ggplotHeatmap: Visualize Heatmap of Multi-Objective Gradients.

In kerschke/mogsa: A Multi-Objective Optimization Algorithm Based on Multi-Objective Gradients

Description

Visualization of the multi-objective gradient landscape by means of a heatmap and on the basis of the ggplot2-package.

Usage

ggplotHeatmap(df, var1 = "x1", var2 = "x2", log.scale = TRUE,
  impute.zero = TRUE, minimalistic.image = FALSE, color.palette,
  legend.position, ...)

Arguments

df	[data.frame] Data frame as returned by computeCumulatedPathLengths.
var1	[character(1L)] Name of the variable indicating the first dimension (default: "x1").
var2	[character(1L)] Name of the variable indicating the second dimension (default: "x2").
log.scale	[logical(1L)] Should the resulting heights be displayed on a log-scale? The default is TRUE.
impute.zero	[logical(1L)] Should height values, which are exactly zero be imputed by a value half the magnitude of the smallest non-zero height? Otherwise ggplot will automatically color the corresponding tiles by a color representing NA values (usually grey). Note that this parameter is only relevant in case of log.scale = TRUE. The default is TRUE.
minimalistic.image	[logical(1L)] Should all information surrounding the image (axes, legends, background, etc.) be discarded? The default is FALSE.
color.palette	[character] Vector of colors used for visualizing the different heights of the landscape. By default, this function tries to use the color palettes from fields::tim.color or viridis. However, if neither of these packages is installed, it will use terrain.colors.
legend.position	[character(1L)] On which side of the plot, should the legend be located? If this information is not provided and minimalisitic.image = FALSE, the legend will be placed on the right side.
...	[any] Further arguments to be passed to the geom_tile function of ggplot.

Value

[ggplot]
A ggplot object displaying the multi-objective gradient landscape.

Examples

# Define two single-objective test problems and a grid of points:
fn1 = function(x) sum((x - c(0.2, 1))^2)
fn2 = function(x) sum((x - c(0.5, 0.5))^2)
points = as.matrix(expand.grid(x1 = seq(0, 0.7, 0.005), x2 = seq(0, 1.25, 0.005)))

# Compute the corresponding gradients and the cumulated path lengths:
gradients = computeGradientField(points, fn1, fn2)
x = computeCumulatedPathLengths(points, gradients)

# Finally, we can visualize the resulting multi-objective "landscape":
ggplotHeatmap(x)

# If one is only interested in the image itself, but not in any labels, legends, etc.
# one should set minimalistic.image = TRUE:
ggplotHeatmap(x, minimalistic.image = TRUE)

kerschke/mogsa documentation built on July 11, 2019, 11:52 p.m.