keel.Algorithms.Genetic_Rule_Learning.UCS

Class GA

java.lang.Object

keel.Algorithms.Genetic_Rule_Learning.UCS.GA

public class GA
extends java.lang.Object

This is the main class of the genetic algorithm. It has an object of crossover and selection, that can take different selection and crossover models.

Since:

JDK1.2

Version:

1.1

Author:

Written by Albert Orriols (La Salle University Ramón Lull, Barcelona) 28/03/2004, Modified by Xavi Solé (La Salle University Ramón Lull, Barcelona) 03/12/2008

Constructor Summary

Constructors

Constructor and Description
GA() Constructs the GA

Method Summary

Modifier and Type	Method and Description
void	doGASubsumption(Population correctSet, Classifier parent1, Classifier parent2, Classifier child1, Classifier child2, boolean actionMutated1, boolean actionMutated2, double[] envState) It applies the GA subsumption.
void	runGA(Population correctSet, double[] envState, int classOfExample, int tStamp) It makes an iteration of the genetic algorithm.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

GA

public GA()

Constructs the GA

Method Detail

runGA

public void runGA(Population correctSet,
                  double[] envState,
                  int classOfExample,
                  int tStamp)

It makes an iteration of the genetic algorithm. It selects two parents, performs crossover, mutate the children and introduces them to the population, accordingly with the doGASubsumption parameter.

Parameters:

correctSet - is the action set where the GA has to be applied.

envState - is the environmental state.

classOfExample - is the class of the input example

tStamp - is the time stamp

doGASubsumption

public void doGASubsumption(Population correctSet,
                            Classifier parent1,
                            Classifier parent2,
                            Classifier child1,
                            Classifier child2,
                            boolean actionMutated1,
                            boolean actionMutated2,
                            double[] envState)

It applies the GA subsumption. It has to see if the parents subsume the children. If they do not, it has to search in the population for a subsumer. Otherwise, it has to insert the new classifier in the population.

Parameters:

correctSet - is the action set where the GA subsumption has to be applied.

parent1 - is the first parent.

parent2 - is the second parent.

child1 - is the first child.

child2 - is the second child.

actionMutated1 - indicates if the action 1 has been mutated

actionMutated2 - indicates if the action 2 has been mutated

envState - is the environmental state.