Efficient metaheuristics to solve the intermodal terminal location problem

Determining the optimal number and location of intermodal transshipment terminals is a decision that strongly influences the viability of the intermodal transportation alternative. In this paper, we develop a model and an optimization method that provides policy makers with a tool to help them take these decisions. The objective of the terminal location problem described in this paper is to determine which of a set of potential terminal locations to use and which not and how to route the supply and demand of a set of customers (representing zones of supply and demand) through the network (by both uni- and intermodal transport) so as to minimize the total cost. We develop two different metaheuristic procedures that both consist of two phases: a solution construction phase and a solution improvement phase. The first metaheuristic constructs solutions using a GRASP procedure, the second one uses the relatively unknown attribute based hill climber (ABHC) heuristic. Innovative in our approach is the integration of a fast heuristic procedure to approximate the total cost given the set of open terminals. Both metaheuristics are compared to the results of an MIP solver. A thorough performance assessment uncovers that both metaheuristics generate close-to-optimal solutions in very short computing times. An argument in favor of the ABHC approach is that it is parameter-free and hence more transparent and likely to be accepted in a business or policy environment.