A method using machine learning to automatically produce controller tuning algorithms is described. The method constitutes a decision tree which selects from a set of tuning rules the rule which is best able to improve controller characteristics. The decision tree is constructed using a training set of example systems. Evaluations of the resulting tuning algorithm are performed using a large independently generated set of example systems
Published in:
Computer-Aided Control System Design, 1992. (CACSD), 1992 IEEE Symposium on
Date of Conference: 17-19 Mar 1992