Modelling the global maritime container network

Abstract

Travel demand modelling has a long history going back to 50s when the conventional four-step modelling structure was introduced and developed for Chicago and Detroit, USA. However, the first travel demand models for freight movements were not developed until 40 years after the first model for passenger cars. Freight models are yet limited to studies looking at goods movements by truck and rail. This paper explores the effectiveness of the conventional travel demand modelling techniques for maritime container movements. Two approaches for modelling the movement of trade in the global maritime container network are discussed. The conventional methods of trip generation and distribution are applied in a sequential model, and compared against an alternative joint model methodology. Results show that the sequential methodology achieves high accuracy, while the joint methodology reveals more detailed trade relationships. Significant relationships are revealed, such as the varying influence of airports, the negative impact of coastline length and the impact of being an island, on containerised trade volumes. The findings of this paper provide a basis for modelling the container network from a transportation discipline viewpoint.