A Regularized Approach to Sparse Optimal Policy in Reinforcement Learning.
Wenhao YangXiang LiZhihua ZhangPublished in: NeurIPS (2019)
Keyphrases
- optimal policy
- reinforcement learning
- markov decision processes
- state space
- decision problems
- finite horizon
- markov decision process
- state dependent
- infinite horizon
- long run
- dynamic programming
- finite state
- control policies
- initial state
- sufficient conditions
- least squares
- bayesian reinforcement learning
- model free
- lost sales
- average reward
- temporal difference
- markov decision problems
- policy iteration
- average cost
- partially observable
- function approximation
- multistage
- partially observable markov decision processes
- machine learning
- inventory level
- reward function
- learning algorithm