SimulationResult Class Reference

SimulationResult stores the results from simulating a joint policy, the obtained rewards in particular. More...

#include <SimulationResult.h>

Public Member Functions
void	AddReward (double r)
	Add a sampled reward to the results set. More...
double	GetAvgReward (void)
	The average of the stored reward samples. More...
double	GetReward (Index i) const
	Get a sampled reward from the results set. More...
std::vector< double >	GetRewards (void)
	Get the full set of stored reward samples. More...
void	Load (std::string filename)
	Load reward samples from file. More...
void	Print (void)
	Print out the stored reward samples. More...
void	PrintSummary (void)
	Print out a summary. More...
void	Save (std::string filename)
	Save the reward samples to disk. More...
	SimulationResult ()
	(default) Constructor More...
	SimulationResult (int horizon, int random_seed, int nrRuns)
	Construct an object with certain parameters. More...
	~SimulationResult ()
	Destructor. More...

Private Member Functions
void	UpdateStatistics ()

Private Attributes
double	_m_avg_reward
unsigned int	_m_horizon
unsigned int	_m_nr_stored
int	_m_random_seed
std::vector< double >	_m_rewards

Detailed Description

SimulationResult stores the results from simulating a joint policy, the obtained rewards in particular.

At the moment only applies to DecPOMDPs.

Constructor & Destructor Documentation

SimulationResult::SimulationResult

(

)

(default) Constructor

SimulationResult::SimulationResult	(	int	horizon,
		int	random_seed,
		int	nrRuns
	)

Construct an object with certain parameters.

Parameters

horizon	planning horizon
random_seed	seed used for sampling
nrRuns	number of runs simulated

SimulationResult::~SimulationResult

(

)

Destructor.

Member Function Documentation

void SimulationResult::AddReward

(

double

r

)

Add a sampled reward to the results set.

Referenced by SimulationTOIDecPOMDPDiscrete::RunSimulations(), SimulationDecPOMDPDiscrete::RunSimulations(), and SimulationFactoredDecPOMDPDiscrete::RunSimulationsRandomActions().

double SimulationResult::GetAvgReward ( void  )

inline

The average of the stored reward samples.

Referenced by DICEPSPlanner::ApproximateEvaluate(), SimulationTOIDecPOMDPDiscrete::RunSimulations(), and SimulationDecPOMDPDiscrete::RunSimulations().

double SimulationResult::GetReward

(

Index

i

)

const

Get a sampled reward from the results set.

vector< double > SimulationResult::GetRewards

(

void

)

Get the full set of stored reward samples.

void SimulationResult::Load

(

std::string

filename

)

Load reward samples from file.

void SimulationResult::Print

(

void

)

Print out the stored reward samples.

void SimulationResult::PrintSummary

(

void

)

Print out a summary.

void SimulationResult::Save

(

std::string

filename

)

Save the reward samples to disk.

Referenced by SimulationTOIDecPOMDPDiscrete::RunSimulations(), and SimulationDecPOMDPDiscrete::RunSimulations().

void SimulationResult::UpdateStatistics ( )

private

Member Data Documentation

double SimulationResult::_m_avg_reward

private

unsigned int SimulationResult::_m_horizon

private

unsigned int SimulationResult::_m_nr_stored

private

int SimulationResult::_m_random_seed

private

std::vector<double> SimulationResult::_m_rewards

private