Compact fuzzy models through complexity reduction and evolutionary optimization

Genetic Algorithms (GAs) and other evolutionary optimization methods to design fuzzy rules from data for systems modeling and classification have received much attention in recent literature. We show that different tools for modeling and complexity reduction can be favorably combined in a scheme with GA-based parameter optimization. Fuzzy clustering, rule reduction, rule base simplification and constrained genetic optimization are integrated in a data-driven modeling scheme with low human intervention. Attractive models with respect to compactness, transparency and accuracy, are the result of this symbiosis.