Selection Methods
In the following table is shown the selection methods that you can be chosen to configurie the evolutionary algorithms skeletons (except CHC).
Selection Methods
|
Id.
|
Default Parameter
|
Name
|
Parameter Description
|
0
|
-
|
Random
|
-
|
1
|
1
|
Tournament
|
Size of sample
|
2
|
-
|
Roulette_Wheel
|
-
|
3
|
100
|
Ranking
|
Percent of population that the method uses.
|
4
|
0
|
Best (Rank-based)
|
The position of an individual in the ordered population
|
5
|
0
|
Worst (Rank-based)
|
The position of an individual in the reverse-ordered
population
|
These methods can be used in selection section of the cofiguration file and in the configuration of the migration operator.
In the CHC skeleton, the selection methods are not optional. In the selection section of the CHC configuration file, you must choose the parameter (percent of population that will restarted) of replace method and the diverge operator that will used to restart the population whenever convergence is detected.
Intra-population Operators
Intra-population operators:
Id. Name Parameters* Others 0 Crossover Application probability HUX in CHC 1 Mutation Application probability bit-flip in CHC 2 Hybrid operator Application probability only in newGASA 3- Other operators Application probability -
Inter-population Operators
Intra-population operators are the operators that are applied between sub-populations. Currently, the skeletons only have got a single inter-population operator, the migration operator. In order to configure this operator, you must define:
- operator number (0)
- operator rate
- number of individuals
- selection method (and its parameters) of individual to send
- and replace method (and its parameters) of individual to send .