Skip to content
forked from DEAP/deap

Distributed Evolutionary Algorithms in Python - Entirely Reworked

License

Notifications You must be signed in to change notification settings

aabmets/deap-er

 
 

Repository files navigation

DEAP-ER

DEAP-ER is a complete rewrite of the original DEAP library for Python 3.10 and up, which includes features such as:

  • Genetic algorithms using any imaginable containers like:
    • List, Array, Set, Dictionary, Tree, Numpy Array, etc.
  • Genetic programming using prefix trees
    • Loosely typed, Strongly typed
    • Automatically defined functions
  • Evolution Strategies (Covariance Matrix Adaptation)
  • Multi-objective optimisation (SPEA-II, NSGA-II, NSGA-III, MO-CMA)
  • Co-evolution (cooperative and competitive) of multiple populations
  • Parallelization of evolution processes using multiprocessing or with Ray
  • Records to track the evolution and to collect the best individuals
  • Checkpoints to persist the progress of evolutions to disk
  • Benchmarks to test evolution algorithms against common test functions
  • Genealogy of an evolution, that is also compatible with NetworkX
  • Examples of alternative algorithms:
    • Symbolic Regression,
    • Particle Swarm Optimization,
    • Differential Evolution,
    • Estimation of Distribution Algorithm

Documentation

See the Documentation for the complete guide to using this library.

Contributing

Please read the CONTRIBUTING.md file before submitting pull requests.

About

Distributed Evolutionary Algorithms in Python - Entirely Reworked

Resources

License

Stars

Watchers

Forks

Languages

  • Python 100.0%