keel.Algorithms.Fuzzy_Rule_Learning.Genetic.Shared.Algorithms

Class SimulatedAnnealing

java.lang.Object

keel.Algorithms.Fuzzy_Rule_Learning.Genetic.Shared.Algorithms.GeneticAlgorithm

keel.Algorithms.Fuzzy_Rule_Learning.Genetic.Shared.Algorithms.SimulatedAnnealing

public class SimulatedAnnealing
extends GeneticAlgorithm

SimulatedAnnealing is the simulated annealing evolutionary algorithm and programming (SAP) algorithm as detailed in "Combining GP operators with SA search to evolve fuzzy rule based classifiers", Sanchez, Couso, Corrales Information Sciences 136 (2001). This class is an specification of GeneticAlgorithm.

Since:

JDK1.4

Version:

1.0

Author:

Written by Luciano Sanchez (University of Oviedo) 20/01/2004, Modified by J.R. Villar (University of Oviedo) 19/12/2008

Constructor Summary

Constructors

Constructor and Description
SimulatedAnnealing(GeneticIndividual initialIndividual, double pPROBMUTAGA, double deltaFit, double p0, double p1, double KM, int pNSUB, Randomize r, int pGAMutationID, int pGPMutationID, int pNUMOL, int pIDOL, double pLOptProb) Class constructor with the following parameters:

Method Summary

Modifier and Type	Method and Description
GeneticIndividual	evolve(int MAXITER) this method is intended for evolving the algorithm for a given number of iterations with an SAP algorithm.

Methods inherited from class java.lang.Object

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

Constructor Detail

SimulatedAnnealing

public SimulatedAnnealing(GeneticIndividual initialIndividual,
                          double pPROBMUTAGA,
                          double deltaFit,
                          double p0,
                          double p1,
                          double KM,
                          int pNSUB,
                          Randomize r,
                          int pGAMutationID,
                          int pGPMutationID,
                          int pNUMOL,
                          int pIDOL,
                          double pLOptProb)

Class constructor with the following parameters:

Parameters:

initialIndividual - a GeneticIndividual to start the search process with the desired type of individual

deltaFit - a double with the coefficient for the fitness error following the simulate annealing evolution rules

pPROBMUTAGA - a double with the mutation probability

p0 - the desired initial temperature

p1 - the desired final temperature

KM - The mean temperature coefficient to allow mutation operations

pNSUB - an int with the number of individuals to be analized in each iteration

r - the Randomize object

pGAMutationID - the genetic algorithm mutation operation used attending the the current Genotype

pGPMutationID - the genetic programming mutation operation used attending the the current Genotype

pNUMOL - an int with the number of iterations in the local optimization method

pIDOL - an int with the local identification method identidication

pLOptProb - a double with the local optimization method probability

Method Detail

evolve

public GeneticIndividual evolve(int MAXITER)
                         throws invalidMutation,
                                invalidOptim

this method is intended for evolving the algorithm for a given number of iterations with an SAP algorithm.

Specified by:

evolve in class GeneticAlgorithm

Parameters:

MAXITER - an integer with the number of iterations torun in the evolucion

Returns:

the best GeneticIndividual found

Throws:

invalidMutation - in case of unsupported mutation.

invalidOptim - in case of local optimization operations.