Non‐Markovian State‐Space Models in Dependability Evaluation

The purpose of this paper is to provide an up‐to‐date treatment of advanced analytic, state‐space based techniques to study dependability models with non‐exponential distributions. We first provide an overview of different techniques for the solution of non‐Markovian state‐space based models, including phase‐type expansion and a general framework which allows us to deal with renewal, semi‐Markov, and Markov‐regenerative processes, trying to characterize them on dependability contexts. In the last part of the paper, we illustrate these techniques by means of some examples dealing with common non‐exponential reliability behaviors. Our aim is to provide a reference for practising engineers, researchers, and students in state‐space dependability modeling and evaluation. Copyright © 2012 John Wiley & Sons, Ltd.