Abstract
This paper investigates the effects of relaxation of the nearest facility assumption, which is usually adopted in many location models, on optimal facility location. We formulate two new models based on p-median model: one is the model in which a user selects from the k nearest facilities, and the other is the model in which a user selects facilities so as to minimize its tour distance. We also build mixed models with the original p-median model. The optimal solutions show that facilities form clusters if the nearest facility assumption is significantly relaxed.
