Information flow and irreversibility in self-organised active matter

点击翻译按钮获取中文摘要

A lot of recent research is devoted to unravel the stochastic thermodynamics of active matter; specifically quantifying the time- reversal symmetry breaking of active dynamics by the fluctuating entropy production, and connecting the entropy production to energetic quantities, like the heat dissipation and energetic cost of the 'activity'. For individual active particles, first studies have demonstrated that time-irreversibility and energetics can be linked with each other by using information-theoretical measures. But so far this perspective has not been applied to collective active systems. With this proposal we aim to put forward the information- thermodynamics of collective active matter. To this end, we analyse in two generic models (a particle-based and an active field model) what kind of information flows occur, which spatial direction they have, and how they reflect the self-organised collective states as well as phase transitions between order and disorder. We further aim to connect the information flow with energetic and thermodynamic measures, specifically the local entropy production, by employing and extending the framework of information- thermodynamics that was previously developed to describe the physics of systems subject to feedback control. In particular, we aim at establishing generalised second laws with continuous information flow for the local entropy production of active matter. By combining concepts from stochastic thermodynamics, information theory, control theory, active matter, and phase transitions, this proposal targets in an innovative way open research questions concerning the stochastic thermodynamics of active matter, as well as active self-organization, e.g., swarm formation. It may further pave the way for a novel type of model-independent investigation of collective dynamics. The results will thus also be of importance for collective nonequilibrium systems in other fields, such as biophysics and computer science.