#include <agrum/FMDP/learning/fmdpLearner.h>

Inheritance diagram for gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >:

Collaboration diagram for gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >:

Public Member Functions
Constructor & destructor.
	FMDPLearner (double learningThreshold, bool actionReward, double similarityThreshold=0.05)
	Default constructor. More...

	~FMDPLearner ()
	Default destructor. More...

Initialization
void	initialize (FMDP< double > *fmdp)
	Initializes the learner. More...

MultiDimFunctionGraph< double > *	_instantiateFunctionGraph_ ()
	Initializes the learner. More...

MultiDimFunctionGraph< double > *	_instantiateFunctionGraph_ (Int2Type< IMDDILEARNER >)
	Initializes the learner. More...

MultiDimFunctionGraph< double > *	_instantiateFunctionGraph_ (Int2Type< ITILEARNER >)
	Initializes the learner. More...

VariableLearnerType *	_instantiateVarLearner_ (MultiDimFunctionGraph< double > target, Set< const DiscreteVariable > &mainVariables, const DiscreteVariable *learnedVar)
	Initializes the learner. More...

VariableLearnerType *	_instantiateVarLearner_ (MultiDimFunctionGraph< double > target, Set< const DiscreteVariable > &mainVariables, const DiscreteVariable *learnedVar, Int2Type< IMDDILEARNER >)
	Initializes the learner. More...

VariableLearnerType *	_instantiateVarLearner_ (MultiDimFunctionGraph< double > target, Set< const DiscreteVariable > &mainVariables, const DiscreteVariable *learnedVar, Int2Type< ITILEARNER >)
	Initializes the learner. More...

RewardLearnerType *	_instantiateRewardLearner_ (MultiDimFunctionGraph< double > target, Set< const DiscreteVariable > &mainVariables)
	Initializes the learner. More...

RewardLearnerType *	_instantiateRewardLearner_ (MultiDimFunctionGraph< double > target, Set< const DiscreteVariable > &mainVariables, Int2Type< IMDDILEARNER >)
	Initializes the learner. More...

RewardLearnerType *	_instantiateRewardLearner_ (MultiDimFunctionGraph< double > target, Set< const DiscreteVariable > &mainVariables, Int2Type< ITILEARNER >)
	Initializes the learner. More...

Incremental methods
bool	addObservation (Idx actionId, const Observation *obs)
	Gives to the learner a new transition. More...

void	updateFMDP ()
	Starts an update of datastructure in the associated FMDP. More...

Miscelleanous methods
double	_rmax_
	learnerSize More...

double	_modaMax_
	learnerSize More...

Size	size ()
	learnerSize More...

const IVisitableGraphLearner *	varLearner (Idx actionId, const DiscreteVariable *var) const
	extractCount More...

virtual double	rMax () const
	learnerSize More...

virtual double	modaMax () const
	learnerSize More...

Detailed Description

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>
class gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >

Definition at line 59 of file fmdpLearner.h.

Member Typedef Documentation

◆ RewardLearnerType

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

typedef LearnerSelect< LearnerSelection, IMDDI< RewardAttributeSelection, true >, ITI< RewardAttributeSelection, true > >::type gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::RewardLearnerType

private

Definition at line 67 of file fmdpLearner.h.

◆ VariableLearnerType

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

typedef LearnerSelect< LearnerSelection, IMDDI< VariableAttributeSelection, false >, ITI< VariableAttributeSelection, false > >::type gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::VariableLearnerType

private

Definition at line 63 of file fmdpLearner.h.

◆ VarLearnerTable

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

typedef HashTable< const DiscreteVariable*, VariableLearnerType* > gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::VarLearnerTable

private

Definition at line 69 of file fmdpLearner.h.

Constructor & Destructor Documentation

◆ FMDPLearner()

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::FMDPLearner	(	double	learningThreshold,
		bool	actionReward,
		double	similarityThreshold = `0.05`
	)

Default constructor.

Definition at line 46 of file fmdpLearner_tpl.h.

                                                           :
       _actionReward_(actionReward),
       _learningThreshold_(lT), _similarityThreshold_(sT) {
     GUM_CONSTRUCTOR(FMDPLearner);
     _rewardLearner_ = nullptr;
   }

◆ ~FMDPLearner()

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::~FMDPLearner ( )

Default destructor.

Definition at line 61 of file fmdpLearner_tpl.h.

References gum::Set< Key, Alloc >::emplace().

                     {
     for (auto actionIter = _actionLearners_.beginSafe(); actionIter != _actionLearners_.endSafe();
          ++actionIter) {
       for (auto learnerIter = actionIter.val()->beginSafe();
            learnerIter != actionIter.val()->endSafe();
            ++learnerIter)
         delete learnerIter.val();
       delete actionIter.val();
       if (_actionRewardLearners_.exists(actionIter.key()))
         delete _actionRewardLearners_[actionIter.key()];
     }
 
     if (_rewardLearner_) delete _rewardLearner_;
 
     GUM_DESTRUCTOR(FMDPLearner);
   }

Here is the call graph for this function:

Member Function Documentation

◆ _instantiateFunctionGraph_() [1/3]

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

MultiDimFunctionGraph< double >* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_instantiateFunctionGraph_ ( )

inline

Initializes the learner.

Definition at line 102 of file fmdpLearner.h.

                                                                   {
       return _instantiateFunctionGraph_(Int2Type< LearnerSelection >());
     }

◆ _instantiateFunctionGraph_() [2/3]

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

MultiDimFunctionGraph< double >* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_instantiateFunctionGraph_ ( Int2Type< IMDDILEARNER > )

inline

Initializes the learner.

Definition at line 106 of file fmdpLearner.h.

                                                                                           {
       return MultiDimFunctionGraph< double,
                                     ExactTerminalNodePolicy >::getReducedAndOrderedInstance();
     }

◆ _instantiateFunctionGraph_() [3/3]

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

MultiDimFunctionGraph< double >* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_instantiateFunctionGraph_ ( Int2Type< ITILEARNER > )

inline

Initializes the learner.

Definition at line 111 of file fmdpLearner.h.

                                                                                         {
       return MultiDimFunctionGraph< double, ExactTerminalNodePolicy >::getTreeInstance();
     }

◆ _instantiateRewardLearner_() [1/3]

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

RewardLearnerType* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_instantiateRewardLearner_	(	MultiDimFunctionGraph< double > *	target,
		Set< const DiscreteVariable * > &	mainVariables
	)

inline

Initializes the learner.

Definition at line 150 of file fmdpLearner.h.

                                                                                                   {
       return _instantiateRewardLearner_(target, mainVariables, Int2Type< LearnerSelection >());
     }

◆ _instantiateRewardLearner_() [2/3]

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

RewardLearnerType* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_instantiateRewardLearner_	(	MultiDimFunctionGraph< double > *	target,
		Set< const DiscreteVariable * > &	mainVariables,
		Int2Type< IMDDILEARNER >
	)

inline

Initializes the learner.

Definition at line 155 of file fmdpLearner.h.

                                                                             {
       return new RewardLearnerType(target,
                                    _learningThreshold_,
                                    _similarityThreshold_,
                                    mainVariables);
     }

◆ _instantiateRewardLearner_() [3/3]

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

RewardLearnerType* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_instantiateRewardLearner_	(	MultiDimFunctionGraph< double > *	target,
		Set< const DiscreteVariable * > &	mainVariables,
		Int2Type< ITILEARNER >
	)

inline

Initializes the learner.

Definition at line 164 of file fmdpLearner.h.

                                                                           {
       return new RewardLearnerType(target, _learningThreshold_, mainVariables);
     }

◆ _instantiateVarLearner_() [1/3]

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

VariableLearnerType* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_instantiateVarLearner_	(	MultiDimFunctionGraph< double > *	target,
		Set< const DiscreteVariable * > &	mainVariables,
		const DiscreteVariable *	learnedVar
	)

inline

Initializes the learner.

Definition at line 119 of file fmdpLearner.h.

                                                                                               {
       return _instantiateVarLearner_(target,
                                      mainVariables,
                                      learnedVar,
                                      Int2Type< LearnerSelection >());
     }

◆ _instantiateVarLearner_() [2/3]

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

VariableLearnerType* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_instantiateVarLearner_	(	MultiDimFunctionGraph< double > *	target,
		Set< const DiscreteVariable * > &	mainVariables,
		const DiscreteVariable *	learnedVar,
		Int2Type< IMDDILEARNER >
	)

inline

Initializes the learner.

Definition at line 128 of file fmdpLearner.h.

                                                                            {
       return new VariableLearnerType(target,
                                      _learningThreshold_,
                                      _similarityThreshold_,
                                      mainVariables,
                                      learnedVar);
     }

◆ _instantiateVarLearner_() [3/3]

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

VariableLearnerType* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_instantiateVarLearner_	(	MultiDimFunctionGraph< double > *	target,
		Set< const DiscreteVariable * > &	mainVariables,
		const DiscreteVariable *	learnedVar,
		Int2Type< ITILEARNER >
	)

inline

Initializes the learner.

Definition at line 139 of file fmdpLearner.h.

                                                                          {
       return new VariableLearnerType(target, _learningThreshold_, mainVariables, learnedVar);
     }

◆ addObservation()

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

bool gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::addObservation	(	Idx	actionId,
		const Observation *	obs
	)

virtual

Gives to the learner a new transition.

Parameters

actionId	: the action on which the transition was made
obs	: the observed transition

Returns: true if learning this transition implies structural changes (can trigger a new planning)

Implements gum::ILearningStrategy.

Definition at line 142 of file fmdpLearner_tpl.h.

References gum::Set< Key, Alloc >::emplace().

                                                              {
     for (SequenceIteratorSafe< const DiscreteVariable* > varIter = _fmdp_->beginVariables();
          varIter != _fmdp_->endVariables();
          ++varIter) {
       _actionLearners_[actionId]->getWithDefault(*varIter, nullptr)->addObservation(newObs);
       _actionLearners_[actionId]->getWithDefault(*varIter, nullptr)->updateGraph();
     }
 
     if (_actionReward_) {
       _actionRewardLearners_[actionId]->addObservation(newObs);
       _actionRewardLearners_[actionId]->updateGraph();
     } else {
       _rewardLearner_->addObservation(newObs);
       _rewardLearner_->updateGraph();
     }
 
     _rmax_ = _rmax_ < std::abs(newObs->reward()) ? std::abs(newObs->reward()) : _rmax_;
 
     return false;
   }

Here is the call graph for this function:

◆ initialize()

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

void gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::initialize ( FMDP< double > * fmdp )

virtual

Initializes the learner.

Implements gum::ILearningStrategy.

Definition at line 90 of file fmdpLearner_tpl.h.

References gum::Set< Key, Alloc >::emplace().

                                       {
     _fmdp_ = fmdp;
 
     _modaMax_ = 0;
     _rmax_    = 0.0;
 
     Set< const DiscreteVariable* > mainVariables;
     for (auto varIter = _fmdp_->beginVariables(); varIter != _fmdp_->endVariables(); ++varIter) {
       mainVariables.insert(*varIter);
       _modaMax_ = _modaMax_ < (*varIter)->domainSize() ? (*varIter)->domainSize() : _modaMax_;
     }
 
     for (auto actionIter = _fmdp_->beginActions(); actionIter != _fmdp_->endActions();
          ++actionIter) {
       // Adding a Hashtable for the action
       _actionLearners_.insert(*actionIter, new VarLearnerTable());
 
       // Adding a learner for each variable
       for (auto varIter = _fmdp_->beginVariables(); varIter != _fmdp_->endVariables(); ++varIter) {
         MultiDimFunctionGraph< double >* varTrans = _instantiateFunctionGraph_();
         varTrans->setTableName("ACTION : " + _fmdp_->actionName(*actionIter)
                                + " - VARIABLE : " + (*varIter)->name());
         _fmdp_->addTransitionForAction(*actionIter, *varIter, varTrans);
         _actionLearners_[*actionIter]->insert(
            (*varIter),
            _instantiateVarLearner_(varTrans, mainVariables, _fmdp_->main2prime(*varIter)));
       }
 
       if (_actionReward_) {
         MultiDimFunctionGraph< double >* reward = _instantiateFunctionGraph_();
         reward->setTableName("REWARD - ACTION : " + _fmdp_->actionName(*actionIter));
         _fmdp_->addRewardForAction(*actionIter, reward);
         _actionRewardLearners_.insert(*actionIter,
                                       _instantiateRewardLearner_(reward, mainVariables));
       }
     }
 
     if (!_actionReward_) {
       MultiDimFunctionGraph< double >* reward = _instantiateFunctionGraph_();
       reward->setTableName("REWARD");
       _fmdp_->addReward(reward);
       _rewardLearner_ = _instantiateRewardLearner_(reward, mainVariables);
     }
   }

Here is the call graph for this function:

◆ modaMax()

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

virtual double gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::modaMax ( ) const

inlinevirtual

learnerSize

Returns

Implements gum::ILearningStrategy.

Definition at line 226 of file fmdpLearner.h.

References gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_modaMax_.

226 { return _modaMax_; }

gum::FMDPLearner::_modaMax_

double _modaMax_

learnerSize

Definition: fmdpLearner.h:229

◆ rMax()

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

virtual double gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::rMax ( ) const

inlinevirtual

learnerSize

Returns

Implements gum::ILearningStrategy.

Definition at line 220 of file fmdpLearner.h.

References gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_rmax_.

220 { return _rmax_; }

gum::FMDPLearner::_rmax_

double _rmax_

learnerSize

Definition: fmdpLearner.h:223

◆ size()

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

Size gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::size ( )

virtual

learnerSize

Returns

Implements gum::ILearningStrategy.

Definition at line 170 of file fmdpLearner_tpl.h.

References gum::Set< Key, Alloc >::emplace().

                                                                                                  {
     Size s = 0;
     for (SequenceIteratorSafe< Idx > actionIter = _fmdp_->beginActions();
          actionIter != _fmdp_->endActions();
          ++actionIter) {
       for (SequenceIteratorSafe< const DiscreteVariable* > varIter = _fmdp_->beginVariables();
            varIter != _fmdp_->endVariables();
            ++varIter)
         s += _actionLearners_[*actionIter]->getWithDefault(*varIter, nullptr)->size();
       if (_actionReward_) s += _actionRewardLearners_[*actionIter]->size();
     }
 
     if (!_actionReward_) s += _rewardLearner_->size();
 
     return s;
   }

Here is the call graph for this function:

◆ updateFMDP()

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

void gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::updateFMDP ( )

virtual

Starts an update of datastructure in the associated FMDP.

Implements gum::ILearningStrategy.

Definition at line 195 of file fmdpLearner_tpl.h.

References gum::Set< Key, Alloc >::emplace().

                   {
     for (SequenceIteratorSafe< Idx > actionIter = _fmdp_->beginActions();
          actionIter != _fmdp_->endActions();
          ++actionIter) {
       for (SequenceIteratorSafe< const DiscreteVariable* > varIter = _fmdp_->beginVariables();
            varIter != _fmdp_->endVariables();
            ++varIter)
         _actionLearners_[*actionIter]->getWithDefault(*varIter, nullptr)->updateFunctionGraph();
       if (_actionReward_) _actionRewardLearners_[*actionIter]->updateFunctionGraph();
     }
 
     if (!_actionReward_) _rewardLearner_->updateFunctionGraph();
   }

Here is the call graph for this function:

◆ varLearner()

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

const IVisitableGraphLearner* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::varLearner	(	Idx	actionId,
		const DiscreteVariable *	var
	)		const

inlinevirtual

extractCount

Implements gum::ILearningStrategy.

Definition at line 216 of file fmdpLearner.h.

                                                                                               {
       return _actionLearners_[actionId]->getWithDefault(var, nullptr);
     }

Member Data Documentation

◆ _actionLearners_

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

HashTable< Idx, VarLearnerTable* > gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_actionLearners_

private

Definition at line 238 of file fmdpLearner.h.

◆ _actionReward_

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

bool gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_actionReward_

private

Definition at line 240 of file fmdpLearner.h.

◆ _actionRewardLearners_

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

HashTable< Idx, RewardLearnerType* > gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_actionRewardLearners_

private

Definition at line 241 of file fmdpLearner.h.

◆ _fmdp_

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

FMDP< double >* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_fmdp_

private

The FMDP to store the learned model.

Definition at line 236 of file fmdpLearner.h.

◆ _learningThreshold_

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

const double gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_learningThreshold_

private

Definition at line 244 of file fmdpLearner.h.

◆ _modaMax_

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

double gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_modaMax_

private

learnerSize

Returns

Definition at line 229 of file fmdpLearner.h.

Referenced by gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::modaMax().

◆ _rewardLearner_

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

RewardLearnerType* gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_rewardLearner_

private

Definition at line 242 of file fmdpLearner.h.

◆ _rmax_

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

double gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_rmax_

private

learnerSize

Returns

Definition at line 223 of file fmdpLearner.h.

Referenced by gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::rMax().

◆ _similarityThreshold_

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

const double gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >::_similarityThreshold_

private

Definition at line 245 of file fmdpLearner.h.

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

agrum/FMDP/learning/fmdpLearner.h
agrum/FMDP/learning/fmdpLearner_tpl.h

Public Member Functions

Miscelleanous methods

Detailed Description

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection> class gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >

Member Typedef Documentation

◆ RewardLearnerType

◆ VariableLearnerType

◆ VarLearnerTable

Constructor & Destructor Documentation

◆ FMDPLearner()

◆ ~FMDPLearner()

Member Function Documentation

◆ _instantiateFunctionGraph_() [1/3]

◆ _instantiateFunctionGraph_() [2/3]

◆ _instantiateFunctionGraph_() [3/3]

◆ _instantiateRewardLearner_() [1/3]

◆ _instantiateRewardLearner_() [2/3]

◆ _instantiateRewardLearner_() [3/3]

◆ _instantiateVarLearner_() [1/3]

◆ _instantiateVarLearner_() [2/3]

◆ _instantiateVarLearner_() [3/3]

◆ addObservation()

◆ initialize()

◆ modaMax()

◆ rMax()

◆ size()

◆ updateFMDP()

◆ varLearner()

Member Data Documentation

◆ _actionLearners_

◆ _actionReward_

◆ _actionRewardLearners_

◆ _fmdp_

◆ _learningThreshold_

◆ _modaMax_

◆ _rewardLearner_

◆ _rmax_

◆ _similarityThreshold_

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>
class gum::FMDPLearner< VariableAttributeSelection, RewardAttributeSelection, LearnerSelection >