private List <BeliefState> SimulateTrial(Policy p, int cMaxSteps)
{
BeliefState bsCurrent = m_dDomain.InitialBelief, bsNext = null;
State sCurrent = bsCurrent.sampleState(), sNext = null;
Action a = null;
Observation o = null;
List <BeliefState> lBeliefs = new List <BeliefState>();
while (!m_dDomain.IsGoalState(sCurrent) && lBeliefs.Count < cMaxSteps)
{
a = p.GetAction(bsCurrent);
sNext = sCurrent.Apply(a);
o = sNext.RandomObservation(a);
bsNext = bsCurrent.Next(a, o);
bsCurrent = bsNext;
lBeliefs.Add(bsCurrent);
sCurrent = sNext;
}
return(lBeliefs);
}
private void SimulateTrial(Policy p, MazeViewer viewer)
{
BeliefState bsCurrent = InitialBelief, bsNext = null;
State sCurrent = bsCurrent.sampleState(), sNext = null;
Action a = null;
Observation o = null;
viewer.CurrentState = (MazeState)sCurrent;
viewer.CurrentBelief = bsCurrent;
while (!IsGoalState(sCurrent))
{
a = p.GetAction(bsCurrent);
sNext = sCurrent.Apply(a);
o = sNext.RandomObservation(a);
bsNext = bsCurrent.Next(a, o);
bsCurrent = bsNext;
sCurrent = sNext;
viewer.CurrentState = (MazeState)sCurrent;
viewer.CurrentBelief = bsCurrent;
viewer.CurrentObservation = (MazeObservation)o;
Thread.Sleep(500);
}
}
