Systems in practice (manufacturing, transport, communication, supplying, etc.) have not only a complex structure but they also comprehend subsystems of different fundamentals: continuous, discrete-event and hybrid (different combinations of previous ones). Therefore, their control is complicated too. It is necessary to decentralize control by means of intelligent subsystems – agents. Cooperation of autonomous agents by a properly designed supervisor can ensure attainment of the global goal of the complex system. This is the main objective of this project.
Petri nets (PN) of different kinds - place/transition PN (P/T PN), timed, hybrid – will be used to model DES and hybrid systems. P/T PN will be used also for the supervisor synthesis, namely their invariants (especially P-invariants), state vectors of places and transitions, possibly the Parikh's vector too, but also other parameters and properties. Constraints imposed on the supervisor synthesis will reflect the real goals to be achieved (often formulated only verbally) in analytical terms.