Lower bounds on entropy production from dynamical correlation functions
Abstract
Entropy production is a key property in stochastic thermodynamics. For partially observed and coarse-grained systems, its inference is challenging and typically rests on proven lower bounds. We derive two versions of such bounds based on the asymmetry of experimentally accessible two-time correlation functions of coarse-grained state observables. For non-equilibrium steady states, the bound is valid for arbitrary correlation lag. For time-dependent processes, it requires the limit of vanishing lag. These bounds hold true for any system that follows either a Markovian dynamics or a coupled set of overdamped Langevin equations on some underlying, unobservable level of description. We illustrate the bounds for both types of dynamics and discuss their optimization and potential tightness.
Related
- Finite temperature Fredholm determinants for dynamical correlation functions
- Asymptotic analysis of a Fredholm determinant occurring in the description of the dynamical correlation functions of the Lieb--Liniger Bose gas
- Maximum-Entropy Extension of Quantum Correlation Functions from Short Real-Time Dynamics
- Jensen-Mercer two-sided bounds for entropy production rate
- Mpemba effect as an entropy-production race:Finite-amplitude bounds on crossing times in Markovian relaxation
BibTeX
@misc{raux2026lower,
title = {{Lower bounds on entropy production from dynamical correlation functions}},
author = {Paul Raux and Alexander M. Maier and Udo Seifert},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2608.03619},
eprint = {2608.03619},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.03619},
}
RIS
TY - UNPB
TI - Lower bounds on entropy production from dynamical correlation functions
AU - Paul Raux
AU - Alexander M. Maier
AU - Udo Seifert
PY - 2026
DA - 2026/10/12
PB - arXiv
DO - 10.48550/arxiv.2608.03619
UR - https://arxiv.org/abs/2608.03619
AB - Entropy production is a key property in stochastic thermodynamics. For partially observed and coarse-grained systems, its inference is challenging and typically rests on proven lower bounds. We derive two versions of such bounds based on the asymmetry of experimentally accessible two-time correlation functions of coarse-grained state observables. For non-equilibrium steady states, the bound is valid for arbitrary correlation lag. For time-dependent processes, it requires the limit of vanishing lag. These bounds hold true for any system that follows either a Markovian dynamics or a coupled set of overdamped Langevin equations on some underlying, unobservable level of description. We illustrate the bounds for both types of dynamics and discuss their optimization and potential tightness.
ER -