Computer Engineering and Applications ›› 2017, Vol. 53 ›› Issue (13): 78-82.DOI: 10.3778/j.issn.1002-8331.1607-0130

Previous Articles     Next Articles

System-level diagnosis algorithm based on Malek model

GUI Weixia, LIU Cui   

  1. School of Computer, Electronics and Information, Guangxi University, Nanning 530004, China
  • Online:2017-07-01 Published:2017-07-12


归伟夏,刘  翠   

  1. 广西大学 计算机与电子信息学院,南宁 530004

Abstract: In multi machine system, each node(processor) in the communication process is extremely easy to failure, therefore, it is very important to choose effective diagnosis algorithm for judging system fault set fast and accurately. The traditional PMC model is based on the test results between nodes, and the test results on faulty node testing others are doubtful, that lead to the results of this diagnosis model is relatively unstable. In view of this situation, taking the Malek diagnosis model instead of the traditional PMC model, with the characteristics of genetic algorithm, the complex network is simplified as binary encoding, and the population search direction is determined according to the fitness function value, which can improve the search efficiency. The algorithm is built by designing constraint equation upon the Malek model, putting forward a new adaptation function, and optimizing mutation operator. The experimental results show that the improved algorithm shortens the CPU time that is required to judge the fault set. Besides that, under the fault symptoms, the probability of this algorithm deciding the target fault set is higher, so it is proved that the Malek model is more efficient than the PMC model.

Key words: Malek model, genetic algorithm, system-level diagnosis

摘要: 在多机系统中,各个结点(处理器)在通信过程中极易发生故障,因此选择有效的诊断算法,快速、准确地判断出系统故障集十分重要。传统的PMC模型以结点相互测试的结果为基础,而故障结点的测试结果不唯一,导致该模型诊断结果相对不稳定。针对这种情况,采用Malek诊断模型代替传统的PMC模型,借助遗传算法特性,将复杂的网络拓扑图简化为二进制编码,并按照适应度函数值确定种群搜索方向,提高搜索效率。该算法根据Malek模型设计约束方程,提出新的适应度函数,优化变异算子。实验表明,算法改进后,缩短了判断故障集所需的CPU时间,同时,算法根据故障症候判断出目标故障集的概率更高,从而证明了用Malek模型代替PMC模型的高效性。

关键词: Malek模型, 遗传算法, 系统故障诊断