The article introduces the GradNav algorithm to enhance exploration of potential energy surfaces. It addresses challenges in molecular simulations, such as escaping deep potential wells and reducing sensitivity to initial conditions. The algorithm iteratively runs short simulation segments, updating starting points based on observation density gradients. Evaluation metrics DWEF and SSIR demonstrate its effectiveness in escaping wells and reducing initialization dependency. Applications include Langevin dynamics simulations and molecular dynamics of the Fs-Peptide protein.
To Another Language
from source content
arxiv.org
Key Insights Distilled From
by Janghoon Ock... at arxiv.org 03-18-2024
https://arxiv.org/pdf/2403.10358.pdfDeeper Inquiries