burlap.behavior.singleagent.interfaces.rlglue.common

Class RLGlueSARSALambdaFactory

java.lang.Object

burlap.behavior.singleagent.interfaces.rlglue.common.RLGLueQlearningFactory

burlap.behavior.singleagent.interfaces.rlglue.common.RLGlueSARSALambdaFactory

All Implemented Interfaces:

RLGlueLearningAgentFactory

public class RLGlueSARSALambdaFactory
extends RLGLueQlearningFactory

This class is used to setup a BURLAP SARSA-lambda algorithm with RLGlue. It can be launched from a main method and have various SARSA-lambda parameters specified by command line options. Use the "--help" option to see full list.

Author:

James MacGlashan

Field Summary

Fields

Modifier and Type	Field and Description
protected double	lambda The lambda value with default value of 0.5

Fields inherited from class burlap.behavior.singleagent.interfaces.rlglue.common.RLGLueQlearningFactory

learningPolicy, learningRate, qInitFunction

Constructor Summary

Constructors

Constructor and Description
RLGlueSARSALambdaFactory()

Method Summary

Methods

Modifier and Type	Method and Description
LearningAgent	generateAgentForRLDomain(Domain domain, double discount, RewardFunction rf, TerminalFunction tf) Returns a BURLAP learning agent for the given RLGlue generated domain, discount, reward function, and terminal function.
double	getLambda()
static void	main(java.lang.String[] args)
void	setLambda(double lambda)

Methods inherited from class burlap.behavior.singleagent.interfaces.rlglue.common.RLGLueQlearningFactory

getLearningPolicy, getLearningRate, getqInitFunction, setLearningPolicy, setLearningRate, setqInitFunction

Methods inherited from class java.lang.Object

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

Field Detail

lambda

protected double lambda

The lambda value with default value of 0.5

Constructor Detail

RLGlueSARSALambdaFactory

public RLGlueSARSALambdaFactory()

Method Detail

generateAgentForRLDomain

public LearningAgent generateAgentForRLDomain(Domain domain,
                                     double discount,
                                     RewardFunction rf,
                                     TerminalFunction tf)

Description copied from interface: RLGlueLearningAgentFactory

Returns a BURLAP learning agent for the given RLGlue generated domain, discount, reward function, and terminal function.

Specified by:

generateAgentForRLDomain in interface RLGlueLearningAgentFactory

Overrides:

generateAgentForRLDomain in class RLGLueQlearningFactory

Parameters:

domain - a BURALP domain that wraps the RLGlue environment and task

discount - the RLGlue specified discount factor

rf - the RLGlue specified reward function

tf - the RLGlue specified terminal function

Returns:

a BURLAP learning agent

getLambda

public double getLambda()

setLambda

public void setLambda(double lambda)

main

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