Approximate Parameter Synthesis for Probabilistic Time-Bounded Reachability

This paper proposes a technique to synthesize parametric rate values in continuous-time Markov chains that ensure the validity of bounded reachability properties. Rate expressions over variables indicate the average speed of state changes and are expressed using the polynomials over reals. The key contribution is an algorithm that approximates the set of parameter values for which the stochastic real-time system guarantees the validity of bounded reachability properties. This algorithm is based on discretizing parameter ranges together with a refinement technique. This paper describes the algorithm, analyzes its time complexity, and shows its applicability by deriving parameter constraints for a real-time storage system with probabilistic error checking facilities.