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 CAbstractQFunction Class Reference Interface for all Q-Functions. More... #include <cqfunction.h> Inheritance diagram for CAbstractQFunction: List of all members. Public Member Functions int  getType () bool  isType (int type) void  addType (int Type)   CAbstractQFunction (CActionSet *actions)   Creates a QFunction, handling Q-Values for all actions in the actionset. virtual  ~CAbstractQFunction () virtual void  saveData (FILE *file) virtual void  loadData (FILE *file) virtual void  printValues () virtual void  resetData () void  getActionValues (CStateCollection *state, CActionSet *actions, double *actionValues, CActionDataSet *data=NULL)   Writes the Q-Values of the specified actions in the actionValues array. virtual CAction *  getMax (CStateCollection *state, CActionSet *availableActions, CActionDataSet *data=NULL)   Calculates the best action from a given action set. virtual double  getMaxValue (CStateCollection *state, CActionSet *availableActions)   Returns the best action value from a given action set. virtual void  getStatistics (CStateCollection *state, CAction *action, CActionSet *actions, CActionStatistics *statistics)   Returns the statistics for a given action. virtual void  updateValue (CStateCollection *, CAction *, double, CActionData *=NULL)   Interface for updating a Q-Value. virtual void  setValue (CStateCollection *state, CAction *action, double qValue, CActionData *data=NULL)   Interface for setting a Q-Value. virtual double  getValue (CStateCollection *state, CAction *action, CActionData *data=NULL)=0   Interface for getting a Q-Value. virtual CAbstractQETraces *  getStandardETraces () Public Attributes bool  mayDiverge Protected Attributes int  type Detailed Description Interface for all Q-Functions. Q-Functions depend on the state and the action you can choose in the current state, so the policy is able to decide which action is best (usually the action with the highest Q-Value). CAbstractQFunction is the base class of all Q-Functions. It just provides the interface for getting, setting and updating (adding a value to the current value) Q-Values. These functions are again getValue, setValue and updateValue, now there is always an action object as additional parameter. The class maintains an action set of all actions which are stored in the Q-Function. In addition the class provides one function to retrieve the action with the best Q-Value (getMax), one function for retrieving the value of these best action (getMaxValue) and there is also a function which writes all Q-Values of a state (i.e. for all actions) in a double array (getActionValues). The class also maintains a gamma factor which should be the same as the gamma factor of its value functions. In addition the interface provides functions for storing and loading the Values of the QFunction (storeValues and loadValues). See also: CQFunction Constructor & Destructor Documentation CAbstractQFunction::CAbstractQFunction (  CActionSet *  actions  )    Creates a QFunction, handling Q-Values for all actions in the actionset. virtual CAbstractQFunction::~CAbstractQFunction (   )  [virtual] Member Function Documentation void CAbstractQFunction::addType (  int  Type  )  void CAbstractQFunction::getActionValues (  CStateCollection *  state, CActionSet *  actions, double *  actionValues, CActionDataSet *  data = NULL )    Writes the Q-Values of the specified actions in the actionValues array. so the size of the array has to be at least the size of the action set. Reimplemented in CContinuousTimeQFunctionFromTransitionFunction. virtual CAction* CAbstractQFunction::getMax (  CStateCollection *  state, CActionSet *  availableActions, CActionDataSet *  data = NULL )  [virtual]   Calculates the best action from a given action set. Returns the best action from the availableActions action set. If several actions have the same best Q-Value, the first action which has this value in the action set is choosen. Reimplemented in CContinuousActionQFunction. virtual double CAbstractQFunction::getMaxValue (  CStateCollection *  state, CActionSet *  availableActions )  [virtual]   Returns the best action value from a given action set. Returns the best action value from the availableActions action set. virtual CAbstractQETraces* CAbstractQFunction::getStandardETraces (   )  [inline, virtual]   Reimplemented in CCALinearFAQFunction, CExplorationQFunction, CQFunctionSum, CGradientQFunction, CQFunction, CQFunctionFromStochasticModel, CComposedQFunction, CQFunctionFromTransitionFunction, and CContinuousTimeQFunctionFromTransitionFunction. virtual void CAbstractQFunction::getStatistics (  CStateCollection *  state, CAction *  action, CActionSet *  actions, CActionStatistics *  statistics )  [virtual]   Returns the statistics for a given action. Reimplemented in CComposedQFunction. int CAbstractQFunction::getType (   )  virtual double CAbstractQFunction::getValue (  CStateCollection *  state, CAction *  action, CActionData *  data = NULL )  [pure virtual]   Interface for getting a Q-Value. Implemented in CContinuousActionQFunction, CExplorationQFunction, CQFunctionSum, CQFunction, CQFunctionFromStochasticModel, CComposedQFunction, CQFunctionFromTransitionFunction, CContinuousTimeQFunctionFromTransitionFunction, CUnknownDataQFunction, and CUnknownDataQFunctionFromLocalRBFRegression. bool CAbstractQFunction::isType (  int  type  )  virtual void CAbstractQFunction::loadData (  FILE *  file  )  [virtual]   Implements CLearnDataObject. Reimplemented in CGradientQFunction, CQFunction, and CComposedQFunction. virtual void CAbstractQFunction::printValues (   )  [inline, virtual]   Reimplemented in CQFunction, and CComposedQFunction. virtual void CAbstractQFunction::resetData (   )  [inline, virtual]   Implements CLearnDataObject. Reimplemented in CGradientQFunction, CQFunction, CComposedQFunction, CQFunctionFromGradientFunction, CUnknownDataQFunction, and CRegressionTreeQFunction. virtual void CAbstractQFunction::saveData (  FILE *  file  )  [virtual]   Implements CLearnDataObject. Reimplemented in CGradientQFunction, CQFunction, and CComposedQFunction. virtual void CAbstractQFunction::setValue (  CStateCollection *  state, CAction *  action, double  qValue, CActionData *  data = NULL )  [virtual]   Interface for setting a Q-Value. Reimplemented in CContinuousActionQFunction, CExplorationQFunction, CQFunction, CQFunctionFromStochasticModel, CComposedQFunction, CQFunctionFromTransitionFunction, and CContinuousTimeQFunctionFromTransitionFunction. virtual void CAbstractQFunction::updateValue (  CStateCollection *  , CAction *  , double  , CActionData *  = NULL )  [inline, virtual]   Interface for updating a Q-Value. Reimplemented in CContinuousActionQFunction, CExplorationQFunction, CGradientQFunction, CQFunction, CQFunctionFromStochasticModel, CComposedQFunction, CQFunctionFromTransitionFunction, and CContinuousTimeQFunctionFromTransitionFunction. Member Data Documentation bool CAbstractQFunction::mayDiverge int CAbstractQFunction::type [protected] The documentation for this class was generated from the following file: cqfunction.h