burlap.behavior.singleagent.learning.modellearning.rmax

Class PotentialShapedRMax.PotentialShapedRMaxRF

java.lang.Object

burlap.behavior.singleagent.learning.modellearning.rmax.PotentialShapedRMax.PotentialShapedRMaxRF

All Implemented Interfaces:

RewardFunction

Enclosing class:

PotentialShapedRMax

protected class PotentialShapedRMax.PotentialShapedRMaxRF
extends java.lang.Object
implements RewardFunction

This class is a special version of a potential shaped reward function that does not remove the potential value for transitions to states with uknown action transitions that are followed. This is accomplished by returning a value of zero when the fictious RMax state is recached, rather than subtracting off the previous states potential.

Author:

James MacGlashan

Field Summary

Fields

Modifier and Type	Field and Description
protected PotentialFunction	potential The state potential function
protected RewardFunction	sourceRF The source reward function

Constructor Summary

Constructors

Constructor and Description
PotentialShapedRMax.PotentialShapedRMaxRF(RewardFunction sourceRF, PotentialFunction potential) Initializes.

Method Summary

Methods

Modifier and Type	Method and Description
double	reward(State s, GroundedAction a, State sprime) Returns the reward received when action a is executed in state s and the agent transitions to state sprime.

Methods inherited from class java.lang.Object

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

Field Detail

sourceRF

protected RewardFunction sourceRF

The source reward function

potential

protected PotentialFunction potential

The state potential function

Constructor Detail

PotentialShapedRMax.PotentialShapedRMaxRF

public PotentialShapedRMax.PotentialShapedRMaxRF(RewardFunction sourceRF,
                                         PotentialFunction potential)

Initializes.

Parameters:

sourceRF - the source reward function to which the potential is added.

potential - the state potential function

Method Detail

reward

public double reward(State s,
            GroundedAction a,
            State sprime)

Description copied from interface: RewardFunction

Returns the reward received when action a is executed in state s and the agent transitions to state sprime.

Specified by:

reward in interface RewardFunction

Parameters:

s - the state in which the action was executed

a - the action executed

sprime - the state to which the agent transitioned

Returns:

the reward received when action a is executed in state s and the agent transitions to state sprime.