基于互连网络系统故障的新型自适应诊断算法

Novel adapted algorithm for interconnection network

互连网络的故障诊断是网络系统可靠性分析的重要内容。PMC模型是一种重要的网络故障模型。针对具有哈密顿环的互连网络(也称做哈密顿网络),利用分治回环思想,提出了一种新的基于PMC故障模型自适应的诊断算法。其核心思想是,对哈密顿网络进行序列划分,然后对得到的每个01序列的结节进行回环诊断,最后利用回环诊断的结果对非01序列的节点进行诊断。对于一个具有多个01序列的互连网络,该算法通过有限次轮回的测试,能准确地定位系统中的故障节点,对于正确节点的诊断可靠度能无限接近100%。当系统中存在的回测边越多时,该算法的诊断效果越好。

The fault diagnosis of intereonneetion network has played an important role in the system reliability analysis. PMC model is a classical system fault self-diagnostic model. This paper presented a novel algorithm which called loop-Back diagnostic algorithm to deal with the problem of self-adaptive diagnosis of Hamilton networks. The main idea was as follows ： firstly, it partitinned the Hamilton network into different 01 sequence. Secondly, it diagnosed these 01 sequences with loop-back diag- nostic algorithm. Finally, it arranged the identified units in 01 sequence to identify other units. Using this new algorithm, the testing result credibility of 01 sequence could access to 1 by limited iterations. The performance analysis shows that this algo-rithm can identify a faulty unit absolutely correct and identify the fault-free units with high credibility. Furthermore, credibilily of the algorithm increases with the increasing number of the return testing links.