中国铁道科学2018，Vol.39Issue(2)：119-127,9.DOI:10.3969/j.issn.1001-4632.2018.02.15
城市轨道交通新站开通初期实时进出站客流量预测
Real-Time Forecast of Entrance and Exit Passenger Flow for Newly Opened Station of Urban Rail Transit at Initial Stage
摘要
Abstract
To solve the problems such as the lack of historical data and great fluctuation of passenger flow in real-time passenger flow forecast for newly opened stations of urban rail transit,an approach for the real-time forecast of entrance and exit passenger flow was proposed based on improved K nearest neighbor non-parametric regression for newly opened stations at initial stage.The change rule of the passenger flow of newly opened stations at initial stage and its correlation with the land use around station were analyzed considering the generation mechanism of passenger flow for urban rail transit.Based on the correlation,the mapping relationship between the new station and similar existing station was constructed through clustering analysis,and a construction method for history database was proposed to support the passenger flow forecast of new station.On this basis,combined with the characteristics of real-time passenger flow,the non-parametric regression algorithm was improved,and an approach for the real-time forecast of entrance and exit passenger flow for newly opened stations at initial stage was proposed.The passenger flow data collected from Guangzhou Metro were used for the case analysis of the real-time forecast of entrance passenger flow for newly opened stations at initial stage.Results show that the proposed approach has better prediction accuracy with the mean absolute error not more than 16 passengers,which can satisfy the demand for the refined operation and management of urban rail transit.关键词
城市轨道交通/新站开通/进出站客流量/实时客流量预测/非参数回归Key words
Urban rail transit/Newly opened station/Entrance and exit passenger flow/Real-time forecast of passenger flow/Non-parametric regression分类
交通工程引用本文复制引用
姚恩建,周文华,张永生..城市轨道交通新站开通初期实时进出站客流量预测[J].中国铁道科学,2018,39(2):119-127,9.基金项目
北京市自然科学基金重点资助项目(8171003) （8171003）
