Estimating Optimal Policy Value in Linear Contextual Bandits Beyond Gaussianity.
Jonathan LeeWeihao KongAldo PacchianoVidya MuthukumarEmma BrunskillPublished in: Trans. Mach. Learn. Res. (2024)
Keyphrases
- optimal policy
- markov decision processes
- finite horizon
- decision problems
- dynamic programming
- state space
- long run
- infinite horizon
- state dependent
- reinforcement learning
- multistage
- markov decision process
- average cost
- serial inventory systems
- sufficient conditions
- policy iteration
- lost sales
- markov decision problems
- average reward
- control policies
- bayesian reinforcement learning
- reward function
- optimal control