Algorithmic procedures for mean-variance optimality in Markov decision chains. Abstract

0410869 - UTIA-B 20020083 CZ eng A - Abstrakt
Sladký, Karel - Sitař, Milan
Algorithmic procedures for mean-variance optimality in Markov decision chains. Abstract.
Prague: Institute of Information Theory and Automation, 2002. Abstracts of the 24th European Meeting of Statisticians & 14th Prague Conference on Information Theory, Statistical Decision Functions and Random Processes. - (Janžura, M.; Mikosch, T.). s. 322
[EMS 2002. 19.08.2002-23.08.2002, Prague]
Grant CEP: GA ČR GA402/02/1015; GA ČR GA402/01/0539
Výzkumný záměr: CEZ:AV0Z1075907
Klíčová slova: Markov decision chains * mean-variance * policy iteration
Kód oboru RIV: BB - Aplikovaná statistika, operační výzkum

We investigate how the mean-variance selection rule, originally proposed for portfolio selection problems, can work in Markovian decision models. We consider a Markov decision chain with finite state and action spaces, however, instead of average expected reward or average expected variance optimality we consider mean variance optimality, square mean variance optimality or weighted difference of average expected rewards and variances. Optimality conditions and algorithmic procedures are presented.
Trvalý link: http://hdl.handle.net/11104/0130956