Horizon-free Learning for Markov Decision Processes and Games: Stochastically Bounded Rewards and Improved Bounds.
Shengshi LiLin YangPublished in: ICML (2023)
Keyphrases
- markov decision processes
- reinforcement learning
- stochastic games
- model based reinforcement learning
- partially observable
- finite state
- state space
- action sets
- transition matrices
- optimal policy
- dynamic programming
- learning algorithm
- finite horizon
- reward function
- policy iteration
- average reward
- planning under uncertainty
- discounted reward
- long run
- average cost
- learning agent
- multi agent
- machine learning