摘要
网络规模逐渐增大,路由节点呈现出了较大的随机性分布,节点的位置关联性被打破,很难建立准确的定位模型,造成传统的基于特征匹配的节点故障定位方法很难描述较大的随机性变化特性,造成故障定位的不准确。为了解决上述问题,提出一种随机游走的网络故障节点检测算法,通过把检测到的故障特征作为随机游走的初值种子点和节点故障配准的特征点,将故障特征配准和位置随机变化检测结合起来,提高了变化检测的效率。改进方法采用了一种最小化特征权值策略来提取故障特征的不透明度,能检测到大范围的节点异常变化,包括一些细微的变化。仿真结果表明,改进方法能够提高故障节点定位的准确性。
Research the accurate localization of network fault node. The paper put forward a random migration network fault node detection algorithm. Through the examination, the algorithm takes the failure characteristics as the initial seed point of random walk and the feature point of node fault registration, combines the failure characteristics registration and position random change detection. The algorithm adopts a minimize feature weights strategy to extract fault features opacity, thus can detect a wide range of node abnormal changes. The simulation results show that the algorithm can improve the accuracy of fault node localization.
出处
《计算机仿真》
CSCD
北大核心
2013年第6期271-274,共4页
Computer Simulation
关键词
网络
节点定位
随机游走
Network
The node localization
Random walk
