Batched Bandit Problems

Author(s)
Perchet, Vianney; Rigollet, Philippe; Chassang, Sylvain; Snowberg, Erik
Thumbnail
DownloadRigollet_Batched bandit.pdf (326.4Kb)
OPEN_ACCESS_POLICY
Terms of use
Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/
Metadata
Show full item record
Abstract
Motivated by practical applications, chiefly clinical trials, we study the regret achievable for stochastic bandits under the constraint that the employed policy must split trials into a small number of batches. Our results show that a very small number of batches gives close to minimax optimal regret bounds. As a byproduct, we derive optimal policies with low switching cost for stochastic bandits.
Date issued
2015-09-24
URI
http://hdl.handle.net/1721.1/98879
Department
Massachusetts Institute of Technology. Department of Mathematics
Journal
forthcoming in Annals of Statistics
Publisher
Institute of Mathematical Statistics
Citation
Perchet, Vianney, Philippe Rigollet, Sylvain Chassang, and Erik Snowberg. "Batched Bandit Problems." Annals of Statistics (2015).
Version: Author's final manuscript
ISSN
0090-5364
Collections