Combining Facility-Location Approaches for Public Schools Expansion

This study proposes determining optimal locations for expanding a higher education system by using populational and social criteria. With this aim, this work evaluates single objective location models in determining the optimal distribution of higher education facilities in Amazonas State, Brazil. Three optimization options are evaluated and made available to decision makers: 1) prioritize cities with a lower United Nation Human Development Index; 2) prioritize cities with a higher population, 3) favor both criteria. Also, the location must equalize student distribution between the regions of Amazonas State. With this aim, three discrete location models were evaluated: p-center model, p-median model, and p-dispersion model. The location models were designed using a Genetic Algorithm metaheuristic. A state-of-art implementation of the Genetic Algorithm that optimizes the solution and converge time was used. The expansions proposed here present lower mean values of United Nation Human Development Index compared with 0.619 from the existing distribution of campuses. The best results were obtained with the p-median model.