Reliability information fusion based on Bayesian generalized mean operator

Abstract

Reliability assessment is difficult for such complex systems as aerospace and aircraft in the case that the field test sample is small for limited money or time. Otherwise, during the development of such complex systems, a lot of information related to reliability can be easily got, such as test data of subsystem and components, test data of similarity products, expert experience etc. In order to improve the confidence level of statistical inference, we should make full use of such reliability information. This paper proposes a fusion approach based on Generalized Mean Operator(GMO), which can effectively represents the redundancy and complementary among multi-source information. The unknown parameters of the fusion model are estimated by the second maximum likelihood estimation method (ML-II).To illustrate the efficiency of the approach, a simulation example is given. © 2012 Springer-Verlag GmbH.