PMC模型下的一个贪婪诊断算法  

A greedy diagnosis algorithm for PMC model

作　　者：宣恒农[1] 张润驰 何涛 刘凌波[1] 

机构地区：[1]南京财经大学信息工程学院,江苏南京210046 [2]亚信科技(南京)有限公司,江苏南京210036

出　　处：《计算机工程与科学》2015年第8期1430-1435,共6页Computer Engineering & Science

基　　金：国家自然科学基金重大研究计划资助项目(90718008);国家自然科学基金重点项目(61133015);江苏省自然科学基金资助项目(2004119)

摘　　要：提出了一种PMC模型下基于矩阵运算的贪婪诊断算法——MGFD算法。算法结合作者曾经提出的"绝对故障基"思想,首先剔除绝对故障基,得到一个维度减小的矩阵,之后根据该矩阵求得集团。在文献[10]提出的四个贪婪诊断算法的基础上,提出集团的内贪婪因子、外贪婪因子、综合贪婪因子等概念,设计了新的贪婪准则。论证了MGFD算法的正确性,并对算法进行了实验仿真。实验结果表明,MGFD算法相比文献[10]提出的贪婪诊断算法,具有较高的诊断正确率。We propose a greedy diagnosis algorithm based on the matrix operations called Matrix based Greedy Fault Diagnosis （MGFD） algorithm under the PMC model. With the ＂absolute fault base＂ concept put forward by the authors in paper [11], we remove the absolute fault units to obtain a matrix of reduced dimensions and then get the groups accordingly. On the base of the four existing greedy diagnostic algorithms proposed by paper[10], we define the concepts of inner greed factors, outer greed factors, comprehensive greed factors of each group, and we also design some new greedy criteria. We demonstrate the correctness of the MGFD algorithm, and design several simulation experiments for the algorithm. Experimental results show that the MGFD algorithm has a higher diagnostic accuracy in comparison with the existing greedy diagnostic algorithm proposed by paper [10].

关 键 词：系统级故障诊断 PMC模型 绝对故障基 集团 贪婪诊断算法 综合贪婪因子 MGFD算法 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]