This paper proposes a methodology for the synthesis of optimal supervisors for automated manufacturing facilities. The methodology is based on local modular supervisory control, an extension of the Supervisory Control Theory with the goal of avoiding states explosion in the modeling and synthesis of supervisors. The modeling consider that manufacturing systems are characteristically formed by several concurrent subsystems that should be synchronized in way to obey a series of specifications. The synthesis process exploits modularity of specifications, as well as the natural decentralized structure of manufacturing systems, to generate minimally restrictive supervisors with reduced number of states. The proposed methodology is applied to a hypothetical example of integrated manufacturing system. The results show considerable reduction of the computational complexity and of the size of supervisors.
Supervisory control; manufacturing systems; discrete-event systems; control system synthesis; decentralized control
