Model-Based Reinforcement Learning for Infinite-Horizon Discounted Constrained Markov Decision Processes.
Aria HasanzadeZonuzyDileep M. KalathilSrinivas ShakkottaiPublished in: IJCAI (2021)
Keyphrases
- markov decision processes
- model based reinforcement learning
- infinite horizon
- optimal policy
- finite horizon
- reinforcement learning
- average cost
- finite state
- dynamic programming
- policy iteration
- state space
- partially observable
- markov decision process
- average reward
- planning under uncertainty
- reinforcement learning algorithms
- dec pomdps
- action space
- decision processes
- inventory level
- long run
- markov chain
- fixed point
- markov decision problems
- stationary policies
- optimal control
- reward function