Learning fuzzy control with hybrid symbolic, connectionist networks

Abstract

The problem of deriving membership functions as a means for describing linguistic variables (for some control process) and the choice of fuzzy inference operators and connectives is at the heart of developing fuzzy control systems. To this date much of the selection process is under the control of the system engineer, and dependent on his or her ability to make the right choice for the right application. In this paper, it will be shown by means of a real world example for controlling a steam engine, how hybrid learning systems can be employed for automating the design of fuzzy controllers. Deriving the necessary linguistic variables and accompanying membership functions from raw data by use of machine learning is addressed. Finally, the viability of such a system is emphasized to act not only as a fuzzy controller, but more importantly independent of human intervention, automatically derive acceptable control strategies.