nojhan/paradiseo
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

feat: separate estimators for init and in loop in edoAlgoAdaptive

Browse source 
Allows for easier reseting or editing of distribution during
restarts.
This commit is contained in:
Johann Dreo 2020-03-24 16:31:49 +01:00
commit e0e1cb8bc9
1 changed files with 48 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

54
edo/src/edoAlgoAdaptive.h
View file

@ -34,7 +34,7 @@ Authors:
#include "edoAlgo.h"
#include "edoEstimator.h"
#include "edoModifierMass.h"
// #include "edoModifierMass.h"
#include "edoSampler.h"
#include "edoContinue.h"
@ -67,7 +67,44 @@ public:
public:
/*!
Takes algo operators, all are mandatory
Takes all operators.
\param distrib A distribution to use, if you want to update this parameter (e.gMA-ES) instead of replacing it (e.g. an EDA)
\param evaluator Evaluate a population
\param selector Selection of the best candidate solutions in the population
\param estimator Estimation of the distribution parameters
\param estimator_init Estimation of the distribution parameters, called before the main loop.
\param sampler Generate feasible solutions using the distribution
\param replacor Replace old solutions by new ones
\param pop_continuator Stopping criterion based on the population features
\param distribution_continuator Stopping criterion based on the distribution features
*/
edoAlgoAdaptive(
D & distrib,
eoPopEvalFunc < EOType > & evaluator,
eoSelect< EOType > & selector,
edoEstimator< D > & estimator,
edoEstimator< D > & estimator_init,
edoSampler< D > & sampler,
eoReplacement< EOType > & replacor,
eoContinue< EOType > & pop_continuator,
edoContinue< D > & distribution_continuator
) :
_distrib(distrib),
_evaluator(evaluator),
_selector(selector),
_estimator(estimator),
_estimator_init(estimator_init),
_sampler(sampler),
_replacor(replacor),
_pop_continuator(pop_continuator),
_dummy_continue(),
_distribution_continuator(distribution_continuator)
{}
/*!
Without the initialization estimator (set the same estimator everywhere).
\param distrib A distribution to use, if you want to update this parameter (e.gMA-ES) instead of replacing it (e.g. an EDA)
\param evaluator Evaluate a population
@ -92,6 +129,7 @@ public:
_evaluator(evaluator),
_selector(selector),
_estimator(estimator),
_estimator_init(estimator),
_sampler(sampler),
_replacor(replacor),
_pop_continuator(pop_continuator),
@ -125,6 +163,7 @@ public:
_evaluator(evaluator),
_selector(selector),
_estimator(estimator),
_estimator_init(estimator),
_sampler(sampler),
_replacor(replacor),
_pop_continuator(pop_continuator),
@ -141,18 +180,18 @@ public:
{
assert(pop.size() > 0);
eoPop< EOType > current_pop;
eoPop< EOType > selected_pop;
// update the extern distribution passed to the estimator (cf. CMA-ES)
// OR replace the dummy distribution for estimators that do not need extern distributions (cf. EDA)
_distrib = _estimator(pop);
_distrib = _estimator_init(pop);
// Evaluating a first time the candidate solutions
// The first pop is not supposed to be evaluated (@see eoPopLoopEval).
// _evaluator( current_pop, pop );
do {
eoPop< EOType > current_pop;
eoPop< EOType > selected_pop;
// (1) Selection of the best points in the population
_selector(pop, selected_pop);
assert( selected_pop.size() > 0 );
@ -205,6 +244,9 @@ protected:
//! A EOType estimator. It is going to estimate distribution parameters.
edoEstimator<D> & _estimator;
//! A EOType estimator. Called before the main loop.
edoEstimator<D> & _estimator_init;
//! A D sampler
edoSampler<D> & _sampler;