Efficient sampling for ab initio Monte Carlo simulation of molecular clusters using an interpolated potential energy...

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Quantum Chemistry

Background:

• Ab initio Monte Carlo and path integral Monte Carlo methods are crucial for molecular simulations.
• Efficient sampling in configuration space is a key challenge for these methods.
• Accurate potential energy surfaces are computationally expensive to obtain.

Purpose of the Study:

• To develop an enhanced sampling technique for ab initio Monte Carlo (AIMC) and ab initio path integral Monte Carlo (aPIMC) calculations.
• To improve the efficiency and reduce the computational cost of simulating molecular clusters.
• To decrease statistical errors in AIMC and aPIMC simulations.

Main Methods:

• Utilizing an approximate potential, derived from the moving least-squares (MLS) method, to guide efficient movement in configuration space.
• Applying the developed scheme to simulations of a water molecule and small protonated water clusters (e.g., H3O+, H5O2+).
• Implementing a dynamic, automatic scheme to update the interpolation reference data set with ab initio data during simulations.

Main Results:

• Statistical errors were reduced by approximately a factor of 3 across most calculations.
• Computational cost was reduced by an order of magnitude due to enhanced sampling efficiency.
• The dynamic update scheme further accelerated convergence, improving simulation speed.

Conclusions:

• The proposed approximate potential-guided sampling method significantly enhances the efficiency of AIMC and aPIMC calculations for molecular clusters.
• This approach offers a substantial reduction in computational resources and statistical uncertainty.
• The dynamic data assimilation scheme provides a pathway for further optimization and faster convergence in complex molecular simulations.