#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 69 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 64 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 72 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 47 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 65 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

◆ 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 159 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 96 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:

◆ 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 107 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 112 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 118 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 165 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 173 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 183 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 128 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 138 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 150 of file fmdpLearner.h.

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

◆ 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 246 of file fmdpLearner.h.

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

246 { return modaMax__; }

gum::FMDPLearner::modaMax__

double modaMax__

learnerSize

Definition: fmdpLearner.h:249

◆ 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 240 of file fmdpLearner.h.

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

240 { return rmax__; }

gum::FMDPLearner::rmax__

double rmax__

learnerSize

Definition: fmdpLearner.h:243

◆ size()

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

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

virtual

learnerSize

Returns

Implements gum::ILearningStrategy.

Definition at line 193 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 222 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 235 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 258 of file fmdpLearner.h.

◆ actionReward__

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

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

private

Definition at line 260 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 261 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 256 of file fmdpLearner.h.

◆ learningThreshold__

template<TESTNAME VariableAttributeSelection, TESTNAME RewardAttributeSelection, LEARNERNAME LearnerSelection>

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

private

Definition at line 264 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 249 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 262 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 243 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 265 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

◆ addObservation()

◆ initialize()

◆ 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]

◆ 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 >