On-Policy Deep Reinforcement Learning for the Average-Reward Criterion.
Yiming ZhangKeith W. RossPublished in: ICML (2021)
Keyphrases
- average reward
- reinforcement learning
- optimality criterion
- optimal policy
- markov decision processes
- state and action spaces
- model free
- policy iteration
- semi markov decision processes
- discounted reward
- total reward
- state action
- stochastic games
- rl algorithms
- state space
- long run
- policy gradient
- actor critic
- markov decision process
- finite state
- reinforcement learning algorithms
- function approximation
- decision problems
- temporal difference
- reward function
- dynamic programming
- policy evaluation
- partially observable markov decision processes
- infinite horizon
- hierarchical reinforcement learning
- learning algorithm
- transfer learning
- action space
- action selection
- td learning
- sufficient conditions
- markov chain
- policy gradient reinforcement learning
- markov decision problems
- approximate dynamic programming
- optimal control
- multi agent