考虑需求变化与时变路网的同城货运路径优化  

Vehicle routing optimization of intra-city freight considering demand change and time-varying road network

作　　者：习江鹏 张佳蕊[2] 董红霞 王宁 XI Jiang-peng;ZHANG Jia-rui;DONG Hong-xia;WANG Ning(Shaanxi Transportation Holding Group Co.Ltd.,Xi'an 710065,Shaanxi,China;School of Transportation Engineering,Chang'an University,Xi'an 710064,Shaanxi,China)

机构地区：[1]陕西交通控股集团有限公司,陕西西安710065 [2]长安大学运输工程学院,陕西西安710064

出　　处：《长安大学学报（自然科学版）》2023年第4期137-150,共14页Journal of Chang’an University(Natural Science Edition)

基　　金：国家自然科学基金项目(71971030);榆林市科技计划项目(CXY-2020-025)。

摘　　要：为了使客户得到更加优质的服务,设计科学合理的算法进行车-货匹配并优化车辆路径,结合同城货运平台订单需求实时变化特点,以车辆使用成本、燃油成本、延误或提前到达货主位置产生的违约成本、为避免道路拥堵或产生过大违约成本产生的等待成本最小为目标函数,建立考虑车-货匹配、城市时变路网等约束的车辆路径优化模型。设计一种混合量子粒子群优化算法,决策最优货物匹配方案、车辆路径及车辆运行时间计划方案。研究结果表明:设计改进的量子粒子群算法得到的小规模算例优化解与Cplex优化软件得到最优解的相对偏差为3.7%,但所设计算法的平均求解时间仅需54.84 s,而Cplex的平均求解时间则需8640.80 s;在合理规划周期内,所提出的路径规划模型可通过调整车辆出发时间有效地将违约成本转化为代价较低的等待成本,或通过牺牲油耗成本降低违约和等待成本从而降低平台运营总成本;当单位等待成本为单位违约成本的30%时,考虑时变路网可有效降低平台运营总成本,当单位等待成本为单位违约成本的20%时,考虑时变路网可降低0.7%的平台运营总成本,当不考虑等待成本时,考虑时变路网可降低10.6%的平台运营总成本。In order to provide customers with higher quality services,a scientifically and reasonably designed algorithm was utilized to match vehicles with cargo and optimize vehicle routes.The real-time changes in orders from the local freight platform were combined comprehensive factors such as vehicle utilization costs,fuel costs,default costs incurred by delays or early arrivals at the consignee's location,and waiting costs to avoid road congestion or excessive default costs.Aimed to minimize the objective function,a vehicle path optimization model considering vehicle-cargo matching,time-varying urban road networks and other constraints was built.A hybrid quantum-behaved particle swarm optimization algorithm was designed to determine the optimal cargo matching scheme,vehicle routes,and vehicle operation time schedule.The results show that the improved quantum particle swarm algorithm produces optimized solutions for small-scale cases with a relative deviation of 3.7%,compared to the optimal solutions obtained by theCplex optimization software.However,the average solving time for the proposed algorithm is only54.84 s,while the average solving time for Cplex is 8640.80 s.Within a reasonable planning period,the proposed route planning model can effectively convert default costs into lower-cost waiting costs by adjusting the departure time of vehicles.Alternatively,it can reduce default and waiting costs by sacrificing fuel consumption costs,thereby reducing the overall operational costs of the platform.When unit waiting costs account for 30%of unit default costs,considering time-varying road networks can effectively reduce the overall operational costs of the platform.When unit waiting costs account for 20%of unit default costs,considering time-varying road networks can reduce the platform's operational costs by 0.7%.When waiting costs are not considered,considering time-varying road networks can reduce the platform's operational costs by 10.6%.11 tabs,11 figs,28 refs.

关 键 词：交通工程 车辆路径 量子粒子群算法 动态需求 城市零担运输 时变路网 

分 类 号：U492.3[交通运输工程—交通运输规划与管理]