Fuzzy Knowledge Base Generation and Optimization Using a Genetic Algorithm

The need of an expert to build the knowledge base of fuzzy decision support systems, herein called the fuzzy knowledge base (FKB), is a strong limitation to the expansion of their use in the industry. This paper presents a genetic algorithm that automatically constructs the FKB and discusses the influences of the optimization and selection criteria on its performances. These criteria allow satisfying two contradictory objectives, i.e., the minimization of the approximation error and the complexity level of FKB. The technique has been successfully applied to the generation of an FKB allowing the prediction of the cutting force from a measure of the feed rate and cutting depth during turning operations. The FKB exhibits a better prediction accuracy than the standard Taylor prediction model.