private double CalculateValue(
GridWorldState state,
IGridWorldPolicy policy,
IGridWorldRewarder gridWorldRewarder)
{
var newValue = 0.0;
foreach (var action in _world.AvailableActions(state))
{
var nextState = _world.NextState(state, action);
var reward = gridWorldRewarder.Reward(state, action);
newValue += policy.PAction(state, action) * (reward + Value(nextState));
}
return(newValue);
}
private static GridWorldAction FindBestAction(
GridWorld world,
GridWorldState state,
GridWorldValueTable gridWorldValueTable,
IGridWorldRewarder gridWorldRewarder)
{
var max = double.MinValue;
var maxAction = GridWorldAction.Down;
foreach (var action in world.AvailableActions(state))
{
var nextState = world.NextState(state, action);
var nextStateValue = gridWorldValueTable.Value(nextState);
var reward = gridWorldRewarder.Reward(state, action);
if (reward + nextStateValue > max)
{
max = reward + nextStateValue;
maxAction = action;
}
}
return(maxAction);
}
public double PAction(GridWorldState state, GridWorldAction action)
{
return(action == _actions[state] ? 1 : 0);
}
public double Value(GridWorldState state)
{
return(_values[state.Position1D]);
}