Policy Evaluation with Variance Related Risk Criteria in Markov Decision Processes
Aviv TamarDotan Di CastroShie MannorPublished in: CoRR (2013)
Keyphrases
- markov decision processes
- policy evaluation
- policy iteration
- variance reduction
- least squares
- optimal policy
- reinforcement learning
- finite state
- state space
- monte carlo
- dynamic programming
- model free
- partially observable
- average cost
- sample size
- decision processes
- temporal difference
- reinforcement learning algorithms
- markov decision process
- planning under uncertainty
- infinite horizon
- reward function
- action space
- fixed point
- decision problems
- average reward
- markov chain
- function approximation
- markov decision problems
- search algorithm