Approaches for the container loading problem

The container loading problem consists of arranging items (boxes) of different sizes inside large objects (containers), in such a way as to optimize an objective function, e.g., maximizing the volume loaded. In this paper we are interested in the special case of arranging the maximum volume of a cargo, composed of low density boxes, in a single container, satisfying stability constraints for the cargo loading. We review approaches known in the literature, such as the two-phase procedures of loading boxes in either horizontal layers or vertical stacks, the application of dynamic programming techniques and, in particular, search methods based on an and/or-graph representation of the solution space. Some of them were implemented in a microcomputer to solve a large real-life example of 784 boxes, and a hard small example of only 17 boxes, whose optimal solution is known to be of nonguillotine type.