progress measure for checking invariants. Explicit-state model-checkers usually exploit search strategies such as depth-first search and breadth-first search to explore the transitions. We introduce several new search strategies that take the probabilities associated with the transitions into account. We compare the amount of progress made by the different search strategies.