网络流量模型的非线性特征量的提取及分析

本文基于相空间重构理论,在高维相空间中对网络流量的宏观和微观特性进行研究分析.首先,提取网络流量的宏观非线性特征量,如关联维数、Kolmogorov熵和最大Lyapunov指数,实现了网络流量时序非线性动力学特性的定量分析.然后,通过对四种典型突发性流量模型的多重分形谱的计算,揭示了流量模型不同层次的行为特征,并给出了刻画突发性流量的有效微观参数.为进一步利用混沌动力学理论对网络行为的控制和建模奠定了基础.

Analysis on Network Flow Time Sequences and Extraction of Nonlinear Characteristic Quantities

Many efficient approaches and analysis techniques are applied to analyze the macro and micro characterizes of network flow data.The attractors are reconstructed by making use of time-delay coordinates.Then the macro nonlinear characteristic quantities of the network flow time sequences such as the Fractal dimension、Kolmogorov entropy and the largest Lyapunov exponents are extracted in this multi-dimension phase-space.The study on the temporal characteristics of these three parameters discovered that the network flow is featured by some chaotic behaviors.The multifractal spectrums of the four difference network flow data are calculated in order to characterize and recognize the dynamic structure of network flow data more deeply,and thus can be effectively exploited for the controlling and modeling of the network behaviors.