com.github.chen0040.rl.learning.qlearn

Class QLearner

java.lang.Object

com.github.chen0040.rl.learning.qlearn.QLearner

All Implemented Interfaces:

Serializable, Cloneable

Direct Known Subclasses:

QLambdaLearner

public class QLearner
extends Object
implements Serializable, Cloneable

Created by xschen on 9/27/2015 0027. Implement temporal-difference learning Q-Learning, which is an off-policy TD control algorithm Q is known as the quality of state-action combination, note that it is different from utility of a state

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected QModel	model

Constructor Summary

Constructors

Constructor and Description
QLearner()
QLearner(int stateCount, int actionCount)
QLearner(int stateCount, int actionCount, double alpha, double gamma, double initialQ)
QLearner(QModel model, ActionSelectionStrategy actionSelectionStrategy)

Method Summary

Modifier and Type	Method and Description
void	copy(QLearner rhs)
boolean	equals(Object obj)
String	getActionSelection()
QModel	getModel()
QLearner	makeCopy()
protected double	maxQAtState(int stateId, Set<Integer> actionsAtState)
IndexValue	selectAction(int stateId)
IndexValue	selectAction(int stateId, Set<Integer> actionsAtState)
void	setActionSelection(String conf)
void	setModel(QModel model)
void	update(int stateId, int actionId, int nextStateId, double immediateReward)
void	update(int stateId, int actionId, int nextStateId, Set<Integer> actionsAtNextStateId, double immediateReward)

Methods inherited from class java.lang.Object

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

Field Detail

model

protected QModel model

Constructor Detail

QLearner

public QLearner()

QLearner

public QLearner(int stateCount,
                int actionCount)

QLearner

public QLearner(QModel model,
                ActionSelectionStrategy actionSelectionStrategy)

QLearner

public QLearner(int stateCount,
                int actionCount,
                double alpha,
                double gamma,
                double initialQ)

Method Detail

makeCopy

public QLearner makeCopy()

copy

public void copy(QLearner rhs)

equals

public boolean equals(Object obj)

Overrides:

equals in class Object

getModel

public QModel getModel()

setModel

public void setModel(QModel model)

getActionSelection

public String getActionSelection()

setActionSelection

public void setActionSelection(String conf)

maxQAtState

protected double maxQAtState(int stateId,
                             Set<Integer> actionsAtState)

selectAction

public IndexValue selectAction(int stateId,
                               Set<Integer> actionsAtState)

selectAction

public IndexValue selectAction(int stateId)

update

public void update(int stateId,
                   int actionId,
                   int nextStateId,
                   double immediateReward)

update

public void update(int stateId,
                   int actionId,
                   int nextStateId,
                   Set<Integer> actionsAtNextStateId,
                   double immediateReward)