基于卡尔曼滤波和小波的网络流量预测算法研究  

Application of Kalman Filter and Wavelet in Traffic Prediction

作　　者：李捷[1] 候秀红[1] 韩志杰[1] 

机构地区：[1]河南大学计算机与信息工程学院,开封475001

出　　处：《电子与信息学报》2007年第3期725-728,共4页Journal of Electronics & Information Technology

基　　金：国家自然科学基金重点项目(No.60434020);河南省自然科学基金(0411014100);河南大学校内重点基金(XK03YBSW0138)资助课题

摘　　要：流量预测是流量工程,拥塞控制和网络管理的核心问题。该文针对网络流量的特点,将卡尔曼滤波和小波分析混合的预测算法引入到网络流量预测领域中,对其进行了理论证明。仿真结果表明,该算法与传统的算法相比,具有较高的预测精度和较好的实时性与广谱性。Traffic prediction is the core of network quality of service problems, such as traffic engineering and congestion control, etc. According to the characters of traffic, a novel network traffic prediction algorithm in which Kalman filter and wavelet are mixed is presented and proved abstractly. The simulation results show that the proposed algorithm can guarantee higher precision and better real-time processing compared with traditional algorithm.

关 键 词：流量预测 小波 卡尔曼滤波 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]