Point process-based approaches for the reliability analysis of systems modeled by costly simulators

Guillaume Perrin (1, 2)

1 GdR MASCOT-NUM - Méthodes d'Analyse Stochastique des Codes et Traitements Numériques
2 COSYS-LISIS - Laboratoire Instrumentation, Simulation et Informatique Scientifique

This paper addresses the issue of guaranteeing the good functioning of physical systems using expensive simulators. More precisely, it is interested in the construction of bounds allowing to majorise with a specified confidence the probability of occurrence of undesired events. In this context, this paper presents two algorithms: a first one allowing to build a bound higher than this probability at a fixed number of simulator evaluations; a second one allowing to reduce as much as possible this bound by adding in an optimized way new simulator evaluations. The efficiency of these algorithms is finally illustrated through the analysis of several test functions.

Mots clés: risk analysis; surrogate modeling; Gaussian process regression; quantile estimation; error control

Domaines: Méthodologie [stat.ME]; Applications [stat.AP]; Machine Learning [stat.ML]

Guillaume Perrin. Point process-based approaches for the reliability analysis of systems modeled by costly simulators. Reliability Engineering and System Safety, 2021, 214, pp.107799. ⟨10.1016/j.ress.2021.107799⟩. ⟨hal-03228196⟩