Related Experiment Video
Updated: Apr 29, 2026
10:52
Multiscale Sampling of a Heterogeneous Water/Metal Catalyst Interface using Density Functional Theory and Force-Field Molecular Dynamics
Published on: April 12, 2019
14.1K
Using multiscale preconditioning to accelerate the convergence of iterative molecular calculations
Jeremy O B Tempkin1, Bo Qi1, Marissa G Saunders1
1Department of Chemistry, The University of Chicago, Chicago, Illinois 60637, USA.
The Journal of Chemical Physics
|May 17, 2014
Summary
This study introduces a preconditioning scheme to accelerate molecular model optimization. Using a less expensive model guides exploration of a more expensive model, speeding up the discovery of stable states.
Area of Science:
- Computational Chemistry
- Molecular Modeling
- Materials Science
Background:
- Optimizing molecular models is computationally intensive.
- Slow convergence hinders the discovery of stable molecular states.
- Accurate representation of molecular features requires significant resources.
Purpose of the Study:
- To develop a preconditioning scheme for accelerating molecular model optimization.
- To enable the use of less expensive models to guide more expensive ones.
- To speed up the discovery of locally stable states in complex molecular systems.
Main Methods:
- Introduced a preconditioning scheme for multilevel simulations.
- Applied the scheme to energy minimization.
- Utilized the scheme with the string method for transition pathway analysis.
Main Results:
- The preconditioning scheme effectively guides exploration of complex energy landscapes.
- Accelerated discovery of locally stable states for more expensive molecular models.
- Demonstrated applicability in both energy minimization and transition pathway searches.
Conclusions:
- The proposed preconditioning scheme offers a significant speedup in molecular modeling.
- This approach facilitates efficient exploration of energy landscapes.
- The method shows promise for various computational chemistry applications.
