Reinforcement Learning for Non-Stationary Markov Decision Processes: The Blessing of (More) Optimism.
Wang Chi CheungDavid Simchi-LeviRuihao ZhuPublished in: ICML (2020)
Keyphrases
- non stationary
- markov decision processes
- reinforcement learning
- optimal policy
- reinforcement learning algorithms
- state space
- finite state
- state and action spaces
- policy iteration
- model based reinforcement learning
- partially observable
- dynamic programming
- markov decision process
- planning under uncertainty
- reachability analysis
- function approximation
- decision theoretic planning
- transition matrices
- random fields
- infinite horizon
- action space
- policy evaluation
- approximate dynamic programming
- average cost
- finite horizon
- reward function
- action sets
- factored mdps
- state abstraction
- average reward
- multi agent
- discounted reward
- model free
- machine learning
- temporal difference
- markov decision problems
- decision problems
- stochastic games
- transfer learning
- empirical mode decomposition
- learning algorithm