【摘要】 有关时齐向量值马氏决策模型，（简记为ＶＭＤＰ），胡齐英［３］讨论了时齐ＶＭＤＰ的（ε１，…，εｎ）最优策略问题，张升等［４］给出了一存在最优策略的充分条件。本文继续讨论报酬函数满足一类绝对平均相对有界条件下的时齐ＶＭＤＰ，将时齐标量值模型的主要结果（存在最优策略的充要条件，最优方程，平稳、策略优势、ε最优策略等）均在此作了推广。更多还原

【Abstract】 On the stationary vector valued Markovian decision model,(Abbreviated as VMDP),Zhang Sheng etal have given out a sufficient condition for the existence of optimal policy,Hu Qi ying has discussed the problem of ( ε 1,…,ε n ) optimal plicy. In this paper,a stationary discounted vector valued Markovian decision model is investigated under the absolute average relatively bounded vector valued reward function is discussed.Almost all the major results in scalar valued Markovian decision model (the optimal equation,the necessary and suffcient conditions for a policy to be optimal,the dominating property of stationary poicy) have been extended in this paper.更多还原