基于遗传算法的BP神经网络短时交通流预测  

The prediction of short-term traffic flow based on the genetic algorithm and BP neural network

作　　者：王成宝[1] 任传祥[1] 尹唱唱[1] 陈梦[1] 

机构地区：[1]山东科技大学交通信息研究所,山东青岛266590

出　　处：《山东交通科技》2012年第5期5-7,12,共4页

基　　金：山东科技大学研究生科技创新基金资助;项目编号:YCA120426;山东省优秀中青年科学家科研奖励基金计划资助;项目编号:BS2012DX031

摘　　要：对BP神经网络模型、方法及遗传算法的基本原理进行了分析,将遗传算法与标准BP神经网络算法相结合构建了基于遗传算法的神经网络算法;建立了基于BP神经网络的短时交通流预测模型,对BP神经网络算法及基于遗传算法的BP神经网络算法进行设计,并将其应用于短时交通流预测模型的仿真计算,仿真结果表明基于遗传算法的BP神经网络算法具有更快的计算速度及更好的适应能力,该方法可以较好地应用于短时交通流预测。In the paper, the model and method of the BP neural network and the principle of genetic algorithm are analyzed, and also the BP neural network algorithm based on the genetic algorithm is set up by combining the genetic algorithm and the BP neural network. The short -term traffic flow prediction model is given based on the BP neural network and the BP neural network algorithm and the BP neural network algorithm based on the genetic al- gorithm are designed and applied to optimize the predic- tion model. Finally, simulation is done and the results indicate the BP neural network algorithm based on the ge- netic algorithm has faster convergence speed and better a- daptability than standard BP neural network algorithm, also is a good method to the short -term traffic flow prediction problem.

关 键 词：短时交通流预测 遗传算法 BP神经网络 

分 类 号：U495[交通运输工程—交通运输规划与管理]