Fluctuation-dissipation relations for a general class of master equations

TL;DR

This paper derives a generalized fluctuation-dissipation relation for master equation models, revealing an additional function needed for non-equilibrium conditions, which depends on the system's temporal evolution and is not directly observable.

Contribution

It introduces a new function $\xi(t,t_w)$ into the fluctuation-dissipation relation for master equations, extending understanding of response in non-equilibrium stochastic systems.

Findings

$\xi(t,t_w)$ vanishes at equilibrium.

Models with certain properties show $\xi(t,t_w)$ approaching zero.

The relation applies to systems with prescribed variable-state connections.

Abstract

The fluctuation-dissipation relation is calculated for a class of stochastic models obeying a master equation. The transition rates are assumed to obey detailed balance also in the presence of a field. It is shown that in general the linear response cannot be expressed via time-derivatives of the correlation function alone, but an additional function $\xi(t,t_w)$, which has been rarely discussed before is required. This function depends on the two times also relevant for the response and the correlation and vanishes under equilibrium conditions. It can be expressed in terms of the propagators and the transition rates of the master equation but it is not related to any physical observable in an obvious way. Instead, it is determined by inhomogeneities in the temporal evolution of the distribution function of the stochastic variable under consideration. $\xi(t,t_w)$ is considered for some examples of stochastic models, some of which exhibit true non-equilibrium dynamics and others approach equilibrium in the long term. In particular, models in which a relevant variable, e.g. a magnetization, is related in a prescribed way to the states of the system are considered as projections from a composite Markov process. From these model calculations, it is conjectured that $\xi(t,t_w)$ vanishes when one is concerned with measurements of the analogue of a structure factor for large wave-vectors or if every transition among the states randomizes the value of the stochastic variable considered.