gum::IDecisionStrategy Class Reference

<agrum/FMDP/SDyna/IDecisionStrategy.h> More...

#include <IDecisionStrategy.h>

Inheritance diagram for gum::IDecisionStrategy:

Collaboration diagram for gum::IDecisionStrategy:

Public Member Functions
Constructor & destructor.
virtual	~IDecisionStrategy ()
	Destructor (virtual and empty since it's an interface) More...
Initialization
virtual void	initialize (const FMDP< double > *fmdp)
	Initializes the learner. More...

Incremental methods
const MultiDimFunctionGraph< ActionSet, SetTerminalNodePolicy > *	optPol_
ActionSet	allActions_
virtual void	checkState (const Instantiation &newState, Idx actionId)=0
void	setOptimalStrategy (const MultiDimFunctionGraph< ActionSet, SetTerminalNodePolicy > *optPol)
virtual ActionSet	stateOptimalPolicy (const Instantiation &curState)

Detailed Description

<agrum/FMDP/SDyna/IDecisionStrategy.h>

Interface for manipulating decision center

Definition at line 51 of file IDecisionStrategy.h.

Constructor & Destructor Documentation

~IDecisionStrategy()

virtual gum::IDecisionStrategy::~IDecisionStrategy ( )

inlinevirtual

Destructor (virtual and empty since it's an interface)

Definition at line 60 of file IDecisionStrategy.h.

{}

Member Function Documentation

checkState()

virtual void gum::IDecisionStrategy::checkState ( const Instantiation &  newState, Idx  actionId  )

pure virtual

Implemented in gum::AdaptiveRMaxPlaner, gum::E_GreedyDecider, gum::StatisticalLazyDecider, gum::LazyDecider, and gum::RandomDecider.

initialize()

virtual void gum::IDecisionStrategy::initialize ( const FMDP< double > *  fmdp )

inlinevirtual

Initializes the learner.

Reimplemented in gum::AdaptiveRMaxPlaner, and gum::E_GreedyDecider.

Definition at line 72 of file IDecisionStrategy.h.

                                                        {
      optPol_ = nullptr;
      if (allActions_.size() == 0)
        for (auto actionIter = fmdp->beginActions(); actionIter != fmdp->endActions(); ++actionIter)
          allActions_ += *actionIter;
    }

setOptimalStrategy()

void gum::IDecisionStrategy::setOptimalStrategy ( const MultiDimFunctionGraph< ActionSet, SetTerminalNodePolicy > *  optPol )

inline

Definition at line 89 of file IDecisionStrategy.h.

                                                                                                   {
      optPol_ = const_cast< MultiDimFunctionGraph< ActionSet, SetTerminalNodePolicy >* >(optPol);
    }

stateOptimalPolicy()

virtual ActionSet gum::IDecisionStrategy::stateOptimalPolicy ( const Instantiation &  curState )

inlinevirtual

Reimplemented in gum::E_GreedyDecider, and gum::RandomDecider.

Definition at line 93 of file IDecisionStrategy.h.

                                                                        {
      return (optPol_ && optPol_->realSize() != 0) ? optPol_->get(curState) : allActions_;
    }

Member Data Documentation

allActions_

ActionSet gum::IDecisionStrategy::allActions_

protected

Definition at line 102 of file IDecisionStrategy.h.

optPol_

const MultiDimFunctionGraph< ActionSet, SetTerminalNodePolicy >* gum::IDecisionStrategy::optPol_

protected

Definition at line 99 of file IDecisionStrategy.h.

---

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

• agrum/FMDP/SDyna/Strategies/IDecisionStrategy.h