Data-driven molecular modeling with the generalized Langevin equation

• Published in: Journal of computational physics Vol. 418; p. 109633
• Main Authors: Grogan, Francesca, Lei, Huan, Li, Xiantao, Baker, Nathan A.
• Format: Journal Article
• Language: English
• Published: Cambridge Elsevier Inc 01.10.2020; Elsevier Science Ltd
• Subjects: Approximation; Coarse-grained models; Complexity; Computational physics; Data-driven parametrization; Dimension reduction; Generalized Langevin equation; Granulation; Mathematical analysis; Molecular dynamics; Molecular dynamics; Coarse-grained models; Dimension reduction; Generalized Langevin equation; Data-driven parametrization
• ISSN: 0021-9991; 1090-2716
• DOI: 10.1016/j.jcp.2020.109633

•Data-driven GLE approximation balances computational cost and accuracy.•Accuracy tunable by adjusting order of memory kernel approximation.•Approximate GLE predicts non-equilibrium properties, like autocorrelation, well. The complexity of molecular dynamics simulations necessitates dimension reduction and coarse-graining techniques to enable tractable computation. The generalized Langevin equation (GLE) describes coarse-grained dynamics in reduced dimensions. In spite of playing a crucial role in non-equilibrium dynamics, the memory kernel of the GLE is often ignored because it is difficult to characterize and expensive to solve. To address these issues, we construct a data-driven rational approximation to the GLE. Building upon previous work leveraging the GLE to simulate simple systems, we extend these results to more complex molecules, whose many degrees of freedom and complicated dynamics require approximation methods. We demonstrate the effectiveness of our approximation by testing it against exact methods and comparing observables such as autocorrelation and transition rates.