基于遗传算法的序贯诊断测试策略生成.pdfVIP

Vol. 16 No. 4 系 统 仿 真 学 报 April 2004 JOURNAL OF SYSTEM SIMULATION • 833 • 基于遗传算法的序贯诊断测试策略生成 于劲松，徐 波，李行善 （北京航空航天大学自动化科学与电气工程学院, 北京 100083 ） 摘 要：针对序贯故障诊断问题，提出一种基于遗传算法的诊断测试策略生成方法。采用变长度 染色体表示有效测试策略，根据测试费用排序进行适应度分配，设计了适于顺序编码的遗传操作， 竟过进化计算获得期望费用最小的测试策略。对飞船发射前测试进行的仿真结果表明，本文方法 与基于信息增益最大的测试策略相比，测试费用显著降低，适于解决大规模诊断测试问题。 关键词：故障诊断；测试策略；测试费用；遗传算法 文章编号：1004-731X (2004) 04-0833-04 中图分类号：TP206.3 文献标识码：A Generation of Test Strategy for Sequential Fault Diagnosis Based on Genetic Algorithms YU Jin-song ，XU Bo ，LI Xing-shan (School of Automation Science and Electrical Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100083, China) Abstract ：A novel approach based on genetic algorithms is proposed to generate the effective and efficient test strategy for sequential fault diagnosis. The effective test strategy is presented by an unfixed length chromosome, the fitness assignment of which is based on its rank of test costs among the whole population. Some new genetic operators suited for sequencing coding are also designed. Finally, the test strategy with minimization of test costs is obtained through evolutionary computation. The results of numerical simulations on the test case of airship pre-launch checkout show that the proposed approach significantly decreases test costs and is more adaptable to solve large-scale test sequencing problems for diagnosis when compared with that based on the information gain maximization. Keywords ：fault diagnosis; test strategy; test cost; genetic algorithms 本文提出基