不确定箱重下内河集装箱班轮航线配载决策

与集装箱海运相比内河集装箱班轮运输具有其独特性，同时对于内贸箱而言，货主订舱时箱重信息的不确定性导致其航线配载决策变得更加复杂.本文考虑不确定箱重影响，以最小化航线班轮堆栈占用数量为目标，构建内河集装箱班轮航线配载决策的随机规划模型.为实现求解，基于随机规划理论，采用机会约束描述随机约束，将随机规划模型转化为随机机会约束规划模型，并设计混合邻域搜索算法求解.算法由蒙特卡罗随机模拟、神经元网络训练及邻域搜索启发式3个部分组成.算例研究表明，混合邻域搜索算法的鲁棒性较好，可实现配载计划对不确定因素的有效吸收.

Inland Container Liner Route Stowage Planning Decision with Uncertain Container Weight

The inland container liner shipping has its particularity comparing with the maritime container shipping. For the domestic trade containers, the uncertainty of weight information which is provided by the cargo owners when booking the shipping space has made the route stowage planning decision more complex. Considering the uncertain container weight, the stochastic programming model for inland container liner route stowage planning decision is built with the objective of minimizing the ship stack occupancy number over the full route. The stochastic programming model is translated into the stochastic chance-constrained programming model based on the stochastic programming theory by describing the stochastic constraints with chance constraints. The hybrid neighborhood search algorithm consisting of Monte Carlo stochastic simulation, neural network training, and neighborhood search heuristics is designed for solving the proposed model. Numerical examples show the hybrid neighborhood search algorithm has a good robustness as it can make the stowage plan to absorb the uncertainty effectively.