布谷鸟搜索算法优化BP神经网络的网络流量预测  

Network flow predicting model based on cuckoo search algorithm optimizing neural network

作　　者：杜振宁[1] 

机构地区：[1]杨凌职业技术学院信息工程学院,陕西杨凌712100

出　　处：《电子技术应用》2015年第3期82-85,共4页Application of Electronic Technique

摘　　要：为了提高预测精度,提出一种布谷鸟搜索算法优化BP神经网络的网络流量预测模型(Cuckoo Search BP neural network Flow Prediction,CS-BPNN)。根据混沌理论建立网络流量学习样本,采用BP神经网络对学习样本进行训练,将模型参数当一个鸟巢,通过模拟布谷鸟寻窝产卵的行为找到最优模型参数,最后采用网络流量数据进行仿真实验,测试模型性能。仿真实验表明:所提出模型较好的解决了BP神经参数优化问题,能够获得更加理想的网络流量预测结果。In order to improve the predicting precision, a novel network flow predicting model based on cuckoo searc algorithm optimizing neural network was proposed in this paper. Firstly, the learning samples were obtained by phase space reconstruction. Secondly, the samples were input to BP neural network to learn, and the parameters were encoded as cuckoo, the optimal parameters were obtained by simulating the cuckoo＇s finding the nest and producing eggs. Finally, the network flow predicting model was built and the simulation experiments were carried out on network flow data. The results show that the proposed model had solved the parameters optimization problem of BP neural network and obtained good predicting results of network flow.

关 键 词：布谷鸟搜索算法 网络流量 神经网络 参数优化 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]