Diagnostic Strategy Optimization Based on Particle Swarm Algorithm

• Published in: Applied Mechanics and Materials Vol. 215-216; pp. 555 - 560
• Main Authors: Jing, Jian Hui, Zhang, Yan Sheng, Qiao, Zhong Tao
• Format: Journal Article
• Language: English
• Published: Zurich Trans Tech Publications Ltd 01.11.2012
• Subjects: Design for Testability; Diagnostic Strategy; Particle Swarm Algorithm; Test Sequence
• ISBN: 3037855010; 9783037855010
• ISSN: 1660-9336; 1662-7482; 1662-7482
• DOI: 10.4028/www.scientific.net/AMM.215-216.555

Fault diagnostic strategy profoundly influences diagnostic efficiency and cost. Diagnostic strategy optimization is a Non-Polynomial optimizing problem, which is a hard one. Applying conventional methods to resolving it, there are some difficulties: more complex to implementing algorithm, more time to diagnosing and more difficult to attaining global optimum. Particle swarm algorithm (PSA) is a new intelligent optimization algorithm, and applied to optimizing diagnostic strategy. Function of all-in cost is constructed by state probability, isolating matrix and test cost, serving as objective function. Test sequences are directly put into particle codes. Particle speeds are transformed to learning probability towards the best one in the swarm. Given proper parameters in PSA, the method can search the global optimum in a little time. At last, an example shows the approach is feasible and effective.