Overcoming the minimum image constraint using the closest point search.

Area of Science:

• Computational physics
• Molecular dynamics simulations
• Algorithm development

Background:

• Periodic boundary conditions are crucial for molecular dynamics (MD) simulations to mimic bulk systems.
• Efficiently finding nearest neighbors is essential for linear-time computation of pairwise distances.
• Existing neighbor-listing algorithms face challenges with large cutoff distances relative to cell size.

Purpose of the Study:

• To develop novel algorithms for generating region-region interaction lists in n-dimensional space.
• To overcome the minimum image restriction in periodic boundary condition calculations.
• To provide efficient solutions for molecular dynamics simulations with large cutoffs.

Main Methods:

• Refined closest vector problem algorithms for direct region-region list generation.
• Lattice reduction methods for efficient higher-dimensional searches.
• Implementation details for n-dimensional space searches free from minimum image restriction.

Main Results:

• Two direct and efficient methods for generating region-region interaction lists are detailed.
• The algorithms effectively handle cutoff distances exceeding half the shortest cell length.
• Lattice reduction criteria offer guidelines for lattice compaction in MD simulations.

Conclusions:

• The proposed methods provide efficient solutions for nearest image searches in MD simulations.
• These algorithms enhance the performance of molecular dynamics by addressing limitations with large cutoffs.
• The work contributes to the development of more scalable and accurate simulation techniques.