Dresden 2026 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
DY: Fachverband Dynamik und Statistische Physik
DY 43: Poster: Statistical Physics
DY 43.16: Poster
Mittwoch, 11. März 2026, 15:00–18:00, P5
Linear coupling reduces entropy production in nonequilibrium multicomponent systems — •Vansh Kharbanda1,2, Anton Burnet1,2, and Benedikt Sabass1,2 — 1Faculty of Physics and Center for NanoScience, Ludwig-Maximilians-Universität München, 80752 Munich, Germany — 2Fakultät Physik, Technische Universität Dortmund, 44227 Dortmund, Germany
Many biological and synthetic systems operate far from equilibrium and pay an energetic cost quantified by the steady-state entropy production rate (EPR). We investigate how linear coupling between subsystems modifies this dissipation in a class of Hurwitz-stable Langevin dynamics. For networks of diffusively coupled Ornstein-Uhlenbeck processes, we derive closed-form expressions for the EPR of n-coupled two-dimensional units and show that any attractive symmetric coupling reduces the EPR, which monotonically approaches the single-unit value in the strong-coupling limit. This behaviour is traced to a suppression of probability currents, driving the collective closer to detailed balance. We then generalize to higher dimensions and arbitrary symmetric coupling topologies and conjecture a structural criterion: commuting coupling and diffusion matrices together with negative semidefinite coupling are sufficient for a monotonic EPR decrease. Violating this criterion generates mismatch-induced irreversible currents, which can increase EPR. Finally, we illustrate this mechanism in nonlinear biological models of bacterial chemotaxis and inner-ear hair bundles, where coupling substantially lowers dissipation without hindering performance metrics.
Keywords: Entropy production; Stochastic thermodynamics; Non-linear stochastic systems