MultiAgentDecisionProcess
|
RewardModelMappingSparse represents a discrete reward model. More...
#include <RewardModelMappingSparse.h>
Public Member Functions | |
virtual RewardModelMappingSparse * | Clone () const |
Returns a pointer to a copy of this class. More... | |
double | Get (Index s_i, Index ja_i) const |
Returns R(s,ja) More... | |
RewardModelMappingSparse (size_t nrS=1, size_t nrJA=1, const std::string &s_str="s", const std::string &ja_str="ja") | |
default Constructor nrS - number of states nrJA - number of joint actions s_str - how to call a state (For example you can use this class to create a mapping from observation histories and ja's to reals. More... | |
void | Set (Index s_i, Index ja_i, double rew) |
Sets R(s_i,ja_i) More... | |
std::string | SoftPrint () const |
Prints a description of this to a string. More... | |
~RewardModelMappingSparse () | |
Copy constructor. More... | |
Public Member Functions inherited from RewardModel | |
size_t | GetNrJointActions () const |
size_t | GetNrStates () const |
RewardModel (size_t nrS=1, size_t nrJA=1) | |
default Constructor More... | |
virtual | ~RewardModel () |
Destructor. More... | |
Public Member Functions inherited from RewardModelDiscreteInterface | |
void | Print () const |
Print this to cout. More... | |
RewardModelDiscreteInterface () | |
default Constructor More... | |
virtual | ~RewardModelDiscreteInterface () |
Destructor. More... | |
Public Member Functions inherited from QTableInterface | |
virtual | ~QTableInterface () |
Private Types | |
typedef boost::numeric::ublas::compressed_matrix < double > | SparseMatrix |
Private Attributes | |
std::string | _m_ja_str |
SparseMatrix | _m_R |
std::string | _m_s_str |
Friends | |
class | RGet_RewardModelMappingSparse |
RewardModelMappingSparse represents a discrete reward model.
|
private |
RewardModelMappingSparse::RewardModelMappingSparse | ( | size_t | nrS = 1 , |
size_t | nrJA = 1 , |
||
const std::string & | s_str = "s" , |
||
const std::string & | ja_str = "ja" |
||
) |
default Constructor nrS - number of states nrJA - number of joint actions s_str - how to call a state (For example you can use this class to create a mapping from observation histories and ja's to reals.
Then this argument could be "joh") ja_str - idem for the joint actions
References _m_ja_str, and _m_s_str.
Referenced by Clone().
RewardModelMappingSparse::~RewardModelMappingSparse | ( | ) |
Copy constructor.
Destructor.
|
inlinevirtual |
Returns a pointer to a copy of this class.
Implements RewardModelDiscreteInterface.
References RewardModelMappingSparse().
Returns R(s,ja)
Implements RewardModelDiscreteInterface.
References _m_R.
Referenced by SoftPrint().
Sets R(s_i,ja_i)
Index ja_i, Index s_i, are indices of the state and taken joint action. r is the reward. The order of events is s, ja, so is the arg. list.
Implements RewardModelDiscreteInterface.
References _m_R, and Globals::REWARD_PRECISION.
|
virtual |
Prints a description of this to a string.
Implements RewardModelDiscreteInterface.
References _m_ja_str, _m_s_str, Get(), RewardModel::GetNrJointActions(), and RewardModel::GetNrStates().
|
friend |
|
private |
Referenced by RewardModelMappingSparse(), and SoftPrint().
|
private |
|
private |
Referenced by RewardModelMappingSparse(), and SoftPrint().