Abstract: The inland container liner route stowage planning needs to make the dynamic decision as the quantity of foreign trade containers can be changed by the customs inspection continuously. Based on the rolling scheduling strategy, the stowage planning of current port is divided into multiple stages according to the random event, namely, customs inspection. The stowage planning model for a single stage in current port is developed with the objective of minimizing the ship stack occupancy number and the differences of stowage plans between adjacent stages. Then the liner route stowage planning decision is made by rolling. The matheuristic algorithm which contains the integer programming, the destructor and the rebuilder is proposed to make the multiple stages rolling stowage planning for each port. Numerical examples based on real-world shipping scenarios show the difference between the proposed model and algorithm for minimizing the ship stack occupancy number is small. And for the differences of stowage plans between adjacent stages, the results from the algorithm are better than the model. Therefore, the proposed model and algorithm can be utilized to assist in making the dynamic stowage planning decision for inland container liner route shipping with uncertain container quantity. They can both make the stowage plans to absorb the uncertain container quantity robustly and the algorithm has a better performance than the model.

Key words: container shipping, route stowage planning, uncertain container quantity, matheuristic, rolling scheduling, dynamic decision

摘要： 内河集装箱班轮运输中海关抽检可导致外贸箱箱量不断发生变化,班轮航线配载需要动态决策。基于滚动调度策略,将当前港口的配载决策按随机事件划分为多个阶段,以最小化班轮堆栈占用数量和相邻阶段间配载计划偏差为目标,构建单港口单阶段的配载决策模型,进而滚动实现班轮航线动态配载决策。基于大邻域搜索思想设计一种包含整数规划、破坏器与修复器的精确启发式算法,实现港口多阶段滚动配载。基于真实场景的算例研究表明,在优化堆栈占用数量方面,模型与算法之间差异不大,但在考虑相邻阶段间配载计划偏差时,算法的求解结果要优于模型。因此,模型与算法可用来辅助实现不确定箱量下内河集装箱班轮航线动态配载决策,且算法表现更优,可实现配载计划对不确定箱量的鲁棒吸收。

关键词: 集装箱运输, 航线配载, 不确定箱量, 精确启发式, 滚动调度, 动态决策