Interatomic potential parameterization using particle swarm optimization: Case study of glassy silica

Rasmus Christensen, Søren Strandskov Sørensen, Han Liu, Kevin Li, Mathieu Bauchy*, Morten Mattrup Smedskjær*

*Corresponding author for this work

Research output: Contribution to journalJournal articleResearchpeer-review

8 Citations (Scopus)
45 Downloads (Pure)

Abstract

Classical molecular dynamics simulations of glassy materials rely on the availability of accurate yet computationally efficient interatomic force fields. The parameterization of new potentials remains challenging due to the non-convex nature of the accompanying optimization problem, which renders the traditional optimization methods inefficient or subject to bias. In this study, we present a new parameterization method based on particle swarm optimization (PSO), which is a stochastic population-based optimization method. Using glassy silica as a case study, we introduce two interatomic potentials using PSO, which are parameterized so as to match structural features obtained from ab initio simulations and experimental neutron diffraction data. We find that the PSO algorithm is highly efficient at searching for and identifying viable potential parameters that reproduce the structural features used as the target in the parameterization. The presented approach is very general and can be easily applied to other interatomic potential parameterization schemes.

Original languageEnglish
Article number134505
JournalJournal of Chemical Physics
Volume154
Issue number13
ISSN0021-9606
DOIs
Publication statusPublished - 7 Apr 2021

Fingerprint

Dive into the research topics of 'Interatomic potential parameterization using particle swarm optimization: Case study of glassy silica'. Together they form a unique fingerprint.

Cite this