The Hub Location Problem with Allocation Constraints

We deal with the capacitated hub location problem in which the capacity of a hub restricts the amount of flow transiting through the hub and each node can only be allocated to a certain number of hubs. In hub networks, all hubs are interconnected and the direct flows between non-hubs are not al owed. In order to benefit from the adoption of hub networks, the challenge is to design the hub networks such that the total cost is minimized. In this research a hybrid heuristic, based on the simulated annealing method, tabu lists, and improvement procedures, is proposed to determine the number of hubs, to properly locate the hubs, and to determine the path for each origin-destination pair. Computational characteristics of the proposed hybrid heuristic are evaluated through computational experiments using the revised CAB data set. Computational experience indicates that the proposed hybrid heuristic is capable of obtaining optimal solutions for these tested problems.