A multi-step Q-learning algorithm was employed to overcome the slow convergence rate of standard Q-learning, and CMAC neural network was used to generalize the continuous state space.

• 为解决连续过程的学习问题，采用CMAC神经网络对连续状态空间进行泛化。