plot.optim_nmsa: Plot an optim_nmsa Object

In optimization: Flexible Optimization of Complex Loss Functions with State and Parameter Space Constraints

Plot an optim_nmsa Object

Description

Creates convergence or contour plots for visual inspection of the optimization result. Note that 'trace' must be activated for this function.
In case of a bivariate optimization, the 'contour' plot gives an overview of the parameter development over time in the entire state space. This is useful for the evaluation of the algorithm settings and therefore helps improving the performance. The development of the response can be visualized via the 'convergence' plot.

Usage

## S3 method for class 'optim_nmsa'
plot(x, type = 'convergence', lower = NA, upper = NA, ...)

Arguments

x	Object of type 'optim_nmsa' to be plotted. The 'trace' entry must not be empty.
type	Character string which determines the plot type. Either 'convergence' or 'contour' is possible.
lower	Vector containing the lower limits of the variables in the plot. Only useful for 'contour' plots.
upper	Vector containing the upper limits of the variables in the plot. Only useful for 'contour' plots.
...	Further arguments for the generic plot function.

Author(s)

Kai Husmann, Alexander Lange

See Also

optim_nm, optim_sa

Examples

# S3 method for class 'optim_nlme'

# Himmelblau's function
hi <- function(x){(x[1]**2 + x[2] - 11)**2 + (x[1] + x[2]**2 -7)**2}

out_nm <- optim_nm(hi, k = 2, trace = TRUE)
out_sa <- optim_sa(fun = hi, start = c(runif(2, min = -1, max = 1)),
                   trace = TRUE, lower = c(-4, -4) ,upper=c(4, 4),
                   control = list(t0 = 1000, nlimit = 1500,r = 0.8))

# Examples for optimization results via 'Nelder-Mead' method.
plot(out_nm)
plot(out_nm, type = "contour", lower = c(-4, -4), upper = c(4, 4))

# Examples for optimization results via 'Simulated Annealing' method.
plot(out_sa)
plot(out_sa, type = "contour")

optimization documentation built on March 18, 2022, 7:41 p.m.