Algoritmos para el problema de localización de plantas y centros de distribución maximizando beneficio

Abstract

This paper presents a particular case of the facility location problem. The characteristic of this problem is that it contains a set of clusters or nodes that represent the accumulation of clients who demand a single product. It is desired to locate P distribution centers (DC) where the demand of clusters is only satisfied if we located a distribution center in this cluster. Moreover, the number of capacitated plants and its location must be determined in order to supply the different distributions centers. The objective is to maximize the utility by considering, in one hand, the incomes that are obtained for satisfying the demand of different clusters and in the other hand, the installation costs incurred when locating the Distribution Centers and the plants as well as the costs of material transport from the plants to the distribution center. Other applications of this problem are appreciated in telecommunications networks, electrical networks, etc. The paper also present a new model of integer programming that allows to solve small instances, up to 300 clusters, in an acceptable CPU time, the problems were programmed with AMPL and solved with CPLEX 11.0. For two larger instances the authors present heuristics, which allow obtaining solutions with a GAP least than 1%.