Preliminary System Safety Analysis with Limited Markov Chain Generation
Résumé
Markov chains are a powerful and versatile tool to calculate reliability indicators. However, their use is limited for two reasons: the exponential blow-up of the size of the model, and the di culty to design models. To overcome this second di culty, a solution consists in generating automatically the Markov chain from a higher level description, e.g. a stochastic Petri net or an AltaRica model. These higher level models describe the Markov chain implicitly. In this article, we propose an algorithm to generate partial Markov chains. The idea is to accept a little loss of accuracy in order to limit the size of the generated chain. The cornerstone of this method is a Relevance Factor associated to each state of the chain. This factor enables the selection of the most representative states. We show on an already published test case, that our method provides very accurate results while reducing dramatically the complexity of the assessment. It is worth noticing that the proposed method can be used with different high-level modeling formalisms.
Domaines
Automatique / Robotique
Origine : Fichiers produits par l'(les) auteur(s)
Loading...