On the Design of Adaptive Supervisors for Discrete Event Systems

The supervised control of complex event-driven Discrete Event Systems (DESs) such as those present in manufacturing systems, or the communication processes involved therein, continue to pose a challenge to system designers. To a certain extent this complexity can be reduced by applying existing modular control approaches to large-scale DES design. These solutions divide the system into different sections in such a way that its overall behavior is given by a suitable arrangement of the different sections. However, if the system is reconfigured frequently, the overall plant models and control specifications computed earlier would no longer be valid. Thus a new controlled system will have to be computed. We propose a new methodology, for ensuring that the new controlled plant will meet any valid control specifications taken from the existing modules. Being built on the framework of Supervisory control theory, this method is guaranteed to work even as the system is being dynamically reconfigured.