摘要
Inthispaperwediscussthediscrete,timenon--homogeneousdiscountedMarkoviandecisionprogramming,wherethestatespaceandallactionsetsarecountable.Supposethattheoptimumvaluefunctionisfinite.Wegivethenecessaryandsufficientconditionsfortheexistenceofanoptimalpolicy.Supposethattheabsolutemeanofrewardsisrelativelybounded.Wealsogivethenecessaryandsufficientconditionsfortheexistenceofanoptimalpolicy.
出版日期
1990年04月14日(中国Betway体育网页登陆平台首次上网日期,不代表论文的发表时间)