# Fileset

[Supplemental.pdf](https://mdr.nims.go.jp/filesets/9be7e110-77e4-4e9d-90fe-11aa262b688a/download)

## Creator

Hiroyuki Tanaka, Kentaro Kutsukake, Kota Asakura, Takuto Kojima, Xin Liu, Noritaka Usami

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[A design methodology of crystal growth furnace and process aided by two-step optimization using machine learning models and genetic algorithm](https://mdr.nims.go.jp/datasets/0bfe60fa-8ad3-424d-9e8a-48d84b8392f8)

## Fulltext

Supplemental  A design methodology of crystal growth furnace and process aided by two-step optimization using machine learning models and genetic algorithm  Hiroyuki Tanakaa, Kentaro Kutsukakea,b,c, Kota Asakuraa, Takuto Kojimad, Xin Liub, and Noritaka Usamia,b,e*  aGraduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan bInstitute of Materials and Systems for Sustainability, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan  c Center of Advanced Intelligence Project, RIKEN, Tokyo 103-0027, Japan d National Institute of Advanced Industrial Science and Technology, Tsukuba 305-8561, Japan e Institutes of Innovation for Future Society, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan  *E-mail: usa@material.nagoya-u.ac.jp    mailto:usa@material.nagoya-u.ac.jpTable S1. DNN accuracy of Molde A   MSE MAE R2 fold1 T[K] t (min) 4.639 2.769 0.998 log10[N/(/m2)] 0.110 0.059 0.872 fold2 T[K] t (min) 4.347 2.707 0.998 log10[N/(/m2)] 0.066 0.037 0.944 fold3 T[K] t (min) 6.416 3.533 0.996 log10[N/(/m2)] 0.125 0.053 0.818 fold4 T[K] t (min) 4.918 2.663 0.997 log10[N/(/m2)] 0.118 0.053 0.881 fold5 T[K] t (min) 4.473 2.843 0.998 log10[N/(/m2)] 0.092 0.044 0.888    Table S2. DNN accuracy of Molde B            MSE MAE R2 fold1 T (K) 19.012 13.224 0.793 fold2 T (K) 15.828 11.481 0.854 fold3 T (K) 19.23 14.077 0.788 fold4 T (K) 21.29 14.194 0.742 fold5 T (K) 14.638 10.902 0.862  Figure S1. The progress of the three objective functions in the genetic algorithm for Model A. The solution converged sufficiently within the 200 generations.    Figure S2. The progress of the three objective functions in the genetic algorithm for Model B. The solution converged sufficiently within the 200 generations.