Output details
11 - Computer Science and Informatics
University of Oxford
Output title
Trading Performance for Stability in Markov Decision Processes
Type
E - Conference contribution
Name of conference/published proceedings
28th Annual IEEE/ACM Symposium on Ligic in Computer Science
Volume number
-
Issue number
-
First page of article
331
ISSN of proceedings
1043-6871
Year of publication
2013
Number of additional authors
3
Additional information
<11>
This paper solves the problem of optimising the expected value and variance of mean-payoff reward in finite-state MDPs. The problem was studied in the past (K-J. Chung, "Mean-variance tradeoffs in an undiscounted MDP: The unichain case" and M. J. Sobel, "Mean-variance tradeoffs in an undiscounted MDP") but only solutions for special cases were given. Our work gives a complete solution, and also identifies that alternative definitions of the problem are needed in certain scenarios. We formalise these alternative definitions, and provide verification algorithms for them.
Interdisciplinary
-
Cross-referral requested
-
Research group
None
Citation count
0
Proposed double-weighted
No
Double-weighted statement
-
Reserve for a double-weighted output
No
Non-English
No
English abstract
-