论文部分内容阅读
当代光网络趋于复杂化,一个故障将引发多个告警事件。同时,同一告警可能由不同的故障导致。本文研究了基于数据挖掘的光网络告警相关性分析,我们从动态网络资源与设备中挖掘关联规则,充分利用和维护原有规则知识,使网络结构和规则库都能快速更新,并提出了新型的动态模糊关联规则挖掘算法IDFARM。同时运用模糊逻辑将数值型告警属性转化为逻辑语言变量,当网络中有新的未知告警发生时,我们对模糊关联规则运用模糊推理来进行故障诊断,这将缩短网络恢复时间,有利于提高光网络故障管理性能。仿真实验验证了文章算法的正确性和有效性。
Contemporary optical networks tend to be complicated, with one failure triggering multiple alarm events. Meanwhile, the same alarm may be caused by different faults. In this paper, we analyze the correlation analysis of optical network alarms based on data mining. We excavate association rules from dynamic network resources and devices, make full use of and maintain the original knowledge of rules so that both network structure and rule base can be rapidly updated, Dynamic Fuzzy Association Rules Mining Algorithm IDFARM. At the same time, fuzzy logic is used to convert numerical alarm attributes into logical language variables. When there are new unknown alarms in the network, we use fuzzy reasoning to diagnose the fuzzy association rules, which will shorten the network recovery time and improve the light Network fault management performance. Simulation experiments verify the correctness and validity of the algorithm.