A Novel Testability Optimization Algorithm Counting the Reliability of Test Points

• Published in: 2019 Prognostics and System Health Management Conference (PHM-Paris) pp. 338 - 342
• Main Authors: Hou, Wenkui, Liu, Liangli, Li, Pengyu
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-05-2019
• Subjects: Bayes methods; Fault diagnosis; Finite impulse response filters; Genetic-PSO; Greedy algorithm; Mathematical model; Optimization; Testability; Uncertainty
• ISSN: 2166-5656
• DOI: 10.1109/PHM-Paris.2019.00064

The traditional testability mathematical model is attributed with inaccurate when applied in real industry occasions for it ignores the reliability of the test points (usually considered fully convinced). In this paper, we devise a novel testability optimization algorithm regarding with the reliability of test points. First, the D-matrix of uncertainty is acquired based on the Bayes-learning. Then, quantizing the loss function with the information entropy and utilizing the global searching ability of Genetic-PSO algorithm and the efficiency of the Greedy algorithm to form the test group. The proposed algorithm is validated with test data of avionics. The experiment result shows the method is able to select the optimal test group considering the uncertainty.