R/plot_pareto.R

In caRamel: Automatic Calibration by Evolutionary Multi Objective Algorithm

#' Plotting of a population of objectives and Pareto front
#'
#' Plots graphs the population regarding each couple of objectives and emphasizes the Pareto front
#'
#' @param MatObj : matrix of the objectives [NInd, nobj]
#' @param nobj : number of objectives (optional)
#' @param objnames : vector, length nobj, of names of the objectives (optional)
#' @param maximized : vector of logical, length nobj, TRUE if objective need to be maximized, FALSE if minimized
#'
#' @examples
#' # Definition of the population
#' Pop <- matrix(runif(300), 100, 3)
#'
#' # Definition of objectives to maximize (Obj1, Obj2) and to minimize (Obj3)
#' maximized <- c(TRUE, TRUE, FALSE)
#'
#' # Call the function
#' plot_pareto(MatObj = Pop, maximized = maximized)
#'
#' @author Celine Monteil

plot_pareto <- function(MatObj, nobj = NULL, objnames = NULL, maximized = NULL) {

  if (is.null(nobj)) nobj <- length(MatObj[1, ])
  if (is.null(objnames)) objnames <- paste("Obj", as.character(c(1:nobj)), sep = "")
  if (is.null(maximized)) maximized <- rep(TRUE, nobj)

  MatTest <- MatObj
  MatTest[, !maximized] <- -MatObj[, !maximized]
  is_pareto <- pareto(MatTest)

  # Graphs
  nb_fen <- choose(n = nobj, k = 2)
  if (nb_fen <= 4) {
    par(mfrow = c(ceiling(nobj / 2), 2 - (nb_fen == 1)))
  } else {
    par(mfrow = c(3, ceiling(nb_fen / 3)))
  }
  l <- seq(1, nobj)
  for (i_fig in 1:(nobj - 1)) {
    l_tmp <- l[-i_fig]
    l_tmp <- l_tmp[l_tmp > i_fig]
    for (i_fig2 in l_tmp) {
      plot(MatObj[, i_fig], MatObj[, i_fig2], xlab = objnames[i_fig], ylab = objnames[i_fig2])
      points(MatObj[is_pareto, i_fig], MatObj[is_pareto, i_fig2], col = 2, pch = 16, cex = 1.2)
    }
  }

}