摘要
针对船舶制造中由于钢板出入库时间段的分散性而引起钢板堆场大量倒垛的问题，建立了以减少计划期内倒垛量为目标的多时段钢板堆场作业计划优化模型。模型考虑钢板入库、倒垛和出库作业过程，以及时间约束和空间约束。根据对钢板入库、倒垛和出库作业流程及堆放原则的研究，提出并实现一种由动态规划启发式算法和变邻域搜索算法组成的两阶段求解方法。最后以某船厂的实际数据为例，分别在不同钢板规模情况下与固定参数的遗传算法及人工决策对比，结果表明，新的方法运算时间短，在较大规模问题时可减少47％～75％的倒垛量。
To deal with the shuffling problems, which are caused by dispersion of different stacking time and pick-up time of steel plates in the stockyard of shipbuilding, an optimization model of multistage operation scheduling for a steel plate stockyard is formulated, whose objective is to minimize the shuffling operations. The model includs three basic operations, such as stacking, shuffling, and pick-up, while considering the temporal and spatial con-straint of the stockyard. According to the research of these three basic operations＆#39; process and the stacking princi-ple, a new two-stage method consisting of a dynamic programming based heuristics algorithm and variable neighbor-hood search algorithm （ VNS） was proposed and realized. Finally, application data were obtained from a shipyard to validate and verify the proposed method. The results of numerical experiment conducted at different steel plate scales show that the new method has a shorter run time and on large scale cases can reduce 47%-75% shuffling operations, by contrasting with fixed parameters genetic algorithm and artificial decision making.
出处
《哈尔滨工程大学学报》
EI
CAS
CSCD
北大核心
2015年第5期638-643,共6页
Journal of Harbin Engineering University
基金
国家自然科学基金资助项目(70872076)
上海科技创新行动计划基金资助项目(11dz1121803)
江苏省南通市2010AA科技创新计划(工业)基金资助项目(AA2010048)
关键词
钢板堆场
多时间段
堆场作业计划
动态规划启发式算法
变邻域搜索算法
steel plate stockyard
multistage
steel plate operation scheduling
dynamic programming based heuris-tics
variable neighborhood search algorithm
