Probabilistic Model Checking for Dependability Properties of Stochastic Systems

Nawal Addouche (1) , Christian Antoine (1) , Jacky Montmain (1)

1 LGI2P - Laboratoire de Génie Informatique et Ingénierie de Production

The paper proposes a dependability analysis method based on probabilistic model checking. Using the profile DAMRTS (Dependability Analysis Models for Real-Time Systems), stochastic real-time systems are modelled with stochastic and probabilistic information. In this profile, static model of the system includes depen-dability information excerpted from fault trees. Behavioural UML models are given with combined collaboration statecharts diagrams. A method of translating these models to continuous time Markov chains (CTMCs) is proposed. The CTMCs are widely used in the context of performance and reli-ability evaluation of various systems. A formal approach is proposed to verify temporal probabilistic proper-ties related to dependability of stochastic systems. It consists to use a probabilistic model checker which sup-ports the CTMC models and then to specify properties with the suitable temporal logic.

Domaines: Automatique; Intelligence artificielle [cs.AI]

Nawal Addouche, Christian Antoine, Jacky Montmain. Probabilistic Model Checking for Dependability Properties of Stochastic Systems. European Safety and Reliability Conference (ESREL 05), 2005, Tri City, Poland. ⟨hal-00354029⟩