Dynamical Networks, Uncertainty and Control
Description
A dynamical network is a dynamical system, in which the behaviour is chiefly determined by its structure. This property is typical of chemical, biological and cellular models, in which different species interact, in order to generate the system's functionality. The type of research taken into account is mainly aimed at identifying structural properties. A property can be defined as structural for a class of systems if it is verified regardless of the numerical values of the parameters. The main goal is to explain, in mathematical terms, the reason why certain structures turn out to be extremely robust, and work adequately even in very diverse conditions, in the presence of strong uncertainties, and despite enormous variations of the parameters.
The problem of the decentralised control of dynamical networks arises in actual problems, such as coordinating autonomous robots, vehicle platooning, controlling communication and distribution networks, where the management has to be decentralized. Typically, this problem arises when several agents have to coordinate the system through local actions in order to guarantee a suitable global behaviour (local actions – global behaviour).
The decentralised control of networks and the structural analysis of networks are complementary problems. Natural phenomena can be explained, and nature itself can teach how to manage complex systems.
There are ongoing collaborations with the Universities of Lund (Sweden), Riverside (California) and the Polytechnic University of Milan.
Research subjects
- Structural analysis of networks
- Influence analysis
- Network-decentralised control
ERC panels
- PE7_1 Control engineering
- PE7_5 (Micro- and nano-) electronic, optoelectronic and photonic components
Tags
- Analisi strutturale di sistemi dinamici, controllo di reti decentralizzato
- Funzioni di Lyapunov, modelli biochimici, agenti autonomi, analisi di reti