应用小波变换改进网络流量的相空间重构  

Improve phase space reconstruction for network traffic flow based on wavelet transform

作　　者：戴悦[1] 刘渊[1] 曹建华[1] 

机构地区：[1]江南大学信息工程学院,江苏无锡214122

出　　处：《计算机工程与设计》2008年第5期1138-1140,共3页Computer Engineering and Design

基　　金：国防基础研究基金项目(A1420061266)

摘　　要：流量预测是流量工程、拥塞控制和网络管理的核心问题,网络流量非线性变化部分有混沌特性的存在。现有的网络流量研究方法只是单一采用线性或者非线性的方法进行处理,而忽略了对混沌特性的研究。针对网络流量的混沌特点,利用小波变换和非线性动力学方法研究网络流量的混沌特性,并改进相空间重构方法。将混沌吸引子投影于小波滤波器向量所张的空间中,并充分利用了小波变换的去噪优点,给网络流量预测提供了行之有效的新方法。Distributed applications use predictions of network traffic to sustain their performance by adapting their behaviors. It is recognized that the network traffic of non-linear part has chaos characteristics. However, existent network traffic prediction study only utilize either linear or non-linear methods to solve the problem and neglect chaos characteristics. The chaos characteristics of network traffic flow are studied based on warelet transform and nonlinearity dynamics method, and phase space reconstrunction is improved. Chaotic attractor is projected in the space of wavelet filter vector, and is fully utilized with warelet denoising. And an affective new method of network traffic prediction is provided.

关 键 词：网络流量 小波变换 混沌 预测 相空间重构 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]