Sensitivity and importance analysis of Markov models using perturbation analysis: Applications in reliability studies

Phuc Do Van (1) , Anne Barros (1) , Christophe Bérenguer (1)

1 LM2S - Laboratoire Modélisation et Sûreté des Systèmes

Sensitivity (or importance analysis) has been first defined for "static systems", i.e. systems described by combinatorial reliability models (fault or event trees) and several measures, both structural and probabilistic, have been proposed to assess component importance. For dynamic systems including inter-component and functional dependencies (cold spare, shared load, shared resources, ....), and described by Markov models or, more generally, by discrete events dynamic systems models (DEDS), the problem of sensitivity analysis remains widely open. In this paper we propose to use the estimation method developed by Cao in (Cao & Chen 1997) in the framework of Perturbation Analysis, to formalize several sensitivity measures in case of dynamic systems. We show with numerical examples why this method offers a promising tool for steady state sensitivity analysis of Markov Processes in reliability studies.

Mots clés: Sensitivity analysis; perturbation analysis; markov model

Domaines: Automatique / Robotique

Phuc Do Van, Anne Barros, Christophe Bérenguer. Sensitivity and importance analysis of Markov models using perturbation analysis: Applications in reliability studies. European Safety and Reliability Conference (ESREL 2006), Sep 2006, Estoril, Portugal. pp.1769-1775. ⟨hal-00705234⟩