%0 Journal Article
%T Node Grouping in System-Level Fault Diagnosis
%A Zhang Dafang
%A Xie Gaogang
%A MIN Yinghua
%A <br>张大方
%A 闵应骅
%J 计算机科学技术学报
%D 2001
%I 
%X With the popularization of network applications and multiprocessor systems, dependability of systems has drawn considerable attention. This paper presents a new technique of node grouping for system-level fault diagnosis to simplify the complexity of large system di-agnosis. The technique transforms a complicated system to a group network, where each group may consist of many nodes that are either fault-free or faulty. It is proven that the transformation leads to a unique group network to ease system diagnosis. Then it studies systematically one-step t-faults diagnosis problem based on node grouping by means of the concept of hide-pendent point sets and gives a simple sufficient and necessary condition. The paper presents a diagnosis procedure for t-diagnosable systems. Furthermore, an efficient probabilistic diagnosis algorithm for practical applications is proposed based on the belief that most of the nodes in a system are fault-free. The result of software simulation shows that the probabilistic diagnosis provides high probability of correct diagnosis and low diagnosis cost, and is suitable for systems of any kind of topology.
%K system-level fault diagnosis
%K one-step t-diagnosable system
%K node grouping
%K diagnosis algorithm
%K probabilistic diagnosis<br>微处理器系统
%K 错误诊断
%K 错误处理器
%U http://www.alljournals.cn/get_abstract_url.aspx?pcid=5B3AB970F71A803DEACDC0559115BFCF0A068CD97DD29835&cid=8240383F08CE46C8B05036380D75B607&jid=F57FEF5FAEE544283F43708D560ABF1B&aid=20C5F338045312ABD65A86C03EBDCAC8&yid=14E7EF987E4155E6&vid=7801E6FC5AE9020C&iid=94C357A881DFC066&sid=2B25C5E62F83A049&eid=2B25C5E62F83A049&journal_id=1000-9000&journal_name=计算机科学技术学报&referenced_num=7&reference_num=8