gal::Population< T > Class Template Reference

Population is the center of an optimisation run: it contains all Chromosome members, the Problem at hand, and evolves the population through many generations. More...

#include <Population.h>

List of all members.

Public Member Functions
	Population (Problem< T > *problem)
	Constructs a new Population without initialising.
	Population (Problem< T > *problem, int size, int chromoLength)
	Constructs a new Population and fills and initialises it.
	~Population ()
	Destroys this Population, freeing all used resources.
void	addMember (Chromosome< T > *g)
	Adds a Chromosome as a member to this Population.
int const	size ()
	The number of Chromosomes in this Population.
void	initialise (int size, int chromoLength)
	Initialises this population with a number of random members.
void	setEliteSize (int nrElite)
	Sets the size of the elite subgroup.
void	setProbabilities (double probMutate, double probCrossover)
	Sets probabilities for genetic evolution.
void	evaluateFitness ()
	Evaluates the fitness of the current population.
void	markElite ()
	Selects the new elite members for the current population.
void	mutate ()
	Performs mutation on the population members.
int	selectByRoulette ()
	Selects a new Chromosome from the current population by roulette wheel selection.
void	selectAndReproduce ()
	Evolves the current population (generation) into the next generation with the following techniques: 1.
void	nextGeneration ()
	Evolutes the population into a new generation.
void const	print (int generationNr, int printbits, int printmode, std::ostream &outstr)
	Prints population details on screen.
int	testConvergence (int nrequal)
	Tests whether the Population has 'converged'.
void	writePopulation (std::ostream &outstr)
	Writes all members in current population to a file (or other output stream).
void	readPopulation (std::istream &instr)
	Reads population data from a file (or other input stream) as produced by writePopulation and updates the population with the chromosome data.

---

Detailed Description

template<class T>
class gal::Population< T >

Population is the center of an optimisation run: it contains all Chromosome members, the Problem at hand, and evolves the population through many generations.

The Population type is a template type and needs to be instantiated with the base type of the search domain. For a function f: R -> R this type is simply 'double'.

See also:

initialise()

nextGeneration()

---

Constructor & Destructor Documentation

template<class T> gal::Population< T >::Population (  Problem< T > *  problem  )  [inline]

Constructs a new Population without initialising. Parameters: problem  The Problem specifying what needs to be optimised. See also: initialise()

template<class T> gal::Population< T >::Population (  Problem< T > *  problem, int  size, int  chromoLength )  [inline]

Constructs a new Population and fills and initialises it. Parameters: problem  The Problem specifying what needs to be optimised. size  The number of chromosomes in this population. chromoLength  The chromosome length of members. See also: initialise()

template<class T> gal::Population< T >::~Population (   )  [inline]

Destroys this Population, freeing all used resources.

---

Member Function Documentation

template<class T> void gal::Population< T >::addMember (  Chromosome< T > *  g  )  [inline]

Adds a Chromosome as a member to this Population. Parameters: g  The Chromosome to be added.

template<class T> void gal::Population< T >::evaluateFitness (   )  [inline]

Evaluates the fitness of the current population. It uses the evaluateObjective function from the Problem object and then normalizes all values, so that they sum up to 1. Note that a Population aims at MAXIMISING the objective.

template<class T> void gal::Population< T >::initialise (  int  size, int  chromoLength )  [inline]

Initialises this population with a number of random members. Parameters: size  The number of chromosomes in this population. chromoLength  The chromosome length of members.

template<class T> void gal::Population< T >::markElite (   )  [inline]

Selects the new elite members for the current population. The selected member Chromosomes are marked as elite, and stored in a Population variable as well.

template<class T> void gal::Population< T >::mutate (   )  [inline]

Performs mutation on the population members. See also: setProbabilities()

template<class T> void gal::Population< T >::nextGeneration (   )  [inline]

Evolutes the population into a new generation. This is done by mutation, selection and crossover. See also: mutate() selectAndReproduce()

template<class T> void const gal::Population< T >::print (  int  generationNr, int  printbits, int  printmode, std::ostream &  outstr )  [inline]

Prints population details on screen. printmodes for bitstrings: PRINTINTERP, PRINTBITSTRING, PRINTVALUE, PRINTOBJECTIVE, PRINTFITNESS These modes may be added in any combination, e.g. PRINTBITSTRING | PRINTVALUE printmodes/selectionmodes for Population members: PRINTALL, PRINTBEST, PRINTWORST, PRINTAVG, ONLYIFBETTER These modes can not be combined, except for ONLYIFBETTER with the others. Parameters: generationNr  The number ('year') of this generation. printbits  The form in which the bitstring should be printed. printmode  The form in which Population members should be printed. outstr  The stream to which the output should be written.

template<class T> void gal::Population< T >::readPopulation (  std::istream &  instr  )  [inline]

Reads population data from a file (or other input stream) as produced by writePopulation and updates the population with the chromosome data. NOTE! Current population should be initialised with same settings for problem, size, elite count, etc.! Parameters: instr  The input stream to read from. See also: writePopulation

template<class T> void gal::Population< T >::selectAndReproduce (   )  [inline]

Evolves the current population (generation) into the next generation with the following techniques: 1. roulette wheel selection to select good chromosomes with high probability. 2. crossover on selected chromosomes to produce offspring.

template<class T> int gal::Population< T >::selectByRoulette (   )  [inline]

Selects a new Chromosome from the current population by roulette wheel selection. Don't forget to disable selected chromosome (at the call site) to prevent double selection. Returns: The *index* in _members vector.

template<class T> void gal::Population< T >::setEliteSize (  int  nrElite  )  [inline]

Sets the size of the elite subgroup. The next generation will be the first to include elite members.

template<class T> void gal::Population< T >::setProbabilities (  double  probMutate, double  probCrossover )  [inline]

Sets probabilities for genetic evolution. Parameters: probMutate  Probability for a single Chromosome bit to mutate. probCrossover  Probability for two Chromosomes to perform crossover. See also: Chromosome::mutate() Chromosome::crossover()

template<class T> int const gal::Population< T >::size (   )  [inline]

The number of Chromosomes in this Population.

template<class T> int gal::Population< T >::testConvergence (  int  nrequal  )  [inline]

Tests whether the Population has 'converged'. Conversion is defined as an unchanged objective value over several generations. Parameters: nrequal  Number of generations that should have same objective value.

template<class T> void gal::Population< T >::writePopulation (  std::ostream &  outstr  )  [inline]

Writes all members in current population to a file (or other output stream). This can serve as input later for a 'resume'. Parameters: outstr  The output stream to write to (probably an ofstream). See also: readPopulation

---

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

• Population.h

---