基于蚁群算法的电子设备多值测试故障诊断策略  

Research on Fault Diagnosis Strategy of Electronic Systems Multi-value Test Based on Ant Colony Algorithm

作　　者：张峻宾[1] 蔡金燕[1] 孟亚峰[1] 李丹阳[1] 

机构地区：[1]军械工程学院,石家庄050003

出　　处：《火力与指挥控制》2014年第9期112-116,共5页Fire Control & Command Control

基　　金：国家自然科学基金资助项目(61271153;61372039)

摘　　要：测试序列优化设计是故障诊断中的重要组成部分,最优测试序列能提高故障诊断的效率,常见的故障诊断系统均基于二值属性,而多值属性系统的测试优化问题研究的较少。针对多值属性系统测试序列优化问题的特点,提出了一种改进蚁群算法,结合二值属性系统和多值属性系统的关系,设计了适应于多值属性优化的状态转移规则和信息素更新机制。针对在等测试费用和故障先验概率的情况下不能寻优的问题,制定了对应的优化标准,符合实际的测试需求。通过对比试验,证明了其能解决电子系统多值属性系统的序列优化问题,扩展了多值属性电子系统的测试优化策略。Test sequencing optimization design is an important part of fault diagnosis. The optimized test sequence can improve the ability of fault diagnosis. The binary attribute systems were researched widely,while there are fewer researches on multi-value attribute systems. According to the multi-value attribute system test sequencing problems,a novel Ant Colony Algorithm was advanced. Combining the relation between binary attribute systems and multi-value attribute systems,multi-value attribute optimization state transition rule and pheromone update mechanism were designed. Under the same fault prior probability and test cost,a novel optimization standard was advanced,it satisfies test requirements. The contrast experiments prove that the proposed scheme can solve multi-value attribute electronic systems fault diagnosis and the strategy has been extended.

关 键 词：多值测试 多值属性 故障诊断 蚁群算法 测试序列优化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]