iterator
Class BasicRealEncodedGA<T extends BoundConstrainedProblem>

java.lang.Object
  iterator.AbstractIterator
      iterator.AbstractOptimizer<T>
          iterator.AbstractInitialSampleBasedOptimizer<T>
              iterator.AbstractPopulationOptimizer<T>
                  iterator.BasicRealEncodedGA<T>

Type Parameters:

T -

All Implemented Interfaces:

DataListener, InitialSamplesSettable, Iterable, LoggableIterator, RepeatableLoggableIterator

Direct Known Subclasses:

NSGA2, PCGA

public abstract class BasicRealEncodedGA<T extends BoundConstrainedProblem>
extends AbstractPopulationOptimizer<T>

Base class for real-encoded GAs. Uses an n-point crossover such as that in PCGA.

Author:

dgorur

Field Summary

protected DataPointComparer	comparator
protected double	extrapolationFactor
protected double	interpolationFactor
protected cern.colt.matrix.DoubleMatrix1D	xMutation

Fields inherited from class iterator.AbstractPopulationOptimizer

currentBestG, lastBestG, nextBlock

Fields inherited from class iterator.AbstractInitialSampleBasedOptimizer

uniform

Fields inherited from class iterator.AbstractOptimizer

doubleFormat, funEvalCount, logWriter

Fields inherited from class iterator.AbstractIterator

diagnosticString, iter

Constructor Summary

BasicRealEncodedGA()

Method Summary

double	getExtrapolationFactor() Returns the extrapolation factor.
double	getInterpolationFactor() Returns the interpolation factor.
Hypercube	getMutationBox() Returns the mutation box.
double	getMutationFraction() Returns the mutation fraction.
double	getMutationProbability() Returns the probability of mutation.
DataPoint[]	getPopulation()
void	init() Initialization routine
protected void	initializePopulation() Creates an initial population.
protected void	makeChildren(cern.colt.matrix.DoubleMatrix1D xBetter, cern.colt.matrix.DoubleMatrix1D xWorse, cern.colt.matrix.DoubleMatrix1D xChild1, cern.colt.matrix.DoubleMatrix1D xChild2) Creates children from the given parents and performs mutations, restricted to problem bounds.
void	setExtrapolationFactor(double extrapolationFactor) Sets the extrapolation factor.
void	setInterpolationFactor(double interpolationFactor) Sets the interpolation factor.
void	setMutationBox(Hypercube mutationBox) Sets the mutation box.
void	setMutationFraction(double fraction) Sets the mutation fraction.
void	setMutationProbability(double mutationProbability) Sets the mutation probability.

Methods inherited from class iterator.AbstractPopulationOptimizer

computeNextPoints, evaluatePoints, getPopulationSize, setInitialSamples, setInitialSamples, setInitialSamples, setPopulationSize, singleIteration

Methods inherited from class iterator.AbstractInitialSampleBasedOptimizer

getRandomSeed, makeInitialSamples, setRandomSeed

Methods inherited from class iterator.AbstractOptimizer

clear, dataAdded, dataAdded, dataRemoved, dataRemoved, getDataHandler, getFunEvalCount, getLogFileName, getMaxFunEval, getProblem, isTerminated, iterate, setDataHandler, setFunEvalCount, setLogFileName, setMaxFunEval, setProblem, updateDiagnostics, writeInitialRunLog

Methods inherited from class iterator.AbstractIterator

isVerbose, printDiagnostics, setVerbose

Methods inherited from class java.lang.Object

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

Methods inherited from interface iterator.LoggableIterator

getLogFileName, setLogFileName

Methods inherited from interface iterator.Iterable

isTerminated, iterate

Field Detail

comparator

protected DataPointComparer comparator

extrapolationFactor

protected double extrapolationFactor

interpolationFactor

protected double interpolationFactor

xMutation

protected cern.colt.matrix.DoubleMatrix1D xMutation

Constructor Detail

BasicRealEncodedGA

public BasicRealEncodedGA()

Method Detail

getExtrapolationFactor

public double getExtrapolationFactor()

Returns the extrapolation factor. To be used in

x_child = extrapolationFactor*x_father + (1-extrapolationFactor)*x_Mother.

Returns:

the extrapolation factor.

getInterpolationFactor

public double getInterpolationFactor()

Returns the interpolation factor. To be used in

x_child = interpolationFactor*x_father + (1-interpolationFactor)*x_Mother.

Returns:

the interpolation factor.

getMutationBox

public Hypercube getMutationBox()

Returns the mutation box. Mutations are performed in an axis-aligned hypercube centered around a population member, and this box describes the extents of that hypercube..

Returns:

the mutation box.

getMutationFraction

public double getMutationFraction()

Returns the mutation fraction. This is a nice way of specifying a mutation box, each of whose sides is a fixed fraction of the corresponding problem bounds.

Returns:

the mutation fraction.

getMutationProbability

public double getMutationProbability()

Returns the probability of mutation.

Returns:

the mutation probability.

getPopulation

public DataPoint[] getPopulation()

init

public void init()

Description copied from interface: Iterable

Initialization routine

Specified by:

init in interface Iterable

Overrides:

init in class AbstractPopulationOptimizer<T extends BoundConstrainedProblem>

setExtrapolationFactor

public void setExtrapolationFactor(double extrapolationFactor)

Sets the extrapolation factor.

Parameters:

extrapolationFactor - the given extrapolation factor.

setInterpolationFactor

public void setInterpolationFactor(double interpolationFactor)

Sets the interpolation factor.

Parameters:

interpolationFactor - the given interpolation factor.

setMutationBox

public void setMutationBox(Hypercube mutationBox)

Sets the mutation box.

Parameters:

mutationBox - the given mutation box.

setMutationFraction

public void setMutationFraction(double fraction)

Sets the mutation fraction.

Parameters:

fraction - the given mutation fraction.

setMutationProbability

public void setMutationProbability(double mutationProbability)

Sets the mutation probability.

Parameters:

mutationProbability - the given mutation probability.

initializePopulation

protected void initializePopulation()

Creates an initial population.

makeChildren

protected void makeChildren(cern.colt.matrix.DoubleMatrix1D xBetter,
                            cern.colt.matrix.DoubleMatrix1D xWorse,
                            cern.colt.matrix.DoubleMatrix1D xChild1,
                            cern.colt.matrix.DoubleMatrix1D xChild2)

Creates children from the given parents and performs mutations, restricted to problem bounds.

Parameters:

xBetter - better parent.

xWorse - worse parent.