Skip to content

Home

EvoBandits

EvoBandits is a cutting-edge optimization algorithm that merges genetic algorithms and multi-armed bandit strategies to efficiently solve stochastic problems.

Build & Test

EvoBandits (Evolutionary Multi-Armed Bandits) is an innovative optimization algorithm designed to tackle stochastic problems efficiently. EvoBandits offers a reinforcement learning-based approach to solving complex, large-scale optimization issues by combining genetic algorithms with multi-armed bandit mechanisms. Whether you're working in operations research, machine learning, or data science, EvoBandits provides a robust, scalable solution for optimizing your stochastic models.

Usage

To install EvoBandits:

pip install evobandits

from evobandits import GMAB

def test_function(number: list) -> float:
    # your function here

if __name__ == '__main__':
    bounds = [(-5, 10), (-5, 10)]
    algorithm = GMAB(test_function, bounds)
    n_trials = 10000
    result = algorithm.optimize(n_trials)
    print(result)

Contributing

Pull requests are welcome. For major changes, please open a discussion first to talk about what you'd like to change.

License

EvoBandits is licensed under the Apache-2.0 license (LICENSE or https://opensource.org/licenses/apache-2-0).

Credit

Deniz Preil wrote the initial EvoBandits prototype in C++, which Timo Kühne and Jonathan Laib rewrote. Timo Kühne ported to Rust, which is now the backend. Felix Würmseher added the Python frontend.

Citing EvoBandits

If you use EvoBandits in your research, please cite the following paper:

Preil, D., & Krapp, M. (2024). Genetic Multi-Armed Bandits: A Reinforcement Learning Inspired Approach for Simulation Optimization. IEEE Transactions on Evolutionary Computation.