Reduction of Stochastic Petri Nets for Reliability Analysis
Wang Xiaoli, Chen Guangju, Zhao Qiang, Guo Zhongping · 2007
The model types used for reliability modeling are classified into two categories: combinatorial model types and state-space model types. Stochastic Petri nets (SPN), belonging to one of state-space types, can represent the dependencies, imperfect coverage, correlated failures, and repair dependencies that realistically occur in genuine life scenarios, but the combinatorial models such as reliability block diagrams, fault trees and reliability graphs can't. So, SPNs as a powerful modeling and analyzing tool have received considerably more attention in system reliability analysis.But the problem of state-space explosion of SPN is so serious that limits its ability to analyze complex and large-scale systems. In this paper, we bring forward a method of equivalent reduction to simplify SPN model before analyzing it. The method utilizes four structure-reduction rules: sequence, concurrence, choice and loop structures of elementary reliability model to simplify SPN model and overcome state-space explosion.