Statistical Verification of Probabilistic Properties with Unbounded Until
We consider statistical (sampling-based) solution methods for verifying probabilistic properties with unbounded until. Statistical solution methods for probabilistic verification use sample execution trajectories for a system to verify properties with some level of confidence. The main challenge with properties that are expressed using unbounded until is to ensure termination in the face of potentially infinite sample execution trajectories. We describe two alternative solution methods, each one with its own merits. The first method relies on reachability analysis, and is suitable primarily for large Markov chains where reachability analysis can be performed efficiently using symbolic data structures, but for which numerical probability computations are expensive. The second method employs a termination probability and weighted sampling. This method does not rely on any specific structure of the model, but error control is more challenging. We show how the choice of termination probability—when applied to Markov chains—is tied to the subdominant eigenvalue of the transition probability matrix, which relates it to iterative numerical solution techniques for the same problem.
KeywordsMarkov Chain Model Check Probabilistic Property Reachability Analysis Symbolic Model Check
Unable to display preview. Download preview PDF.
- 6.Clarke, E.M., Grumberg, O., Peled, D.A.: Model Checking. The MIT Press, Cambridge (1999)Google Scholar
- 8.Etessami, K., Rajamani, S.K. (eds.): CAV 2005. LNCS, vol. 3576. Springer, Heidelberg (2005)Google Scholar
- 14.Hermanns, H., Meyer-Kayser, J., Siegle, M.: Multi terminal binary decision diagrams to represent and analyse continuous time Markov chains. In: Proc. 3rd International Workshop on the Numerical Solution of Markov Chains, pp. 188–207, Prensas Universitarias de Zaragoza (1999)Google Scholar
- 20.Monniaux, D.: An abstract monte-carlo method for the analysis of probabilistic programs. In: Proc. 28th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, pp. 93–101. Association for Computing Machinery (2001)Google Scholar
- 27.Zapreev, I.S.: Model checking Markov chains: Techniques and tools. PhD thesis, University of Twente (2008)Google Scholar