Issues in the development of approaches to container loading

The paper argues that existing approaches to container loading problems are each applicable only to a narrow part of the spectrum of situations encountered in practice and that there are many scenarios for which there are no adequate methodologies. A number of examples are given. Two approaches aimed at addressing some of the present limitations are discussed. One is designed to produce stable, evenly distributed packing patterns; the other caters for multi-drop loads. A comprehensive performance analysis using both published and randomly generated test problems is presented. The procedure used to create the latter is described in full to enable replication. It is shown that the two methods, when combined, provide a powerful and versatile tool.