Abstract
The agent-based approach is used to model and supervise manufacturing systems (MSs). Because most MSs have hybrid character (i.e. contain both discrete and continuous parts), hybrid agents will be used here. A hybrid agent makes a cooperation of continuous and discrete event devices possible. To model autonomous hybrid agents, First Order Hybrid Petri Nets (FOHPNs) are used here. The cooperation of hybrid agents is synthesized by means of supervision methods known in DESs (Discrete-Event Systems) control theory. Two principal methods of supervision are used. While first of them utilizes place invariants (P-invariants) of Place/Transition Petri Nets (P/T PNs), the second one uses both places and transitions of P/T PNs. The proposed approach is illustrated in details by the cooperation of hybrid production lines in a real continuous manufacturing system. The convenience of the approach consists in the fact that it can be described in analytical terms. Thus, the approach is perspicuous also in more complicated cases.