setMBOControlMultiPoint: Set multipoint proposal options.

In pday2/mlrMBO_mP: Bayesian Optimization and Model-Based Optimization of Expensive Black-Box Functions

Description

Extends an MBO control object with options for multipoint proposal.

Usage

setMBOControlMultiPoint(
  control,
  method = NULL,
  cl.lie = NULL,
  moimbo.objective = NULL,
  moimbo.dist = NULL,
  moimbo.selection = NULL,
  moimbo.maxit = NULL,
  moimbo.sbx.eta = NULL,
  moimbo.sbx.p = NULL,
  moimbo.pm.eta = NULL,
  moimbo.pm.p = NULL
)

Arguments

control	[MBOControl] Control object for mbo.
method	[character(1)] Method used for proposal of multiple infill points, for parallel batch evaluation. Possible values are: “cb”: Proposes points by optimizing the confidence bound “cb” criterion, propose.points times. Each lambda value for “cb” is drawn randomly from an exp(1)-distribution, so do not define infill.opt.cb.lambda. The optimizer for each proposal is configured in the same way as for the single point case, i. e., by specifying infill.opt and related stuff. “moimbo”: Proposes points by multi-objective infill criteria via evolutionary multi-objective optimization. The EA is a (mu+1) type of algorithm and runs for moimbo.maxit generations. The population size is set to propose.points. The selection criterion is moimbo.selection. If this method is selected the infill criterion in setMBOInfillCrit is ignored. “cl”: Proposes points by constant liar strategy. Only meaningful if infill.crit == "cb" In the first step the kriging model is fitted based on the real data and the best point is calculated according to the regular EI-criterion. Then, the function value of the best point is simply guessed by the worst seen function evaluation. This lie is used to update the model in order to propose the subsequent point. The procedure is applied until the number of best points achieves propose.points. Default is cb.
cl.lie	[function] Function used by constant liar method for lying. Default is min.
moimbo.objective	[character(1)] Objectives which are optimized in multi-objective approach. Possible values are: “mean.dist”, “ei.dist”, “mean.se”, “mean.se.dist”. Default is “ei.dist”.
moimbo.dist	[character(1)] Distance function used in multi-objective EA. Possible values are: “nearest.neighbor”, “nearest.better”. Default is “nearest.better”.
moimbo.selection	[character(1)] Method used for selecting 1 element for removal from the population in each iteration of the multi-objective EA. Possible values are: “hypervolume”: Non-dominated sorting + hypervolume contribution. “crowdingdist”: Non-dominated sorting + crowding distance based ranking. “first”: Non-dominated sorting + first objective of moimbo.objective as criterion. “last”: Non-dominated sorting + last objective of moimbo.objective as criterion. Default is hypervolume.
moimbo.maxit	[integer(1)] Number of generations for multi-objective EA. Default is 100.
moimbo.sbx.eta	[numeric(1)] Distance parameter of crossover distribution, see sbx_operator. Default is 15.
moimbo.sbx.p	[numeric(1)] Probability of 1-point crossover, see sbx_operator. Default is 1.
moimbo.pm.eta	[numeric(1)] Distance parameter of mutation distribution, see pm_operator. Default is 15.
moimbo.pm.p	[numeric(1)] Probability of 1-point mutation, see pm_operator. Default is 1.

Value

[MBOControl].

See Also

Other MBOControl: makeMBOControl(), setMBOControlInfill(), setMBOControlMultiObj(), setMBOControlTermination()

pday2/mlrMBO_mP documentation built on July 21, 2020, 12:16 a.m.