DYVERSE Networks: pushing the boundaries of formal verification
The controlled motion of swarm satellites; the cooperation of several robotic systems to meet a common goal; the operation of large-scale electrical power networks with renewable generation sources; the automation of public transport scheduling systems in metropolitan areas; the synchronisation patterns of beta-cells in your pancreas; the self-organisation of cells: these are examples of interdependent, interacting, distributed and networked systems. Their operation is usually safety critical and demands the preservation of stability despite uncertainties and the recovery from contingencies minimising losses. Since they exhibit different transitions and switching behaviour patterns, they can be seen as hybrid systems. That is, systems which combine continuous and discrete, smooth and abrupt dynamics (discontinuities, discrete events).
To check that these systems satisfy some desired properties requires the application of methods beyond classical control engineering procedures or stability analysis techniques. To this end, the combination of formal verification methods from computer science and dynamical analysis techniques seems to be promising. This is one of the main aims of DYVERSE. DYVERSE is the DYnamical-driven VERification of Systems with Energy considerations. Here, energy refers to the abstract energy of dynamical systems, which is studied by means of the dissipativity theory. Dissipativity theory analyses dynamical systems behaviour by means of the exchange of energy with the environment. This project aims to expand the already-existing results on formal verification of hybrid systems within the project DYVERSE to networks.
The framework proposed here is general and applicable to a broad class of physical, biological and engineering systems. Depending on the student's interests, different application domains can be explored. This research would be part of the project DYVERSE (DYnamical-driven VERification of Systems with Energy considerations).