geneticengine.algorithms.gp.adaptive

Attributes

T

Classes

ParameterlessPopulationInitializer
AjustPopulationSizeStep
FeedbackParallelStep
GenericAdaptiveMutationStep
GenericAdaptiveCrossoverStep
AdaptiveGeneticProgramming	A Genetic Programming version that automatically adjusts population size, operator probabilities and weights between alternative operators.

Functions

generate_random_population_size(random)
time_for_initialization(budget)
best_of_population(population, problem)

Module Contents

geneticengine.algorithms.gp.adaptive.generate_random_population_size(random)

Parameters:

random (geneticengine.random.sources.RandomSource)

Return type:

int

geneticengine.algorithms.gp.adaptive.time_for_initialization(budget)

Parameters:

budget (geneticengine.evaluation.budget.SearchBudget)

Return type:

Optional[float]

class geneticengine.algorithms.gp.adaptive.ParameterlessPopulationInitializer(budget, tracker)

Bases: geneticengine.algorithms.gp.structure.PopulationInitializer

Parameters:

• budget (geneticengine.evaluation.budget.SearchBudget)

• tracker (geneticengine.evaluation.tracker.ProgressTracker)

budget

tracker

initialize(problem, representation, random, target_size, **kwargs)

Parameters:

• problem (geneticengine.problems.Problem)

• representation (geneticengine.representations.api.Representation)

• random (geneticengine.random.sources.RandomSource)

• target_size (int)

Return type:

Iterator[geneticengine.solutions.individual.PhenotypicIndividual]

class geneticengine.algorithms.gp.adaptive.AjustPopulationSizeStep(pgp)

Bases: geneticengine.algorithms.gp.operators.combinators.IdentityStep

last_best: geneticengine.problems.Fitness

pgp

first_iteration = True

last_improvement = -1

post_iterate(problem, evaluator, representation, random, population, target_size, generation)

Parameters:

• problem (geneticengine.problems.Problem)

• evaluator (geneticengine.evaluation.api.Evaluator)

• representation (geneticengine.representations.api.Representation)

• random (geneticengine.random.sources.RandomSource)

• population (Iterator[geneticengine.solutions.individual.PhenotypicIndividual])

• target_size (int)

• generation (int)

Return type:

None

class geneticengine.algorithms.gp.adaptive.FeedbackParallelStep(tracker, steps, weights=None)

Bases: geneticengine.algorithms.gp.operators.combinators.ParallelStep

Parameters:

• tracker (geneticengine.evaluation.tracker.ProgressTracker)

• steps (list[geneticengine.algorithms.gp.structure.GeneticStep])

• weights (list[float] | None)

tracker

iterate(problem, evaluator, representation, random, population, target_size, generation)

Parameters:

• problem (geneticengine.problems.Problem)

• evaluator (geneticengine.evaluation.api.Evaluator)

• representation (geneticengine.representations.api.Representation)

• random (geneticengine.random.sources.RandomSource)

• population (Iterator[geneticengine.solutions.individual.PhenotypicIndividual])

• target_size (int)

• generation (int)

Return type:

Iterator[geneticengine.solutions.individual.PhenotypicIndividual]

geneticengine.algorithms.gp.adaptive.T

geneticengine.algorithms.gp.adaptive.best_of_population(population, problem)

Parameters:

• population (Iterable[T])

• problem (geneticengine.problems.Problem)

Return type:

T

class geneticengine.algorithms.gp.adaptive.GenericAdaptiveMutationStep(probability=1)

Bases: geneticengine.algorithms.gp.operators.mutation.GenericMutationStep

Parameters:

probability (float)

last_fitness: geneticengine.problems.Fitness

first = True

post_iterate(problem, evaluator, representation, random, population, target_size, generation)

Parameters:

• problem (geneticengine.problems.Problem)

• evaluator (geneticengine.evaluation.api.Evaluator)

• representation (geneticengine.representations.api.Representation)

• random (geneticengine.random.sources.RandomSource)

• population (Iterator[geneticengine.solutions.individual.PhenotypicIndividual])

• target_size (int)

• generation (int)

Return type:

None

class geneticengine.algorithms.gp.adaptive.GenericAdaptiveCrossoverStep(probability=1)

Bases: geneticengine.algorithms.gp.operators.crossover.GenericCrossoverStep

Parameters:

probability (float)

last_fitness: geneticengine.problems.Fitness

first = True

post_iterate(problem, evaluator, representation, random, population, target_size, generation)

Parameters:

• problem (geneticengine.problems.Problem)

• evaluator (geneticengine.evaluation.api.Evaluator)

• representation (geneticengine.representations.api.Representation)

• random (geneticengine.random.sources.RandomSource)

• population (Iterator[geneticengine.solutions.individual.PhenotypicIndividual])

• target_size (int)

• generation (int)

Return type:

None

class geneticengine.algorithms.gp.adaptive.AdaptiveGeneticProgramming(problem, budget, representation, random=None, tracker=None)

Bases: geneticengine.algorithms.gp.gp.GeneticProgramming

A Genetic Programming version that automatically adjusts population size, operator probabilities and weights between alternative operators.

Parameters:

• problem (geneticengine.problems.Problem)

• budget (geneticengine.evaluation.budget.SearchBudget)

• representation (geneticengine.representations.api.Representation)

• random (geneticengine.random.sources.RandomSource)

• tracker (geneticengine.evaluation.tracker.ProgressTracker | None)

population_initializer

population_size

mutation_tournament

mutation

crossover_tournament

crossover

step