An efficient multi-scale Green's function reaction dynamics scheme.

Area of Science:

• Computational chemistry
• Biophysics
• Chemical physics

Background:

• Particle-based simulation methods are crucial for understanding reaction-diffusion dynamics.
• Traditional methods like molecular dynamics (MD) are computationally expensive, especially at higher concentrations.
• Existing multiscale approaches, such as MD-GFRD, offer improved efficiency but can be further optimized.

Purpose of the Study:

• To present an efficient scheme for multiscale Molecular Dynamics-Green's Function Reaction Dynamics (MD-GFRD) simulations.
• To investigate the impact of propagation domain optimization on computational performance.
• To extend the concentration range where MD-GFRD is more efficient than traditional methods.

Main Methods:

• Developed an efficient coupling scheme for multiscale MD-GFRD simulations.
• Implemented domain optimization based on local concentration.
• Proposed a minimal domain size criterion for efficient switching between simulation scales.

Main Results:

• The optimized MD-GFRD scheme demonstrates superior efficiency compared to brute-force Brownian dynamics up to 10³ μM.
• The method is up to an order of magnitude more efficient than previous MD-GFRD schemes.
• Optimized domain construction significantly impacts computational performance.

Conclusions:

• The presented MD-GFRD scheme offers a significant computational advantage for simulating particle dynamics and reaction-diffusion systems.
• Domain optimization is critical for maximizing the efficiency of multiscale simulations.
• This method extends the applicability of efficient particle-based simulations to higher concentrations.