多值测试诊断策略优化设计  

Optimal Design for Diagnosis Strategy Based on Multi—valued Test

作　　者：赵文俊[1] 张强[1] 匡翠婷 宋家友[2] 

机构地区：[1]空军第一航空学院,河南信阳464000 [2]郑州大学信息工程学院,郑州450000

出　　处：《计算机测量与控制》2015年第12期3936-3939,共4页Computer Measurement &Control

基　　金：军内计划项目(KJ2012255)

摘　　要：针对电子系统中多值测试条件下的测试排序问题,提出了一种诊断策略优化方法;首先,以相关性矩阵模型为基础,将已有的二值测试的优化算法同多值测试问题相结合,提出了适用于多值测试的基于霍夫曼编码的启发式函数;其次,将平均测试代价最小和平均测试步骤最少作为优化目标,采用与或树启发式搜索算法生成诊断树,得到多值测试的诊断策略,并给出了诊断策略优化方法的具体实现步骤;最后,将其应用到航空设备的实例中;结果表明提出的基于霍夫曼编码的与或树启发式搜索算法是可行的,其生成的诊断树是最优的,比基于信息熵的与或树启发式搜索算法具有更小的平均测试代价和更少的平均测试步骤。Aiming at the test sequencing problem with multi--valued tests of electronic system, an optimal method for diagnosis strategy based on multi--valued test is presented. At first, on the basis of the dependency matrix, combined traditional binary test--based algorism with multi--valued test--based problem, a heuristic function based on H uffman Code suitable for multi--valued test is discussed. Then, with minimum test cost and steps for the optimal goal, AND/OR tree heuristic search algorithm is used for diagnosis generation, the diagno- sis strategy based on multi--valued test is improved, and the specific implementation steps of optimal method for fault diagnosis strategy is offered. Finally, apply it to aviation equipment. The result shows that the AND/OR heuristic search algorithm based on Huffman Code is feasible, the diagnosis tree it generate is optimal, has less average test cost and steps than the AND/OR heuristic search algorithm based on information entropy.

关 键 词：诊断策略 多值测试 霍夫曼编码 启发式函数 诊断树 

分 类 号：TP206[自动化与计算机技术—检测技术与自动化装置]