VERSIONE IN ITALIANO
My research activity as a theoretical physicist at the Department of Physics, University of Catania, is mainly aimed in three directions.
- The first direction covers areas such as Chaos Theory, Statistical Physics, Dynamics of Systems with Long Range Interactions, Complex Networks, Synchronization, Self Organized Criticality and Sociophysics.
On this front, in the last years I have devoted primarily to the study of dynamics and thermodynamics of the so called Hamiltonian mean-field model (HMF). It is a model of planar rotators (XY) with long range interactions, exactly solvable in canonical ensemble and showing a second order equilibrium phase transition. In this respect it is paradigmatic of a large class of non-extensive systems, such as self-gravitating systems, multifragmented nuclear systems, atomic clusters, etc.
Starting from out of equilibrium initial conditions, microcanonical simulations of HMF model show many dynamical anomalies in a given range of energy density. In this situation the system does not relax immediatly towards Boltzmann-Gibbs canonical equilibrium but remains trapped in metastable quasi-stationary states (QSS) whose lifetime diverges with the size of the system. During this transient the system shows negative specific heat, vanishing Lyapunov exponents (weak mixing), anomalous diffusion, non-gaussian velocity distributions, clusters formation, dynamical frustration and aging, power-law decay of correlations and many other anomalies.
With other colleagues at Catania University, we tried to explain the nature of QSS by using two conceptual frameworks, strictly related with the dynamical anomalies of the HMF model: the nonextensive statistical mechanics proposed by C.Tsallis, whose formalism well adapt to the study of long-range system far from equilibrium, and Spin-Glass models, characterized by aging, frustration and weak ergodicity breaking. Both these approaches seem to be applicable to the HMF context, since during the metastable QSS regime the system is not free to explore all the a-priori available phase space but results to be trapped into narrow fractal-like phase-space regions thus violating ergodicity prescription (see my PHD thesis for an overview on HMF model). All these features drive the HMF model towards a violation of the Central Limit Theorem, in agreement with a recent generalization of the latter theorem proposed by Tsallis, as we showed in some papers published with Tsallis himself.
My interest is also related to synchronization of coupled oscillators, for example in the context of the celebrate Kuramoto Model. Through extensive numerical simulations we found QSS-like metastable states in a given range of the coupling parameter, just below the second order phase transition.
We are working also on synchronization over complex networks, developing models of dynamical systems able to identify modular structures through the progressive desynchronization (dynamical clustering) of oscillators coupled with the nodes of the network. These researches find interesting applications in many fields, from genetics to neurobiology, from ecology to social science.
Recently we started to work on consensus formation and opnion dynamics in the context of the so called "Sociophysics", a new branch of statistical physics which deals with emergent phenomena in social science. We explored existing models, like the Heigselmann e Krause one, but also introduced a new model, inspired to Kuramoto and to the concept of "bounded confidence", the so called OCR (Opinion Changing Rate) model. Here opinions are represented as coupled oscillators with different natural frequencies, which represent different tendencies to change. Despite of these differences, opinions are able to synchronize - again above a given treshold of the coupling parameters - in order to reach consensus.
In the context of non-extensive statistical mechanics we worked on the SOC (Self-Organized Criticality) interpetation of the sysmic activity of the terrestrial crust, showing that a small-world version of the del OFC (Olami-Feder-Christensen) simulated model is able to reproduce fat tails observed in the PDF of the real earthquakes energy differences (returns).
My scientific interests address also to Neural Networks, computational models of biological nervous systems, able to learn from experience. In my degree thesis I worked on Hopfield Networks, a particular kind of attractor neural networks strictly related to physical frustrated systems like Spin-Glasses. Actually I am trying to develop learning algorithms able to run on the GRID and based on darwinian evolution (random modification of synaptic weights and natural selection) of Feed-Forward networks.
A recent work, published on PRE, try to put in evidence the relationships between sincrony, noise and correlations in systems at the edge of caos, in the context of nonextensive statistical mechanics.
- The second direction towards which, more recently, I'm turning my interest is that one of the Agent Based Models simulations .
In this respect, in the last period I have been investigating the use of the Java platform and in particular of the software NetLogo. All the simulations you can find on this web site and on the Simulab section of the Cactus Group web site have been realized by us with this powerful software. NetLogo represents the ideal develop environment for the simulation of physical, biological or social systems, since it offers meny tools for treating complex systems made of many interacting units and for visualizing their significant parameters in real time.
From the intuitive graphics-interface of Netlogo (see my pdf presentation) you can easily give instructions to hundred or thousand of independent agents which compete or cooperate one with each other inside limited or not-limited virtual world, thus exploring connections between micro-level of individuals and macro-level of patterns emergent from their reciprocal interaction (Netlogo provides as default a rich "models library" containig many interesting simulations).
There are numerous collaborations born in these years based on agent based simulations, particularly with the INFN, with the National Laboratories of the South, and with the Department of Civil and Environmental Engineering and the Department of Mathematics and Computer Science, both at the University of Catania, with whom we have developed many researches and simulations on NetLogo platform concerning traffic management, pedestrian evacuation from public buildings, dynamics of movement in confined environments, simulations of Grid Computing, dynamical techniques based on synchronization to find structures in complex networks, etc.. (here the Master, Doctoral Thesis or made by our students).
Still using the NetLogo platform, in collaboration with colleagues physicists, sociologists and economists, we recently realized agent based simulations of social hierarchical organizations, demonstrating how the application of random promotion strategies allows, under certain conditions, to circumvent the so-called "Peter Principle" and to improve the efficiency of the system. These researches has had an unexpected success in the international community and were rewarded with the 2010 IgNobel prize for Management. Even more recently we open another line of research, once again exploiting the constructive role of chance in social processes by applying it to the dynamics of political parties in a Parliament. In this case, simulations show how the introduction of a certain percentage, analytically determined, of legislators selected at random among the candidates, allows to increase the efficiency of the Parliament in terms of both passed laws and "social welfare" ensured (see related article). Another recent application of random strategies to social systems concerns the discovery that random trading in financial markets is competitive with technical trading and could also serve to dampen the size of financial crisis or bubbles (see the link).
- The third direction of my research line concerns the epistemological implications of both Quantum Mechanics (see my answers to Ulisse - in italian -, on the SISSA website) and Relativity, with particular regard to alternative interpretations of quantum theory, like Bohm-Vigier theory or SIMQ (Stochastic Interpretation of Quantum Mechanics), and to the role of a basic-level noise, intrinsic to natural phenomena, which could be fundamental for the emergence of complexity at higher physical levels.
On this topic I and my colleagues have recently investigated the idea that quantum fluctuations might reflect the existence of an 'objective randomness', i.e. a basic property of the vacuum state (the zero-point energy field in Quantum Fields Theory, considered as a turbulent ether) which is independent of any experimental accuracy of the observations or limited knowledge of initial conditions. For more technical details see for example M.Consoli, A.Pluchino, A.Rapisarda, "Basic Randomness of Nature and Ether-drift Experiments", http://arxiv.org/abs/1106.1277v1. Further our works in this direction suggest a possible role of the new generation of "ether drift" experiments in clarifying the turbulent nature of vacuum (see for example this REVIEW and other publications).
In the same direction can be viewed my research about Plasmas and Screening Effect, performed in collaboration with colleagues of Laboratori Nazionali del Sud (LNS) in Catania. Through numerical simulations we reproduced, starting from simple and original hypothesis, some dynamical properties sperimentally observed at LNS about a plasma generated by laser ablation in nanosecond domain on a pure Aluminum target (see S.Mascali et al., http://arxiv.org/abs/1112.1235). Within such a scenario we are trying to quantify a possible role played by the fluctuations of a turbulent ether on the plasma properties, in order to give an alternative explanation of the screening effect.