基于混沌算子网络模型的网络流量预测研究

摘要

采用混沌算子构造预测网络,对网络流量数据进行预测分析.结合相空间重构理论将已知数据构造成训练样本,利用遗传算法对混沌算子参数进行训练调节,从而改变网络的动力学特性,使之逐渐逼近被预测时间序列的动力学特性,并保持与之变化一致该方法可对各种网络流量数据序列进行有效的预测分析.仿真实验结果表明:与传统的预测方法相比,该方法具有更好的预测趋势.%A new network traffic prediction method is proposed based on chaotic operator network. Training samples are constructed by known data according to the phase space reconstruction theory. The control parameters of the chaotic operators are trained by genetic algorithm. Thus, the dynamic characteristics of the prediction network are changed to approach to the dynamic characteristics of the system predicted. The prediction method can be applied to predict kinds of network traffics series. Simulation results show that the method has better predicting trends than conventional methods.