[9a-W321-1~11] 23.1 Joint Session N "Informatics"

[9a-W321-5] Optimization of atomic arrangement in crystal structures using Naïve Bayes

〇Yusuke Noda¹, Tatsuya Yokoi¹, Katsuyuki Matsunaga¹ (1.Nagoya Univ.)

Keywords:combinatorial optimization, crystal structure, atomic arrangement

In this study, we develop a data-driven combinatorial optimization approach using Naïve Bayes and apply it to the optimization problem of atomic arrangements. We set a combinatorial optimization problem for Si₅₀C₅₀ three-dimensional crystal structures as a case study, and validate the performance to predict a perfect silicon carbide SiC. It is confirmed that the proposed Naïve Bayes-based approach is better at the prediction performance than a genetic algorithm approach.