Abstract: Under the condition of sharing vehicles and orders, considering the simultaneous delivery and pickup scenario with soft time window, this study establishes the MJVRPSDPTW model. Then a hybrid genetic algorithm based on large neighborhood search is designed. The virtual center is set and optimized by wholeness method. The “destruction” and “repair” operators in large neighborhood searchare integrated to enhance the optimization ability of the algorithm. Through several groups of comparative experiments, it is proved that the algorithm is superior to the two-stage programming method and the classical genetic algorithm. Finally, based on the optimized network, the profit distribution of logistics enterprises under different alliances is carried out by using Shapley value method, which proves the effectiveness of the model.

Key words: joint pick-up and delivery, vehicle routing problem, hybrid genetic algorithm, profit distribution, Shapley value

摘要： 共享模式是电商逆向物流的新发展趋势，在共享车辆与客户订单的条件下，考虑带时间窗约束的客户同时取送货情景，建立了MJVRPSDPTW（multi-centers joint vehicle routing problem with simultaneous delivery and pick-up and time window）模型。接着设计一种基于大邻域搜索的混合遗传算法进行求解，针对“多对多网络”设置虚拟中心并利用整体法优化，同时融合大邻域搜索算法中的“破坏”与“修复”算子，增强算法寻优能力。通过多组算例对比实验，证明该算法优于两阶段规划法与经典遗传算法。基于优化后的网络，利用Shapley值法对不同联盟情况下的各物流企业进行利润分配，结果证明大联盟最稳定且共同利润最大。

关键词: 联合取送货, 车辆路径优化, 混合遗传算法, 利润分配, Shapley值法