ABSTRACT
The uncapacitated multiple allocation hub location problem (UMAHLP) is a decision problem with the hub network structure. In hub networks, all hubs, which act as transshipment points for internodal flows, are interconnected and none of the non-hubs are directly connected. The critical factors for designing an economical hub-and-spoke network are to determine the optimal number of hubs, to properly locate hubs, and allocate the non-hubs to the hubs. In this paper an approach to derive an upper bound for the optimal number of hubs along with an effective heuristic based on the simulated annealing method, tabu list, and improvement procedure are presented to resolve the UMAHLP. Computational experiences indicate that by applying the derived upper bound for the number of hubs the proposed heuristic is capable of obtaining the optimal solution for the widely tested CAB and small-sized AP problems very efficiently.