基于径向基神经网络的交通出行预测  

Trip Genearation Forecasting Based on Radial Basis Function Neural Network

作　　者：芦有鹏[1] 杨菊 LU You-peng;YANG Ju(School of Traffic and Transportation,Lanzhou Jiaotong University,Lanzhou 730070,China)

机构地区：[1]兰州交通大学交通运输学院,甘肃兰州730070

出　　处：《兰州交通大学学报》2018年第2期27-30,42,共5页Journal of Lanzhou Jiaotong University

基　　金：国家自然科学基金(71671079;71361018)

摘　　要：实现交通供需平衡是解决城市交通拥堵问题的关键,而分析交通出行结构则是平衡供需矛盾的基础,针对传统交通出行预测模型的误差大、考虑因素单一等缺陷,对多个非定量指标因素进行归纳,进而利用神经网络在非线性关系映射方面的优势,建立了径向基神经网络交通出行预测模型,并通过实例验证了方法的可行性和可靠性.最后利用建立好的神经网络模型对各个影响因素对交通出行的影响程度大小进行了分析和探讨,并得出了与实际相符的结论.Balance between total traffic demand and supply is the key to the problem of urban traffic congestion,while analysis of traffic travel structure is the basis of keeping the balance between supply and demand.Considering the defects of traditional forecasting models in large forecasting errors and consideration of single factors,we sumed up a number of non quantitative indicators,and then established a radial basis function neural network model for forecasting by making use of the advantages of nueral network in nonlinear relation.The feasibility and reliability of the method were verified by a specific example.We finally analyzed and discussed the influence degree of each factor on resident trip by using the model established,and arrived in conclusions coincident with the fact.

关 键 词：管理科学与工程 神经网络 径向基函数 出行预测 

分 类 号：U491[交通运输工程—交通运输规划与管理]