CS算法优化BP神经网络短时交通流量预测_高述涛.pdfVIP

106 2013 ，49 （9 ） Computer Engineering and Applications 计算机工程与应用 CS 算法优化BP 神经网络的短时交通流量预测 高述涛 GAO Shutao 湖南外贸职业学院 服务外包学院，长沙 410014 School of Service Outsourcing, Hunan International Business Vocational College, Changsha 410014, China GAO Shutao. Short time traffic flow prediction model based on neural network and cuckoo search algorithm. Computer Engineering and Applications, 2013, 49 （9 ）：106-109. Abstract ：In order to improve the prediction accuracy of short time traffic flow ，this paper proposes a network traffic prediction model based on Cuckoo Search algorithm and BP Neural Network （CS-BPNN ）.The time series of short time traffic flow is recon- structed to form a multidimensional time series based on chaotic theory ，and then the time series are input into BP neural net- work to learn which parameters of BP neural network are optimized by cuckoo search algorithm to find the optimal parameters and establish the short time traffic flow prediction model.The performance of CS-BPNN is tested by the simulation experiments. The simulation results show that the proposed model improves the prediction accuracy of short time traffic flow and can more describe network traffic complex trend compared with reference models. Key words ：short time traffic flow; phase space reconstruction; cuckoo search algorithm; Gaussian disturbance; Back Propaga- tion （BP ）neural network 摘 要：为了提高短时交通流量的预测精度，提出一种布谷鸟搜索算法优化BP 神经网络参数的短时交通流量预测模型 （CS-BPNN ）。基于混沌理论对短时交通流量时间序列进行相空间重构，将重构后的时间序列输入到BP 神经网络进行学习， 采用布谷鸟搜索算法找到BP 神经网络最优参数，建立短时交通流量预测模型，通过具体实例对CS-BPNN 性能进行测试。 仿真结果表明，相对于对比模型，CS-BPNN 提高了短时交通流量的预测精度，更加准确反映了短时交通流量的变化趋势。 关键词：短时交通流量；相空间重构；布谷鸟搜索算法；高斯扰动；反向传播（BP ）神经网络 文献标志码：A 中图分类号：TP393 doi ：10.3778/j.issn. 1002-8331.1212-0188 [7] 1 引言 型对未来时刻的短时交通流量值进行预测 。BP 神经网络 随着城市化进程加快和交通流量增加，交通堵塞日益