//this function extends a previous policy jpolPrevTs for ts-1 with the behavior specified by the policy of the BayesianGame for time step ts (jpolBG).
PartialJointPolicyDiscretePure*
BayesianGameForDecPOMDPStage::ConstructExtendedPolicy(
PartialJointPolicyDiscretePure& jpolPrevTs,
JointPolicyDiscretePure& jpolBG,
vector<size_t>& nrOHts,
vector<Index>& firstOHtsI
)
{
//check policy types
if(jpolPrevTs.GetIndexDomainCategory() != PolicyGlobals::OHIST_INDEX)
throw E("BayesianGameForDecPOMDPStage::ConstructExtendedPolicy --- jpolPrevTs.GetIndexDomainCategory() != PolicyGlobals::OHIST_INDEX ");
if(jpolBG.GetIndexDomainCategory() != PolicyGlobals::TYPE_INDEX)
throw E("BayesianGameForDecPOMDPStage::ConstructExtendedPolicy --- jpolPrevTs.GetIndexDomainCategory() != PolicyGlobals::TYPE_INDEX ");
//construct a policy for the DecPOMDP:
//a copy of jpolPrevTs with extended to this time step (ts) by
//jpolBG
PartialJointPolicyDiscretePure* jpolTs = new
PartialJointPolicyPureVector(jpolPrevTs);
jpolTs->SetDepth( jpolTs->GetDepth()+1 );
for(Index agentI=0; agentI < GetNrAgents(); agentI++)
{
for(Index type = 0; type < nrOHts[agentI]; type++)
{
Index ohI = type + firstOHtsI[agentI];
jpolTs->SetAction(agentI, ohI,
jpolBG.GetActionIndex(agentI, type) );
}
}
return(jpolTs);
}
void PlanningUnitFactoredDecPOMDPDiscrete::
ComputeFSAOHDist(
FactoredStateAOHDistribution* fsaoh,
const PartialJointPolicyDiscretePure& pJPol) const
{
Index depth = pJPol.GetDepth();
//cout << "computing factored FSAOHDist for policy of depth"<< depth<<endl;
fsaoh->InitializeFromISD( GetFDPOMDPD()->GetFactoredISD() );
fsaoh->SanityCheck();
Index lastStage;
if(depth > 0)
//depth 1 specifies action for t=0 -> compute FSAOH for t=1
lastStage = depth;
else
return;
//cerr<<"starting updates from t=0 to t<"<<lastStage<<endl;
for(Index t=0 ; t < lastStage; t++)
fsaoh->Update(pJPol);
}
