| EvalGeneStoch {xegaSelectGene} | R Documentation |
Evaluates a gene in a stochastic problem environment.
Description
EvalGeneStoch evaluates a gene in
a stochastic problem environment.
Usage
EvalGeneStoch(gene, lF)
Arguments
gene |
A gene. |
lF |
The local configuration of the genetic algorithm. |
Details
In a stochastic problem environment, the expected fitness
is maximized. The computation of the expectation is
done by incrementally updating the mean.
For this, need the number of evaluations of the gene
($obs of the gene).
In addition, we compute the incremental variance
of the expected fitness
stored in $var.
The standard deviation is then gene$var/gene$obs.
If the evaluation of the fitness function of the
problem environment fails, we catch the error and
return NA for the first evaluation of the gene.
If the gene has been evaluated, we return the old gene.
Value
A gene with the elements
-
$evaluated: Boolean. -
$evalFail: Boolean. -
$fit: Mean fitness of gene. -
$gene1: Gene. -
$obs: Number of evaluations of gene. -
$var: Variance of fitness. -
$sigma: Standard deviation of fitness.
See Also
Other Evaluation Functions:
EvalGeneDet(),
EvalGeneR(),
EvalGeneU(),
EvalGene()
Examples
DeJongF4<-DeJongF4Factory()
lF<-NewlFevalGenes(DeJongF4)
g1<-list(evaluated=FALSE, evalFail=FALSE, fit=0, gene1=c(1.0, -1.5))
g1
g2<-EvalGeneStoch(g1, lF)
g2
g3<-EvalGeneStoch(g2, lF)
g3
g4<-EvalGeneStoch(g3, lF)
g4
g5<-EvalGeneStoch(g4, lF)
g5