From centralized to decentralized approach for optimal Controller of Discrete Manufacturing Systems

This paper deals with a comparison of centralized and decentralized approaches to obtain an optimal controller for discrete manufacturing systems. It is based on a modular modeling of the plant to avoid combinatory explosion found in centralized structure and a synthesis algorithm. From the local Plant Elements, local constraints are defined to build local supervisors. Local constraints restrict the system behavior within a desired specification. Global constraints are added to establish high level supervisors. The resulting automata are translated in a normalized language for implementation in a Programmabl