Semi-stochastic full configuration interaction quantum Monte Carlo: Developments and application.

Area of Science:

• Quantum Chemistry
• Computational Physics
• Materials Science

Background:

• Full Configuration Interaction Quantum Monte Carlo (FCIQMC) is a powerful method for electronic structure calculations.
• Recent advancements include semi-stochastic adaptations to improve efficiency.
• Generating the deterministic space often requires prior knowledge of the wave function.

Purpose of the Study:

• To present an alternate, efficient semi-stochastic adaptation for FCIQMC.
• To investigate its performance across molecular and lattice systems.
• To enhance the calculation of reduced density matrices and obtain accurate energies.

Main Methods:

• Developed a novel method for deterministic space generation without a priori wave function knowledge.
• Implemented an efficient semi-stochastic FCIQMC algorithm.
• Utilized replica sampling for reduced density matrix calculations.
• Applied the method to explicitly correlated corrected FCIQMC energy calculations.

Main Results:

• Demonstrated stochastic efficiencies for various molecular and lattice systems.
• Showcased the positive impact of the adaptation on parallel performance in FCIQMC.
• Observed significant efficiency gains in reduced density matrix calculations using replica sampling.
• Achieved wavenumber accuracy in stochastic errors for beryllium dimer energy calculations.

Conclusions:

• The semi-stochastic FCIQMC adaptation offers substantial computational advantages.
• The method enhances accuracy and efficiency for electronic structure calculations.
• It enables high-precision energy calculations for challenging systems like the beryllium dimer.