摘要
在有限马尔可夫决策过程的线性规划求解方法以及神经网络算法的基础上提出了运用神经网络求解有限马尔可夫决策问题的方法.并通过算例验证了该方法的有效性.
Based of the method of linear programming of finite Markov decision process and neural network algorithm, a method for the solution of finite Markov decision problems has been introduced, the efficiency of the method being explained with examples.
关键词
马尔可夫决策过程
神经网络
线性规划
Markov Decision Process (MDP)
Neural Network (NN)
Linear Programming (LP)
