R/HAR.TLineSeg.R

In TML: Tropical Geometry Tools for Machine Learning

#'Uniformly sample from a max-plus tropical line segment
#'
#'This function uses a hit-and-run sampler to uniformly sample from a max-plus
#'tropical line segment
#'
#'@param D1 point in the tropical projective torus
#'@param D2 point in the tropical projective torus
#'@param tadd string; max indicates max-plus addition, min indicates min-plus
#'  addition. Defaults to max
#'@return point on the line segment defined by D1 and D2
#'@author Ruriko Yoshida \email{ryoshida@@nps.edu}
#'@references Yoshida, Ruriko, Keiji Miura and David Barnhill (2022). Hit and
#'  Run Sampling from Tropically Convex Sets.
#'@export
#' @examples
#'D1 <-c(0,4,2)
#'D2 <- c(0,7,-1)
#'HAR.TLineSeg(D1, D2,tadd=max)
#'HAR.TLineSeg(D1, D2,tadd=min)

HAR.TLineSeg <- function(D1, D2,tadd=max){
  if(length(D1) != length(D2))
    warning("dimension is wrong!")

  d <- length(D1)
  is_min <- identical(tadd, base::min)

  index <- order(D2 - D1, decreasing=is_min)
  lambda <- (D2 - D1)[index]

  if (is_min) l <- runif(1, min=lambda[d], max=lambda[1])
  else l <- runif(1, min=lambda[1], max=lambda[d])

  x1 <- rep(0, d)

  for(i in 1:d){
    x1[i] <- tadd(l + D1[i], D2[i])
  }

  return(x1)
}