摘要
为了提高互联网的管理和控制水平,进而优化配置网络资源,一种新的估计网络内部参数的方法"网络层析成像"得到了广泛关注。提出一种基于递归神经网络的非平稳网络丢包层析成像方法,利用递归多层感知器求解非平稳网络丢包模型。采用NS2仿真工具进行实验,证明了该算法能够自适应非平稳网络丢包率随时间变化而产生的波动,以实时追踪网络内部链路的丢包率。
In order to optimize the performance of Intemet, we need promote the management level of our network. In recent years, a new technology named network tomography was introduced. An algorithm based on the recurrent multilayer pereeptron was proposed to solve the nonstaionary network loss tomography problem. Simulations were carried out using NS2 to demonstrate the accuracy of this estimation procedure, which also show the ability of tracking the loss rate of links in real time.
出处
《计算机应用》
CSCD
北大核心
2008年第B06期116-119,122,共5页
journal of Computer Applications
基金
国家自然科学基金资助项目(60572092)
关键词
网络层析成像
非平稳网络
丢包模
递归多层感知器
network tomography
nonstationary network
loss model
Recurrent MultiLayer Pereeptron （RMLP）
