基于数据驱动的电动出租车充电站规划方法研究

Research on Data-driven Electric Taxi Charging Station Planning Method

为了规划城市道路网上电动出租车的充电基础设施,以满足整个城市电动出租车的充电需求,本文利用出租车乘客起讫点大数据,基于M/M/K排队模型提出电动出租车在充电站处排队模型,建立行驶距离最小的充电站选址优化模型及遗传求解算法。以苏州市出租车GPS数据为例,对模型进行验证。研究结果表明:随着充电速率的增加,将减少电动出租车充电需求的单位成本,从而增加充电站设站数量,由此将缩短电动出租车充电的行驶总距离;基于不同工作日GPS大数据求解的充电站选址在道路网中呈簇状分布,模型结果具有稳健性,这表明可以在邻近区域内寻找可布设点建设充电站。研究结果可为出租车充电站规划和运营提供决策依据。

In order to plan the charging infrastructure of electric taxis on urban roads to meet the charge needs of electric taxis in the entire city, utilizing the big data of taxi passengers, a queueing model of electric taxis charging at charging station based on the M/M/K queuing model and a location optimal model of electric taxi charging stations for minimum total travel distance and models’ algorithm are all proposed. The GPS data of Suzhou taxi is used to verify the model. The research results show that with the increase of the charging rate, the unit cost of the electric taxi charging demand will be reduced, while the number of charging stations increases, thus reducing the total distance of electric taxis. The locations of charging stations are distributed in clusters in the road network, thus the model results are robust. This means that applicable sites can be found to build charging stations in adjacent areas. The research results can provide decision basis for the planning and operation of taxi charging stations.