STPGA (version 5.2.1)

GenAlgForSubsetSelectionMONoTest: Genetic algorithm for subset selection no given test with multiple criteria for Multi Objective Optimized Experimental Design.

Description

It uses a nondominated selection genetic algorithm to find the solutions on the frontier that optimizes several design criteria at the same time. The test set is taken as the complement of the training individuals.

Usage

GenAlgForSubsetSelectionMONoTest(Pcs = NULL, Dist = NULL, Kernel = NULL, Candidates,
                 ntoselect, selectionstats, selectionstatstypes,
                 plotdirections, npopGA = 100, mutprob = 0.8,
                 mutintensity = 1, nitGA = 500, lambda = 1e-06,
                 plotiters = FALSE, mc.cores = 1, InitPop = NULL, C =
                 NULL, axes.labels = NULL)

Arguments

Pcs

Principal components matrix for the individuals

Dist

Distance matrix for the individuals

Kernel

Kernel matrix for the individuals

Candidates

The set of individuals from which the training set is selected.

ntoselect

number of individuals to select in the training set.

selectionstats

a vector of design optimization criteria

selectionstatstypes

a vector describing the type of optimality criteria used in selectionstats.

plotdirections

A vector that is used to change the sign of statistics while plotting. This doesnt effect the optimization.

npopGA

genetic algorithm parameter, number of solutions at each iteration

mutprob

genetic algorithm parameter, probability of mutation for each generated solution.

mutintensity

genetic algorithm parameter, mean of the poisson variable that is used to decide the number of mutations for each cross.

nitGA

genetic algorithm parameter, number of iterations.

lambda

scalar shrinkage parameter (\(\lambda>0\)).

plotiters

plot the convergence: TRUE or FALSE. Default is TRUE.

mc.cores

number of cores to use.

InitPop

a list of initial solutions

C

Contrast Matrix.

axes.labels

Labels for the axes for plotting iterations

Value

A list of length 2. The first item in the list is the list of solutions found by the algorithm. The second item is a matrix of criteria values for the solutions in the first list.

Examples

# NOT RUN {
library(STPGA)
library(GenomicMating)

data(WheatData)


Msvd<-svd(scale(Wheat.M, scale=F, center=T), nu=50, nv=50)
Dgeno<-as.matrix(dist(scale(Wheat.M, scale=F, center=T)))^2
P<-Wheat.M%*%Msvd$v
dim(P)
rownames(Dgeno)<-colnames(Dgeno)<-rownames(P)<-rownames(Wheat.M)
test<-sample(rownames(P), 25)
candidates<-setdiff(rownames(P), test)
outnewprog<-GenAlgForSubsetSelectionMONoTest(Pcs=P,Dist=Dgeno,
Candidates=candidates,ntoselect=75, 
selectionstats=list("DOPT", "neg_dist_in_train2", "dist_to_test2"),
selectionstatstypes=c("Pcs", "Dist", "Dist"),
plotdirections=c(1,1,1),npopGA=300,
mutprob=1, mutintensity=2, nitGA=100,
plotiters=TRUE, mc.cores=1, InitPop=NULL)

#####Best solution according to ideal solution concept
outnewprog[[1]][[which.min(disttoideal(outnewprog[[2]]))]]

# }