大规模光通信网络流量的非线性建模与预测  

Nonlinear Modeling and Forecasting of Large Scale Optical Communication Networks

作　　者：王槐源[1] 

机构地区：[1]琼州学院电子信息工程学院,海南三亚572022

出　　处：《激光杂志》2016年第7期101-104,共4页Laser Journal

基　　金：海南省自然科学基金项目(613170)

摘　　要：流量建模与预测对了解光通信网络变化状态具有重要意义,为了有效描述光通信网络流量的混沌性,提出一种大规模光通信网络流量的非线性建模与预测方法。首先利用相空间重构理论对光通信网络流量数据进行预处理,然后采用最小二乘支持向量机描述光通信网络流量的变化特点,并采用遗传算法选择核函数参数和惩罚参数,最后采用光通信网络流量预测实例对其性能进行实证分析,并与其它方法进行对比实验。实验结果表明,本文方法具有较高的光通信网络流量预测精度,可以描述光通信网络流量的混沌性,具有一定推广价值。Traffic modeling and prediction has important significance to know the changes of optical communication network. In order to effectively describe the chaos in optical communication network, a new method for nonlinear modeling and prediction of large scale optical communication network traffic is proposed. First of all, the theory of phase space reconstruction is used to preprocess optical communication network traffic data, and then least square support vector machine is used to describe the variation of optical communication network traffic, and genetic algorithm is used to adaptively select the kernel parameters and penalty coefficient, finally, a practical example is used to analyze the performance of the method, and the other methods are compared with the experimental results. Experimental results show that this method has high accuracy in traffic prediction, and can be described by the chaos of optical communication network traffic, has certain promotion value.

关 键 词：大规模光通信网络 网络流量 建模与预测 非线性变化特征 

分 类 号：TN249[电子电信—物理电子学]