Despite the use of genetic algorithms in many optimization problems, many new versions were proposed since them, from distributed versions of the canonical genetic algorithm (GA) to more restructured evolutions like the Biased Random-Key Genetic Algorithm (BRKGA). Aiming to explore the best of both techniques, in this paper, a novel approach was proposed, resulting in a Distributed BRKGA (D-BRKGA) with a stratified migration policy. To compare the performance of the Distributed Genetic Algorithm (DGA) and the D-BRKGA, some functions of the CEC 2013 Benchmark set they were chosen because of their high complexity and greater dimensionality. The analysis of the results aimed to explore three aspects: quality of the final solutions, population diversity and convergence curve of both approaches. The results point out to a superior performance of D-BRKGA, proving to be efficient and scalable in relation to the number of distributions, in addition to maintaining a high population diversity.
