R/update_pool.R

In garylin2099/GA: Genetic Algorithm for Variable Selection in Regression

#' Updating Chromosome Pool
#'
#' \code{updatePool} is used to update the chromosome pool in each iteration of the genetic algorithms. It returns a vector of length P*C, which is loaded by all genes of next generation chromosomes.
#'
#' @param pool a matrix of booleans representing the chromosome pools
#' @param fitness a vector indicating the fitness score of each chromosome in the pool
#' @param tournamentSelection a logical indicating whether to use tournament selection. If FALSE, rank-based selection is used.
#' @param groupNum number of groups to be partitioned in each round of the tournament selection. Value must be between 2 and poolSize / 2 (inclusive) to ensure meaningful tournaments
#' @param oneParentRandom a logical applying to rank-based selection. If TRUE, then we select one parent with the probability being its fitness score and the other randomly, otherwise, select both parents according to their fitness score.
#' @param numCrossoverSplit number of crossover points.
#' @param mutationRate constant mutation rate. Value must be between 0 and 1.
#' @param maxMutationRate the mutation rate expected by the user when iteration limit is reached. If specified, then the mutation rate is linearly increasing in each iteration
#' @param iterCounter an integer specifying which iteration in the  the current update
#' @param maxIter maximum number of iterations of updating the generation.
#'
#' @export

updatePool <- function(pool, fitness, tournamentSelection, groupNum, oneParentRandom, numCrossoverSplit,
                       mutationRate, maxMutationRate, iterCounter, maxIter) {
  poolSize <- nrow(pool)
  chromoSize <- ncol(pool)
  # selection and crossover
  if (!tournamentSelection) {
    childVector <- rankSelect(pool, fitness, oneParentRandom, numCrossoverSplit)
  } else {
    childVector <- tournamentSelect(pool, fitness, numCrossoverSplit, groupNum)
  }
  # mutation
  if (is.null(maxMutationRate)) {
    childVector <- mutate(childVector, mutationRate)
  } else {
    childVector <- mutate(childVector, currMutationRate(mutationRate, maxMutationRate, iterCounter, maxIter))
  }

  return(matrix(childVector[1:(poolSize*chromoSize)], nrow = poolSize, byrow = TRUE))
}