A practical perspective on the implementation of hyperdynamics for accelerated simulation.

Abstract

Consideration is given to several practical issues arising during the implementation of hyperdynamics, a methodology that extends the time scale of the conventional molecular dynamics simulation potentially by orders of magnitude. First, the methodology is reformulated in terms of the transition rate based on the buffer region approach (buffer rate), which can describe transitions in more general contexts than the transition state theory (TST). It will be shown that hyperdynamics can exactly preserve the buffer rate as well as the TST rate, which broadens the scope of the method. Next, the originally proposed scheme to compute the boost factor on-the-fly is reviewed and some alternative methods, one of which uses the umbrella sampling method, are presented. Finally, the methodology is validated in the context of a 1-dimensional example potential and a 3-dimensional simulation of the motion of an atomic force microscope tip moving along a surface.