selection

Functions for selecting individuals from populations in EA algorithms.

Package Contents

Functions

multiple_unique(→ numpy.typing.NDArray[numpy.float_])

Select multiple distinct individuals from a population using the provided selection function.

pareto_frontier(→ list[int])

Return individuals based on their respective frontier values and their domination order.

topn(→ list[int])

Get indices of the top n genotypes sorted by their fitness.

tournament(→ int)

Perform tournament selection and return the index of the best individual.
multiple_unique(selection_size: int, population: list[TIndividual], fitnesses: list[TFitness], selection_function: Callable[[list[TIndividual], list[TFitness]], int]) → numpy.typing.NDArray[numpy.float_]

Select multiple distinct individuals from a population using the provided selection function.

Parameters:

  • selection_size – Amount of of individuals to select.

  • population – List of individuals to select from.

  • fitnesses – Fitnesses of the population.

  • selection_function – Function that select a single individual from a population. ([TIndividual], [TFitness]) -> index.

Returns:

Indices of the selected individuals.

pareto_frontier(frontier_values: list[list[TValues]], frontier_order: list[bool], to_take: int) → list[int]

Return individuals based on their respective frontier values and their domination order.

For mor information on the pareto frontier check: https://en.wikipedia.org/wiki/Pareto_front.

Parameters:

  • frontier_values – Lists of values that are used for the frontier. The order of the list represents the importance of a value in descending order. These values need to be numeric.

  • frontier_order – List of orders for the values used in frontier selection. True = ascending, False = descending.

  • to_take – The amount of individuals to return from the frontier.

Returns:

The index of the individuals that were selected. Returned in descending order, wrt. frontier values and domination order (Best is first).

topn(n: int, genotypes: list[Genotype], fitnesses: list[Fitness]) → list[int]

Get indices of the top n genotypes sorted by their fitness.

Parameters:

  • n – The number of genotypes to select.

  • genotypes – The genotypes. Ignored, but argument kept for function signature compatibility with other selection functions/

  • fitnesses – Fitnesses of the genotypes.

Returns:

Indices of the selected genotypes.

tournament(rng: numpy.random.Generator, fitnesses: list[Fitness], k: int) → int

Perform tournament selection and return the index of the best individual.

Parameters:

  • rng – Random number generator.

  • fitnesses – List of finesses of individuals that joint the tournamente.

  • k – Amount of individuals to participate in tournament.

Returns:

The index of te individual that won the tournament.