Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

DY: Fachverband Dynamik und Statistische Physik

DY 21: Stochastic Thermodynamics

DY 21.9: Vortrag

Dienstag, 10. März 2026, 12:00–12:15, ZEU/0114

Thermodynamic optimal control out of equilibrium: insights from active and driven systems — •Kristian Olsen1, Rémi Goerlich2,1, Yael Roichman2,3, and Hartmut Löwen11Institut für Theoretische Physik II - Weiche Materie, Heinrich-Heine-Universität Düsseldorf, D-40225 Düsseldorf, Germany — 2Raymond & Beverly Sackler School of Chemistry, Tel Aviv University, Tel Aviv 6997801, Israel — 3Raymond & Beverly Sackler School of Physics & Astronomy, Tel Aviv University, Tel Aviv 6997801, Israel

Optimal control far from equilibrium raises intriguing questions in stochastic thermodynamics and offers a route to design microscopic engines that operate out of equilibrium. We study thermodynamically optimal protocols for a harmonically trapped particle driven by arbitrary time-dependent forces, including those from active matter or external fields [1]. The resulting protocols are able to harness non-equilibrium forces to extract a net work. We provide exact solutions for the optimal protocol and associated work for arbitrary forces and protocol duration. We also derive a quasistatic work bound that splits into three parts: an information-geometric term capturing energy stored in an initial non-equilibrium state, the work extracted from time-averaged forces, and extra work from fast dynamical modes. Finally, we analyze the energetic cost of adding boundary constraints, giving insights into the cost of precision in these protocols.

1. Harnessing non-equilibrium forces to optimize work extraction, Kristian Stølevik Olsen, Rémi Goerlich, Yael Roichman and Hartmut Löwen, In press Nature Communications, arXiv: 2504.07049, 2025.

Keywords: Stochastic thermodynamics; Optimal control; Active matter

100% | Bildschirmansicht | English Version | Kontakt/Impressum/Datenschutz
DPG-Physik > DPG-Verhandlungen > 2026 > Dresden