Controlled Markov Decision Processes with AVaR Criteria for Unbounded Costs

Published: 2016/11/27

Applications - OR and Management Sciences average value at risk, markov decision processes, optimal control Short URL: https://optimization-online.org/?p=14293

In this paper, we consider the control problem with the Average-Value-at-Risk (AVaR) criteria of the possibly unbounded L 1 -costs in infinite horizon on a Markov Decision Process (MDP). With a suitable state aggregation and by choosing a priori a global variable s heuristically, we show that there exist optimal policies for the infinite horizon problem for possibly unbounded costs.

Controlled Markov Decision Processes with AVaR Criteria for Unbounded Costs

Citation

Article