Selector::TimingSelectionMetrics Class Reference

Decorator on SelectionMetrics class to add timing informations. More...

#include <timingSelectionMetrics.h>

Inheritance diagram for Selector::TimingSelectionMetrics:

Public Member Functions
virtual	~TimingSelectionMetrics ()=default
	Default destructor.
double	getAgentTime () const
	Return the execution time of learning agent.
double	getLeTime () const
	Return the execution time of learning environment.
size_t	getNbActions () const
	Return the number of actions performed.
double	getScore () const override
double	getUtility () const override
	TimingSelectionMetrics (std::shared_ptr< SelectionMetrics > obj)
	Constructor for the SelectionMetrics with timing.
void	extractMetricsStep (const TPG::TPGVertex *agent, const std::vector< double > actionValues, const Learn::LearningEnvironment &learningEnvironment) override
	Forward to wrapped method.
void	extractMetricsEpisodeWithTiming (const TPG::TPGVertex *agent, size_t nbStepsExecuted, const Learn::LearningEnvironment &learningEnvironment, double agentTimeEpisode, double leTimeEpisode)
	Specialization of extractMetricsEpisode to add timings.
void	weightedSum (std::shared_ptr< SelectionMetrics > other, size_t nbEvaluation, size_t nbEvaluationOther) override
	Specialization of weightedSum to add timings and nbActions.
Public Member Functions inherited from Selector::SelectionMetrics
	SelectionMetrics ()=default
	Default constructor.
virtual	~SelectionMetrics ()=default
	Default destructor.
	SelectionMetrics (double score, double utility=0)
	Constructor with score and utility initialization.
virtual void	initMetrics (const TPG::TPGVertex *agent, const Learn::LearningEnvironment &learningEnvironment)
	Init the metrics for the agent in the learning environment.
virtual void	extractMetricsEpisode (const TPG::TPGVertex *agent, size_t nbStepsExecuted, const Learn::LearningEnvironment &learningEnvironment)
	Extract metrics from the agent in the learning environment.

Protected Attributes
std::shared_ptr< SelectionMetrics >	wrapped
	wrapped SelectionMetrics to time.
double	agentTime = 0
	Execution time of learning agent.
double	leTime = 0
	Execution time of learning environment.
size_t	nbActions = 0
	Number of actions performed.
Protected Attributes inherited from Selector::SelectionMetrics
double	score = 0
double	utility = 0

Additional Inherited Members
Static Public Member Functions inherited from Selector::SelectionMetrics
template<class T >
static T	weightedSum (T value, T valueOther, size_t nbEvaluation, size_t nbEvaluationOther)
	Perform a weighted sum between 2 values.

Detailed Description

Decorator on SelectionMetrics class to add timing informations.

Constructor & Destructor Documentation

TimingSelectionMetrics()

Selector::TimingSelectionMetrics::TimingSelectionMetrics ( std::shared_ptr< SelectionMetrics > obj )

inlineexplicit

Constructor for the SelectionMetrics with timing.

Parameters

obj	SelectionMetrics being timed.

Member Function Documentation

extractMetricsEpisodeWithTiming()

void Selector::TimingSelectionMetrics::extractMetricsEpisodeWithTiming	(	const TPG::TPGVertex *	agent,
		size_t	nbStepsExecuted,
		const Learn::LearningEnvironment &	learningEnvironment,
		double	agentTimeEpisode,
		double	leTimeEpisode )

Specialization of extractMetricsEpisode to add timings.

Parameters

[in]	agent	the TPGVertex representing the agent
[in]	nbStepsExecuted	number of steps executed at the end of the episode
[in]	learningEnvironment	the learning environment in which the agent is evaluated.
[in]	agentTimeEpisode	execution time of learning agent.
[in]	leTimeEpisode	execution time of learning environment.

extractMetricsStep()

void Selector::TimingSelectionMetrics::extractMetricsStep ( const TPG::TPGVertex * agent, const std::vector< double > actionValues, const Learn::LearningEnvironment & learningEnvironment )

overridevirtual

Forward to wrapped method.

Reimplemented from Selector::SelectionMetrics.

getAgentTime()

double Selector::TimingSelectionMetrics::getAgentTime ( ) const

inline

Return the execution time of learning agent.

Returns

Execution time of learning agent.

getLeTime()

double Selector::TimingSelectionMetrics::getLeTime ( ) const

inline

Return the execution time of learning environment.

Returns

Execution time of learning environment.

getNbActions()

size_t Selector::TimingSelectionMetrics::getNbActions ( ) const

inline

Return the number of actions performed.

Returns

Number of actions performed.

getScore()

double Selector::TimingSelectionMetrics::getScore ( ) const

inlineoverridevirtual

Return the score of the agent.

Reimplemented from Selector::SelectionMetrics.

getUtility()

double Selector::TimingSelectionMetrics::getUtility ( ) const

inlineoverridevirtual

Return the utility of the agent.

Reimplemented from Selector::SelectionMetrics.

weightedSum()

void Selector::TimingSelectionMetrics::weightedSum ( std::shared_ptr< SelectionMetrics > other, size_t nbEvaluation, size_t nbEvaluationOther )

overridevirtual

Specialization of weightedSum to add timings and nbActions.

Reimplemented from Selector::SelectionMetrics.

---

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

• D:/a/gegelati/gegelati/gegelatilib/include/selector/timingSelectionMetrics.h
• D:/a/gegelati/gegelati/gegelatilib/src/selector/timingSelectionMetrics.cpp