Multi-Decision Supervisory Control: Parallel Decentralized Architectures Cooperating for Controlling Discrete Event Systems

This paper deals with decentralized supervisory control, where a set of local supervisors cooperate in order to achieve a given global specification by controlling a discrete event system. We propose a new framework, called multi-decision control, whose basic principle consists in using several decentralized supervisory control architectures working in parallel and whose decisions are combined disjunctively or conjunctively. We have identified sufficient conditions that make a decentralized architecture eligible to be used in the multi-decision framework. We have studied the generic framework consisting of several eligible architectures running in parallel, and we have also considered in detail the particular case of several inference-based Inf_Nj-supervisors running in parallel.