PyGA

PyGA is an extensible toolkit for Genetic Algorithms (GA) in Python.

The library aims to provide a high-level declarative interface which ensures that GAs can be implemented and customised with ease. PyGA features an extensible framework which allows researchers to provide custom implementations which interface with existing functionality.

License: MIT
Python Versions: 3.6+

Features:

High-level module for Genetic Algorithms.
Extensible API for implementing new functionality.

Installation:

To install PyGA, run this command in your terminal:

$ pip install pygaopt

Basic Usage:

PyGA aims to provide a high-level interface for Genetic Algorithms - the code below demonstrates just how easy running an optimisation procedure can be.

import pyga
from pyga.utils.functions import single_objective as fx


bounds = {
    'x0': [-1e6, 1e6],
    'x1': [-1e6, 1e6],
    'x2': [-1e6, 1e6]
}

optimiser = pyga.SOGA(bounds, n_individuals=30, n_iterations=100)
optimiser.optimise(fx.sphere)

History:

The optimisation history is written to a History data structure to allow the user to further investigate the optimisation procedure upon completion. This is a powerful tool, letting the user define custom history classes which can record whichever data the user desires.

Tracking the history of the optimisation process allows for plotting of the results, an example demonstration is seen in the plot_fitness_history function - this can be further customised through the designation of a PlotDesigner object which provides formatting instructions for the graphing tools.

Constraints:

PyGA allows the user to define a set of constraints for the optimisation problem - this is achieved through inheriting a template class and implementing the designated method. An example of which is demonstrated below:

from pyga.constraints.base_constraints import PositionConstraint


class UserConstraint(PositionConstraint):

    def constrain(self, position):
        return position['x0'] > 0 and position['x1'] < 0


optimiser.constraint_manager.register_constraint(UserConstraint())

This provides the user with a large amount of freedom to define the appropriate constraints and allows the ConstraintManager to deal with the relevant constraints at the appropriate time.

Customisation:

Though the base SOGA will work for many, there maybe aspects that one may want to change, such as the selection / recombination methods. A common interface has been designed for these, this ensures that the user can alter the functionality at will and researchers can implement additional functionality with ease.

Attributes of the SOGA instance can be modified to implement alternative methods, this is demonstrated below:

# using 'uniform crossover' as the crossover method
from pyga.utils.crossovers import UniformCrossover
optimiser.crossover = UniformCrossover(p_swap=0.25)

# using 'fitness-proportionate selection' as the selection method
from pyga.utils.selections import FitnessProportionateSelection
optimiser.selection = FitnessProportionateSelection()

It is also possible to define alternative termination criteria through implementation of a TerminationManager class, a couple of examples are demonstrated below:

# using elapsed time as the termination criteria
from pyga.utils.termination_manager import TimeTerminationManager
optimiser.termination_manager = TimeTerminationManager(t_budget=10_000)

# using error as the termination criteria
from pyga.utils.termination_manager import ErrorTerminationManager
optimiser.termination_manager = ErrorTerminationManager(
    optimiser, target=0.0, threshold=1e-3
)

danielkelshaw/PyGA