| public static RandomElement ( int a ) : int | ||
| a | int | |
| Résultat | int | |
override public Tuple <int, int> GeneratePerceptAndUpdate()
{
int observation = Utils.RandomElement(this.Environment.ValidObservations);
int reward = Utils.RandomElement(this.Environment.ValidRewards);
return(new Tuple <int, int>(observation, reward));
}
public override Tuple <int, int> PerformAction(int action)
{
//note: here is little confussion because my actions are his observations and vice versa
Debug.Assert(this.IsValidAction(action), "non-valid action used " + action);
this.Action = action;
//Bias in environment: if we won playing rock, we repeat it:
if ((this.Observation == Rock) && (this.Reward == RLose))
{
this.Observation = Rock;
}
else
{
this.Observation = Utils.RandomElement(this.ValidObservations);
}
if (action == this.Observation)
{
this.Reward = this.RDraw;
}
else if ((action == Rock && Observation == Paper) ||
(action == Paper && Observation == Scissors) ||
(action == Scissors && Observation == Rock))
{//Agent lost; env won
this.Reward = RLose;
}
else
{//Agent won
this.Reward = RWin;
}
return(new Tuple <int, int>(this.Observation, this.Reward));
}
override public int Search()
{
return(Utils.RandomElement(this.Environment.ValidActions));
}
public int GenerateRandomReward()
{
return(Utils.RandomElement(this.Environment.ValidRewards));
}
public int GenerateRandomAction()
{
return(Utils.RandomElement(this.Environment.ValidActions));
}