Crystal structure prediction of materials with high symmetry using differential evolution

Crystal structures influence material properties. Existing methods rely on contact maps and global optimization. However, high-symmetry structures pose unique challenges due to dimensional mismatches in contact maps. Prior research has shown that contact maps can guide crystal reconstruction. This gap motivated the development of new strategies for high-symmetry crystal prediction. Known approaches struggle with symmetry-related constraints. No prior work had resolved the dimensional inconsistency issue. This study addresses that limitation with a novel algorithmic approach.

This paper aims to improve crystal structure prediction for high-symmetry materials. The specific problem is the mismatch in contact map dimensions during optimization. The motivation comes from the need to predict stable structures with high symmetry. Existing global optimization algorithms fail in this context. The study proposes a solution using differential evolution. The goal is to generate chemically valid structures under symmetry constraints. The approach combines PyXtal with contact map optimization. This addresses a critical limitation in current crystal prediction methods.

The study uses PyXtal to generate crystal structures with specified symmetry. Random structures are filtered based on chemical formulas and space groups. Contact maps are used as optimization targets. Differential evolution algorithms replace traditional global optimization. Non-special coordinates at Wyckoff positions are searched. The algorithm is named CMCrystalHS. The method avoids dimensional mismatches in contact maps. This approach enables high-symmetry crystal prediction.

The proposed CMCrystalHS algorithm successfully predicts high-symmetry crystal structures. Contact map inconsistencies are resolved through differential evolution. The method generates chemically valid structures with given symmetry. Experimental results confirm the algorithm's effectiveness. The predicted structures match target contact maps. The approach outperforms prior optimization strategies. Dimensional mismatches are no longer a barrier. This demonstrates the algorithm's potential for material discovery.

The study concludes that differential evolution improves high-symmetry crystal prediction. The CMCrystalHS algorithm addresses dimensional inconsistencies. Contact maps remain central to the optimization process. The method enables accurate structure prediction under symmetry constraints. The findings suggest that PyXtal integration is beneficial. The results support the use of differential evolution over genetic algorithms. The approach is effective for materials with high symmetry. The authors propose that this method advances crystal structure prediction.