Induction of decision trees for rule-based modelling and control

Abstract

The modeling of nonlinear dynamical systems is addressed. In the special case where the system being modeled is an existing feedback controller, the approach becomes that of control systems synthesis. The focus here is on the case in which the controller is a skilled human manually controlling a complex dynamical system. The automatic machine induction of control rules from past records of skilled human behavior is considered. The aim is to install the induced rules as an automatic control program, based on induction of decision trees from examples. The algorithms used are a product of the machine learning sub-field of artificial intelligence research. A formalism is developed whereby the modeling and control of general dynamical systems are transformed into classification problems, and therefore become amenable to processing by the induction algorithms. Simulation results are presented. >