Skip to contents

Visualize objective as interactive surface

VisualizerSurfaceObj.Rd

This class is used to create interactive surface visualizations and animations of optimization traces for 2D objectives using plotly.

Super classes

vistool::Visualizer -> vistool::VisualizerSurface -> VisualizerSurfaceObj

Public fields

objective

(Objective)
The objective function.

Methods

Public methods

Inherited methods

Method new()

Creates a new instance of this R6 class.

Usage

VisualizerSurfaceObj$new(
  objective,
  x1_limits = NULL,
  x2_limits = NULL,
  padding = 0,
  n_points = 100L
)

Arguments

objective

(Objective)
The objective to optimize.

x1_limits

(numeric(2))
The x1 limits.

x1_limits

(numeric(2))
The x1 limits.

x2_limits

(numeric(2))
The x2 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.

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.

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

VisualizerSurfaceObj$add_optimization_trace(
  opt,
  line_color = NULL,
  mcolor_out = "black",
  line_width = NULL,
  npoints = NULL,
  npmax = NULL,
  name = NULL,
  offset = NULL,
  add_marker_at = 1,
  marker_shape = "circle",
  marker_color = NULL,
  line_type = 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.

line_width

(numeric(1))
Width of the optimization trace line. If NULL, uses theme$line_width.

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 uses only opt$id (optimizer 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.

line_type

(character(1)) Line type for the optimization trace (e.g., "solid", "dashed", "dotted", "dash", "dot", "dashdot"). Mapped to Plotly's line$dash. If NULL, defaults to solid.

...

Further arguments passed to add_trace(...).

Method plot()

Renders the surface plot with all added layers.

Usage

VisualizerSurfaceObj$plot(...)

Arguments

...

(any)
Additional arguments passed to the parent plot method.

Returns

The plotly object.

Method add_taylor()

Add a Taylor approximation (for 1 and 2 degrees).

Usage

VisualizerSurfaceObj$add_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 add_hessian()

Add two "arrows" as eigenvectors of the Hessian.

Usage

VisualizerSurfaceObj$add_hessian(x0, x1length = 0.1, x2length = 0.1, ...)

Arguments

x0

(numeric(2))
The point at which the Hessian is calculated.

x1length

(numeric(1))
The length of the first eigenvector.

x2length

(numeric(1))
The length of the second eigenvector.

...

(any)
Additional arguments passed to add_trace.

Method animate()

Create an animation of $plot().

Usage

VisualizerSurfaceObj$animate(
  dir = "animation",
  nframes = 10L,
  view_start = list(x = 1, y = 1, z = 1),
  view_end = list(x = 1, y = 1, z = 1),
  fext = "png",
  stops = NULL,
  ...
)

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 $add_optimization_trace() 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(...).

Method clone()

The objects of this class are cloneable with this method.

Usage

VisualizerSurfaceObj$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.