burlap.domain.singleagent.mountaincar

Class MountainCar

java.lang.Object

burlap.domain.singleagent.mountaincar.MountainCar

All Implemented Interfaces:

DomainGenerator

public class MountainCar
extends java.lang.Object
implements DomainGenerator

A domain generator for the classic mountain car domain with default dynamics follow those implemented by Singh and Sutton [1]. In this domain you can change the parameters for min/max position and velocity, the scale of the cosine curve on which the car travels, the force of gravity, acceleration, and the amount of time that elapses between simulation/decision steps. If you previously generated a Domain, changing the physics parameters of this DomainGenerator will not affect how the previously generated domain behaves, only future generated ones.

1. Singh, Satinder P., and Richard S. Sutton. "Reinforcement learning with replacing eligibility traces." Machine learning 22.1-3 (1996): 123-158.

Author:

James MacGlashan

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	MountainCar.ClassicMCTF A Terminal Function for the Mountain Car domain that terminates when the agent's position is >= the max position in the world (0.5 default).
static class	MountainCar.MCModel
static class	MountainCar.MCPhysicsParams

Field Summary

Fields

Modifier and Type	Field and Description
static java.lang.String	ACTION_BACKWARDS A constant for the name of the backwards action
static java.lang.String	ACTION_COAST A constant for the name of the coast action
static java.lang.String	ACTION_FORWARD A constant for the name of the forward action
static java.lang.String	ATT_V A constant for the name of the velocity attribute
static java.lang.String	ATT_X A constant for the name of the x attribute
MountainCar.MCPhysicsParams	physParams The physics parameters for mountain car.
protected RewardFunction	rf
protected TerminalFunction	tf

Constructor Summary

Constructors

Constructor and Description
MountainCar()

Method Summary

Modifier and Type	Method and Description
SADomain	generateDomain() Returns a newly instanced Domain object
RewardFunction	getRf()
TerminalFunction	getTf()
static void	main(java.lang.String[] args) Will launch a visual explorer for the mountain car domain that is controlled with the a-s-d keys.
void	setRf(RewardFunction rf)
void	setTf(TerminalFunction tf)
MCState	valleyState()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

ATT_X

public static final java.lang.String ATT_X

A constant for the name of the x attribute

See Also:

Constant Field Values

ATT_V

public static final java.lang.String ATT_V

A constant for the name of the velocity attribute

See Also:

Constant Field Values

ACTION_FORWARD

public static final java.lang.String ACTION_FORWARD

A constant for the name of the forward action

See Also:

Constant Field Values

ACTION_BACKWARDS

public static final java.lang.String ACTION_BACKWARDS

A constant for the name of the backwards action

See Also:

Constant Field Values

ACTION_COAST

public static final java.lang.String ACTION_COAST

A constant for the name of the coast action

See Also:

Constant Field Values

physParams

public MountainCar.MCPhysicsParams physParams

The physics parameters for mountain car.

rf

protected RewardFunction rf

tf

protected TerminalFunction tf

Constructor Detail

MountainCar

public MountainCar()

Method Detail

getTf

public TerminalFunction getTf()

setTf

public void setTf(TerminalFunction tf)

getRf

public RewardFunction getRf()

setRf

public void setRf(RewardFunction rf)

generateDomain

public SADomain generateDomain()

Description copied from interface: DomainGenerator

Returns a newly instanced Domain object

Specified by:

generateDomain in interface DomainGenerator

Returns:

the newly instantiated Domain object.

valleyState

public MCState valleyState()

main

public static void main(java.lang.String[] args)

Will launch a visual explorer for the mountain car domain that is controlled with the a-s-d keys.

Parameters:

args - empty arguments.