Abordagens multiobjetivo para alguns problemas de localização de facilidades

Abstract

This work presents solution approaches for the Multi-objective and Multi-level Capacitated Oil Platform Location Problem (MMCOPLP) and for the MultiObjective Support Units Location Problem for Road Traffic Surveys with Multistages (MSULPRTS). For the MMCOPLP, the objective functions aim to minimize drilling, installation and construction costs, maximize the oil production and minimize costs with possible environmental damages. For the MSULPRTS, the objective functions aim to minimize travel costs and minimize the total amount of support units selected. The exact -Constraint method was used to find the Pareto Optimal for both problems. Considering the MMCOPLP, we implemented the NSGA-II evolutionary algorithm, in its original form, as an approximate method and its results were compared against the ones found by the exact method and the ones provided by a GRASP metaheuristic. For the MSULPRTS, we propose three versions of the NSGA-II with specific strategies for the problem, mainly related to the structure of chromosomes and to the crossover and mutation operators. Our results show that these versions provide good solutions for large instances.