A simple, flexible genetic algorithm library for Python. Easily define your own chromosomes, fitness functions, and evolutionary strategies to solve optimization and search problems.
- Minimal, intuitive API
- Customizable chromosomes and fitness functions
- Population management and evolution
pip install pygenaOr clone and install from source:
git clone https://github.com/atasoglu/pygena.git
cd pygena
pip install .Here's a minimal example of evolving a list of numbers to sum to a target:
from pygena import Chromosome, Population
import random
target = 100
chromosome_size = 10
population_size = 10
mutation_rate = 0.05
iterations = 100
def random_int(gene=None):
return random.randint(-100, 100)
def random_list():
return [random_int() for _ in range(chromosome_size)]
def fitness_fn(chromosome):
diff = abs(sum(chromosome.genes) - target)
return 1 / (diff + 1e-5)
random.seed(42)
population = Population(
chromosomes=[Chromosome(random_list()) for _ in range(population_size)],
mutation_rate=mutation_rate,
mutation_fn=random_int,
)
for i, local_best, global_best in population.run(iterations, fitness_fn):
print(f"Iteration {i}: {global_best.genes} (sum: {sum(global_best.genes)})")
if sum(global_best.genes) == target:
print(f"Target reached in {i} iterations.")
breakFor more examples, see the examples/ directory.
You can simply run an example script: python3 -m pygena.examples.text. Run with the --help flag to see the full list of arguments.
Contributions are welcome! Feel free to open issues or submit pull requests to help improve pygena.
MIT License. See LICENSE.