Análise comparativa de algoritmos evolutivos multiobjetivo aplicados no problema de restabelecimento de energia

Abstract

7 ABSTRACTSOUZA, Luciana. M. X. De. Comparative analysis of applied multi-objective evolutionary algorithms in energy recovery problem. 2015. 88f. Trabalho de conclusão de curso – Graduação em Engenharia Elétrica, Universidade Tecnológica Federal do Paraná. Cornélio Procópio, 2015. Finding a configuration that reduces the number of necessary operational switch to restore the power after the occurrence of a fault it is considered to be a problem combinatorial nature, and usually has characteristics that make the use of mathematical programming techniques. Evolutionary algorithms have been increasingly used because they present satisfactory results for the restoration of electricity in the distribution system. This work is carried out in-depth analysis of the solution sets of multi-objective evolutionary algorithms that address issues of power restoration. So this detailed analysis is intended to check the quality of multi-objective solutions and see if there is a difference or not the solutions found in the number of switching operations obtained for the system reconfiguration process. To develop this analysis is employed some classic approaches of the literature on multi-objective evolutionary algorithms and the algorithms AEMT, AEMT-SP and the NSGA-II, and this evaluation was carried out through comparative metrics like hypervolume, Epsilon, R2 and R3 and also through the Wilcoxon test. These methods are implemented in softwareR in order to find the algorithm which has better performance and greater accuracy adding the solution of choice by the electric system operator. The main advantage of the proposed analysis is to know which one best algorithm is the distribution system.