Learning and Planning in Average-Reward Markov Decision Processes.
Yi WanAbhishek NaikRichard S. SuttonPublished in: ICML (2021)
Keyphrases
- markov decision processes
- average reward
- stochastic games
- reinforcement learning
- partially observable
- optimal policy
- state space
- decision theoretic planning
- policy iteration
- finite state
- actor critic
- semi markov decision processes
- learning algorithm
- discounted reward
- long run
- dynamic programming
- planning under uncertainty
- state abstraction
- average cost
- hierarchical reinforcement learning
- factored mdps
- state action
- real time dynamic programming
- policy gradient
- reinforcement learning algorithms
- rl algorithms
- decision processes
- learning agent
- markov decision process
- decision theoretic
- learning tasks