Parametric NdRFT for the derivation of optimal repair strategies

Abstract

Non deterministic Repairable Fault Trees (NdRFT) are a recently proposed modeling formalism for the study of optimal repair strategies: they are based on the widely adopted Fault Tree formalism, but in addition to the failure modes, NdRFTs allow to define possible repair actions. In a previous pa per the formalism has been introduced together with an analysis method and a tool allowing to automatically derive the best repair strategy to be applied in each state. The analysis technique is based on the generation and solution of a Markov Decision Process. In this paper we present an extension, ParNdRFT, that allows to exploit the presence of redundancy to reduce the complexity of the model and of the analysis. It is based on the translation of the ParNdRFT in to a Markov Decision Well-Formed Net, i.e. a model specified by means of an High Level Petri Net formalism. The translated model can be efficiently solved thanks to existing algorithms that generate a reduced state space automatically exploiting the model symmetries.