MultiAgentDecisionProcess
|
MonahanPOMDPPlanner implements Monahan's (1982) POMDP algorithm, which basically generates all possible next-step alpha vectors, followed by pruning. More...
#include <MonahanPOMDPPlanner.h>
Public Member Functions | |
MonahanPOMDPPlanner (const PlanningUnitDecPOMDPDiscrete *pu, bool doIncPrune=true) | |
(default) Constructor More... | |
MonahanPOMDPPlanner (const boost::shared_ptr< const PlanningUnitDecPOMDPDiscrete > &pu, bool doIncPrune=true) | |
std::string | SoftPrintBrief () const |
~MonahanPOMDPPlanner () | |
Destructor. More... | |
Public Member Functions inherited from MonahanPlanner | |
size_t | GetNrVectors () const |
Returns the total number of alpha-vectors stored (summed over all time steps). More... | |
double | GetQ (Index jaohI, Index jaI) const |
double | GetQ (const JointBeliefInterface &b, Index jaI) const |
double | GetQ (const JointBeliefInterface &b, Index t, Index jaI) const |
void | Load (const std::string &filename) |
MonahanPlanner (const PlanningUnitDecPOMDPDiscrete *pu, bool doIncPrune=true) | |
(default) Constructor More... | |
MonahanPlanner (const boost::shared_ptr< const PlanningUnitDecPOMDPDiscrete > &pu, bool doIncPrune=true) | |
virtual void | Plan () |
virtual void | PlanWithCache (const std::string &filenameCache, bool computeIfNotCached=true) |
void | Save (const std::string &filename) const |
void | SetMaxNrAlphas (const std::vector< size_t > &maxNrAlphas) |
void | SetResultsFilename (const std::string &filename) |
virtual | ~MonahanPlanner () |
Destructor. More... | |
Public Member Functions inherited from AlphaVectorPlanning | |
AlphaVectorPlanning (const PlanningUnitDecPOMDPDiscrete *pu) | |
(default) Constructor More... | |
AlphaVectorPlanning (const boost::shared_ptr< const PlanningUnitDecPOMDPDiscrete > &pu) | |
AlphaVectorPlanning (const PlanningUnitFactoredDecPOMDPDiscrete *pu) | |
AlphaVectorPlanning (const boost::shared_ptr< const PlanningUnitFactoredDecPOMDPDiscrete > &pu) | |
GaoVectorSet | BackProject (const ValueFunctionPOMDPDiscrete &v) const |
Back projects a value function. More... | |
GaoVectorSet | BackProject (const VectorSet &v) const |
Back projects a value function, represented as a VectorSet. More... | |
VectorSet | CrossSum (const VectorSet &A, const VectorSet &B) const |
Compute the cross-sum of two vector sets. More... | |
void | ExportPOMDPFile (const std::string &filename) const |
Exports the POMDP to file named filename. More... | |
size_t | GetAcceleratedPruningThreshold () const |
ValueFunctionPOMDPDiscrete | GetImmediateRewardValueFunction () const |
Returns the value function induced by the reward model. More... | |
const PlanningUnitDecPOMDPDiscrete * | GetPU () const |
Returns a ref to the PlanningUnit. More... | |
void | Initialize () |
QFunctionsDiscrete | Prune (const QFunctionsDiscrete &Q) const |
Prune a Q-valuefunction. More... | |
ValueFunctionPOMDPDiscrete | Prune (const ValueFunctionPOMDPDiscrete &V) const |
Prune a POMDP valuefunction. More... | |
VectorSet | Prune (const VectorSet &V) const |
Prune a VectorSet. More... | |
BeliefSet | SampleBeliefs (const ArgumentHandlers::Arguments &args) const |
Sample a belief set according to the arguments. More... | |
BeliefSetNonStationary | SampleBeliefsNonStationary (const ArgumentHandlers::Arguments &args) const |
Sample a non-stationary belief set according to the arguments. More... | |
BeliefSetNonStationary | SampleBeliefsNonStationary (int nrBeliefs=10, int uniqueBeliefs=0, int resetAfter=0, int useQMDPforSamplingBeliefs=0, double QMDPexploreProb=0.1) const |
Sample a non-stationary belief set specifying each argument. More... | |
void | SetAcceleratedPruningThreshold (size_t acceleratedPruningThreshold) |
VectorSet | Union (const VectorSet &A, const VectorSet &B) const |
Computes the union of two vector sets. More... | |
QFunctionsDiscrete | ValueFunctionToQ (const ValueFunctionPOMDPDiscrete &V) const |
virtual | ~AlphaVectorPlanning () |
Destructor. More... | |
Public Member Functions inherited from TimedAlgorithm | |
void | AddTimedEvent (const std::string &id, clock_t duration) |
Adds event of certain duration, e.g., an external program call. More... | |
std::vector< double > | GetTimedEventDurations (const std::string &id) |
Returns all stored durations (in s) for a particular event. More... | |
void | LoadTimers (const std::string &filename) |
Load timing info from file filename. More... | |
void | PrintTimers () const |
Print stored timing info. More... | |
void | PrintTimersSummary () const |
Sums data and prints out a summary. More... | |
void | SaveTimers (const std::string &filename) const |
Save collected timing info to file filename. More... | |
void | SaveTimers (std::ofstream &of) const |
Save collected timing info to ofstream of. More... | |
void | StartTimer (const std::string &id) const |
Start to time an event identified by id. More... | |
void | StopTimer (const std::string &id) const |
Stop to time an event identified by id. More... | |
TimedAlgorithm () | |
(default) Constructor More... | |
virtual | ~TimedAlgorithm () |
Destructor. More... | |
Private Member Functions | |
virtual QFunctionsDiscrete | BackupStage (const QFunctionsDiscrete &Q, size_t maxNrAlphas=0) |
Compute a backup stage. More... | |
void | Initialize () |
void | MonahanCrossSum (const GaoVectorSet &G, QFunctionsDiscrete &Q, Index a, bool doIncPrune, size_t maxNrAlphas=0) const |
Additional Inherited Members | |
Static Public Member Functions inherited from AlphaVectorPlanning | |
static void | CrossSum (const std::vector< AlphaVector > &A, const std::vector< AlphaVector > &B, std::vector< AlphaVector > &output) |
Compute the cross-sum of two vector sets. More... | |
static bool | EqualVS (const VectorSet &VS1, const VectorSet &VS2) |
static void | ExportBeliefSet (const BeliefSet &B, const std::string &filename) |
Exports a belief set to a file. More... | |
static void | ExportPOMDPFile (const std::string &filename, const DecPOMDPDiscreteInterface *decpomdp) |
Exports the POMDP represented by pu to file named filename. More... | |
static void | ExportValueFunction (const std::string &filename, const ValueFunctionPOMDPDiscrete &V, bool includeBGindices=true) |
Exports a value function V to file named filename. More... | |
static void | ExportValueFunction (const std::string &filename, const QFunctionsDiscrete &Q, bool includeBGindices=true) |
Exports a Q functions Q to file named filename. More... | |
static void | ExportValueFunction (const std::string &filename, const QFunctionsDiscreteNonStationary &Q, bool includeBGindices=true) |
Exports a Q functions Q to file named filename. More... | |
static std::vector< int > | GetDuplicateIndices (const VectorSet &V) |
Returns a vector<int> which for each vector k in V specifies whether it is unique, in which case vector<int>[k] is set to -1, or whether it is a duplicate of another vector l in V, in which case vector<int>[k] is set to the index of l, where 0 <= l < V.size()). More... | |
static ValueFunctionPOMDPDiscrete | GetImmediateRewardValueFunction (const PlanningUnitDecPOMDPDiscrete *pu) |
Returns the value function induced by the reward model of pu. More... | |
static ValueFunctionPOMDPDiscrete | ImportValueFunction (const std::string &filename) |
Imports a value function from a file named filename. More... | |
static QFunctionsDiscreteNonStationary | ImportValueFunction (const std::string &filename, size_t nr, size_t nrA, size_t nrS) |
static ValueFunctionPOMDPDiscrete | QFunctionsToValueFunction (const QFunctionsDiscrete &Q) |
Takes all the vectors from the Q-function Q and throws them together as a 'V' function. More... | |
static std::string | SoftPrint (const VectorSet &VS) |
static QFunctionsDiscrete | ValueFunctionToQ (const ValueFunctionPOMDPDiscrete &V, size_t nrA, size_t nrS) |
static VectorSet | ValueFunctionToVectorSet (const ValueFunctionPOMDPDiscrete &V) |
static bool | VectorIsDominated (const AlphaVector &alpha, const ValueFunctionPOMDPDiscrete &V) |
static bool | VectorIsDominated (Index i, const VectorSet &V, const std::vector< bool > &vectorsInVtoConsider) |
static bool | VectorIsInValueFunction (const AlphaVector &alpha, const ValueFunctionPOMDPDiscrete &V) |
static VectorSet * | VectorOfVectorsToVectorSet (const std::vector< std::vector< double > > &vectors) |
static ValueFunctionPOMDPDiscrete | VectorSetToValueFunction (const VectorSet &VS, Index a=0, AlphaVector::BGPolicyIndex betaI=-1) |
Protected Member Functions inherited from MonahanPlanner | |
void | CheckMaxNrVectors (size_t maxNrAlphas, size_t nrAlphas) const |
int | GetTimeStep () const |
Protected Attributes inherited from MonahanPlanner | |
bool | _m_alreadyComputed |
bool | _m_doIncPrune |
bool | _m_initialized |
std::vector< size_t > | _m_maxNrAlphas |
QFunctionsDiscreteNonStationary | _m_qFunction |
std::string | _m_resultsFilename |
MonahanPOMDPPlanner implements Monahan's (1982) POMDP algorithm, which basically generates all possible next-step alpha vectors, followed by pruning.
The pruning is simply checking the values at each corner of the belief simplex (Eagle's algorithm). It can do a simple form of Incremental Pruning.
MonahanPOMDPPlanner::MonahanPOMDPPlanner | ( | const PlanningUnitDecPOMDPDiscrete * | pu, |
bool | doIncPrune = true |
||
) |
(default) Constructor
MonahanPOMDPPlanner::MonahanPOMDPPlanner | ( | const boost::shared_ptr< const PlanningUnitDecPOMDPDiscrete > & | pu, |
bool | doIncPrune = true |
||
) |
MonahanPOMDPPlanner::~MonahanPOMDPPlanner | ( | ) |
Destructor.
|
privatevirtual |
Compute a backup stage.
Implements MonahanPlanner.
References MonahanPlanner::_m_doIncPrune, AlphaVectorPlanning::BackProject(), MonahanPlanner::CheckMaxNrVectors(), PlanningUnitMADPDiscrete::GetNrJointActions(), MonahanPlanner::GetNrVectors(), AlphaVectorPlanning::GetPU(), MonahanCrossSum(), and AlphaVectorPlanning::QFunctionsToValueFunction().
|
privatevirtual |
Implements MonahanPlanner.
References MonahanPlanner::_m_initialized, and AlphaVectorPlanning::Initialize().
|
private |
References MonahanPlanner::CheckMaxNrVectors(), AlphaVectorPlanning::CrossSum(), PlanningUnitMADPDiscrete::GetNrJointObservations(), PlanningUnitMADPDiscrete::GetNrStates(), MonahanPlanner::GetNrVectors(), AlphaVectorPlanning::GetPU(), AlphaVectorPlanning::Prune(), AlphaVector::SetAction(), AlphaVector::SetValue(), and ArgumentHandlers::size().
Referenced by BackupStage().
|
inlinevirtual |
Implements MonahanPlanner.