It is the official implementation of "Multi-BOWS: Multi-Fidelity Multi-Objective Bayesian Optimization with Warm Starts for Nanophotonic Structure Design," which is published in Digital Discovery.
- src: a directory for source files
- materials: a directory for material definitions
Most of the packages required for this project are described in requirements.txt.
Along with the packages described in requirements.txt, our project requires lumapi, which is provided by Ansys Lumerical. The specifics of lumapi are described in the following link. Note that the Lumerical license is needed to run this Python API.
The simulations conducted in our paper can be run with the following commands in the src directory.
python run_bo_ours.py --str_fun threelayer --num_iter_low -1 --num_iter_high -1 --seed 42
python run_bo_ours.py --str_fun matched --num_iter_low -1 --num_iter_high -1 --seed 42
python run_bo_ours.py --str_fun unmatched --num_iter_low -1 --num_iter_high -1 --seed 42
python run_bo_ours.py --str_fun automatic --num_iter_low -1 --num_iter_high -1 --seed 42A set of two options --num_iter_low -1 --num_iter_high -1 indciates the use of the time budget declared in constants.py.
@article{KimJ2024dd,
title={{Multi-BOWS}: Multi-Fidelity Multi-Objective {Bayesian} Optimization with Warm Starts for Nanophotonic Structure Design},
author={Kim, Jungtaek and Li, Mingxuan and Li, Yirong and G\'{o}mez, Andr\'{e}s and Hinder, Oliver and Leu, Paul W.},
journal={Digital Discovery},
volume={3},
number={2},
pages={381--391},
year={2024}
}
This software is under the MIT license.