OpenAI Gym CartPole - Deep Q-Learning (dqn store transition)

dqn_store_transition.py

def store_transition(self, state, action, reward, next_state):
    # 打包 experience
    transition = np.hstack((state, [action, reward], next_state))

    # 存進 memory；舊 memory 可能會被覆蓋
    index = self.memory_counter % self.memory_capacity
    self.memory[index, :] = transition
    self.memory_counter += 1