基于组合模型的网络业务流多重分形建模  

Network Traffic Multi-fractal Modeling Based on Combination Model

作　　者：徐兴[1] 徐胜[2] 王卫星[1] 岳学军[1] 林涛[1] 

机构地区：[1]华南农业大学工程学院,广州510642 [2]广东工业大学物理与光电工程学院,广州510642

出　　处：《科学技术与工程》2012年第18期4422-4425,4431,共5页Science Technology and Engineering

基　　金：广东高校优秀青年创新人才培育项目(LYM10035);公益性行业(农业)科研专项经费项目(20090323-06);广东省省部产学研结合项目(2011B090400359)资助

摘　　要：研究表明网络业务流具有长相关、短相关特性以及小尺度上的多重分形特性的特点。将可以描述网络业务流波动情况的GARCH模型和可以描述网络业务流长相关特性和短相关特性的FARIMA模型引入网络业务流的多重分形建模中。提出了一种FARIMA/GARCH组合模型。数值仿真对比结果表明提出的FARIMA/GARCH组合模型对网络业务流多重分形谱描述更为准确,实现了对网络业务流多重分形特性的有效建模。Recent research has discovered that network traffic model has the property of long-range dependence,short-range dependence and multi-fractal.It is known that GARCH model could analyze the volatility of time series,and FARIMA model could describe the long range dependent and short range dependent of time series.So a FARIMA / GARCH combination model is developed to the modeling of network traffic.Numerical results prove that the combination model could effectively describe the multi-fractal nature of network traffic comparing with other models.

关 键 词：FARIMA GARCH 多重分形 建模 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]