Computational Methods for Long-Timescale Atomistic Simulations

Abstract

Because of their considerable predictive power, atomistic simulations are extremely powerful tools in the computational materials scientist’s toolbox. This power however comes at a significant computational price that rather strongly limits the accessible simulation space, especially in terms of the timescales that can be directly simulated. Specialized methods specifically designed to overcome timescale limitations while still faithful to the underlying dynamical behavior of the system are hence essential in order to bridge the gap with experiments. This section summarizes the most recent advances in a class of open-ended long-timescale atomistic simulation techniques that include accelerated molecular dynamics and kinetic Monte Carlo methods. The different chapters introduce the basics of these methods as well as a review of their most recent developments.