Combinatorial Optimization via Memory Metropolis: Template Networks for Proposal Distributions in Simulated Annealing applied to Nanophotonic Inverse Design

Code for the paper Combinatorial Optimization via Memory Metropolis: Template Networks for Proposal Distributions in Simulated Annealing applied to Nanophotonic Inverse Design at AI4MAT workshop at NeurIPS 2023

Installation

Install requirements using pip pip -r requirements.txt. The simulation backend simFrame will be released soon. For now a dummy backend class is implemented instead.

Execution

python main.py --config configurations/algorithms/parallelMeme.yml --device_config configurations/devices/mdm.yml