Visualize Objective
Visualizer2DObjective.RdThis class is used to create visualizations and animations of optimization traces.
Super class
vistool::Visualizer2D -> Visualizer2DObjective
Public fields
objective(`Objective`)
The objective which was optimized. This object is used to generate the surface/contour lines.
Methods
Method new()
Creates a new instance of this [R6][R6::R6Class] class.
Usage
Visualizer2DObjective$new(
objective,
x1_limits = NULL,
x2_limits = NULL,
padding = 0,
n_points = 100L
)Arguments
objective(`Objective`)
The objective which was optimized. This object is used to generate the surface/contour lines.x1_limits(`numeric(2)`)
The x1 limits.x2_limits(`numeric(2)`)
The x2 limits.padding(`numeric(1)`)
A margin that is added to x1limits and x2limits. The x1 margin is calculated by `max(x1lmits) - min(x1limits) * padding`.n_points(`integer(1)`)
The number of generated point per dimension. Note that a grid of `npoints^2` values is generated and evaluated by `objective$eval(x)` to plot the surface.
Method add_optimization_trace()
Add an optimization trace.
Usage
Visualizer2DObjective$add_optimization_trace(
opt,
line_color = colSampler(),
mcolor_out = "black",
npoints = NULL,
npmax = NULL,
name = NULL,
offset = NULL,
add_marker_at = 1,
marker_shape = "circle",
marker_color = NULL,
...
)Arguments
opt(`Optimizer`)
The optimizer from which the archive is extracted and used to plot the trace.line_color(`character(1)`)
The color of the trace.mcolor_out(`character(1)`)
The outer line color of the marker.npoints(`integer(1)`)
The number of used points from the archive. Default is `NULL` which means that all points are used. If set, a sequence from 1 to `nrow(opt$archive)` is created.npmax(`integer(1)`)
The number of points used from the sequence `seq_len(nrow(opt$archive))[seq_len(npmax)]`name(`character(1)`)
The name of the trace in the legend. Default is `NULL` which means that the name is pasted from `opt$id` and `objective$id`.offset(`numeric(3)`)
Trace shift in direction (x, y, z).add_marker_at(`integer()`)
Vector of iterations at which a marker is added.marker_shape(`character()`)
Vector indicating the shape of the markers. If `length(marker_shape) == 1`, all markers get the same shape. The other option is to specify all markers individually by passing a vector of `length(add_marker_at)`. For a list of all shapes see `schema(F)$traces$XXX$attributes$marker$symbol$values` with `XXX` one of `scatter` or `scatter3d`. If `marker_shape = NA`, no marker are added.marker_color(`character()`)
The colors for the markers....Further arguments passed to `add_trace(...)`.
Method add_layer_taylor()
Add a Taylor approximation (for 1 and 2 degrees).
Usage
Visualizer2DObjective$add_layer_taylor(
x0,
degree = 2,
x1margin = 0,
x2margin = 0,
npoints_per_dim = 20L,
zlim = NULL,
...
)Arguments
x0(`numeric()) `
The point around which the approximation is done.degree(`integer(1)`)
The degree of the approximation (only 1 and 2 is implemented).x1margin(`numeric(1)`)
The "length" of the hyperplane in direction x1.x2margin(`numeric(1)`)
The "length" of the hyperplane in direction x2.npoints_per_dim(`integer(1)`)
Number of points per dimension for the plotting grid.zlim(`numeric(2)`)
The limits for z. Can be helpful if the hyperplane as a huge z range and therefore the plot looks ugly....(`any`)
Additional parameter passed to `add_surface()`.
Method animate()
Create an animation of `$plot()`.
Arguments
dir(`character(1)`)
The directory in which all the images are saved.nframes(`integer(1)`)
The number of frames.view_start(`list()`)
The start view of the animation.view_end(`list()`)
The end view of the animation.fext(`character(1)`)
The file extension (default is `png`).stops(`integer()`)
The step / iteration in the archives of the optimizers added by `$addLayerOptimizationTrace()` at which a frame is taken. Must have exact the same length as defined in `nframes`. By default, a sequence with equidistant points is generated for `stops`....(`any`)
Additional arguments passed to `$save(...)`.