基于多智能体和神经网络的网络故障诊断及二维可视化(英文)  

Multi-Agent and Neural Network based Network Fault Diagnosis and 2-Dimensional Visualization

作　　者：戚湧[1] 刘凤玉[1] 

机构地区：[1]南京理工大学计算机系,南京210094

出　　处：《系统仿真学报》2005年第9期2171-2174,共4页Journal of System Simulation

基　　金：This work was supported by National Nature Science Foundation of China(69973020)

摘　　要：提出一种基于多智能体和神经网络的网络故障可视化诊断系统。每个被诊断的网络具有一个网络故障征兆收集智能体,负责收集并且预处理该网络的故障征兆信息,该信息被传送到位于诊断中心的网络故障诊断智能体,用作诊断及可视化。该系统的优点是网络故障诊断智能体可以以可视化的方式诊断网络的故障,并且动态监控网络的运行状态。试验结果证明所述系统的有效性和鲁棒性。A network fault diagnosis and visualization system is proposed based on neural network and multi-agent. Each network has a Network Fault Symptom Collection Agent （NFSC Agent）, which collects and preprocesses the network fault symptom information. The collected and preprocessed network fault symptom information will then be sent to the Network Fault Diagnosis Agent （NFD Agent） for diagnosis and visualization. The advantage of the proposed system is that the NFD Agent can simultaneously diagnosis the network fault and dynamically monitor the networks＇ the running status in 2-dimensional visualization. Experimental results demonstrate the effectiveness and robustness of the proposed system.

关 键 词：网络故障诊断 神经网络 多智能体 可视化 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]