A Novel Traceable Evidential Reasoning Method and Its application in Health Assessment of Complex Equipment

• Hongwei Gao,
• Xianlong Wu,
• Bangcheng Zhang,
• Liang Chang,
• Guanyu Hu,

Vol. 19, No. 11, November 30, 2025
10.3837/tiis.2025.11.010, Download Paper (Free):

• Backtracking path search algorithm
• Evidence contribution
• Evidential reasoning method
• multi-attribute fusion
• Traceable evidential reasoning

Abstract

A traceable evidential reasoning (TER) method with backtracking path is proposed to address the challenges of traditional evidential reasoning (ER) methods in ineffective solving the lack of labeled samples leads to the inability to measure the accuracy and trace back the factors of a certain result in health assessment of complex equipment. In TER, backtracking path of result and credibility calculating method are both proposed to provide guidance for model optimization and subsequent maintenance decisions, which enhanced the credibility. Firstly, after fusing evidences by ER, a calculation method of evidence contribution is designed to reflect the degree of influence of evidence on results. Evidence contribution takes into account weight, reliability and credibility. Secondly, a backtracking path search algorithm is proposed based on the evidence contributions, through which the backtracking path that maximizes the sum of the contribution degrees can be found as core factor chain leading to a certain result. Finally, a credibility analysis method is proposed to deeply analyze the result, which can provide a strong basis for the parameters optimization. Case study from practical engineering proves the effectiveness of TER, which solves the credibility and the factor backtracking problem, and expands the scenarios of ER in engineering applications.

Statistics

Cite this article