Authors:
João Ferro
1
;
José Brito
1
;
Robério Santos
2
;
Roberta Lopes
1
and
Evandro Costa
1
Affiliations:
1
Computing Institute, Federal University of Alagoas, Av. Lourival Melo Mota, Maceio, Brazil
;
2
Eixo das Tecnologias, Campus do Sertão, Federal University of Alagoas, Delmiro Gouveia, Brazil
Keyword(s):
Fuzzy Logic, Genetic Algorithms, Uncertainty, Optimization.
Abstract:
This article addresses issues involving two sources of uncertainty in the stochastic search problem based on a genetic algorithm approach. We improve the mutation rate parameter by fuzzifying the population diversity and the individual adaptation value. A relevant aspect of this investment is related to the fact that this parameter, which presents uncertainty of the possibilistic type, directly interferes with the uncertainty of the probabilistic type of the genetic algorithm and also in the convergence and quality of the solution found by the genetic algorithm. Moreover, in parallel, we improve the understanding behavior of selection and replacement methods. Experiments were carried out on the case study with the classic OneMax problem to evaluate the performance of the proposed solution, analyzing aspects such as the convergence time, the quality of the solution, and the diversity of the population. The results obtained through the treatment of uncertainty and its impacts are prese
nted in this article, showing relevant performance for the proposed algorithm, with the respective treatment of uncertainties.
(More)