evox.operators.selection.tournament_selection

Module Contents

Functions

tournament_selection_multifit

Perform tournament selection based on multiple fitness values.

tournament_selection

Perform tournament selection based on single fitness values.

API

evox.operators.selection.tournament_selection.tournament_selection_multifit(n_round: int, fitnesses: List[torch.Tensor], tournament_size: int = 2) torch.Tensor[source]

Perform tournament selection based on multiple fitness values.

Parameters:
  • n_round – Number of rounds of selection (how many solutions to select).

  • fitnesses – A list of 1D tensors, each representing the fitness values of candidates for different objectives.

  • tournament_size – Number of candidates in each tournament. Defaults to 2.

Returns:

Indices of the selected solutions after tournament selection.

This function performs tournament selection by randomly selecting a group of candidates for each round, and selecting the best one from each group based on their fitness values across multiple objectives.

evox.operators.selection.tournament_selection.tournament_selection(n_round: int, fitness: torch.Tensor, tournament_size: int = 2) torch.Tensor[source]

Perform tournament selection based on single fitness values.

Parameters:
  • n_round – Number of rounds of selection (how many solutions to select).

  • fitness – A 1D tensor representing the fitness values of candidates.

  • tournament_size – Number of candidates in each tournament. Defaults to 2.

Returns:

Indices of the selected solutions after tournament selection.

This function performs tournament selection by randomly selecting a group of candidates for each round, and selecting the best one from each group based on their fitness values.