On-line scheduling and control of flexible manufacturing cells using automata theory

An effective flexible-manufacturing-cell (FMC) controller can be synthesized using classical automata and Ramadge–Wonham (R–W) supervisory-control theories. Such a discrete-event-system controller/supervisor would enable or disable (controllable) events for maximum deadlock-free, correct behaviour of the FMC. However, in cases of multiple part-routes, machine redundancy, etc. the R–W supervisor could include states with subsequent multiple controllable events. The question of choice arises at such states necessitating the use of an on-line decision-making agent, which, consequently, would determine the overall performance of the FMC. In the above context, this paper presents a novel methodology for the on-line, deadlock-free scheduling and control of FMCs. A two-phased method is proposed. During the off-line phase, the FMC is first modelled using time-augmented automata and a deadlock-free supervisor is synthesized using R–W control theory. Subsequently, an off-line decision-making plan is constructed. During the on-line phase, based on the latest state of the workcell and the off-line plan, the best-possible scheduling decisions are made using a real-time optimization search technique. The proposed novel approach is illustrated through a typical manufacturing-cell simulation example and compared with a random-based decision-making policy. It is clearly shown that significant improvement is achieved when using the proposed approach.