SinglePoleBalancing Class Reference

In this environment the agent has to control a cart such that the pole mounted at the top of the cart does not pass a defined angle threshold. More...

#include <SinglePoleBalancing.h>

Inheritance diagram for SinglePoleBalancing:

Public Member Functions
	SinglePoleBalancing (bool fullyObservable=true)
virtual bool	actionSpaceContinuous () const
virtual int	actionSpaceDimension () const
virtual int	actionSpaceElements () const
virtual const Action &	actionSpaceLowerBound () const
virtual const Action &	actionSpaceUpperBound () const
virtual double	deltaT () const
	Time between two simulation steps. More...
virtual const Action &	getAction () const
	Get current action. More...
virtual const ActionSpace::A &	getDiscreteActionSpace () const
virtual const StateSpace::S &	getDiscreteStateSpace () const
virtual const State &	getState () const
	Get current state. More...
virtual void	restart ()
	Restart environment. More...
virtual double	reward () const
	Get reward. More...
virtual bool	stateSpaceContinuous () const
virtual int	stateSpaceDimension () const
virtual int	stateSpaceElements () const
virtual const State &	stateSpaceLowerBound () const
virtual const State &	stateSpaceUpperBound () const
virtual void	stateTransition (const Action &action)
	Perform an action. More...
virtual int	stepsInEpisode () const
	Number of steps during the episode. More...
virtual bool	successful () const
	Check if the agent was successful. More...
virtual bool	terminalState () const
	Check for terminal state. More...
Public Member Functions inherited from OpenANN::Environment
virtual	~Environment ()
Public Member Functions inherited from OpenANN::StateSpace
virtual	~StateSpace ()
Public Member Functions inherited from OpenANN::ActionSpace
virtual	~ActionSpace ()

Additional Inherited Members
Public Types inherited from OpenANN::StateSpace
typedef Eigen::VectorXd	State
typedef std::vector< State >	S
Public Types inherited from OpenANN::ActionSpace
typedef Eigen::VectorXd	Action
typedef std::vector< Action >	A

Detailed Description

In this environment the agent has to control a cart such that the pole mounted at the top of the cart does not pass a defined angle threshold.

A more difficult version of this benchmark is DoublePoleBalancing.

Constructor & Destructor Documentation

SinglePoleBalancing::SinglePoleBalancing

(

bool

fullyObservable = true)

Member Function Documentation

bool SinglePoleBalancing::actionSpaceContinuous ( ) const

virtual

Implements OpenANN::ActionSpace.

int SinglePoleBalancing::actionSpaceDimension ( ) const

virtual

Implements OpenANN::ActionSpace.

int SinglePoleBalancing::actionSpaceElements ( ) const

virtual

Implements OpenANN::ActionSpace.

const OpenANN::Environment::Action & SinglePoleBalancing::actionSpaceLowerBound ( ) const

virtual

Implements OpenANN::ActionSpace.

const OpenANN::Environment::Action & SinglePoleBalancing::actionSpaceUpperBound ( ) const

virtual

Implements OpenANN::ActionSpace.

double SinglePoleBalancing::deltaT ( ) const

virtual

Time between two simulation steps.

Reimplemented from OpenANN::Environment.

const OpenANN::Environment::Action & SinglePoleBalancing::getAction ( ) const

virtual

Get current action.

Returns

action

Implements OpenANN::Environment.

const OpenANN::ActionSpace::A & SinglePoleBalancing::getDiscreteActionSpace ( ) const

virtual

Implements OpenANN::ActionSpace.

const OpenANN::StateSpace::S & SinglePoleBalancing::getDiscreteStateSpace ( ) const

virtual

Implements OpenANN::StateSpace.

const OpenANN::Environment::State & SinglePoleBalancing::getState ( ) const

virtual

Get current state.

Returns

state

Implements OpenANN::Environment.

void SinglePoleBalancing::restart ( )

virtual

Restart environment.

Implements OpenANN::Environment.

double SinglePoleBalancing::reward ( ) const

virtual

Get reward.

Returns

reward for the last state transition

Implements OpenANN::Environment.

bool SinglePoleBalancing::stateSpaceContinuous ( ) const

virtual

Implements OpenANN::StateSpace.

int SinglePoleBalancing::stateSpaceDimension ( ) const

virtual

Implements OpenANN::StateSpace.

int SinglePoleBalancing::stateSpaceElements ( ) const

virtual

Implements OpenANN::StateSpace.

const OpenANN::Environment::State & SinglePoleBalancing::stateSpaceLowerBound ( ) const

virtual

Implements OpenANN::StateSpace.

const OpenANN::Environment::State & SinglePoleBalancing::stateSpaceUpperBound ( ) const

virtual

Implements OpenANN::StateSpace.

void SinglePoleBalancing::stateTransition ( const Action &  action)

virtual

Perform an action.

Parameters

action

next action

Implements OpenANN::Environment.

int SinglePoleBalancing::stepsInEpisode ( ) const

virtual

Number of steps during the episode.

Implements OpenANN::Environment.

bool SinglePoleBalancing::successful ( ) const

virtual

Check if the agent was successful.

Returns

was the agent successful?

Implements OpenANN::Environment.

bool SinglePoleBalancing::terminalState ( ) const

virtual

Check for terminal state.

Returns

is the environment in a terminal state?

Implements OpenANN::Environment.

---

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

• examples/polebalancing/SinglePoleBalancing.h
• examples/polebalancing/SinglePoleBalancing.cpp