摘要
The key factors for designing a successful hub-and-spoke network are to determine the optimal number of hubs, to properly locate hubs, and to allocate the non-hubs to the hubs. This paper proposed a hybrid heuristic based on genetic algorithm （GA） to resolve the uncapacitated single allocation hub location problem with fixed cost （ USAHLPFC）, according to the upper bound of the ratio of the number of hubs to the number of nodes. Computational experiments were conducted using the Civil Aeronautics Board （CAB） data set. Results indicate that, applying the proposed heuristic, one can obtain the optimal solutions for all small-scaled problems very efficiently. A comparison between the proposed hybrid heuristic method and the GA presented by Topcuoglu et al. shows that the former outperforms the latter when the discount factor α is less.
The key factors for designing a successful hub-and-spoke network are to determine the optimal number of hubs, to properly locate hubs, and to allocate the non-hubs to the hubs. This paper proposed a hybrid heuristic based on genetic algorithm （GA） to resolve the uncapacitated single allocation hub location problem with fixed cost （ USAHLPFC）, according to the upper bound of the ratio of the number of hubs to the number of nodes. Computational experiments were conducted using the Civil Aeronautics Board （CAB） data set. Results indicate that, applying the proposed heuristic, one can obtain the optimal solutions for all small-scaled problems very efficiently. A comparison between the proposed hybrid heuristic method and the GA presented by Topcuoglu et al. shows that the former outperforms the latter when the discount factor α is less.
