Evaluation of Various Molecular Dynamics Algorithms

Molecular dynamics method is a technique to understand physical phenomena by tracing movements of particles from a microscopic viewpoint. Many algorithms have been proposed in order to solve a lot of physical problems. We investigate the property, accuracy, stability and calculation efficiency of seven representative algorithms : velocity Verlet, leap frog. 4 value Gear, Beeman, Runge Kutta, position Verlet and Tuckerman Berne. The Lenard Jones potential was used in this study. The velocity Verlet algorithm is useful for the problem under high density or low temperature conditions. The position Verlet algorithm is useful under high density or high temperature conditions. The Tuckerman Berne algorithm is useful under low density conditions. The Tuckerman-Berne algorithm has high stability and its calculation time is shortest in the algorithms used in this study. However, its precision is low. Therefore, we need to choose the optimum algorithm in consideration of the state of the focused system.