论文部分内容阅读
提出一种基于自适应学习算法的故障监测智能代理，通过分段测量抽取描述网络正常行为的ＭＩＢ变量值并检测偏差，对学习获得的信息经由贝叶斯图加以组合，从而鉴别未知的或不可预见的故障．实验结果表明，智能监测代理系统能够在故障发生以前检测网络异常行为．
A fault monitoring intelligent agent based on adaptive learning algorithm is proposed. By segmenting and measuring the value of MIB variable which describes the normal behavior of network and detecting the deviation, the learning information is combined via Bayesian graph to identify unknown or not Predicted failure. The experimental results show that the intelligent monitoring agent system can detect the abnormal behavior of the network before the fault occurs.