public void Learn()
{
Random rand = new Random();
learner = new QLearner();
for (int i = 0; i < LearningConstants.LearningRounds; i++)
{
int[] agentPos = new int[] { rand.Next(map.Dimension), rand.Next(map.Dimension) };
State state = new State(map, agentPos);
agent = new Agent(state, learner);
for (int j = 0; j < stepLimit; j++)
{
bool isDone = agent.Act(state, agent.ChooseAction(state), (1 - (double)j / stepLimit));
if (isDone)
{
learningDone = true;
break;
}
}
}
}
public Agent(State initialState, QLearner qLearner)
{
learner = qLearner;
currentState = initialState;
}