NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:4597
Title:Finite-time Analysis of Approximate Policy Iteration for the Linear Quadratic Regulator

The paper studies approximate policy iteration methods in LQR models, contributing to the theoretical results for model-free methods in this setting. The finite-sample results are new and interesting. One issues raised by reviewers is whether this analysis would be generalizable to other more complex RL settings