generateValidationData: Generates validation data for the algorithm portfolio...

In danielhorn/multicrit_result_test: Multi-objective selection of algorithm portfolios

Description

Generates validation data for the algorithm portfolio selection.

Usage

generateValidationData(N, M, split.type = "uniform",
  discretize.type = "deterministic", replications.type = "parameter-noise",
  k = 20L, replications = 10L)

Arguments

N	[integer] Number of algorithms on the common pareto front.
M	[integer] Number of additional algorithms that are not on the commo pareto front.
split.type	[character] Determines whether the split points between the algorithms are chosen in a uniform or non-uniform way.
discretize.type	[character] Determines how the discrete approximation of the pareto front is done. The values deterministic, random, NSGA-II and NSGA-II_g are possible.
replications.type	[character] Determines whether noise is added to the parameters of the functions that form the fronts (parameter-noise) or to the points (point-noise) to get replications of the discrete approximation. It is also possible to add no noise (replications.type = without-noise).
k	[integer] Number of points that is generated for every algorithm.
replications	[integer] Number of replications of the discrete approximation.

Value

[list] List that contains the true (original) pareto landscape (functions that form the pareto fronts, true splitpoints, ...), the names of the algorithms and the validation data.

danielhorn/multicrit_result_test documentation built on May 14, 2019, 4:05 p.m.