burlap.behavior.singleagent.shaping

Class ShapedRewardFunction

java.lang.Object

burlap.behavior.singleagent.shaping.ShapedRewardFunction

All Implemented Interfaces:

RewardFunction

Direct Known Subclasses:

PotentialShapedRF

public abstract class ShapedRewardFunction
extends java.lang.Object
implements RewardFunction

This abstract class is used to define shaped reward functions. Shaped reward functions take the base true objective reward function of a task and add some additional reward value to it that helps suggest useful states. Subclasses of the ShapedRewardFunction must implement a method that specifies the additive reward to the base reward.

Author:

James MacGlashan

Field Summary

Fields

Modifier and Type	Field and Description
protected RewardFunction	baseRF The base objective reward function for the task.

Constructor Summary

Constructors

Constructor and Description
ShapedRewardFunction(RewardFunction baseRF) Initializes with the base objective task reward function.

Method Summary

Modifier and Type	Method and Description
abstract double	additiveReward(State s, Action a, State sprime) Returns the reward value to add to the base objective reward function.
double	reward(State s, Action 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

baseRF

protected RewardFunction baseRF

The base objective reward function for the task.

Constructor Detail

ShapedRewardFunction

public ShapedRewardFunction(RewardFunction baseRF)

Initializes with the base objective task reward function.

Parameters:

baseRF - the objective task reward function.

Method Detail

additiveReward

public abstract double additiveReward(State s,
                                      Action a,
                                      State sprime)

Returns the reward value to add to the base objective reward function.

Parameters:

s - the previous state

a - the action taken the previous state

sprime - the successor state

Returns:

the reward value to add to the base objective reward function.

reward

public double reward(State s,
                     Action 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.