Project

MetaZoo

· Updated

MetaZoo

A zoo of metaheuristic algorithms (please do not feed them) for optimization. MetaZoo is a Python library implementing a variety of metaheuristic algorithms and tooling for learning and experimentation. The project includes interactive resources (a Streamlit “PettingZoo” app and Jupyter Playgrounds) and a lightweight gym-like environment for algorithm evaluation.

⚠️ This library is in preview and is not ready for production use. Expect bugs and missing features.

Highlights

  • Collection of metaheuristic algorithms (genetic algorithms, differential evolution, particle swarm, and more)
  • Modular architecture with a core library, algorithm modules, and application/service layer
  • Learning resources: an interactive Streamlit app (PettingZoo) and Jupyter Playgrounds
  • Lightweight Gym-style environments for running experiments and visualizations

Modules (examples)

  • Genetic Algorithms (GA)
  • Differential Evolution (DE)
  • Particle Swarm Optimization (PSO)
  • Utilities: crossover, mutation, selection operators and evolutionary utilities
  • Environments and application layer for experiments and visualization

Quickstart

  1. Clone the repository:

    git clone https://github.com/roicort/MetaZoo.git

  2. Create a virtual environment and install dependencies (example using pip):

    python -m venv .venv && source .venv/bin/activate

    pip install -r requirements.txt

  3. Run the interactive PettingZoo app (if available):

    streamlit run pettingzoo/app.py (see repository for exact path and commands)

  4. Explore the Playgrounds notebooks: open the playgrounds/ notebooks with JupyterLab or VS Code and experiment with algorithms and parameters.

Architecture

MetaZoo follows a modular design: core library (algorithms and operators), modules (bio/physics/human-inspired categories), a Gym-like environment for evaluation, and application/service wrappers for demos and apps. The README contains a mermaid diagram showing component relationships.

Status & contributing

  • Preview state — actively developed. Check the repository README for the current checklist and issues.
  • Contributions, issues, and experiments are welcome.

Links: Repository · License: see repo