半马氏MDP平均模型

Semi-Markov decision programming with average criterion

研究了半马氏MDP平均模型，提出了新的较弱的假设条件，证明了半马氏MDP平均模型最优方程解的存在性，然后从最优方程出发，证明了存在ε（≥0）－最优平稳策略。

In this paper,Semi - Markov decision programming with average criterion is investigated.Under some conditions, the optimal equations for the Semi- Maarkov decision programming with average criterion is established and the existene of it's soution is proven, From the optimal equations, the existence of ε(≥0) -optimal stationary policies is proven.