Abstract: Container loading is a complicated combinatorial optimization problem in three-dimensional packing. It is a NP-hard problem. Six space partitions are presented and a set of rules for merging spaces is proposed, and stability constraint and load bearing strength are applied in the six space partitions. Through testing a benchmark data set, it compares the influence of application constraint for the space utilization. The results also show the influence of load bearing strength for different space partitions, and which space partition will produce the most of influence, and the testing results also show which space partitioning based on the common-effect of stability constraint and load bearing strength constraints is the best space utilization. Compared to the same domestic research algorithm, the test results show that this algorithm is best than other algorithms at home and abroad at present.

Key words: container loading, space partitioning, application constraint, space utilization

摘要： 在三维装箱问题中，集装箱问题是一个复杂的组合优化问题，是属于NP-hard问题。提出了六种空间划分和不同规则的空间合并，并将稳定性约束和承载力约束加入其中，通过测试标准数据集去比较应用约束对空间利用率的影响。测试结果显示了承载力约束对不同的空间划分的影响，以及对哪一种空间划分的影响最大，以及在基于稳定性约束和承载力约束的共同作用下，哪一种空间划分的空间利用率最好。同时，与国内外相同研究方向的算法进行了对比，测试结果显示了该算法的性能要高于目前国内外的算法。

关键词: 集装箱装入, 空间划分, 应用约束, 空间利用率