Fuzzy extension of discrete event supervisory control

This paper proposes an enhancement of industrial control through a combination of discrete event supervisory control methods and fuzzy techniques. Such a combination enables a control system not only to perform standard tasks of preventing the forbidden states, driving through the shortest path, etc., but also to make more flexible and sophisticated decisions using additional uncertain and/or incomplete information, e.g. on "safer" or "faster" trajectories, "preferable" states, etc. Forms of introducing fuzziness into a discrete event control framework are analysed. Three possible implementations are considered: fuzzy assignment of strategies, fuzzy strategy, and fuzzy synchronisation. Formal considerations are accompanied by illustrative examples from automated manufacturing.