Time for Statistical Model Checking of Real-Time Systems

Abstract

Abstract. We propose the first tool for solving complex (some undecidable) problems of timed systems by using Statistical Model Checking (SMC). The tool monitors several runs of the system, and then relies on statistical algorithms to get an estimate of the correctness of the entire design. Contrary to other existing toolsets, ours relies on i) a natural stochastic semantics for networks of timed systems, ii) an engine capable to solve problems that are beyond the scope of classical model checkers, and iii) a friendly user interface. ContextTimed model checking (TMC) is a technique used to prove the absence of bugs in systems whose behaviors depend on real or discrete time constraints. The approach has been implemented in several tools [4,2,5] capable of handling case studies of industrial size. Unfortunately, many applications are still out of scope of TMC. This is due to the complexity of the timed behaviors, which can even make the problem undecidable.In a recent work [8], we presented Constant Slope Timed Automata (CSTA), that are timed systems in that clocks may have different rates (even potentially negative) in different locations. Such automata are as expressive as linear hybrid automata or priced timed automata, but the addition of features such as input and output modalities allows us to specify complex problems in an elegant manner. Unfortunately most of such problems are either undecidable or too complex to be solved with classical model checking approaches. In [8], we proposed to estimate undecidable problems by using Statistical Model Checking (SMC) [13,9]. SMC consists of monitoring some runs of the system and then uses a statistical algorithm to obtain an estimate for the system. Such simulation-based techniques were applied in other contexts where they outperformed classical model checking techniques with an order of magnitude [13,14,1].

As Soon as Probable: Optimal Scheduling under Stochastic Uncertainty

“…Implementation and experimentation: we currently have a prototype implementation building upon the infra-structure developed in [12] for computing integrals over zones. Once completed, we intend to compare cost/quality tradeoffs of our algorithm with statistical methods such as [8,13] that evaluate and synthesize schedulers based on sampling the duration space. Such experiments will determine whether symbolic techniques can be part of tools for design-space exploration [13].…”

Semantic Importance Sampling for Statistical Model Checking

“…More recently, an estimation approach [2] has been taken in which the goal is to estimate the probability that a system property holds. SMC has been applied to a wide variety of systems, such as stochastic hybrid automata [4], and real time systems [5]. In addition to importance sampling, SMC can also be improved by a method known as "importance splitting" [10].…”

Statistical Model Checking of Simulink Models with Plasma Lab

“…Several model-checking tools have added SMC as a complement to exhaustive model-checking. This includes the model-checker UPPAAL [5] for timed automata, the probabilistic model-checker PRISM [7], and the model-checker Ymer [9] for continuous time Markov chains. Plasma Lab [3] is the first platform entirely dedicated to SMC.…”