convertInitDesignToGrid: Converts an Initial Design into a Cell Mapping Grid

In kerschke/flacco: Feature-Based Landscape Analysis of Continuous and Constrained Optimization Problems

Converts an Initial Design into a Cell Mapping Grid

Description

This function takes an initial design – with rows being the observations and columns standing for the dimensions (plus the corresponding objective) – and adds an additional column to the data.frame. This additional column states the cell ID for each observation.

Usage

convertInitDesignToGrid(init, lower, upper, blocks)

Arguments

init	[data.frame] The initial design, consisting of d + 1 columns (d dimensions and one column for the objective value) and one row per observation.
lower	[numeric or integer] The lower limits per dimension.
upper	[numeric or integer] The upper limits per dimension.
blocks	[integer] The number of blocks per dimension.

Value

[data.frame].
A data.frame, which includes an additional column (cell.ID) compared to the initial design (init). The cell.ID will be a value between 1 and prod(blocks).

Examples

# (1) create an initial design:
X = t(replicate(n = 200, expr = runif(n = 5, min = -10, max = 10)))
f = function(x) sum(x^2)
y = apply(X = X, MARGIN = 1, FUN = f)
init = data.frame(X, y = y)

# (2) compute the cell mapping grid
convertInitDesignToGrid(init = init, lower = -10, upper = 10, blocks = 20)

kerschke/flacco documentation built on Dec. 5, 2022, 12:56 a.m.