Beyond Average Return in Markov Decision Processes.
Alexandre MartheAurélien GarivierClaire VernadePublished in: NeurIPS (2023)
Keyphrases
- markov decision processes
- average cost
- state space
- finite state
- optimal policy
- discounted reward
- dynamic programming
- policy iteration
- model based reinforcement learning
- factored mdps
- transition matrices
- reinforcement learning
- average reward
- decision theoretic planning
- partially observable
- reachability analysis
- reinforcement learning algorithms
- finite horizon
- planning under uncertainty
- risk sensitive
- stationary policies
- action space
- decision processes
- markov decision process
- action sets
- markov decision problems
- policy evaluation
- state and action spaces
- reward function