org.liris.smartgov.lez.politic.policyagent.learning.strategy

Class NNBest

java.lang.Object

org.liris.smartgov.lez.politic.policyagent.learning.strategy.Strategy

org.liris.smartgov.lez.politic.policyagent.learning.strategy.NNBest

public class NNBest
extends Strategy

Deep Reinforcement Learning strategy using tensorflow and Keras.

Author:

Simon

Constructor Summary

Constructors

Constructor and Description
NNBest(java.lang.String id, int stateSize, int nbActions, java.util.List<java.lang.String> labels, java.util.List<PolicyAction> policyActions)

Method Summary

Modifier and Type	Method and Description
PolicyAction	chooseAction()
void	replayModelMessage()
void	saveModelMessage()
void	setLastPerception(FeaturesDouble lastPerception)

Methods inherited from class org.liris.smartgov.lez.politic.policyagent.learning.strategy.Strategy

getLastAnswer, getLastPredictedAction, isStillExploration, setLastAction

Methods inherited from class java.lang.Object

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

Constructor Detail

NNBest

public NNBest(java.lang.String id,
              int stateSize,
              int nbActions,
              java.util.List<java.lang.String> labels,
              java.util.List<PolicyAction> policyActions)

Method Detail

chooseAction

public PolicyAction chooseAction()

Specified by:

chooseAction in class Strategy

setLastPerception

public void setLastPerception(FeaturesDouble lastPerception)

saveModelMessage

public void saveModelMessage()

replayModelMessage

public void replayModelMessage()