GSO-2010: Mikhail Prokopenko
Guided Self-organisation in Complex Networks
Complex Systems such as modern power grids, sensor networks, communication and transport systems, and so on, are characterised by rich interactions among components (hubs, spatial regions, network nodes, agents, neurons, etc.), producing non-trivial information flows, irregular network topologies and nonlinear dynamics. In addition, complex systems evolve over time, making their management and control extremely challenging. An important research area providing solutions to many of these challenges is guided self-organisation. Typically, self-organisation (SO) is defined as the evolution of a system into an organised form in the absence of external pressures. SO within a system brings about several attractive properties, in particular, robustness, adaptability and scalability. Guided self-organisation (GSO) combines task-independent objectives (e.g., maximisation of predictive information within the systems) with task-dependent constraints. As many complex systems are amenable to be described as networks (e.g., genetic regulatory networks, structural or functional cortical networks, metabolism of biological species, power grids, etc.), it becomes increasingly clear that GSO applied at the network level may produce particularly interesting results. Among specific areas of GSO at the network level, this talk will address
* propagation and processing of information within networks analysed as (Shannon and Fisher) information dynamics: such analysis requires to consider not only networks' topology, but also the time-series dynamics at individual nodes;
* phase transitions of Random Boolean Network properties between ordered and chaotic regimes, where information transfer and information storage is often maximised.