基于FrFT的网络流量Hurst参数估计器  

Frft Based Hurst Parameter Estimator for Network Traffic

作　　者：王霁[1] 单佩韦[1] 

机构地区：[1]华东师范大学,上海200241

出　　处：《现代电子技术》2011年第8期33-35,共3页Modern Electronics Technique

摘　　要：实际的网络流量模型采用自相似模型,Hurst参数是序列长相关程度的度量。为方便工程上对网络流量的长相关性进行估计,介绍了网络流量自相似模型和Hurst参数小波分析法,建立了分数阶傅里叶变换(FrFT)与小波分析之间的联系,在此基础上介绍了一种新的基于FrFT的网络流量Hurst参数估计方法,并运用此方法设计了网络流量Hurst参数估计器。通过对白噪声和已知Hurst参数的实际网络流量数据进行估计,本方法可以有效估计随机时间序列的Hurst参数。Self-similar model is used to describe real network traffic model,whose degree of long-range dependence（LRD） is characterized by Hurst parameter.In this paper,the self-similar model of network traffic and wavelet analysis method of Hurst parameter are introduced,and a relation between the fractional Fourier transform（Frft） and wavelet analysis is established in order to conveniently estimate the long-range dependence of network traffic.On the basis of the above achievements,a new Frft-based Hurst parameter estimation method of network traffic is presented.The Frft-based Hurst parameter estimator for network traffic was designed with the new method.It can effectively estimate Hurst parameters of the random time sequence.

关 键 词：分数阶傅里叶变换 HURST参数 长相关 自相似 网络流量 

分 类 号：TN919-34[电子电信—通信与信息系统]