Prototype implementation of CatMADS for the paper CatMADS: MESH ADAPTIVE DIRECT SEARCH FOR CONSTRAINED BLACKBOX OPTIMIZATION WITH CATEGORICAL VARIABLES.
The current code is a proof of concept. While the code is shared for transparency, it is still under development and not intended for direct reuse at this stage. The project remains under active development. It can possibly be used by setting the appropriate paths in CatMADS_built_on_NOMAD/CatMADS/CatMADS.cpp, installing the required Python packages and compiling NOMAD as described in the official guide https://nomad-4-user-guide.readthedocs.io/en/latest/Installation.html
If you are interested in using CatMADS, feel free to contact [email protected]
This repository contains 32 analytical mixed-variable optimization problems, including 16 constrained ones, for benchmarking.