基于共享电动汽车时空分布规律的高峰时段车辆调度研究

Research on Peak Hour Vehicle Scheduling Based on the Temporal and Spatial Distribution Law of Shared Electric Vehicles

研究综合考虑了目前共享电动汽车运营公司在车辆调度方面不能很好的满足用户用车的动态需求的情况和忽略用户还车随机性造成的站点间“有车无位”“有位无车”的现象。因此为了满足用户用车需求的动态性,提出了基于共享电动汽车时空分布规律的高峰时段车辆调度数量预测模型,运用基于历史数据的用户需求预测模型和基于残差平方和的线性组合模型,对未来高峰时段内用户车辆使用数和用户订单使用结束时的区域还车数量进行组合预测。最后以青岛市崂山、李沧、市南和市北四区的共享电动汽车使用数据进行预测。模型求解得出预测车辆调度量比当天实际调度量减少了19.6%,有效减少了站点间车辆无效调度,一定程度上为运营公司节省了运营成本。

The research comprehensively considers the current situation where shared electric vehicle operating companies cannot effectively meet the dynamic needs of users in terms of vehicle scheduling and the phenomenon of"vehicles without available spots"and"spots without vehicles"between stations caused by ignoring the randomness of users returning vehicles.In order to meet the dynamics of user car demand,a prediction model of the number of vehicle dispatches during peak hours based on the spatial and temporal distribution law of shared electric vehicles is proposed,and a linear combination model based on the sum of squares of residuals is used to comprehensively predict the number of vehicles occupied during peak hours and the number of regional returns at the end of order.Finally,the shared electric vehicle usage data of Laoshan,Licang,Shinan and Shibei districts of Qingdao City was used for prediction.Through the model solution,the predicted vehicle scheduling volume is reduced by 19.6%compared with the actual scheduling volume on the same day,which effectively reduces the invalid scheduling of vehicles between stations and saves the operating costs for the operating company to a certain extent.