VOPy: A Framework for Black-box Vector Optimization

What is VOPy?

VOPy is an open-source Python library built to address noisy black-box vector optimization problems, where the user preferences are encoded with a cone order.

What to do with VOPy?

VOPy includes several pre-implemented algorithms, models, orders, and problems from the literature for black-box vector optimization, allowing users to select and utilize components based on their specific needs. Specifically, you can:

• Use existing black-box vector optimization methods for new problems
• Benchmark novel algorithms with the state-of-the-art
• Utilize a wide range of existing tools in black-box vector optimization

How To Start?

Visit our website to see tutorials, examples and API references on how to use VOPy. We recommend starting out with the motivating example.

Setup

Installation using pip:

pip install vopy

Latest (Unstable) Version

To upgrade to the latest (unstable) version, run

pip install --upgrade git+https://github.com/Bilkent-CYBORG/VOPy.git

Manual installation (for development)

If you are contributing a pull request, it is best to perform a manual installation:

git clone https://github.com/Bilkent-CYBORG/VOPy.git
cd VOPy
mamba env create --name vopy --file environment.yml  # To setup a proper development environment
pip install -e .

For all development requirements, see requirements.txt or environment.yml.

Further, installing the pre-commit hooks are highly encouraged.

# Inside the package folder
pre-commit install

Citing

If you use VOPy, please cite the following paper:

@article{yildirim2024vopy,
  title={{VOPy}: A Framework for Black-box Vector Optimization},
  author={Yıldırım, Yaşar Cahit and Karagözlü, Efe Mert and Korkmaz, İlter Onat and Ararat, Çağın and Tekin, Cem},
  journal={arXiv preprint arXiv:2412.06604},
  year={2024}
}

License

VOPy is under MIT license.