Abstract: This paper focuses on the truck scheduling problem of multi- size container tasks in a local area near a terminal under separation mode. During the transportation process, the remaining capacity of the truck changes continuously with the tasks performed by the truck. Considering the matching between the container size of the task and the current state of the container truck, a mixed integer programming model is established with the objective of minimizing the total cost during the use, driving, and wait of all trucks. According to the model characteristics, an ant colony algorithm based on an infeasible arc filtering strategy is designed to improve the solution performance. Numerical experiments are carried out by randomly generating numerical examples with different task types, proportions, and sizes from Solomon data sets. The results prove the correctness of the model and the stability and effectiveness of the algorithm, and then the cost of the separation mode is compared with that of the traditional mode. The results show that the total transportation cost of the separation mode is 45.10% lower than that of the traditional mode. When the task scale increases, the difference in the cost between the two modes becomes more obvious, which shows the advantages of the combination of the separation mode and the multi-size container collection and distribution operations.

Key words: transportation economy, scheduling optimization, ant colony algorithm, multi-size container tasks, separation mode

摘要： 针对甩箱模式下港口堆场与内陆腹地间多箱型任务组合的集卡调度问题，本文根据运输过程中集卡剩余箱位随集卡执行任务动态变化的特点，考虑不同箱型任务需求与集卡当前状态的匹配，以集卡的启用、行驶及等待过程的成本最小化为目标，建立混合整数规划模型。并根据模型特点设计基于不可行弧过滤策略的蚁群算法，以提升算法性能。利用Solomon标准数据集 随机生成任务类型比例和规模大小不同的算例进行数值试验，试验结果证明了模型的正确性与算法的稳定性和有效性，进而对比了甩箱模式与传统模式下的总成本。结果表明：相较于传统模式，甩箱模式下集卡的运输总成本平均减少了45.10%；当任务规模越大时，两种模式之间的成本差距越大，显示了甩箱模式与港口多箱型集疏运任务运输作业结合的显著优势。

关键词: 交通运输经济, 调度优化, 蚁群算法, 多箱型任务组合, 甩箱