genepi

genepi is a genetic algorithm framework written in python.

Documentation

Documentation is being built with Sphinx and will come soon. You can build a partial documentation by installing sphinx and issuing

make html

in the docs folder of genepi source.

Until then, here is a little example.

from genepi.core.ga import GeneticAlgorithm
from genepi.core.factories import protogene_factory
from genepi.core.protogenome import ProtoGenome
from genepi.core.gene import IntGene
from genepi.core.stopcriteria import raw_score_stop
from genepi.core.crossover import single_point_crossover
from genepi.core.selectors import select_from_top


#example fitness function
#fitness is the sum of parameters
def fitness_function(genome):
    score = 0.0
    for value in genome.list_value():
        score += value
    return score


# example genetic algorithm
# we want to minimize the sum of a list of 50 integer numbers from 0 to 10

protogenes = protogene_factory(IntGene, 'x', 50, min_value=0, max_value=10)
protogenome = ProtoGenome(protogenes, mutation_probability = 0.1)

algo = GeneticAlgorithm(protogenome, fitness_function,
    population_size = 200,
    optimization_mode = 'min',
    num_parents = 4,
    crossover_method = single_point_crossover,
    selection_method = select_from_top,
    termination_criteria=[raw_score_stop], stop_score=0)

#main evolution cycle
algo.evolve(debug=True)

bi = algo.best_individual()
print bi.score
print bi.dict_value()

More examples can be found in test/integration_test folder.