Mean variance optimality in Markov decision chains

0411443 - ÚTIA 2010 RIV CZ eng C - Conference Paper (international conference)
Sladký, Karel - Sitař, Milan
Mean variance optimality in Markov decision chains.
[Optimalita prumerne variance v markovskych rozhodovacich procesech.]
Proceedings of the 23rd International Conference Mathematical Methods in Economics 2005. Hradec Králové: Gadeamus, 2005 - (Skalská, H.), s. 350-357. ISBN 978-80-7041-535-1.
[Mathematical Methods in Economics 2005 /23./. Hradec Králové (CZ), 14.09.2005-16.09.2005]
R&D Projects: GA ČR GA402/05/0115
Institutional research plan: CEZ:AV0Z10750506
Keywords : Markov reward processes * expectation and variance of cumulative rewards
Subject RIV: BB - Applied Statistics, Operational Research

In this note, we consider discrete-time Markov decision processes with finite state space. Recalling explicit formulas for the growth rate of expected value and variance of the cumulative (random) reward, algorithmic procedures for finding optimal policies with respect to various mean variance optimality criteria are discussed. Computational experience with large scale numerical examples is reported.

V praci se studuji diskretni markovske rozhodovaci procesy s konecnym stavovym prostorem. Vyuzitim explicitnich vztahu pro rychlost rustu ocekavanych hodnot, jakoz i rozptylu kumulativniho (nahodneho) vynosu, jsou navrzeny algorithnmicke postupy pro nalezeni optimalniho rizeni vzhledem k ruznym kriteriim.
Permanent Link: http://hdl.handle.net/11104/0131524