Short-time Fokker-Planck propagator beyond the Gaussian approximation

TL;DR

This paper introduces a perturbation method for accurately calculating the short-time Fokker-Planck propagator beyond Gaussian approximation, enabling precise expectation value evaluations and improved modeling of stochastic dynamics.

Contribution

The authors develop a perturbation approach that computes the short-time propagator of the Fokker-Planck equation to arbitrary order, surpassing the Gaussian approximation in accuracy.

Findings

Perturbation expansions match analytical solutions for specific systems.

Gaussian approximation errors can be several orders of magnitude larger.

Method enables more accurate stochastic process modeling and analysis.

Abstract

We present a perturbation approach to calculate the short-time propagator, or transition density, of the one-dimensional Fokker-Planck equation, to in principle arbitrary order in the time increment. Our approach preserves probability exactly and allows us to evaluate expectation values of analytical observables to in principle arbitrary accuracy; to showcase this, we derive perturbation expansions for the moments of the spatial increment, the finite-time Kramers-Moyal coefficients, and the mean medium entropy production rate. For an explicit multiplicative-noise system with available analytical solution, we validate all our perturbative results. Throughout, we compare our perturbative results to those obtained from the widely used Gaussian approximation of the short-time propagator; we demonstrate that this Gaussian propagator leads to errors that can be many orders of magnitude larger than those resulting from our perturbation approach. Potential applications of our results include parametrizing diffusive stochastic dynamics from observed time series, and sampling path spaces of stochastic trajectories numerically.