摘要
采用小波神经网络对网络流量数据的时间序列进行建模与预测。针对传统小波神经网络训练算法的不足,提出了自适应量子粒子优化算法——AQPSO,用于训练小波神经网络,优化网络参数,建立基于AQPSO算法优化的小波网络预测模型。实验结果表明,该模型对网络流量的短期预测是有效可行的,并具有良好的收敛性和稳定性。
The time series of network traffic data was modeled and forecasted based on Wavelet Neural Network ( WNN). The limitation of the conventional training algorithm of WNN was introduced and Adaptive Quantum-Behaved Particle Swarm Optimization (AQPSO) algorithm was proposed to train WNN. The parameters of WNN were optimized and the forecast model of WNN based on AQPSO algorithm was built. The experimental results prove that the model is efficient in network traffic prediction with good astringency and stability.
出处
《计算机应用》
CSCD
北大核心
2007年第12期2986-2988,共3页
journal of Computer Applications
基金
国家自然科学基金项目(60675011)
关键词
小波分析
神经网络
粒子群优化
网络流量预测
wavelet analysis
neural network
Particle Swarm Optimization (PSO)
network traffic forecast