mies_filter_offspring: Filter Offspring

View source: R/mies_methods.R

mies_filter_offspringR Documentation

Filter Offspring

Description

Uses a Filtor to extract a subset of individuals from a given set. The individuals are either returned directly (when get_indivs is TRUE) or in form of an index into the given individuals (when get_indivs is FALSE).

Filtors must always select individuals without replacement, so selected individual indices are unique.

Usage

mies_filter_offspring(
  inst,
  individuals,
  lambda,
  filtor = NULL,
  budget_id = NULL,
  fidelity = NULL,
  get_indivs = TRUE
)

Arguments

inst

(OptimInstance)
Optimization instance to evaluate.

individuals

(data.frame | data.table)
Individuals to filter. Must have columns according to filtor$primed_ps, and must have at least filtor$needed_input(lambda) rows.

lambda

(integer(1))
Number of individuals to filter down to.

filtor

(Filtor | NULL)
Filtor operator that filters. When NULL is given, then the FiltorNull operation is performed and the first lambda individuals are taken from individuals.

budget_id

(character(1) | NULL)
Budget component when doing multi-fidelity optimization. This component of the search space is added to individuals according to fidelity. Should be NULL when no multi-fidelity optimization is performed (default).

fidelity

(atomic | NULL)
scalar indicating the value of the budget_id component with which to evaluate individuals to be filtered.
This value must be NULL when no multi-fidelity optimization is performed, but it may also be NULL when the maximum value of the budget_id found in inst$archive should be used (the default).

get_indivs

(logical(1))
Whether to return the data.frame or data.table of selected individuals, or an index into individuals.

Value

If get_indivs is TRUE: a data.frame or data.table (depending on the input type of individuals) of filtered configurations. If get_indivs is FALSE: an integer vector indexing the filtered individuals.


miesmuschel documentation built on Sept. 11, 2024, 8:23 p.m.