Risk-Sensitive Markov Decision Processes with Long-Run CVaR Criterion.
Li XiaPeter W. GlynnPublished in: CoRR (2022)
Keyphrases
- risk sensitive
- long run
- markov decision processes
- optimal policy
- average reward
- optimality criterion
- average cost
- infinite horizon
- finite state
- state space
- expected cost
- finite horizon
- dynamic programming
- reinforcement learning
- queueing networks
- decision problems
- policy iteration
- reinforcement learning algorithms
- state dependent
- markov decision process
- multistage
- markov decision problems
- control policies
- decision processes
- partially observable
- robust optimization
- control policy
- action space
- planning under uncertainty
- reward function
- fixed point
- sufficient conditions
- decision theoretic