edoAlgoAdaptive< D > Class Template Reference

A generic stochastic search template for algorithms that need a distribution parameter. More...

#include <edoAlgoAdaptive.h>

Inheritance diagram for edoAlgoAdaptive< D >:

List of all members.

Public Types
typedef D::EOType	EOType
	Alias for the type EOT.
typedef EOType::AtomType	AtomType
	Alias for the atom type.
typedef EOType::Fitness	Fitness
	Alias for the fitness.
Public Member Functions
	edoAlgoAdaptive (D &distrib, eoPopEvalFunc< EOType > &evaluator, eoSelect< EOType > &selector, edoEstimator< D > &estimator, edoSampler< D > &sampler, eoReplacement< EOType > &replacor, eoContinue< EOType > &pop_continuator, edoContinue< D > &distribution_continuator)
	edoAlgoAdaptive (D &distrib, eoPopEvalFunc< EOType > &evaluator, eoSelect< EOType > &selector, edoEstimator< D > &estimator, edoSampler< D > &sampler, eoReplacement< EOType > &replacor, eoContinue< EOType > &pop_continuator)
	constructor without an edoContinue
void	operator() (eoPop< EOType > &pop)
	Call the algorithm.
Protected Attributes
D &	_distrib
	The distribution that you want to update.
eoPopEvalFunc< EOType > &	_evaluator
	A full evaluation function.
eoSelect< EOType > &	_selector
	A EOType selector.
edoEstimator< D > &	_estimator
	A EOType estimator. It is going to estimate distribution parameters.
edoSampler< D > &	_sampler
	A D sampler.
eoReplacement< EOType > &	_replacor
	A EOType replacor.
eoContinue< EOType > &	_pop_continuator
	A EOType population continuator.
edoDummyContinue< D >	_dummy_continue
	A D continuator that always return true.
edoContinue< D > &	_distribution_continuator
	A D continuator.

---

Detailed Description

template<typename D>
class edoAlgoAdaptive< D >

A generic stochastic search template for algorithms that need a distribution parameter.

An adaptive algorithm will directly updates a distribution, it must thus be instanciated with an edoDistrib at hand. Thus, this distribution object should be instanciated appart. The reference to this distribution is generally also needed by at least one of the algorithm's operator, generally for algorithms that shares the same algorithms across operators and/or iterations.

If you no operator needs to update the distribution, then it is simpler to use an edoAlgoStateless .

Definition at line 55 of file edoAlgoAdaptive.h.

---

Constructor & Destructor Documentation

template<typename D>

edoAlgoAdaptive< D >::edoAlgoAdaptive	(	D &	distrib,
		eoPopEvalFunc< EOType > &	evaluator,
		eoSelect< EOType > &	selector,
		edoEstimator< D > &	estimator,
		edoSampler< D > &	sampler,
		eoReplacement< EOType > &	replacor,
		eoContinue< EOType > &	pop_continuator,
		edoContinue< D > &	distribution_continuator
	)		[inline]

Takes algo operators, all are mandatory

Parameters:

distrib	A distribution to use, if you want to update this parameter (e.gMA-ES) instead of replacing it (e.g. an EDA)
evaluator	Evaluate a population
selector	Selection of the best candidate solutions in the population
estimator	Estimation of the distribution parameters
sampler	Generate feasible solutions using the distribution
replacor	Replace old solutions by new ones
pop_continuator	Stopping criterion based on the population features
distribution_continuator	Stopping criterion based on the distribution features

Definition at line 81 of file edoAlgoAdaptive.h.

template<typename D>

edoAlgoAdaptive< D >::edoAlgoAdaptive	(	D &	distrib,
		eoPopEvalFunc< EOType > &	evaluator,
		eoSelect< EOType > &	selector,
		edoEstimator< D > &	estimator,
		edoSampler< D > &	sampler,
		eoReplacement< EOType > &	replacor,
		eoContinue< EOType > &	pop_continuator
	)		[inline]

constructor without an edoContinue

Takes algo operators, all are mandatory

Parameters:

distrib	A distribution to use, if you want to update this parameter (e.gMA-ES) instead of replacing it (e.g. an EDA)
evaluator	Evaluate a population
selector	Selection of the best candidate solutions in the population
estimator	Estimation of the distribution parameters
sampler	Generate feasible solutions using the distribution
replacor	Replace old solutions by new ones
pop_continuator	Stopping criterion based on the population features

Definition at line 115 of file edoAlgoAdaptive.h.

---

Member Function Documentation

template<typename D>

void edoAlgoAdaptive< D >::operator()

(

eoPop< EOType > &

pop

)

[inline]

Call the algorithm.

Parameters:

pop	the population of candidate solutions

Returns:

void

Definition at line 140 of file edoAlgoAdaptive.h.

References edoAlgoAdaptive< D >::_distrib, edoAlgoAdaptive< D >::_distribution_continuator, edoAlgoAdaptive< D >::_estimator, edoAlgoAdaptive< D >::_evaluator, edoAlgoAdaptive< D >::_pop_continuator, edoAlgoAdaptive< D >::_replacor, edoAlgoAdaptive< D >::_sampler, and edoAlgoAdaptive< D >::_selector.

---

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

• edoAlgoAdaptive.h