Resumen:

Facility Location and Network Design Problems constitute a major area of interest for researchers and practitioners within Operations Research (OR) with several connections with combinatorial and discrete optimization. These problems have shown applicability in several fields such as supply chain management, telecommunication networks, transportation, logistics, health care, among others. Moreover, the mathematical structure of some facility location problems has proven beneficial for the development of solution methodologies that are broadly used today in OR. In this talk we address a general class of facility location problems, called multi-level facility location problems. First, we describe the general version of the problem and discuss the particular cases of interest. We then study some of their fundamental properties for the development of efficient solution methods. In particular we discuss the use of the submodularity property to attain different formulations for the problem and approximation algorithms. We also present the results obtained after developing an exact algorithm based on Benders decomposition. Computational experiments are presented to assess the performance of the various models and algorithms studied.