ARIMA模型在城市轨道交通短期客流预测中的应用研究  

Short-term passenger flow prediction of urban rail transit based on ARIMA model

作　　者：赵晓丽 Zhao Xiaoli

机构地区：[1]天津轨道交通运营集团有限公司,天津300392

出　　处：《现代城市轨道交通》2023年第8期77-82,共6页Modern Urban Transit

摘　　要：城市轨道交通站点的客流监控与短期精准预测对实现高效列车调度、防范安全隐患、降低运营成本具有关键作用。文章通过优化ARIMA模型参数对不同时间粒度下天津地铁某站点周一到周五的AFC客流数据进行预测,预测结果表明ARIMA方法能够准确预测城市轨道交通站点的短期客流。减小客流数据时间粒度能够获得更多的客流数据细节信息,有利于提升预测的精准度。研究发现即便在客流数据能够通过平稳性检验的情况下,对客流数据进行差分处理依然能够明显提升ARIMA模型的预测精度。Passenger flow monitoring and the associated short-term accurate prediction in urban rail transit play a key role in achieving efficient train scheduling,preventing potential safety hazards and reducing operating costs.In this paper,the parameters in ARIMA model are optimized based on the AFC passenger flow data of a Tianjin metro station from Monday to Friday,which are further utilized for the predictions at different time scales.The results demonstrate an accurate prediction on the short-term passenger flow by using the ARIMA method.The reduction of time granularity gives rise to more detailed information of passenger flow,which helps to improve the prediction accuracy.We further found the differential processing on the passenger flow data enhances the prediction accuracy of the ARIMA model significantly even though the passenger flow data can pass the stationarity test.

关 键 词：城市轨道交通 短期客流预测 ARIMA 时间粒度 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]