北京地铁换乘站客流预测模型研究

根据北京地铁全网一票换乘和一卡通无障碍换乘机制,以及早晚高峰出行等特点,对各种地铁客流预测模型进行分析,研究北京地铁换乘站客流预测模型的应用。依据换乘站不同类型的客流,以北京地铁历史客流和实时客流数据为基础,探讨采用历史平均预测法、基于最小二乘支持向量机时间序列预测法、分峰段混合预测法、基于概率树全路网预测方法等对进站客流、出站客流、换乘客流和站内客流进行预测。

Research on Forecast Model of Passenger Flow in Beijing Metro Transfer Station

Based on the system of one ticket transfer in whole Beijing Metro network and One-card barrier free transfer mechanism,as well as the feature of morning and evening peak traffic hours,the analysis of various prediction models on subway passenger flow is conducted to research the application of the prediction model on passenger flow in Beijing Subway transfer stations.According to different types of passenger flow in transfer stations,on the basis of the data of Beijing Metro historical passenger flow and real time passenger flow,the application of prediction method of historical average data,support vector machine time series based on least squares,peak shunting complex,whole metro network prediction based on probability tree are discussed to predict passenger flow going in and out station,as well as the passenger flow of transfer and staying in stations.