一种网络流量预测模型的研究  

Research into a Network Traffic Prediction Model

作　　者：陈南岳[1] 汤永斌[2] 岳淼[3] 滕云[3] Chen Nanyue;Tang Yongbin;Yue Miao;Teng Yun(Modern Educational and Technological Center of North Sichuan Medical College,Nanchong 637000,China;Modern Educational and Technological Center of Nanchong Professinal Technic College,Nanchong 637131,China;School of Computer Science of China West Normal University,Nanchong 637002,China)

机构地区：[1]川北医学院现代教育技术中心,四川南充637000 [2]南充职业技术学院现代教育技术中心,四川南充637131 [3]西华师范大学计算机学院,四川南充637002

出　　处：《科技通报》2017年第7期86-89,共4页Bulletin of Science and Technology

摘　　要：云计算等概念的出现导致网络规模需求不断增大,网络流量不断增多,如何能够更好地预测网络流量成为了网络研究的热点,本文在网络流量预测中分析了小波-BP神经网络的不足,采用改进的下降梯度法优化BP神经网络,另外通过采用实体编码,适应度函数选择,权重系数提高粒子群算法优化小波-BP神经网络。仿真实验说明本文算法能够有效的提高了网络流量的预测精度,该算法具有广泛的运用前景。The emergence of cloud computing concept has led to the constant increase of network scaleand demands as well as network traffic,so how to better predict network traffic has become a focus ofresearches into the Internet.This paper analyzes the deficiency of wavelet-BP neutral network innetwork traffic prediction,and adopts improved descent gradient method to optimize the BP neutralnetwork.In addition,this paper adopts the physical coding,selects fitness function and weight coefficientto improve the particle swarm algorithm and optimize the wavelet-BP neutral network.Simulationexperiment shows that algorithm in this paper can effectively improve the precision of predicting networkflow,and this algorithm has broad application prospect.

关 键 词：网络流量 小波-BP神经网络 下降梯度法 实体编码 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]