Cooperative Coevolutionary Fuzzy Systems

Genetic algorithms have proved to be suited for fuzzy modeling. However, with the number of dimensions increases, most genetic algorithms tend to be inefficient. In this paper, we present a combination of cooperative coevolutionary genetic algorithm (CCGA) with fuzzy system. The flexible coding scheme permits one chromosome to correspond to a rule. For a high dimensional system, all rules can be coevolved in a small scale. One experiment showed its better performance than that of fuzzy system evolved by standard GA. We applied this method to Iris classification problem and obtain a very satisfactory result.