摘要
本文通过对多个ＡＲ１过程（ｏｒｄｅｒ１ａｕｔｏｒｅｇｒｅｓｉｖｅｐｒｏｃｅｓ）的累加（ｓｕｐｅｒｐｏｓｉｔｉｏｎ）过程的研究，发现当有限个参数特定的ＡＲ１过程累加后，该累加序列具有明显的自相似性．这从某种程度上揭示了为何不具备长期相关性的单个信源，经复接、交换等处理后，使得网络中业务流呈自相似性的原因．本文给出了数学分析方法和计算机仿真结果，由此还可得出一种生成自相似序列的快速方法．
In this paper,we analyze the superposition of AR1 process.We discover that when the parameters of these AR1 processes satisfy some specified conditions,the superposition process has the property of self similarity.This can be a physical explanation of the self similar nature of the packet network,and partly answers the question why the short range dependent traffic sources can cause the fractal phenomena in modern network.We present the mathematical analysis and simulation result,and a fast algorithm to generate long run time series is also derived.
出处
《电子学报》
EI
CAS
CSCD
北大核心
1999年第4期6-10,共5页
Acta Electronica Sinica
基金
国家自然科学基金
电科院预研基金
关键词
自相似性
长期相关
AR1模型
网络
业务流
Self similarity,Long range dependence,AR1 superposition process,Fractal Gaussian noise
