Abstract class describing a population of models. More...

#include <IPopulation.h>

Inheritance diagram for IPopulation:

Collaboration diagram for IPopulation:

Public Member Functions
	IPopulation ()

virtual	~IPopulation ()

void	initialize (int n)

void	crossOver ()

void	mutate ()

void	score ()

void	setRandomSeed (int seed)

void	setMutateRate (double rate)

void	setFigureOfMerit (IFigureOfMerit *fom)

IFigureOfMerit *	getFigureOfMerit ()

IModel *	getBestFitted (int rank=0)

int	size ()

double	getScoreMean ()

double	getScoreRMS ()

void	clear ()

Protected Member Functions
virtual void	doInitialize (int n)=0

virtual void	doCrossOver (const std::vector< std::vector< IModel *> > &parents)=0

virtual void	doMutate (IModel *model)=0

virtual void	selectParents (int &p1, int &p2)

void	sort ()

void	checkFigureOfMerit ()

Protected Attributes
double	m_mutateRate
	Stores the mutate rate.

bool	m_sorted
	Stores whether the ranking is valid or needs to be redone.

std::vector< IModel * >	m_individuals
	Stores the individuals of this population.

IFigureOfMerit *	m_fom
	Stores the figure of merit to be used to calculate scores and perform the ranking.

TRandom3 *	m_random
	Stores a random number generator.

double	m_scoreMean
	Stores the mean score for the population.

double	m_scoreRMS
	Stores the score RMS for the population.

std::vector< std::vector< IModel * > >	m_parents
	Stores the list of parents about to be crossed-over.

Detailed Description

Abstract class describing a population of models.

This class should implement the following functionalities:

Initialize the population: usually this use a random number generator. The actual implementation is left to the derived classes.
Selection: decide which members of the population should be cross-overed. The default behavior is to select parents with a probability that is linear with the rank.
Cross-over: implements the logic based on which a child is constructed from its parents.
Mutation: implements the logic based on which a child is altered through random mutations.

Derive from this class by implementing at least these three methods:

doInitialize(): performs the initialization of the population
doCrossOver(): performs the cross-over.
doMutate(): performs the mutation of a model. Additionally, one might override the selectParents() method to change its default behavior.

Definition at line 27 of file IPopulation.h.

Constructor & Destructor Documentation

◆ IPopulation()

IPopulation::IPopulation ( )

Default Constructor

Definition at line 10 of file IPopulation.cxx.

◆ ~IPopulation()

IPopulation::~IPopulation ( )

virtual

Destructor

Definition at line 20 of file IPopulation.cxx.

Member Function Documentation

◆ initialize()

void IPopulation::initialize ( int n )

Initialize a population of a given size.

Parameters

n	Desired size of the population.

Definition at line 28 of file IPopulation.cxx.

◆ crossOver()

void IPopulation::crossOver ( )

Performs population cross-over.

Definition at line 34 of file IPopulation.cxx.

◆ mutate()

void IPopulation::mutate ( )

Performs individual mutations.

Definition at line 56 of file IPopulation.cxx.

◆ score()

void IPopulation::score ( )

Compute the scores for the members of the population.

This function also calculates the mean and RMS for the scores of this population.

Definition at line 93 of file IPopulation.cxx.

◆ setRandomSeed()

void IPopulation::setRandomSeed ( int seed )

Sets the random seed for the random number generator.

Parameters

seed	Random number generator seed.

Definition at line 70 of file IPopulation.cxx.

◆ setMutateRate()

void IPopulation::setMutateRate ( double rate )

Sets the mutation rate.

Parameters

rate	Desired mutation rate.

Definition at line 79 of file IPopulation.cxx.

◆ setFigureOfMerit()

void IPopulation::setFigureOfMerit ( IFigureOfMerit * fom )

Sets the figure of merit to be used to calculate scores and perform the ranking.

Parameters

fom	Pointer to a figure of merit object to be used to calculate scores and perform the ranking.

Definition at line 120 of file IPopulation.cxx.

◆ getFigureOfMerit()

IFigureOfMerit * IPopulation::getFigureOfMerit ( )

Returns the figure of merit that is used to calculate scores and perform the ranking.

Returns: Pointer to the figure of merit object used to calculate scores and perform the ranking.

Definition at line 128 of file IPopulation.cxx.

◆ getBestFitted()

IModel * IPopulation::getBestFitted ( int rank = 0 )

Returns the method at a given rank.

Parameters

rank Rank.

Returns: Model at the given rank.

Definition at line 137 of file IPopulation.cxx.

◆ size()

int IPopulation::size ( )

Returns the size of the population.

Returns: Size of this population.

Definition at line 154 of file IPopulation.cxx.

◆ getScoreMean()

double IPopulation::getScoreMean ( )

Returns the mean score for the population.

Returns: Mean score for this population.

Definition at line 162 of file IPopulation.cxx.

◆ getScoreRMS()

double IPopulation::getScoreRMS ( )

Returns the RMS of the scores for the population.

Returns: RMS of the scores for this population.

Definition at line 170 of file IPopulation.cxx.

◆ clear()

void IPopulation::clear ( )

Resets the population.

Definition at line 175 of file IPopulation.cxx.

◆ doInitialize()

virtual void IPopulation::doInitialize ( int n )

protectedpure virtual

Should implement the actual initialization.

Implemented in ParametricModelPopulation.

◆ doCrossOver()

virtual void IPopulation::doCrossOver ( const std::vector< std::vector< IModel *> > & parents )

protectedpure virtual

Should implement the cross-over.

Implemented in ParametricModelPopulation.

◆ doMutate()

virtual void IPopulation::doMutate ( IModel * model )

protectedpure virtual

Should implement the mutation of a model.

Implemented in ParametricModelPopulation.

◆ selectParents()

void IPopulation::selectParents	(	int &	p1,
		int &	p2
	)

protectedvirtual

Selects parents to be crossed-over

Parameters

p1	Returns the index of the first parent.
p2	Returns the index of the second parent.

Definition at line 221 of file IPopulation.cxx.

◆ sort()

void IPopulation::sort ( )

protected

Performs the ranking from the best to the least fitted.

Definition at line 183 of file IPopulation.cxx.

◆ checkFigureOfMerit()

void IPopulation::checkFigureOfMerit ( )

protected

Makes sure an IFigureOfMerit object is assigned to this population.

Definition at line 209 of file IPopulation.cxx.

The documentation for this class was generated from the following files:

include/IPopulation.h
src/IPopulation.cxx

Public Member Functions

Protected Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

◆ IPopulation()

◆ ~IPopulation()

Member Function Documentation

◆ initialize()

◆ crossOver()

◆ mutate()

◆ score()

◆ setRandomSeed()

◆ setMutateRate()

◆ setFigureOfMerit()

◆ getFigureOfMerit()

◆ getBestFitted()

◆ size()

◆ getScoreMean()

◆ getScoreRMS()

◆ clear()

◆ doInitialize()

◆ doCrossOver()

◆ doMutate()

◆ selectParents()

◆ sort()

◆ checkFigureOfMerit()