New Results on the Differential Importance Measures of Markovian Systems

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

1 LM2S - Laboratoire Modélisation et Sûreté des Systèmes

This paper presents the development of the differential importance measures (DIM), proposed recently for the use in risk-informed decision-marking, in the context of dynamic systems including inter-component, functional dependencies, or more generally, systems described by Markov models at steady state. Several new evaluation formulas of the existing DIMs are provided. The study also proposes an extension one that provides better results than those obtained from these existing measures, and a new calculation formula of the total variation of the system availability provoked by the simultaneous changes of system parameters. A numerical example is introduced to compare the extension proposed and these existing measures. The suggestions for the future research in this area are also presented.

Mots clés: Sensitivity analysis; Differential importance measures; Markov models

Domaines: Automatique / Robotique

Phuc Do Van, Anne Barros, Christophe Bérenguer. New Results on the Differential Importance Measures of Markovian Systems. Ninth International Probabilistic Safety Assessement and Management Conference (PSAM 9), May 2008, hong kong, China. 8p. ⟨hal-00705241⟩