Determining hub port locations and feeder network designs: The case of China-West Africa trade

Considering the impacts of low port distribution (i.e., concentration of ports in a certain geographic area) and liner network development, we study hub port locations, current and future, along the West African coastline. We feel that the relationship between hub port locations and liner network structures should be further researched. To that effect, we apply our modelling to the Asia - West Africa trade, in view of China's increasing economic involvement in the development of the African continent. In formulating our model, we specify the relationship between the path choice behavior of shippers, the liner network and the hub port location. To do this, we split the problem into two associated subproblems: a liner network design with a non-discrete calling port set (LNDNS) and the shippers' path choice problem (SPCP). In contrast to conventional discrete location models, our method relaxes the usual constraint requiring hub port alternatives be listed a priori. This allows us to consider the development of new port infrastructure, in green-field areas, along the West African coast. We believe such an approach—i.e., the identification of ‘greenfield’ sites—could be of particular interest to infrastructure financiers, such as the World Bank and the Asian Infrastructure Investment Bank, as a port policy blueprint, for regions such as West Africa, characterized by low port density. Our results show that Abidjan, Cotonou and Lomé, located in the central part of the West African coastline, are more important for the China–West Africa container transport system, and could potentially be selected as hub ports. Currently, the western and southern parts of the coastline do not appear to be suitable for this function; ports there, however, could be developed to important feeder ports in the future, in a pendulum type of network design, as we demonstrate. Interestingly, we also find that, as the number of routes increases, the structure of the liner network changes from circular to hub-and-spoke. Finally, the networks we design and present in our model show how shippers' path choice behavior affects liner network design and the location of hub ports.