模糊聚类分析在光网络多故障定位中的应用  

Applications of fuzzy clustering analysis in multi-fault location in optical networks

作　　者：宋继恩[1,2] 夏芸[1] 李新[3] 张引[3] 黄善国[3] 赵永利[3] 张杰[3] 

机构地区：[1]华信邮电咨询设计研究院有限公司,浙江杭州310014 [2]浙江大学,浙江杭州310014 [3]北京邮电大学,北京100876

出　　处：《光通信研究》2013年第4期20-23,共4页Study on Optical Communications

摘　　要：利用模糊数学理论对光网络中的多故障定位问题进行分析,得到了光网络中多故障定位的模糊聚类模型。根据由模糊聚类算法得到的专家系统和由此确定的告警和故障之间的隶属函数公式进行故障定位。在格状光网络拓扑下进行多故障定位仿真实验,得出故障定位成功率的对比图和时间性能的对比图。实验结果表明,模糊聚类算法可以高效快速地定位故障。This paper analyzes the issue of multi-fault location in optical networks by using the fuzzy clustering theory and builds up its fuzzy clustering model. Then it performs fault location on the basis of the expert system obtained from the fuzzy cluste- ring algorithm and the membership function formula between alerts and faults thus determined. Finally, it performs multi-fault location simulation experiments in optical mesh network topology and obtains the comparison charts of success rate in fault lo- cation and time performance. The experimental results show that the fuzzy clustering algorithm can locate faults fast and effi- ciently.

关 键 词：模糊聚类 光网络 多故障 故障定位 

分 类 号：TN915.5[电子电信—通信与信息系统]