An Approach to Testability Evaluation Based on Improved D-S Evidence Theory

Abstract

To address the conflict of prior information in traditional D-S evidence theory, an approach to testability assessment based on improved D-S evidence theory is proposed by integrating testability virtual test information, testability expert experience information and testability prediction information. On the basis of such information, density distribution function is developed for equipment testability indicator fault detection rate (FDR), and then mass functions are further constructed. Eventually, testability assessment results are obtained by introducing Lance and Williams distance for information fusion of improved D-S evidence theory.