Continuous Stochastic Logic (CSL) can be interpreted over continuous-time Markov decision processes (CTMDPs) to specify quantitative properties of stochastic systems that allow some external control. Model checking CSL formulae over CTMDPs requires then the computation of optimal control strategies to prove or disprove a formula. The paper presents a conservative extension of CSL over CTMDPs—with rewards—and exploits established results for CTMDPs for model checking CSL. A new numerical approach based on uniformization is devised to compute time bounded reachability results for time dependent control strategies. Experimental evidence is given showing the efficiency of the approach.
|Title of host publication||Computer Aided Verification - 23rd International Conference, CAV 2011, Snowbird, UT, USA, July 14-20, 2011. Proceedings|
|Number of pages||18|
|Publication status||Published - 2011|