模糊理论在网络故障实时诊断中的应用

Application of Fuzzy Theory in Real-time Diagnosis of Network Fault

通信网络故障诊断的核心就是进行告警相关性分析,定位根源告警,从而定位故障。文中将基于数据挖掘的相关性分析方法与模糊理论相结合应用于网络故障实时诊断:将模糊聚类方法应用于网络告警模糊化处理,提出了一种应用于告警模糊关联规则知识库建立的挖掘算法,最后应用模糊聚类和模糊匹配方法对实时收集的新发告警集进行根源告警的模糊推理。模糊理论在通信网络故障诊断中的应用,为网络故障的实时诊断提供了一种崭新思路,对网络故障的及时恢复具有重要意义。实验仿真验证了整个思路的可行性。

The core of the communication network fault diagnosis is the alarm correlation analysis,locate the root alarm for fault. The correlation analysis method based on data mining and fuzzy theory are combined in this article,applying to real-time network fault diagnosis. Fuzzy clustering method is applied to the fuzzy processing of network alarms. After that, a new alarm fuzzy association rules mining algorithm is raised for establishing the knowledge base. At last, fuzzy clustering and fuzzy matching method is applied to infer the root one of real-time collected alarm set. The application of fuzzy theory in the communication network fault diagnosis gives a new method for real-time diagnosis of network failures. It is of great significance for the timely recovery of network failures. Simulation experiments verify the feasibility of the whole idea.