delayedQLearning.cpp

Go to the documentation of this file.

#include "delayedQLearning.h"

using namespace Mdp;

void DelayedQLearning::updateActionValues(state_t previousState, state_t /*nextState*/, action_t previousAction, double reward)
{
    double delta = previousReward + discountFactor * getMaxQ(previousState);
    double av = actionValues->getValue(previousPreviousState, previousPreviousAction);
    delta -= av;
    tabularAv->updateValue(previousPreviousState, previousPreviousAction, av + alpha*delta);
    updateState(previousState, previousAction, reward);
}