private void FixedUpdate()
{
OnStep?.Invoke();
Time.timeScale = TimeScale;
DecayHyperparameters();
int action;
if (Random.Range(0f, 1f) < ExplorationFactor)
{
action = Random.Range(0, 4);
}
else
{
action = QTable[currentState].ToList().IndexOf(QTable[currentState].Max());
}
var envResult = AgentMover.Move(action);
if (envResult.reward == 1)
{
OnEpisodeComplete?.Invoke();
}
float oldQValue = QTable[currentState][action];
var nextMax = QTable[envResult.state].Max();
var newQValue = oldQValue + LearningRate * (envResult.reward + DiscountFactor * nextMax - oldQValue);
QTable[currentState][action] = newQValue;
currentState = envResult.state;
}
private void HandleOnGameStart()
{
_agentMover.Move();
_isGettingReady = false;
_isFinish = false;
}