Análise de algoritmos genéticos e evolução diferencial para otimização de funções não-lineares multimodais

Abstract

Our society has been looking to solve more complex problems every day, with this context in mind, traditional solving methods can, many times, fall short on generating answers with the required speed or quality . For this reason, we studied, on this paper, Genetic Algorithms and Differential Evolution algorithms, both are classified as Evolutionary Algorithms, optimization methods based on the evolution of the species by Natural Selection. Were described 23 different Evolutionary Algorithms strategies and tested them using three different Benchmark functions, the answers were them studied and compared, by their speed of convergence and the quality of their outputs. There were found methods that, not only are an improvement over traditional optimization methods, but also over Evolutionary Algorithms that are currently more known and used.