plotLoss: Visualize loss of optimization procedure

In PhilippScheller/visualDescend: Visulaization of descent method for multi-dimensional optimization functions

Description

This functions visualizes the difference between the true optimum and the optimum in the current iteration (loss) during the optimization procedure. The default upper and lower bounds for the plot are set by the box constraints of the respective optimization function.

Usage

plotLoss(
  f,
  x1.lower,
  x1.upper,
  x2.lower,
  x2.upper,
  xmat,
  trueOptZ = NULL,
  algoName = NULL,
  optimError = FALSE
)

Arguments

f	a (multi-) dimensional function to be eptimized.
x1.lower	the lower boundary for the range of the x1 coordinate displayed.
x1.upper	the upper boundary for the range of the x1 coordinate displayed.
x2.lower	the lower boundary for the range of the x2 coordinate displayed.
x2.upper	the upper boundary for the range of the x2 coordinate displayed.
xmat	an object of class list containing the results of the algorithm.
trueOptZ	is a numeric of the z coordinate of the true optimum.
algoName	list of strings with names of algorithms. Length must equal length of 'xmat'.
optimError	passed argument from optimization. Specifies if error in gradient occured.
n.x	the number (equidistant) points at which f is evaluated w.r.t x1 and x2 coordinate.

PhilippScheller/visualDescend documentation built on Feb. 5, 2020, 4:04 a.m.