Performance of optimal linear-response processes in driven Brownian motion far from equilibrium

Abstract

Considering the paradigmatic driven Brownian motion, we perform extensive numerical analysis on the performance of optimal linear-response processes far from equilibrium. We focus on the overdamped regime where exact optimal processes are known analytically, and most experiments operate. This allows us to compare the optimal processes obtained in linear response and address their relevance to experiments, using realistic parameter values from experiments with optical tweezers. Our results help assess the accuracy of perturbative methods in calculating the irreversible work for cases where an exact solution does not exist. For that, we present a performance metric comparing the approximate optimal solution to the exact one. Our main result is that optimal linear-response processes can perform surprisingly well, even far from where they were expected.