神经网络与四阶段预测法在铁路运量预测中的组合应用研究  

Study on the Joint Application of Neural Network and Four-Stage Predict Methodin Railway Transport Volume Prediction

作　　者：王增兵[1] 李远富[1] 

机构地区：[1]西南交通大学土木工程学院,四川成都610031

出　　处：《铁道运输与经济》2003年第12期43-46,共4页Railway Transport and Economy

摘　　要：利用神经网络与四阶段预测法组合构造出新的交通量预测模型,以胶济铁路提速改造为例,就构造的客运量预测模型进行了应用研究。其中以平均增长率法计算客流量的交通分布;以重力模型法计算诱发客流;依据运输阻力构建的分担率模型计算转移客流;在计算诱发客流时考虑了时间价值。A new transport volume prediction model is constructed bycombining neural networks and four-stage predict method. Taking the reformof speed-raising in Jiao-Ji Railway as example, the paper carries out anapplication study on establishing the prediction model for passenger transportvolume, wherewith the average increase rate method to calculate trafficdistribution of passengers flow; with gravity model method to calculate inducedbroker flow, with sharing rate model that is constructed based on transportresistance to calculate the transfering passenger flow. The time value is alsotaken into consideration while calculating the induced broker flow.

关 键 词：神经网络 四阶段预测法 交通分布 诱发客流 交通分担量 铁路运量预测 

分 类 号：U293.13[交通运输工程—交通运输规划与管理]