基于证据推理的Bayes可靠性评定多源信息融合方法

he Bayesian method is adopted as an important and suitable approach for system reliability evaluation in small sample circumstance. In order to reduce the testing data, all kinds of prior information must be used, so the fusion of information of multiple sources is a key problem in Bayesian analysis. A new method based on the Dempster-Shsfer theory to realize the fusion of information of multiple sources was proposed, which can integrate of more prior distributions of obtained using different convergence criteria. The simulation example shows that the method proposed is effective.