Bayesian Analysis of ARA Imperfect Repair Models

Abstract

This article proposes a Bayesian analysis of a class of imperfect repair models, the ARA models. The choice of prior distributions and the computation of posterior distributions are discussed. The presentation is unified for all ARA models and many kinds of possible priors. A numerical study on the quality of the Bayesian estimators is presented, as well as a comparison with the maximum likelihood estimators. Finally, the approach is applied to a real data set.

Keywords:

• Bayesian inference
• Imperfect repair
• Maintenance efficiency
• MCMC algorithms
• Repairable systems reliability
• Virtual age

Mathematics Subject Classification:

• Primary 62N05
• Secondary 60K10

Acknowledgment

This research is supported by a grant from EDF R&D. The authors thank Emmanuel Remy for his helpful comments.