R/operator.EAMutatorBitflip.R

In jakobbossek/ecr3: Evolutionary Computation in R - Version 3

#' @title
#' Bitplip mutator.
#'
#' @description
#' This operator works only on binary representation and flips each bit
#' with a given probability \eqn{p \in (0, 1)}. Usually it is recommended to
#' set \eqn{p = \frac{1}{n}} where \eqn{n} is the number of bits in the
#' representation.
#'
#' @references
#' [1] Eiben, A. E. & Smith, James E. (2015). Introduction to Evolutionary
#' Computing (2nd ed.). Springer Publishing Company, Incorporated. 52.
#'
#' @param ind [\code{binary}]\cr
#'   Binary vector, i.e., vector with elements 0 and 1 only.
#' @param p [\code{numeric(1)}]\cr
#'   Probability to flip a single bit.
#'   Default is \code{0.1}.
#' @return [\code{binary}]
#' @family mutators
#' @export
EAMutatorBitflip = function(p = 0.1) {
  EAMutator$new(
    name = "Bitflip mutation",
    params = list(p = p),
    representations = "binary",
    fun = function(ind, p = 0.1, ...) {
      #checkmate::assertNumber(p, lower = 0, upper = 1)
      .Call(`_ecr3_EAMutatorBitflipC`, ind, p)
      # n = length(ind)
      # mut.idx = runif(n) < p
      # ind[mut.idx] = 1 - ind[mut.idx]
      # return(ind)
    }
  )
} # EAMutatorBitflip