// function LRTA*-AGENT(s') returns an action
// inputs: s', a percept that identifies the current state
public override IAction Execute(IPercept psPrimed)
{
S sPrimed = ptsFn(psPrimed);
// if GOAL-TEST(s') then return stop
if (problem.testGoal(sPrimed))
{
a = default(A);
}
else
{
// if s' is a new state (not in H) then H[s'] <- h(s')
if (!H.ContainsKey(sPrimed))
{
H.Put(sPrimed, getHeuristicFunction().applyAsDouble(sPrimed));
}
// if s is not null
double min = double.MaxValue;
if (null != s)
{
// result[s, a] <- s'
result.Put(s, a, sPrimed);
// H[s] <- min LRTA*-COST(s, b, result[s, b], H)
// b (element of) ACTIONS(s)
min = double.MaxValue;
foreach (A b in problem.getActions(s))
{
double cost = lrtaCost(s, b, result.Get(s, b));
if (cost < min)
{
min = cost;
}
}
H.Put(s, min);
}
// a <- an action b in ACTIONS(s') that minimizes LRTA*-COST(s', b,
// result[s', b], H)
// Just in case no actions
a = default(A);
foreach (A b in problem.getActions(sPrimed))
{
double cost = lrtaCost(sPrimed, b, result.Get(sPrimed, b));
if (cost < min)
{
min = cost;
a = b;
}
}
}
// s <- s'
s = sPrimed;
if (a == null)
{
// I'm either at the Goal or can't get to it,
// which in either case I'm finished so just die.
SetAlive(false);
}
// return a
return(a);
}
// function ONLINE-DFS-AGENT(s') returns an action
// inputs: s', a percept that identifies the current state
public override IAction Execute(IPercept psPrimed)
{
S sPrimed = ptsFn(psPrimed);
// if GOAL-TEST(s') then return stop
if (problem.testGoal(sPrimed))
{
a = default(A);
}
else
{
// if s' is a new state (not in untried) then untried[s'] <-
// ACTIONS(s')
if (!untried.ContainsKey(sPrimed))
{
untried.Put(sPrimed, problem.getActions(sPrimed));
}
// if s is not null then do
if (null != s)
{
// Note: If I've already seen the result of this
// [s, a] then don't put it back on the unbacktracked
// list otherwise you can keep oscillating
// between the same states endlessly.
if (!(sPrimed.Equals(result.Get(s, a))))
{
// result[s, a] <- s'
result.Put(s, a, sPrimed);
// Ensure the unbacktracked always has a list for s'
if (!unbacktracked.ContainsKey(sPrimed))
{
unbacktracked.Put(sPrimed, CollectionFactory.CreateQueue <S>());
}
// add s to the front of the unbacktracked[s']
unbacktracked.Get(sPrimed).Insert(0, s);
}
}
// if untried[s'] is empty then
if (untried.Get(sPrimed).IsEmpty())
{
// if unbacktracked[s'] is empty then return stop
if (unbacktracked.Get(sPrimed).IsEmpty())
{
a = default(A);
}
else
{
// else a <- an action b such that result[s', b] =
// POP(unbacktracked[s'])
S popped = unbacktracked.Get(sPrimed).Pop();
foreach (Pair <S, A> sa in result.GetKeys())
{
if (sa.GetFirst().Equals(sPrimed) && result.Get(sa).Equals(popped))
{
a = sa.getSecond();
break;
}
}
}
}
else
{
// else a <- POP(untried[s'])
a = untried.Get(sPrimed).Pop();
}
}
if (a == null)
{
// I'm either at the Goal or can't get to it,
// which in either case I'm finished so just die.
SetAlive(false);
}
// s <- s'
s = sPrimed;
// return a
return(a);
}
