Reinforcement Learning Toolbox 2.0 last updated:

General Documentation Manual Tutorial Class Reference Master Thesis Examples Related Papers Downloads Links News mailto:webmaster Main Page     Class Hierarchy   Compound List   File List   Compound Members   File Members cpolicies.h Go to the documentation of this file. // Copyright (C) 2003 // Gerhard Neumann (gneumann@gmx.net) // Stephan Neumann (sneumann@gmx.net) // // This file is part of RL Toolbox. // http://www.igi.tugraz.at/ril_toolbox // // All rights reserved. // // Redistribution and use in source and binary forms, with or without // modification, are permitted provided that the following conditions // are met: // 1. Redistributions of source code must retain the above copyright // notice, this list of conditions and the following disclaimer. // 2. Redistributions in binary form must reproduce the above copyright // notice, this list of conditions and the following disclaimer in the // documentation and/or other materials provided with the distribution. // 3. The name of the author may not be used to endorse or promote products // derived from this software without specific prior written permission. // // THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR // IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES // OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. // IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, // INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT // NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF // THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. #ifndef CTDLEARNERPOLICIES_H #define CTDLEARNERPOLICIES_H #include "cagentcontroller.h" #include "cparameters.h" class CAbstractFeatureStochasticEstimatedModel; class CTransitionFunction; class CAbstractQFunction; class CActionSet; class CFeatureList; class CActionStatistics; class CAbstractVFunction; class CQFunctionFromTransitionFunction; class CStateCollectionImpl; #include "newmat/newmat.h" class CQGreedyPolicy : public CAgentController { protected: CAbstractQFunction *qFunction; CActionSet *availableActions; public: CQGreedyPolicy(CActionSet *actions, CAbstractQFunction *qFunction); ~CQGreedyPolicy(); virtual CAction *getNextAction(CStateCollection *state, CActionDataSet *data = NULL); }; class CActionDistribution : virtual public CParameterObject { public: virtual void getDistribution(CStateCollection *state, CActionSet *availableActions, double *actionFactors) = 0; virtual bool isDifferentiable() {return false;}; virtual void getGradientFactors(CStateCollection *state, CAction *usedAction, CActionSet *actions, double *actionFactors, ColumnVector *gradientFactors); }; class CSoftMaxDistribution : public CActionDistribution { protected: public: CSoftMaxDistribution(double beta); virtual void getDistribution(CStateCollection *state, CActionSet *availableActions, double *values); virtual bool isDifferentiable() {return true;}; virtual void getGradientFactors(CStateCollection *state, CAction *usedAction, CActionSet *actions, double *actionFactors, ColumnVector *gradientFactors); }; class CAbsoluteSoftMaxDistribution : public CActionDistribution { protected: public: CAbsoluteSoftMaxDistribution(double maxAbsValue); virtual void getDistribution(CStateCollection *state, CActionSet *availableActions, double *values); virtual bool isDifferentiable() {return false;}; //virtual void getGradientFactors(CStateCollection *state, CAction *usedAction, CActionSet *actions, double *actionFactors, ColumnVector *gradientFactors); }; class CGreedyDistribution : public CActionDistribution { public: virtual void getDistribution(CStateCollection *state, CActionSet *availableActions, double *values); }; class CEpsilonGreedyDistribution : public CActionDistribution { protected: public: // double epsilon; CEpsilonGreedyDistribution(double epsilon); virtual void getDistribution(CStateCollection *state, CActionSet *availableActions, double *values); }; class CStochasticPolicy: public CAgentStatisticController { protected: double *actionValues; CActionDistribution *distribution; ColumnVector *gradientFactors; CFeatureList *actionGradientFeatures; CActionSet *availableActions; virtual void getActionStatistics(CStateCollection *, CAction *, CActionStatistics *) {}; public: CStochasticPolicy(CActionSet *actions, CActionDistribution *distribution); ~CStochasticPolicy(); virtual void getActionProbabilities(CStateCollection *state, CActionSet *availableActions, double *actionValues, CActionDataSet *actionDataSet = NULL); virtual CAction *getNextAction(CStateCollection *state, CActionDataSet *dataset, CActionStatistics *stat); virtual void getActionValues(CStateCollection *state, CActionSet *availableActions, double *actionValues, CActionDataSet *actionDataSet = NULL) = 0; virtual bool isDifferentiable() {return false;}; virtual void getActionProbabilityGradient(CStateCollection *state, CAction *action, CActionData *data, CFeatureList *gradientState); virtual void getActionProbabilityLnGradient(CStateCollection *state, CAction *action, CActionData *data, CFeatureList *gradientState); virtual void getActionGradient(CStateCollection *state, CAction *action, CActionData *data, CFeatureList *gradientState); }; class CQStochasticPolicy : public CStochasticPolicy { protected: CAbstractQFunction *qfunction; virtual void getActionStatistics(CStateCollection *state, CAction *action, CActionStatistics *stat); public: CQStochasticPolicy(CActionSet *actions, CActionDistribution *distribution, CAbstractQFunction *qfunction); ~CQStochasticPolicy(); virtual void getActionValues(CStateCollection *state, CActionSet *availableActions, double *actionValues, CActionDataSet *actionDataSet = NULL); virtual void getActionGradient(CStateCollection *state, CAction *action, CActionData *data, CFeatureList *gradientState); virtual bool isDifferentiable(); virtual CAbstractQFunction *getQFunction() {return qfunction;}; }; class CQFunctionFromTransitionFunction; class CVMStochasticPolicy : public CQStochasticPolicy { protected: CStateCollectionImpl *nextState; CStateCollectionImpl *intermediateState; CAbstractVFunction *vFunction; CQFunctionFromTransitionFunction *qFunctionFromTransitionFunction; CTransitionFunction *model; CRewardFunction *reward; public: CVMStochasticPolicy(CActionSet *actions, CActionDistribution *distribution, CAbstractVFunction *vFunction, CTransitionFunction *model, CRewardFunction *reward, std::list<CStateModifier *> *modifiers); ~CVMStochasticPolicy(); virtual void getActionGradient(CStateCollection *state, CAction *action, CActionData *data, CFeatureList *gradientState); virtual bool isDifferentiable(); }; #endif