EXISTENCE OF OPTIMAL POLICY FOR TIME NON-HOMOGENEOUS DISCOUNTED MARKOVIAN DECISION PR0GRAMMING

在线阅读 下载PDF 导出详情
摘要 Inthispaperwediscussthediscrete,timenon--homogeneousdiscountedMarkoviandecisionprogramming,wherethestatespaceandallactionsetsarecountable.Supposethattheoptimumvaluefunctionisfinite.Wegivethenecessaryandsufficientconditionsfortheexistenceofanoptimalpolicy.Supposethattheabsolutemeanofrewardsisrelativelybounded.Wealsogivethenecessaryandsufficientconditionsfortheexistenceofanoptimalpolicy.
机构地区 不详
出版日期 1990年04月14日(中国期刊网平台首次上网日期,不代表论文的发表时间)