| mies_prime_operators | R Documentation | 
Prime the given MiesOperators for an optimization run with the given search space.
In its simplest form, MIES optimization only optimizes the search space of the Objective to be optimized. However,
more advanced optimization may handle a "budget" parameter for multi-fidelity optimization differently: It is still selected by Selectors,
but not mutated or recombined and instead handled separately. It is also possible to add additional components to the search space that are
not evaluated by the objective function, but that are used for self-adaption by other operators.
The mies_prime_operators() function uses the information that the user usually has readily at hand – the Objectives search space, the budget parameter, and additional components -- and primes [Mutator], [Recombinator], and [Selector'] objects in the right way:
Selectors are primed on a union of search_space and additional_components
Mutators and Recombinators are primed on the Selector's space with the budget_id Domain removed.
mies_prime_operators() is called with an arbitrary number of MiesOperator arguments; typically one Mutator, one Recombinator and
at least two Selector: one for survival selection, and one parent selection. Supplied MiesOperators are primed by-reference, but
they are also returned as invisible list.
If neither additional components nor multi-fidelity optimization is used, it is also possible to use the $prime() function of hte MiesOperators
directly, although using mies_prime_operators() gives flexibility for future extension.
mies_prime_operators(
  search_space,
  mutators = list(),
  recombinators = list(),
  selectors = list(),
  filtors = list(),
  ...,
  additional_components = NULL,
  budget_id = NULL
)
| search_space | ( | 
| mutators | ( | 
| recombinators | ( | 
| selectors | ( | 
| filtors | ( | 
| ... | (any) | 
| additional_components | ( | 
| budget_id | ( | 
invisible named list with entries $mutators (list of Mutator, primed mutators), $recombinators (list of Recombinator, primed recombinators),
and $selectors (list of Selector, primed selectors).
# Search space of a potential TuningInstance for optimization:
search_space = ps(x = p_dbl(), y = p_dbl())
# Additoinal search space components that are not part of the TuningInstance
additional_components = ps(z = p_dbl())
# Budget parameter not subject to mutation or recombination
budget_id = "y"
m = mut("gauss")
r = rec("xounif")
s1 = sel("best")
s2 = sel("random")
f = ftr("null")
mies_prime_operators(search_space, mutators = list(m),
  recombinators = list(r), selectors = list(s1, s2), filtors = list(f),
  additional_components = additional_components, budget_id = budget_id
)
# contain search_space without budget parameter, with additional_components
m$primed_ps
r$primed_ps
# contain also the budget parameter
s1$primed_ps
s2$primed_ps
f$primed_ps
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.