def __init__(self, worldRef, pos, sigma, agentType, trueBallPos,maxPlayerSpeed, maxBallSpeed , posession = False, epsilon=0.8,alpha=0.2,discount=0.9): agent.__init__(self, worldRef, pos, sigma, agentType, trueBallPos, maxPlayerSpeed, maxBallSpeed, posession) #I'm using these for some stupid hand coded decisions. #delete these when coming up with intelligent agents #TODO: FILL as needed for qlearning self.q_values = collections.defaultdict(int) self.epsilon = float(epsilon) #(exploration prob) self.alpha = float(alpha) #(learning rate) self.discount = float(discount) #(discount rate) #self.start = (self.startx, self.starty) #self.end = (self.goalx, self.goaly) self.isStochastic = True self.isInTraining = False self.oldState = [] self.myAction = False self.action = "HoldBall" self.agent_block_size = 23 self.thresh = 100
def __init__(self, env, start, discount, reward, tprob): agent.__init__(self, env, start, discount)
def __init__(self, name, symbol, gamma_discount_factor=0.9, alpha_learning_rate=0.2, exp_rate= 0.3): agent.__init__(self,name, symbol, gamma_discount_factor, alpha_learning_rate, exp_rate)