عنوان مقاله [English]
The hub location problem is widely used in the real world for areas, such as communications, postal services, transportation, and airline systems. This problem is applied to a network with high flow between nodes. The goal of the hub-location problem is locating some hub facilities on the nodes, and allocation of other nodes to hub facilities, in order to minimize the total cost of service. The hub facilities may face a queue of service from demand nodes, due to limited service capacity, and, hence, considering the queue in the model, may improve the performance of the system.We consider the waiting time to receive the service, as well as the total cost of networks, as the objectives of the problem, and formulate a multi objective mathematical model under a fuzzy environment. We assume the flow rates between nodes and service rates at hub facilities are fuzzy numbers. We use the credibility theory in modeling the problem, which is a new approach in formulating optimization problems in a fuzzy environment. To the best of our knowledge, this is the first attempt to formulate a hub location problem considering multi objectives using the credibility theory.The decision variables in the problem are the location of hub facilities on the network, as well as allocation of demand nodes to hubs. The objective function is to minimize the total waiting time in hubs and to minimize the expected cost in hubs to serve the demand nodes. The constraints of the model are general constraints that are assumed in hub location problems. We assume that the service demand flows from nodes to hubs, and the service rate in hubs, are fuzzy umbers.Therefore the waiting times in hubs are formulated in a fuzzy manner and we use the credibility theory to formulate and solve the problem.To solve the problems in the credibility environment, we need to use the fuzzy simulation approach. Therefore, we propose a hybrid intelligent solution method, integrating fuzzy simulation and the simulated annealing method. A DOE approach is applied to tune the parameters of the solution method and, hence, the performance of the solution method is improved. The computational results show the reasonable performance of the solution method.