pymoo
Latest Version: pymoo==0.3.0

pymoo

The framework is available on PyPi and can be installed with:

pip install -U Cython>=0.29 numpy>=1.15 pymoo

Please note, that the dependencies used in the command above must be fulfilled before compiling some modules of pymoo and can, therefore, not be ensured to be installed prior compilation with the setup script. More details about the installation can be found here.

Getting Started

If you intend to use pymoo for any profit-making purposes, please contact Julian Blank.

We are currently working on a paper. Meanwhile, if you have used our framework for research purposes, please cite us with:

@misc{pymoo,
    author = {Julian Blank and Kalyanmoy Deb},
    title = {pymoo - {Multi-objective Optimization in Python}},
    howpublished = {https://pymoo.org}
}

Features

Algorithms: Genetic Algorithm, Differential Evolution, NSGA-II, R-NSGA-II, NSGA-III, U-NSGA-III, R-NSGA-III, MOEA/D

Performance Indicators: Hypervolume, GD, IGD, R-Metric

Non-Dominated Sorting: Naive, Fast, Best Order

Random Generators: Custom, Python, Numpy

Selection: Random, Tournament Selection

Sampling: Random, Latin Hypercube Sampling

Crossover: Simulated Binary Crossover, Uniform Crossover, Half Uniform Crossover, Differential Crossover, Point Crossover, Exponential Crossover

Mutation: Polynomial Mutation, Bitflip Mutation

About

This framework is developed and maintained by Julian Blank who is affiliated to the Computational Optimization and Innovation Laboratory (COIN) supervised by Kalyanmoy Deb at the Michigan State University in East Lansing, Michigan, USA. Each algorithms is developed as close as possible to the proposed version to the best of our knowledge. NSGA-II and NSGA-III have been develop collaboratively with one of the authors and, therefore, we recommend using them for official benchmarks.

News

April 10, 2019: The framework has reached a new degree of professionality by improving the software documentation regarding tutorial and API.

Indices and tables