Safe On-The-Fly Steady-State Detection for Time-Bounded Reachability

The time-bounded reachability problem for continuoustime Markov chains (CTMCs) amounts to determine the probability to reach a (set of) goal state(s) within a given time span, such that prior to reaching the goal certain states are avoided. Efficient algorithms for time-bounded reachability are at the heart of probabilistic model checkers such as PRISM and ETMCC. For large time spans, on-the-fly steady-state detection is commonly applied. To obtain correct results (up to a given accuracy), it is essential to avoid detecting premature stationarity. This paper gives a detailed account of criteria for steady-state detection in the setting of time-bounded reachability. This is done for forward and backward reachability algorithms. As a spin-off of this study, new results for on-the-fly steady-state detection during CTMC transient analysis are reported. Based on these results, a precise procedure for steady-state detection for time-bounded reachability is obtained. Experiments show the impact of these results in probabilistic model checking.