An Adaptive Heuristic for Feeder Network Design with Optional Transshipment

This paper studies the Feeder Network Design Problem (FNDP), which considers the design of a minimum cost liner shipping network for the transportation of cargo (containers) between a given hub port and a set of feeder ports. In addition to determining the services to operate (i.e., the routes), the FNDP deals with deciding the fleet of vessels to deploy on these services. In contrast to most FNDPs previously studied in the literature, the feeder network can, if found beneficial, be a hub-and-spoke system where cargo can be transshipped at any feeder ports. Thus, we denote our problem as the Feeder Network Design with Optional Transshipment (FND-OT). To solve the FND-OT, we propose a novel adaptive Heuristic with a special data structure for solution representation, which gives significant speed-ups. Furthermore, a new escape algorithm is used to escape from local optima. We show that the adaptive heuristic outperforms an existing solution algorithm in the literature on a set of realistic test instances. We also present results for a new set of instances adapted from a previously published benchmark suite (LINER-LIB) and show that including the possibility of having cargo transshipment in the FNDP can give significant benefits and reduced costs.