Controller dependability analysis by probabilistic model checking

This paper demonstrates how probabilistic model checking, a formal verification method for the analysis of systems which exhibit stochastic behaviour, can be applied to the study of dependability properties of software-based control systems. By using existing formalisms and tool support from this area, it is possible to construct large and complex Markov models from an intuitive high-level description and to take advantage of the efficient implementation techniques which have been developed for these tools. This paper provides an overview of probabilistic model checking and of the tool PRISM which supports these techniques. It illustrates the applicability of the approach through the use of a case study and demonstrates that a wide range of useful dependability properties can be analysed in this way. (c) 2006 Elsevier Ltd. All rights reserved.