Supervised Learning Neural Networks

Abstract

Application of fuzzy inference systems to automatic control was first reported in Mamdani’s paper (Mamdani & Assilian, 1975), where a “fuzzy logic controller” (FLC) was used to emulate a human operator’s control of a steam engine and boiler combination. Since then, “fuzzy logic control” has been recognized as the most significant and fruitful application for fuzzy logic (Kosko, 1992). In the past few years, advances in microprocessors and hardware technologies have created an even more diversified application domain for fuzzy logic controllers, which ranges from consumer electronics to the automobile industry. However, without adaptive capability, the performance of fuzzy systems relies exclusively on two factors: the availability of human experts, and the knowledge acquisition techniques to convert human expertise into appropriate fuzzy rules. These two factors substantially restrict the application domain of fuzzy systems.