[22p-M206-1~16] 23.1 Joint Session N "Informatics"

[22p-M206-15] Application of parallel Monte Carlo Method to Total Reflection High-energy Positron Diffraction (TRHEPD) Experiment -Analysis of Bayesian Posterior Probability Distribution-

〇Takeo Hoshi^1,2, Harumichi Iwamoto¹, Kinoshita Naoki¹, Yuuma Terachi¹ (1.Tottori Univ., 2.KEK-IMSS-SPF)

Keywords:massively parallel Monte Carlo method, Total Reflection High-energy Positron Diffraction (TRHEPD), Bayesian posterior probability distribution

We applied the parallel Monte Carlo method to Total Reflection High-energy Positron Diffraction (TRHEPD) experiment using our recently developed measurement data analysis framework, 2DMAT. As a test system, the structural analysis (determination of atomic coordinates) of the Ge(001)-c4x2 surface structure was performed using the supercomputer Fugaku. In particular, we report on the Bayesian posterior probability distribution with multiple solutions in high-dimensional data space, which was analyzed using principal component analysis and other methods.