Original title: Mean variance models in Markovian decision processes: Optimality conditions
Authors: Sladký, Karel ; Sitař, M.
Document type: Papers
Conference/Event: Mathematical Methods in Economics 2000 /18./, Praha (CZ), 2000-09-13 / 2000-09-15
Year: 2000
Language: eng
Abstract: We consider a discrete-time Markov reward processes with finite state and action spaces. In contrast with the classical models we assume that the (weighted) long run mean variance, i.e. the (weighted) difference of the ratio of long run second to first moments of total expected reward and the long run average return, is minimized. Ideas for finding optimal long-run average return of Markov and semi-Markov decision processes by policy iterations are heavily employed.
Project no.: AV0Z1075907 (CEP), GA402/99/1136 (CEP), GA402/98/0742 (CEP)
Funding provider: GA ČR, GA ČR
Host item entry: Proceedings of the 18th International Conference on Mathematical Methods in Economics 2000

Institution: Institute of Information Theory and Automation AS ČR (web)
Document availability information: Fulltext is available at the institute of the Academy of Sciences.
Original record: http://hdl.handle.net/11104/0130550

Permalink: http://www.nusl.cz/ntk/nusl-34819


The record appears in these collections:
Research > Institutes ASCR > Institute of Information Theory and Automation
Conference materials > Papers