集装箱场地箱位分配问题  

Location assignments for outbound containers in container terminals

作　　者：张灿荣[1] 钟明[1] 缪立新[1] 

机构地区：[1]清华大学深圳研究生院,深圳518055

出　　处：《清华大学学报（自然科学版）》2015年第10期1150-1156,共7页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金项目(71102011;71472108)

摘　　要：为减少集装箱在装船时的翻倒箱次数,针对出口集装箱在堆场贝内箱位分配问题进行研究。以往的研究假设集装箱的重量概率分布在整个集港过程中保持不变,由于该假设与实际情况不相符合,该文考虑了集装箱的重量概率分布随堆存状况可变的情形,使其更加符合实际情况,并构建了一个带约束的随机动态规划模型。在求解算法方面,小规模算例可直接通过动态规划模型求得最优解。针对大规模算例,提出了两阶段的启发式算法:第一阶段基于邻域搜索的启发式算法,设计出各重量组的集装箱在不同堆垛形态下的优先堆放次序;第二阶段设计了基于翻滚策略的箱位堆放局部优化算法。数值计算结果表明:适用于小规模算例的动态规划算法和适用于大规模算例的两阶段启发式算法都能显著改善解质量,减少翻倒箱次数。This study analyzes the location assignments for outbound containers in terminals to reduce the container re-handling during loading operations. Previous studies assumed that the container arrival probability for each weight group remained unchanged during the entire receiving process, which is not true in practice. This study adjusts the probability that the remaining containers have not yet arrived at the terminal. A constrained dynamic programming model is then constructed for this problem. Small-instances can he directly optimized by the dynamic programming model. Large-instances are solved with a two-stage heuristic algorithm. The first stage develops a heuristic to generate the precedence of the stacking patterns for each container weight group. The second stage uses a heuristic algorithm based on the roiling strategy. Numerical calculations show that the dynamic programming method for small-instances and the two-stage heuristic algorithm for large-instances both significantly improve the solution quality and reduce the container re-handling.

关 键 词：集装箱 箱位分配 动态规划 启发式算法 

分 类 号：U169.65[交通运输工程] O221.3[理学—运筹学与控制论]