| Abstract [eng] |
The combinatorial optimization problem considered in this paper is a special three-dimensional packing problem arising especially in the logistics and transportation industries. This problem is that of orthogonally packing a subset of some given rectangular-shaped boxes into a rectangular container of fixed dimensions. It is generally called a container loading problem (CLP). Due to the combinatorial complexity of this problem, it seems impossible to solve it optimally for every instance. For this reason, the experimental behaviour of special heuristic on sets of randomly generated test problems has been analysed. Computational experiments showed that the performance of heuristic algorithm strictly depends on the ranking rules, which are used to select a number M3 of layer depths and a number M2 of strip widths/heights. Twenty-four different ranking rules were analysed and compared by filling ratio, by packing time and by some other rates. In addition to this, the influence on filling ratio and packing time depending on various combination of parameters M3 and M2 and depending on the number of boxes which we are trying to pack is considered. The performance of the heuristic is also experimentally compared for homogeneous, strongly and weakly heterogeneous instances. |
