SelectPropFitDiffOnln {xegaSelectGene} | R Documentation |
Selection proportional to fitness differences O(n ln(n)).
Description
SelectPropFitDiffOnln()
implements selection
proportional to fitness differences. Negative fitness
vectors are shifted to R^+
.
The default of the function lF$Offset()
is 1
.
Holland's schema theorem uses this selection function.
See John Holland (1975) for further information.
Usage
SelectPropFitDiffOnln(fit, lF, size = 1)
Arguments
fit |
Fitness vector. |
lF |
Local configuration. |
size |
Number of selected genes. Default: 1. |
Details
This is a fast implementation which gives exactly the same
results as the functions SelectPropFitDiff()
and SelectPropDiffFitM()
.
Its runtime is O(n . ln(n))
.
An epsilon (lF$Eps()
) is added to the fitness
difference vector. This guarantees numerical stability,
even if all genes in the population have the same fitness.
Value
The index vector of the selected genes.
Credits
The code of this function has been adapted by Fabian Aisenbrey.
Warning
There is a potential slow for-loop in the code.
References
Holland, John (1975): Adaptation in Natural and Artificial Systems, The University of Michigan Press, Ann Arbor. (ISBN:0-472-08460-7)
See Also
Other Selection Functions:
SelectDuel()
,
SelectLRSelective()
,
SelectLinearRankTSR()
,
SelectPropFit()
,
SelectPropFitDiff()
,
SelectPropFitDiffM()
,
SelectPropFitM()
,
SelectPropFitOnln()
,
SelectSTournament()
,
SelectSUS()
,
SelectTournament()
,
SelectUniform()
,
SelectUniformP()
Examples
fit<-sample(10, 15, replace=TRUE)
SelectPropFitDiffOnln(fit, NewlFselectGenes())
SelectPropFitOnln(fit, NewlFselectGenes(), length(fit))