Skip to main content

Genetic Algorithm Library

Project description

Holland

Genetic Algorithm Library for Python

Computer programs that "evolve" in ways that resemble natural selection can solve complex problems even their creators do not fully understand

PyPI Build Coverage Documentation Status License: MIT

Description

Holland is a simple, flexible package for implementing the Genetic Algorithm in Python. The program is designed to act on an arbitrary evaluation function with arbitrary encoding of individuals within a population, both of which are provided by the user.

Installing

Holland is available via the Python Package Index (PyPI) and can be installed with:

pip install holland

Usage

Full Documentation

Hello World!

from holland import Evolver
from holland.library import get_uniform_crossover_function
from holland.utils import bound_value
import random


# Define a fitness function
def fitness_function(genome):
    message = genome["message"]
    target = "Hello World!"
    score = 0
    for i in range(len(message)):
        score += abs(ord(target[i]) - ord(message[i]))
    return score

def mutation_function(value):
    mutated_value = ord(value) * random.random() * 2
    return chr(bound_value(mutated_value, minimum=32, maximum=126, to_int=True))

# Define genome parameters for individuals
genome_params = {
    "message": {
        "type": "[str]",
        "size": len("Hello World!"),
        "initial_distribution": lambda: chr(random.randint(32, 126)),
        "crossover_function": get_uniform_crossover_function(),
        "mutation_function": mutation_function,
        "mutation_rate": 0.15
    }
}

# Define how to select individuals for reproduction
selection_strategy = {"pool": {"top": 10}}

# Run Evolution
evolver = Evolver(
    fitness_function,
    genome_params,
    selection_strategy,
    should_maximize_fitness=False
)
final_population = evolver.evolve(stop_conditions={"target_fitness": 0})

With sample run:

Generation: 0; Top Score: 201:     N~flx.JGcu-*

Generation: 1; Top Score: 98:       Xljlw);mj]f

Generation: 2; Top Score: 64:       =c}kk SmsYf

Generation: 3; Top Score: 37:       Kcjlk$Vms]f

Generation: 4; Top Score: 24:       Cdjkn Smshf

Generation: 5; Top Score: 16:       Idjln Vmshf

Generation: 6; Top Score: 14:       Idjln Voshf

Generation: 7; Top Score: 11:       Hdjln Vmslf

Generation: 8; Top Score: 9:         Hdjln Voslf

Generation: 9; Top Score: 8:         Hdjln Vosle

Generation: 10; Top Score: 7:       Hdmln Vosle

Generation: 11; Top Score: 6:       Hdlln Vosle

Generation: 12; Top Score: 5:       Hdllo Vosle

Generation: 13; Top Score: 4:       Hdllo Vosle!

Generation: 14; Top Score: 3:       Hello Vosle!

Generation: 15; Top Score: 2:       Hello Wosle!

Generation: 16; Top Score: 2:       Hello Wosle!

Generation: 17; Top Score: 1:       Hello Worle!

Generation: 18; Top Score: 1:       Hello Worle!

Generation: 19; Top Score: 1:       Hello Worle!

Generation: 20; Top Score: 0:       Hello World!

Best Genome:

{
    'message': ['H', 'e', 'l', 'l', 'o', ' ', 'W', 'o', 'r', 'l', 'd', '!']
}

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Filename, size & hash SHA256 hash help File type Python version Upload date
holland-0.0.2-py3-none-any.whl (44.2 kB) Copy SHA256 hash SHA256 Wheel py3
holland-0.0.2.tar.gz (32.0 kB) Copy SHA256 hash SHA256 Source None

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page