Skip to content

handcraftsman/GeneticPy

Repository files navigation

GeneticPy

Genetic solver written in Python

Usage

import genetic

best = genetic.getBest(get_fitness, display, minLen, optimalFitness,
        geneSet=None, createGene=None, maxLen=None,
        customMutate=None, customCrossover=None)

Output:

an Individual with attributes:

  • Genes: list of genes
  • Fitness: fitness value for those genes
  • Strategy: strategy that produced the gene sequence (mutate, crossover, or random)

Inputs:

  • get_fitness: takes a list of genes and returns an integer value representing how close that particular candidate comes to the optimal solution. Higher values are better.
  • display: takes an Individual. Called to provide visual output of better candidates as they are discovered.
  • minLen: minimum valid length for an Individual's genes.
  • maxLen: optional maximum valid length for an Individual's genes. Defaults to minLen.
  • optimalFitness: expected fitness value for the best solution.
  • geneSet: optional list of gene values for generating new gene sequences. If not provided, createGene must be provided.
  • createGene: optional function to create a random gene. Called with the gene sequence index into which it will be placed, and the length of that gene sequence. If not provided, geneSet must be provided.
  • customMutate: optional replacement for the built in mutate function. Called with the new child genes. Changes should be made to the child genes.
  • customCrossover: optional replacement for the built in crossover function. Called with the new child genes, and a copy of a second parent's genes. Changes should be made to the child genes.

Sample projects (in order of genetic complexity)

Obsolete samples

The following sample projects included in this repository work but no longer have a related blog post because I've changed my implementation approach.

  • equationGenerationTests.py - equation discovery

  • operationGenerationTests.py - generating OR and XOR with only AND and NOT

e-book

The code and examples in ths repository work as is and you are welcome to use them. In late 2015 I restarted from scratch with a slower feature progression to resolve some performance issues with more advanced genetic programming topics. That has evolved into an e-book entitled Genetic Algorithms with Python that I'm writing on Leanpub. Try the sample chapter.

License

MIT License

Releases

No releases published

Packages

No packages published

Languages