基于最大熵算法网络流量预测模型研究  

Network Traffic Based on Maximum Entropy Algorithm for Prediction Model

作　　者：蔡晓丽[1] 宁慧[2] 陈舜青[1] 

机构地区：[1]常州工学院,江苏常州213002 [2]哈尔滨工程大学,黑龙江哈尔滨150001

出　　处：《计算机仿真》2011年第10期107-110,共4页Computer Simulation

基　　金：国家基金(60975071);863项目(2009AA04Z215)

摘　　要：研究了网络流量准确预测优化问题。由于网络资源的调度和传输速率时变性较强,互联网中对网络流量的不确定影响因素较多,同时由于传统的流量预测模型精度不高等缺陷。为解决上述问题,提出了一种新的基于最大熵算法建立网络流量预测模型。先获得约束的条件,是从对记录误差采用归一化处理方法得到的,而记录误差是模型根据历史来得到的预测结果,然后用最大熵算法原理来推测得到的结果分布,最后通过算法对网络的实际流量进行仿真预测。实验的仿真结果说明,此改进方法在预测精确度上,比传统的集中网络流量预测计算方法更胜一筹,为网络流量预测优化问题提供了依据。Optimization of network traffic prediction is studied.For the uncertain effects of many factors on internet network traffic,and because the accuracy of traditional traffic forecasting model is not high,a new algorithm is proposed based on maximum entropy principle of network traffic forecasting model.First of all,the model carried out normalization based on historical records of the forecast error,then,the prediction results were used as the constraint information,and the maximum entropy algorithm was used to obtain the distribution of predict results.Finally,the real network traffic was predicted by simulation.The simulation results show that this method has higher prediction accuracy than the traditional network traffic prediction algorithm.

关 键 词：最大熵 网络流量 预测 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]