Tutorials

The direct problem solver, sim_trhepd_rheed, is based on the Reflection-High-Energy Electron Diffraction (RHEED, TRHEPD) analysis software developed by Prof. Takashi Hanada at Tohoku University. In TRHEPD, when atomic coordinates are given, diffraction data is given as a simulation result. Therefore, we are dealing with the direct problem from atomic coordinates to diffraction data. On the other hand, in many cases, diffraction data is given experimentally, and the atomic coordinates are required to reproduce the experimental data. These are inverse problems to the above direct problems.

In 2DMAT, the algorithms for solving the inverse problem can be selected as following algorithms:

• minsearch

  Estimating plausible atomic coordinates using the Nealder-Mead method.

• mapper_mpi

  Estimate plausible atomic coordinates by searching the entire search grid for a given parameter.

• bayes

  Estimate plausible atomic coordinates using Bayesian optimization.

• exchange

  Sampling plausible atomic coordinates using a replica exchange Monte Carlo method.

• pamc

  Sampling plausible atomic coordinates using a population annealing Monte Carlo method.

In this tutorial, we will first introduce how to run the sequential problem program, and then how to run minsearch , mapper_mpi, bayes, exchange, and pamc .

• TRHEPD Direct Problem Solver
• Optimization by Nelder-Mead method
• Grid search
• Optimization by Bayesian Optimization
• Optimization by replica exchange Monte Carlo
• Replica Exchange Monte Carlo search with limitation
• Optimization by population annealing
• Addition of a direct problem solver