Fast-forward Langevin dynamics with momentum flips.

Area of Science:

• Computational Physics
• Molecular Dynamics Simulations
• Statistical Mechanics

Background:

• Stochastic thermostats, such as the Langevin equation, are widely used for temperature control in molecular dynamics (MD).
• Traditional Langevin thermostats exhibit sluggish behavior at high friction, limiting their efficiency in strong coupling regimes.
• Nosé-Hoover thermostats offer better performance in strong coupling but can have other limitations.

Purpose of the Study:

• To develop a modified Langevin thermostat that overcomes the high-friction sluggishness of traditional methods.
• To improve the sampling efficiency and ergodicity of molecular dynamics simulations under strong coupling conditions.
• To introduce a simple, easy-to-implement modification to the Langevin algorithm.

Main Methods:

• A novel integration scheme for the Langevin equation, termed the fast-forward Langevin equation, was proposed.
• This method involves reversing particle momentum if the thermostat action causes a direction change.
• The modified algorithm was tested on a 1D harmonic oscillator, water, and Lennard-Jones polymers.

Main Results:

• The fast-forward Langevin equation preserves the momentum distribution, ensuring correct equilibrium sampling.
• It mimics the quadratic behavior of Nosé-Hoover thermostats and performs comparably in the strong coupling limit.
• Sampling efficiency, measured by correlation time, was at least as good as the traditional Langevin thermostat, and significantly better (up to an order of magnitude) in the overdamped regime.

Conclusions:

• The fast-forward Langevin equation provides an effective solution to the sluggishness of traditional Langevin thermostats at high friction.
• This method enhances the performance of molecular dynamics simulations, particularly in strong coupling regimes.
• The proposed modification is simple to implement and maintains the ergodicity and accuracy of simulations.