Path probability ratios for Langevin dynamics -- exact and approximate

TL;DR

This paper derives an exact path probability ratio for Langevin dynamics with a specific integrator, enabling precise reweighting of biased simulations, and shows that an approximate ratio is highly accurate for practical use.

Contribution

The authors derive an exact path probability ratio for Langevin Leapfrog integrator and demonstrate the high accuracy of an existing approximation for dynamic reweighting.

Findings

Exact path probability ratio $M_L$ for Langevin Leapfrog integrator derived.

Approximate ratio $M_{approx}$ differs from exact by $ ext{O}(\xi^4\Delta t^4)$, ensuring high accuracy.

Validated the method using butane simulations to demonstrate efficiency and accuracy.

Abstract

Path reweighting is a principally exact method to estimate dynamic properties from biased simulations - provided that the path probability ratio matches the stochastic integrator used in the simulation. Previously reported path probability ratios match the Euler-Maruyama scheme for overdamped Langevin dynamics. Since MD simulations use Langevin dynamics rather than overdamped Langevin dynamics, this severely impedes the application of path reweighting methods. Here, we derive the path probability ratio $M_L$ for Langevin dynamics propagated by a variant of the Langevin Leapfrog integrator. This new path probability ratio allows for exact reweighting of Langevin dynamics propagated by this integrator. We also show that a previously derived approximate path probability ratio $M_{\mathrm{approx}}$ differs from the exact $M_L$ only by $\mathcal{O}(\xi^4\Delta t^4)$, and thus yields highly accurate dynamic reweighting results. ($\Delta t$ is the integration time step, $\xi$ is the collision rate.) The results are tested and the efficiency of path-reweighting is explored using butane as an example.