private double CalculateValue(
TState state,
IPolicy <TState, TAction> policy,
IRewarder <TState, TAction> rewarder)
{
var newValue = 0.0;
foreach (var action in _problem.AvailableActions(state))
{
foreach (var(nextState, pNextState) in _problem.PossibleStates(state, action))
{
var reward = rewarder.Reward(state, nextState, action);
newValue +=
policy.PAction(state, action)
* pNextState
* (reward + Value(nextState));
}
}
return(newValue);
}
