plotStoSOO: Plot 2-D 'StoSOO' result

In OOR: Optimistic Optimization in R

Plot 2-D StoSOO result

Description

Plot bivariate tree structure obtained when running StoSOO

Usage

plotStoSOO(
  sol,
  lower = rep(0, length(sol$par)),
  upper = rep(1, length(sol$par)),
  levels = NULL,
  add = FALSE,
  cpch = ".",
  lcols = 1,
  ylim = NULL
)

Arguments

sol	outcome of running StoSOO, with control$light set to FALSE
lower, upper	vectors of bounds on the variables.
levels	which levels to print. Default to all levels
add	if TRUE, use existing plot
cpch	points pch code for the centers
lcols	color at each level, or a single color for all levels (default)
ylim	vector of bounds, required in the 1d case

Examples

#------------------------------------------------------------
# Example 1 : Deterministic optimization with SOO, 1-dimensional function
#------------------------------------------------------------
## Define objective
fun1 <- function(x) return(-guirland(x))

## Optimization
Sol1 <- StoSOO(par = NA, fn = fun1, nb_iter = 1000, 
  control = list(type = "det", verbose = 1, light = FALSE))

## Display objective function and solution found
curve(fun1, n = 1001, ylim = c(-1.3, 0))
abline(v = Sol1$par, col = 'red')
plotStoSOO(Sol1, ylim = c(-1.3, -1.1), add = TRUE)

#------------------------------------------------------------
# Example 2 : Deterministic optimization with SOO, 2-dimensional function
#------------------------------------------------------------
set.seed(42)

## 2-dimensional noiseless objective function, defined on [0, pi/4]^2
fun <- function(x){return(-sin1(x[1]) * sin1(1 - x[2]))}

## Optimizing
Sol <- StoSOO(par = rep(NA, 2), fn = fun, upper = rep(pi/4, 2), nb_iter = 1000,
  control = list(type = 'det', light = FALSE))
  
## Display solution
plotStoSOO(Sol, upper = rep(pi/4, 2))

OOR documentation built on Aug. 23, 2023, 1:08 a.m.