Emergency Public Service System Design Using IP-Solver

The paper deals with the public service system design, in which the generalized utility is to be maximized. The generalized utility defined for a public service system assumes that user’s utility comes generally from more than one located service center, on the contrary to the former formulations, where only one service center was taken into account as a source of individual user’s utility. Classical approach to the design makes use of location-allocation model and subsequent usage of a commercial IP-solver. Real instances of the public service system design problem are characterized by big number of possible facility locations, but, in such cases, the location-allocation approach usually fails due to enormous computational time or huge memory demands. This weakness can be overcome by the approximate covering approach based on so called radial model of the problem. Hereby, we study and compare the both approaches from the point of their impact on the solution accuracy and saved computational time. Key-Words:Generalized utility, public service system design, radial model, approximate approach