摘要
为自动获取平均测试费用最少,且能快速实现系统故障检测与隔离的最优测试序列,提出了一种准多步前向搜索算法。该算法结合了一步前向搜索和多步前向搜索的基本特点,并以信息增益为启发策略,获得了计算精度和复杂度的权衡。基于该算法,开发了一个测试序列优化工具,通过对大量实例的计算,证明该工具可以在实用中取得满意结果,且该工具可以用作自动测试设备和便携式维修辅助软件的核心诊断推理机。
To achieve an optimal test sequence which ensures to accomplish the system fault detection and isolation quickly under the least average testing cost, a quasi multi-step look-ahead search algorithm was presented. This algorithm was based on the existing one-step look-ahead search algorithm and multi-step look-ahead search algorithm. It employed the information gain as its heuristics and achieved an efficient tradeoff between precision and computational complexity. Based on this algorithm, a computer-aided Test Sequence Optimization Tool （TSOT） was developed. Computation on a large number of examples with practical data revealed the satisfactory results of this tool in practical applications, and besides, it could be employed as a kernel of a diagnostic reasoning machine in Automatic Test Equipment （ATE） or Portable Maintenance Aid （PMA） software.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2008年第5期1036-1040,共5页
Computer Integrated Manufacturing Systems
基金
国防预研基金资助项目(51317040102)~~
关键词
测试性设计
测试序列优化
计算机辅助设计
准多步前向搜索算法
design for testability
test sequence optimization
computer-aided design
quasi multi-step look-aheadsearch algorithm
