共享电动汽车系统车队规模与停车泊位数优化  

Fleet size and parking capacity optimization of electric carsharing system

作　　者：马舒予 胡路[1,2] 吴佳媛 刘珺 MA Shu-yu;HU Lu;WU Jia-yuan;LIU Jun(School of Transportation and Logistics,Southwest Jiaotong University,Chengdu 611756,China;National Engineering Laboratory of Integrated Transportation Big Data Application Technology,Chengdu 611756,China)

机构地区：[1]西南交通大学,交通运输与物流学院,成都611756 [2]综合交通大数据应用技术国家工程实验室,成都611756

出　　处：《交通运输工程与信息学报》2022年第3期31-42,共12页Journal of Transportation Engineering and Information

基　　金：国家自然科学基金项目(71901183)。

摘　　要：基于流体排队逼近理论,将共享电动汽车系统用户在车站取车和在路径上用车分别描述为与车站、路径状态相关的流体排队模型,从运营商、用户、公众利益三方出发,考虑用户出行需求与拥堵的交互影响,建立以运营商日利润最大化为目标,以用户还车和引入共享车后的道路拥堵为约束的混合整数非线性模型(Mixed-Integer Nonlinear Programming,MINLP),联合优化共享电动汽车系统车队规模和停车泊位数。设计网格自适应搜索算法(Mesh Adaptive Direct Search,MADS)和遗传算法(Genetic Algorithm,GA)对优化模型进行求解,结果显示,MADS算法收敛速度更快,在优化结果和求解效率上更具优势,平均求解时间为2.71 h,仅为遗传算法的1/3。利用成都市出租车行程数据,分析需求、道路拥堵、定价对优化结果的影响,结果表明:(1)系统停车泊位数与车队规模比值处于1.8~2.7时能够解决车站溢出严重的问题,并获得最大利润;(2)忽略需求、拥堵交互影响的系统将全面降低运营商利润、车辆利用率和系统服务率;(3)在一个城市处于中度拥堵时(全网平均道路占有率处于0.68~0.80),考虑交互影响的系统将获得最大利润,而忽略交互影响的系统将最大化高估用户需求。Based on fluid queuing theory,the process of electric carsharing system users in picking up EVs at a station and driving EVs on routes are described as fluid queuing models associate with the station and routes,respectively.From the perspective of operators,users,and the public,a mixedinteger nonlinear programming model with road congestion and station congestion constraints is developed to jointly optimize the fleet size and parking capacity of the electric carsharing system to achieve the maximum daily profits,while considering the synergistic effect of user travel demand and road congestion.The mesh adaptive direct search(MADS)algorithm and genetic algorithm(GA)are designed to solve the model.The average CPU time of the MADS algorithm is 2.71 h,which is only 1/3 that of the GA,thus demonstrating that the MADS algorithm converges faster and affords a higher solving efficiency.Taxi trip data in Chengdu are used to investigate the effects of demand,road congestion,and pricing on the optimization results.The results indicate the following:(1)the system can avoid the station overflow problem and achieve the maximum profit when the ratio of the parking capacity to the fleet size is 1.8~2.7;(2)disregarding the interaction effect between demand and congestion will reduce the operator’s profits,vehicle utilization,and system service rates;(3)the electric carsharing system can maximize the profits by considering the interaction effect while maximizing the overestimation of user demand when moderate congestion(i.e.,the average road occupancy rate of the entire network is 0.68~0.80)occurs in a city.

关 键 词：城市交通 配置优化 流体排队 网格自适应搜索算法 遗传算法 共享电动汽车系统 

分 类 号：U491.17[交通运输工程—交通运输规划与管理]