Abstract: To fulfill the individual requirements for different last-mile delivery methods and time slots in urban distribution,a multi-objective integrated optimization model of reception box locating,time slot allocation and vehicle routing was constructed by considering the demand depending on the last-mile delivery and time slots.The model represented the effect of reception box distance and delivery time deviation to the customers' choice procedure.Multi-objective Particle Swarm Optimization (MOPSO) with modified coding method and initializing swarm construction was applied to solve a Solomon's benchmark problem.The result showed that the solution with the minimum cost and max delivery quantity at the same time did not realized;the solutions in the two objectives should decrease the number of reception box and provide more attended home service as the customer's sensitivity on the reception box distance increased;the solutions in the two objectives could not avoid the delivery quantity decreasing and the trend of the reception box and vehicles quantity was not significant as the customer's sensitivity on the delivery time deviation increased.

Key words: urban distribution, last mile delivery, time slot, reception box locating, routing planning

摘要： 为了满足城市配送中顾客对末端交付方式和服务时间窗的个性化需求,考虑送货上门和自提柜两种末端交付方式,基于自提柜选址及配送路径对配送节点末端交付方式和时间窗分配的制约性,以配送数量最大化和配送成本最小化为目标,构建了自提柜选址—时间窗分配—路径规划多目标联合优化问题模型,刻画了自提柜距离、配送时间误差对顾客配送服务需求的影响,优化设计了多目标粒子群算法的编码和初始种群构建方法,通过Solomon算例库仿真分析发现：Pareto解集中无法分离出配送数量最大化和配送成本最小化均占优的解;随着顾客对自提柜距离敏感度的增加,两种目标偏好都应减少自提柜数量,尽量采用送货上门服务;随着顾客对配送时间误差敏感度的增加,两种目标偏好均无法避免配送数量下降,车辆和自提柜数量变化趋势并不显著。

关键词: 城市配送, 末端交付, 时间窗, 自提柜选址, 路径规划