computeCumulatedPathLengths: Compute the Cumulated Path Lengths.
In kerschke/moPLOT: Visualizing Multi-Objective Optimization Problems

View source: R/computeCumulatedPathLengths.R

computeCumulatedPathLengths

R Documentation

Compute the Cumulated Path Lengths.

Description

Given a grid of points in the search space, along with their corresponding multi-objective gradients, this function will compute (for each point of the grid) the length of the cumulated path from a point towards its attracting local efficient point.

Usage

computeCumulatedPathLengths(
  centers,
  gradients,
  dims,
  local.efficient.ids = numeric(0),
  prec.vector.length = 0.001,
  prec.norm = 1e-06,
  cumulate.gradient.length = TRUE,
  fix.diagonals = FALSE,
  check.data = TRUE
)

Arguments

`centers`	[`matrix`] Matrix containing the points of a grid of cells, for which the lengths of cumulated gradient paths should be computed. Each row corresponds to a single point.
`gradients`	[`matrix`] Matrix containing the multi-objective gradients for each point of `centers`.
`local.efficient.ids`	[`numeric`] Vector of (row) indices that identify the locally efficient points.
`prec.vector.length`	[`numeric`] Precision value (= threshold) for the length of the multi-objective gradient. Based on this threshold, the function will classify the corresponding point as locally efficient (or not). The default value is `1e-3`.
`prec.norm`	[`numeric`] Precision threshold when normalizing a vector. That is, every element of the vector, whose absolute value is below this threshold, will be replaced by 0. The default is `1e-6`.
`cumulate.gradient.length`	[`logical`] Whether to cumulate gradient lengths or only compute locally efficient ids
`fix.diagonals`	[`logical`] Whether a correction should be used that penalizes diagonals by their additional length
`check.data`	[`logical`] Should sanity checks be performed? The default is `TRUE`. Note that the checks should only be turned off (e.g., for a slight speed up), if you are sure that you provide the input data in the correct format.