Abstract: In order to integrate the forward and reverse logistics of logistics enterprises in the e-commerce environment, a vehicle path planning model with soft time windows and the return and replacement is established acording to the actual situation. An ant colony algorithm improved by the variable neighborhood search algorithm is designed to solve such problem. Based on the basic ant colony algorithm, the initial pheromone setting, state transition rules and pheromone update strategy are improved according to the characteristics of the problem. The variable neighborhood search algorithm is combined to improve the search ability of the algorithm. The proposed algorithm has better performance in the experiments of the relevant literature data, standard examples and the real data of enterprises. It is an effective algorithm for solving the proposed problem.

Key words: e-commerce, soft time windows, return and replacement, Vehicle Routing Problem（VRP）, ant colony algorithm

摘要： 为了整合电子商务环境下物流企业的正、逆向物流，依据实际情况建立了带软时间窗和退换货的车辆路径规划模型，并且设计了一种混合变邻域改进蚁群算法来求解此类问题。在基本蚁群算法的基础上，在初始信息素的设置、状态转移规则以及信息素的更新策略上，根据所研究问题的特点做了相应的改进。同时，结合混合变邻域算法提高了算法的搜索能力。通过对相关文献数据、标准算例以及实际企业数据实验，验证了所提出算法具有较好的性能，是求解所提出问题的一种有效算法。

关键词: 电子商务, 软时间窗, 退换货, 车辆路径问题（VRP）, 蚁群算法