A multi-objective evolutionary algorithm for fuzzy modeling
Fernando Jiménez, Antonio Skármeta, Hans Roubos, Robert Babuška · 2002
In this paper a multi-objective evolutionary algorithm with a single run is proposed in order to consider several objectives dealing with transparency and compactness in obtaining a fuzzy model besides the standard accuracy objective. In this way the use of Pareto-optimal solutions within the evolutionary algorithm let us obtain attractive fuzzy models with respect to compactness, transparency and also accuracy. The results of the combination of Pareto-based multi-objective evolutionary algorithms and fuzzy modeling are compared with other approaches in the literature.