CMA-ES Acquisition Function Optimizer — AcqOptimizerCmaes • mlr3mbo

CMA-ES acquisition function optimizer. Calls cmaes() from libcmaesr. The default algorithm is "abipop" with unlimited restarts and a budget of 100 * D^2 function evaluations, where D is the dimension of the search space. The optimization starts from the best point in the archive. For the meaning of the control parameters, see libcmaesr::cmaes_control().

Only fully numeric search spaces (all parameters of type p_dbl) are supported.

Parameters

algo: character(1)
CMA-ES variant to use, see libcmaesr::cmaes_algos. Default is "abipop".
lambda: integer(1)
Number of generated descendants per iteration. Deactivate with NA (Default).
sigma: numeric(1)
Initial sigma for the covariance. Deactivate with NA (Default).
max_restarts: integer(1)
Maximum number of restarts for the IPOP and BIPOP variants. Default is 1e5, i.e., restarts are only limited by the evaluation budget. Deactivate with NA.
tpa: integer(1)
Activates or deactivates the two-point adaptation step-size mechanism. 0 for no, 1 for auto, 2 for yes. Deactivate with NA (Default).
tpa_dsigma: numeric(1)
Value of the two-point adaptation dsigma. Deactivate with NA (Default).
seed: integer(1)
Seed of the random number generator of libcmaes. If NA (Default), the seed is drawn from R and the optimization is therefore reproducible via set.seed().
quiet: logical(1)
Should the output of libcmaes be suppressed? Default is TRUE.
skip_already_evaluated: logical(1)
Should the proposed candidate be rejected if it was already evaluated on the actual bbotk::OptimInstance? If TRUE and the candidate was already evaluated, an error is raised so that the loop_function can propose a randomly sampled point instead. Default is TRUE.

Termination Parameters

The following termination parameters can be used.

max_fevals: integer(1)
Maximum number of function evaluations. Default is 100 * D^2, where D is the dimension of the search space. Deactivate with NA.
max_iter: integer(1)
Maximum number of iterations. Deactivate with NA (Default).
ftarget: numeric(1)
Target function value. Deactivate with NA (Default).
f_tolerance: numeric(1)
Function tolerance. Deactivate with NA (Default).
x_tolerance: numeric(1)
Parameter tolerance. Deactivate with NA (Default).

Super class

AcqOptimizer -> AcqOptimizerCmaes

Public fields

state: (list())
Result of the last libcmaesr::cmaes() run.

Active bindings

print_id: (character)
Id used when printing.
label: (character(1))
Label for this object. Can be used in tables, plot and text output instead of the ID.
man: (character(1))
String in the format [pkg]::[topic] pointing to a manual page for this object.

Methods

Inherited methods

`AcqOptimizerCmaes$new()`

Creates a new instance of this R6 class.

Usage

AcqOptimizerCmaes$new(acq_function = NULL)

Arguments

acq_function: (NULL | AcqFunction).

`AcqOptimizerCmaes$optimize()`

Optimize the acquisition function.

Usage

AcqOptimizerCmaes$optimize()

Returns

data.table::data.table() with 1 row per candidate.

`AcqOptimizerCmaes$reset()`

Reset the acquisition function optimizer.

Clears the state of the previous optimization run.

Usage

AcqOptimizerCmaes$reset()

`AcqOptimizerCmaes$clone()`

The objects of this class are cloneable with this method.

Usage

AcqOptimizerCmaes$clone(deep = FALSE)

Arguments

deep: Whether to make a deep clone.

Examples

if (requireNamespace("libcmaesr")) {
  acqo("cmaes")
}
#> <AcqOptimizerCmaes>: (OptimizerCmaes)
#> * Parameters: algo=abipop, max_restarts=100000,
#>   skip_already_evaluated=TRUE, catch_errors=TRUE