Institute of Transportation, MOTC

      Shippers are quite often confronted with the problem that air freight is too expensive and sea freight is too slow. Sea-air multimodal transport can provide a solution to the problem by allowing shippers to opt for first sea freight and then air freight. It has been shown that efficiently combining the two transport modes can offer a 30 to 50 percent faster service than ocean freight and can have cost savings of 40 to 60 percent compared to air freight. That is, sea-air multimodal transport allows shippers to manage their supply chains more cost-effectively and quickly. However, these advantages accrue only to those who can efficiently integrate these two transport modes. Currently, most shippers rely on freight forwarders to make global shipping as seamless as possible by utilizing multiple transport modes. Besides, some logistics companies, e.g., DHL Express, also provide multimodal transport services to the shippers. Obviously, freight forwarders can only use existing sea and air services to arrange their shipping. On the other hand, the logistics companies may use existing sea and/or air service to develop their new sea-air transport services. This research aims at helping a container shipping company and an air freight company design their cooperative sea-air multimodal freight transport networks, which involve both existing and new developed sea-air transport services. More precisely, the goal of this research is to optimally design the cooperative sea-air multimodal freight transport networks to maximize the volume of sea-air freight that are used to be shipped by sea freight or by air freight. It is well known that it is quite challenging to simultaneously design sea-air system and service networks. Hence, this research tackles this complicated network design problem by first mathematically modeling this NP-hard problem, then proposing efficient matheuristics hybridizing metaheuristics and mathematical programming for its solution. Finally, experiments are conducted to exam the performance of the proposed models and solution algorithms. In addition, this research also applies the proposed models and algorithms to real-life case studies.