机场到港旅客乘坐出租车短时需求预测

Short-term demand forecast of taxis for passengers arriving at airport

对机场旅客陆侧交通需求进行准确预测是机场陆侧交通资源合理分配的重要依据.针对机场到港旅客乘坐出租车短时需求开展研究,从群体角度出发,考虑其他交通方式的影响,基于交通方式的交互直接预测机场乘坐出租车到港旅客流量.采集机场到港旅客出租车订单数据、机场到港飞机数据、机场天气报文数据和机场到港旅客地铁闸门数据,搭建LightGBM模型并根据具体情景采取网格法或贝叶斯优化进行参数标定,预测机场到港旅客乘坐出租车流量,结果优于6个传统经典模型,相比于以往研究,模型考虑乘坐地铁旅客数量对乘坐出租车旅客需求的影响,预测精度更高.研究结果可推广到机场其他陆侧交通方式,完善机场集疏运系统.

Accurate prediction of airport passenger landside transportation demand is an important basis for the rational allocation of airport landside transportation resources.A study was carried out on the short-term demand of passengers arriving by taxi at the airport.From the perspective of the group,considering the influence of other transportation modes,the traffic of arriving passengers by taxi at the airport was directly predicted based on the interaction of transportation modes.The collected the taxi order data of airport arrival passengers,airport arrival aircraft data,airport weather message data and airport arrival passenger subway gate data,build a LightGBM model,and used grid method or Bayesian optimization to calibrate parameters according to specific scenarios to predict the flow of passengers arriving by taxi from the airport,and the results are better than the six traditional classic models.Compared with previous studies,the model considered the impact of the number of passengers taking subways on the demand of passengers taking taxis,and the prediction accuracy was higher.The research results can be extended to other landside transportation modes of the airport and improve the airport collection and distribution system.