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 CNeuralNetworkStateModifier Class Reference State Modifier used for Neural Networks input states. More... #include <cstatemodifier.h> Inheritance diagram for CNeuralNetworkStateModifier: List of all members. Public Member Functions   CNeuralNetworkStateModifier (CStateProperties *originalState)   CNeuralNetworkStateModifier (CStateProperties *originalState, unsigned int *dimensions, unsigned int numDim)   ~CNeuralNetworkStateModifier () virtual void  getModifiedState (CStateCollection *originalState, CState *modifiedState)   Data preprocessing for NN's. void  setPreprocessing (ColumnVector *input_mean, ColumnVector *input_std) Public Attributes bool  normValues Protected Member Functions virtual void  preprocessInput (ColumnVector *input, ColumnVector *norm_input) Static Protected Member Functions static int  getNumInitContinuousStates (CStateProperties *properties, unsigned int *dimensions, unsigned int numDim)   Returns the number of continuous states for the NN-state. Protected Attributes CStateProperties *  originalState unsigned int *  dimensions unsigned int  numDim ColumnVector *  input_mean ColumnVector *  input_std ColumnVector *  buffVector Detailed Description State Modifier used for Neural Networks input states. This state modifier does the input data preprocessing for neural networks. For Non-Periodic continuous state variables, the current state value gets scaled into the intervall [-1,1], periodic state variables get a special treatment. In order to model the periodicity more accurate, the periodic state variable gets splitted into two state variables, one representing the sinus and one the cosinus of the normalized periodic state. So a neural network state has the number of periodic state variables more state variables than the originalstate. The discrete state variables remain the same. The original state can be set in the constructor and is usually the modelstate. Constructor & Destructor Documentation CNeuralNetworkStateModifier::CNeuralNetworkStateModifier (  CStateProperties *  originalState  )  CNeuralNetworkStateModifier::CNeuralNetworkStateModifier (  CStateProperties *  originalState, unsigned int *  dimensions, unsigned int  numDim )  CNeuralNetworkStateModifier::~CNeuralNetworkStateModifier (   )  Member Function Documentation virtual void CNeuralNetworkStateModifier::getModifiedState (  CStateCollection *  originalState, CState *  modifiedState )  [virtual]   Data preprocessing for NN's. For Non-Periodic continuous state variables, the current state value gets scaled into the intervall [-1,1], periodic state variables get a special treatment. In order to model the periodicity more accurate, the periodic state variable gets splitted into two state variables, one representing the sinus and one the cosinus of the normalized periodic state. So a neural network state has the number of periodic state variables more state variables than the originalstate. The discrete state variables remain the same. The original state can be set in the constructor and is usually the modelstate. Implements CStateModifier. static int CNeuralNetworkStateModifier::getNumInitContinuousStates (  CStateProperties *  properties, unsigned int *  dimensions, unsigned int  numDim )  [static, protected]   Returns the number of continuous states for the NN-state. virtual void CNeuralNetworkStateModifier::preprocessInput (  ColumnVector *  input, ColumnVector *  norm_input )  [protected, virtual] void CNeuralNetworkStateModifier::setPreprocessing (  ColumnVector *  input_mean, ColumnVector *  input_std )  Member Data Documentation ColumnVector* CNeuralNetworkStateModifier::buffVector [protected] unsigned int* CNeuralNetworkStateModifier::dimensions [protected] ColumnVector* CNeuralNetworkStateModifier::input_mean [protected] ColumnVector* CNeuralNetworkStateModifier::input_std [protected] bool CNeuralNetworkStateModifier::normValues unsigned int CNeuralNetworkStateModifier::numDim [protected] CStateProperties* CNeuralNetworkStateModifier::originalState [protected] The documentation for this class was generated from the following file: cstatemodifier.h