一种新的P2P网络流量预测模型  

Novel Peer to Peer Network Traffic Prediction Algorithm

作　　者：韩志杰[1] 王汝传[2] 

机构地区：[1]苏州大学计算机学院,苏州21500 [2]南京邮电大学计算机科学与技术系,南京210003

出　　处：《计算机科学》2008年第9期39-41,共3页Computer Science

基　　金：国家自然科学基金(60573141和60773041);江苏省自然科学基金(BK2005146);江苏省高技术研究计划(BG2005038;BG2006001);国家高科技863项目(2006AA01Z219;2006AA01Z439;2007AA01Z404;2007AA01Z478);2006江苏省软件专项;南京市高科技项目(2007软资127);现代通信国家重点实验室基金(9140C1101010603);江苏省计算机信息处理技术重点实验室基金(kjs050001;kjs06006)资助;江苏省高校自然科学研究计划(07KJB520083)资助

摘　　要：当前,对等计算(Peer to Peer)引起了广泛的关注,其典型应用有文件共享,即时通信等等,为了更好地使用P2P资源,合理规划P2P网络资源,建立P2P流量预测模型具有十分重要的理论意义和现实价值。基于小波神经网络提出了一个P2P网络流量预测模型,预测模型共分为输入层、隐含层、输出层,通过使用小波代替传统神经网络的Sigmoid作为神经网络的隐节点激励函数,在Matlab6.5仿真平台中实现,结果显示能够有效提高P2P网络流量预测精度。Recently, there has been a growing interest in the potential use of Peer to Peer computing （P2P） in many applications such as file sharing, instant communication. Therefore, to realize their potential, there is a need of a P2P traffic prediction algorithm that facilities the deployment of a network traffic that is optimized in terms of network bandwidth. This paper focuses on developing a novel P2P traffic prediction using wavelet neural network. The architecture has three layers ： input layer, hidden layer, and output layer, By using the wavelet analysis taken place on the Sigmoid as excitation function of neural network hidden node, the proposed algorithm is implemented in matlab 6. 5 and the simulation result has been proved is viable.

关 键 词：神经网络 小波 P2P 网络流量 

分 类 号：TP393[自动化与计算机技术—计算机应用技术] O211.67[自动化与计算机技术—计算机科学与技术]