Knowledge-based self-adaptive control

In order to manage a self-adaptive controller in an environment far from ideal, a good deal of `experience´ or heuristics is programmed into the controller to manage it and to assure a good (or at least reasonable) behaviour under all operational circumstances. The paper describes some ideas on the management of self-adaptive controllers by borrowing basic concepts from the expert systems area. It gives a brief review of classical approaches to the subject. It discusses new developments using artificial intelligence tools. Future directions of study are indicated