Abstract: Collaborative delivery models, implemented by multiple logistics companies via information platforms, can significantly cut costs and boost efficiency. Yet, as cooperation scales up, fulfilling customer’s expected delivery times proves challenging. Existing time-window allocation strategies often neglect customer satisfaction in favor of cost. In response, this paper explores a multi?owner collaborative vehicle routing problem with time window assignment (MOCVRPTWA), aiming to optimize costs while ensuring customer satisfaction. Because MOCVRPTWA model belongs to NP-hard problem, to reduce the computational complexity to ensure the quality of solution, this paper introduces a hybrid heuristic algorithm using K-means and improved NSGA-II to tackle the complex MOCVRPTWA. This paper also proposes an improved Shapley value method for optimal profit distribution in stable delivery alliances. Experimental Results demonstrate high accuracy and robustness, highlighting the potential of collaborative delivery and time-window allocation to improve resource usage, delivery efficiency, and urban logistics sustainability.

Key words: collaborative delivery, time window assignment, profit distribution, multi-objective optimization

摘要： 多方物流企业依托信息平台实施协同配送模式能够显著降本增效，但随着合作规模扩大，顾客预期时间愈发难以满足。有研究采取了时间窗分配策略，但大多策略从成本角度出发，忽视了顾客满意度的重要性。提出了一个多目标时间窗分配策略下多方协同车辆路径问题（multi-owner collaborative vehicle routing problem with time window assignment，MOCVRPTWA），可以优化成本且保障顾客满意度。由于MOCVRPTWA模型属于NP-hard问题，为降低计算复杂度保证求解质量，设计了一种由K-means聚类算法与改进的NSGA-II算法组成的混合启发式算法来优化MOCVRPTWA模型。提出了一种改进Shapley值法来寻找稳定协同配送联盟的最优利润分配方案和最优联盟序列。实验结果表明，提出的算法在数据验证上表现出了较高的准确率和鲁棒性，MOCVRPTWA模型同时优化了配送成本和顾客满意度，表明协同配送机制和时间窗分配策略可以提高资源利用率和配送效率，有助于城市物流网络的高效运行和可持续发展。

关键词: 协同配送, 时间窗分配, 利润分配, 多目标优化