Bayesian parameter estimation for multi-state components

This paper develops a Bayesian method to estimate transition intensities of multi-state components whose degradation follows a Markov process. Due to lack of sufficient data and vague knowledge from experts´ judgments, transition intensities between any pair of states cannot be precisely known or estimated. The proposed Bayesian method can merge the prior knowledge from experts´ judgments with continuous inspection data or discontinuous inspection data to get the posterior distributions of transition intensities. A numerical example along with comparative studies is presented to demonstrate the effectiveness of the proposed method.