Hub location under capacity constraints

Hub-and-spoke networks are employed in cargo transportation. This paper presents a model for hub location in these kinds of networks. Hubs are considered capacity limited. For that reason, costs due to congestion in hubs are introduced into the model. A Simulated Annealing algorithm has been developed to solve the model. The algorithm includes a balanced module, the purpose of which is to reduce congestion. Each hub is modelled as an M/M/1 queuing system. The algorithm has been tested in randomly created networks. The results show that when a situation of congestion occurs, the proposed algorithm enables better solutions to be found, thereby improving the standard of service. (C) 2006 Elsevier Ltd. All rights reserved.