考虑随机行程时间的电商同城快递取派联合优化模型

Joint Optimization Model for Pickup and Delivery of E-Commerce Intra-City Express Delivery Considering Random Travel Time

摘要: 随着电商订单即时配送需求的激增，城市交通波动会削弱同城快递取派调度的可行性与服务可靠性，在行程时间随机的场景下电商订单延误风险更为突出。在行程时间不确定条件下兼顾电商平台运营成本与订单履约时效，生成稳健的订单分配与配送车路径方案。针对电商末端配送场景，构建了基于蒙特卡洛采样评估随机行程时间的取派调度模型，并设计了变邻域搜索与模拟退火融合的VNS-SA启发式算法进行求解。基于Solomon VRPTW C类基准算例，结合电商订单特征改造生成快递订单与配送车初始位置。实现电商订单全覆盖，启用9辆配送车，总成本2044.04，总里程336.18 km，总订单在途时间833.65 min，时间窗违规值4.05 min。

Abstract: With the surge in demand for instant delivery of e-commerce orders, urban traffic fluctuations undermine the feasibility and service reliability of urban courier pickup-and-delivery dispatch, making the risk of e-commerce order delays more pronounced under stochastic travel time conditions. To balance e-commerce platform operational costs and order fulfillment timeliness under travel time uncertainty, generating robust order assignment and delivery vehicle routing plans. Targeting the e-commerce last-mile delivery scenario, a courier pickup-and-delivery scheduling model was developed that evaluates stochastic travel times via Monte Carlo sampling, and a hybrid Variable Neighborhood Search combined with Simulated Annealing (VNS-SA) heuristic algorithm was designed for solution. Courier orders and initial delivery vehicle locations were generated based on modified Solomon VRPTW C-class benchmark instances, incorporating e-commerce order characteristics. Full e-commerce order coverage was achieved using 9 delivery vehicles, with a total cost of 2044.04, a total travel distance of 336.18 km, a total in-transit time of 833.65 min, and a total time-window violation of 4.05 min.