def getReward(self): Environment.getReward(self) angle = abs(self.sensors[0]) s = abs(self.sensors[2]) reward = 0 if angle < 0.05 and (s < 0.05 or not self.centerCart): reward = +2. elif angle > 0.7 or s > 2.4: reward = 0. else: if self.centerCart: reward = max(+1., 2*(1.-s)) else: reward = +1. return reward