com.github.chen0040.rl.models

Class UtilityModel

java.lang.Object

com.github.chen0040.rl.models.UtilityModel

All Implemented Interfaces:

Serializable

public class UtilityModel
extends Object
implements Serializable

Author:

xschen 9/27/2015 0027. Utility value of a state $U(s)$ is the expected long term reward of state $s$ given the sequence of reward and the optimal policy Utility value $U(s)$ at state $s$ can be obtained by the Bellman equation Bellman Equtation states that $U(s) = R(s) + \gamma * max_a \sum_{s'} T(s,a,s')U(s')$ where s' is the possible transitioned state given that action $a$ is applied at state $s$ where $T(s,a,s')$ is the transition probability of $s \rightarrow s'$ given that action $a$ is applied at state $s$ where $\sum_{s'} T(s,a,s')U(s')$ is the expected long term reward given that action $a$ is applied at state $s$ where $max_a \sum_{s'} T(s,a,s')U(s')$ is the maximum expected long term reward given that the chosen optimal action $a$ is applied at state $s$

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
UtilityModel()
UtilityModel(int stateCount, int actionCount)
UtilityModel(int stateCount, int actionCount, double initialU)

Method Summary

Modifier and Type	Method and Description
void	copy(UtilityModel rhs)
boolean	equals(Object rhs)
int	getActionCount()
int	getStateCount()
Vec	getU()
double	getU(int stateId)
UtilityModel	makeCopy()
void	reset(double initialU)
void	setU(Vec U)

Methods inherited from class java.lang.Object

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

Constructor Detail

UtilityModel

public UtilityModel(int stateCount,
                    int actionCount,
                    double initialU)

UtilityModel

public UtilityModel(int stateCount,
                    int actionCount)

UtilityModel

public UtilityModel()

Method Detail

setU

public void setU(Vec U)

getU

public Vec getU()

getU

public double getU(int stateId)

getStateCount

public int getStateCount()

getActionCount

public int getActionCount()

copy

public void copy(UtilityModel rhs)

makeCopy

public UtilityModel makeCopy()

equals

public boolean equals(Object rhs)

Overrides:

equals in class Object

reset

public void reset(double initialU)