Abstract: A novel mixed integer linear programming mathematical model for the hub-and-spoke container shipping network design in a competitive environment is proposed for the competition problem caused by reorganization and global expansion of container shipping liners. Path-based formulations are adopt as decision variables in the model. A discretized attraction function is introduced to indicate the number of network traffic (customers) of the container shipping alliance that can be attracted to the newcomer container shipping liners through competition. The aim of the model is to maximize the number of network traffic attracted which depends on lower transportation cost and shorter service time - by designing its mixed hub-and-spoke container shipping network. There are 13 constraints including the number of hub ports, routes connection, route transition and network traffic competition to ensure competition factors and network design factors the model should consider. As it is a NP-hard problem, a multi-point crossover genetic algorithm is given to solve the problem. By comparing the genetic algorithm allowed infeasible solutions, the results show that the multi-point crossover genetic algorithm which can avoid infeasible solutions has a better solution. Finally, the example of a container shipping market of Asia-Europe routes is analyzed, and the major data sources are obtained from CI (Containerization International)-Online, 100allin, Seagle and BLM-Shipping. Hub-and-spoke container shipping networks are designed for the cases of the various demands of customers and the countermeasure of container shipping alliance. The calculation results show that (1) Small and medium enterprises should consider that the costs reduction offered by the scale economy for transferring origin-destination (O-D) flows among the reasonable set of hub ports. (2) Since direct routes between the origin-destination (O-D) ports are relatively fixed and can avoid excessive costs and time generated by the detour, small and medium enterprises should consider that cutting costs when entering the container shipping market. (3) Small and medium enterprises should focus on the customer demands from the container shipping market not covered by container shipping alliance, in order to design the hub-and-spoke container shipping network by reasonable hub port selection strategy.

Key words: hub-and-spoke network, competition, hub location, genetic algorithm, container shipping