/FFEAT

Framework for Evolutionary Algorithms in Torch

Primary LanguagePythonOtherNOASSERTION

FFEAT

Framework For Evolutionary Algorithms in Torch

This library implements various evolutionary algorithms, specifically

  • Genetic Algorithms in ffeat.genetic module.
  • Real-Coded Evolutionary Algorithms in ffeat.strategies module.
  • Evolution Strategies in ffeat.strategies module.
  • Particle Swarm Optimisation in ffeat.pso module.

The algorithms are fully vectorized and can run on GPU.

Each module consists of selection, crossover, and mutation submodule implementing relevant operators (with the exception of PSO algorithm). The operators may be arbitrarily combined.

See examples for more information on how to use the library.

This library was developed as part of my master thesis: https://github.com/PatrikValkovic/MasterThesis. You can find more information about the implementation there.

Author: Patrik Valkovič

License: MIT