摘要
讨论Markov决策问题中样本与不可观测状态之间的关系问题。引进了避错特征函数的概念 ,利用隐Markov过程的结构和Markov决策模型建立了隐Markov决策模型。并进一步讨论了状态估计和阶段收益函数。
A discussion is made of the relationship between the unobservable status and sample value of Markov decision processes. With a definition of error avoiding characteristic function introduced, a hidden Markov decision model is built by using the structure of hidden Markov processes and Markov decision models. The status estimation and the stage income function are also discussed.
出处
《西南交通大学学报》
EI
CSCD
北大核心
2002年第4期467-469,共3页
Journal of Southwest Jiaotong University
