Second Order Optimality in Transient and Discounted Markov Decision Chains

0448938 - ÚTIA 2016 RIV CZ eng C - Conference Paper (international conference)
Sladký, Karel
Second Order Optimality in Transient and Discounted Markov Decision Chains.
Procedings of the 33rd International Conference Mathematical Methods in Economics MME 2015. Plzeň: University of West Bohemia, Plzeň, 2015, s. 731-736. ISBN 978-80-261-0539-8.
[Mathematical Methods in Economics 2015 /33./. Cheb (CZ), 09.09.2015-11.09.2015]
R&D Projects: GA ČR GA13-14445S; GA ČR GA15-10331S
Institutional support: RVO:67985556
Keywords : dynamic programming * discounted and transient Markov reward chains * reward-variance optimality
Subject RIV: BC - Control Systems Theory
http://library.utia.cas.cz/separaty/2015/E/sladky-0448938.pdf

The article is devoted to second order optimality in Markov decision processes. Attention is primarily focused on the reward variance for discounted models and undiscounted transient models (i.e. where the spectral radius of the transition probability matrix is less than unity). Considering the second order optimality criteria means that in the class of policies maximizing (or minimizing) total expected discounted reward (or undiscounted reward for the transient model) we choose the policy minimizing the total variance. Explicit formulae for calculating the variances for transient and discounted models are reported along with sketches of algoritmic procedures for finding second order optimal policies.
Permanent Link: http://hdl.handle.net/11104/0250633