带二维装箱约束的客货滚装船配载决策

客货滚装船长期处于人工配载状态，其独特性导致其无法直接采用已有滚装船配载方法.为提高客货滚装港口配载的决策水平，针对配载过程的两阶段和二维装箱特性，以船舱面积利用率最大为目标，构建客货滚装船配载决策模型.为实现大规模求解，设计偏随机密钥混合遗传算法求解，算法由多阶段启发式嵌入偏随机密钥遗传算法构成.多阶段启发式首先通过首层配载、主体配载和补充配载三段启发式完成初始解构造，其次采用横纵倾优化策略调整船舶稳性.算例研究表明，偏随机密钥混合遗传算法表现良好，且在大规模应用实验中鲁棒性较好，可实现客货滚装船的有效配载.

Passenger-cargo RORO Ships Stowage Planning Decision with Two-dimensional Packing Constraints

The stowage planning for the passenger-cargo RORO(Roll-on/roll-off) ship still stays in the manual decision stage. Its particularity makes it impossible to adopt existing stowage methods for RORO ships. In order to improve the stowage planning decision of passenger-cargo RORO ports, the stowage planning decision model is constructed with the maximization of ship cabin area utilization considering two- phase and two- dimensional packing characteristics of stowage planning process. The biased random key hybrid genetic algorithm which composed of a multi- phase heuristic and biased random key genetic algorithm is designed to find large- scale solutions. The initial solution is constructed firstly through the multi- phase heuristic including the first layer stowage planning, main stowage planning and supplementary stowage planning. And then the ship’s stability is modified through the heeling and trimming moment optimization strategy. Numerical example shows that the algorithm performs well and has a good robustness for large- scale application experiments. It can achieve the effective stowage planning for the passenger-cargo RORO ship.