Batched bandit problems
Statistics Theory
2016-03-30 v3 Statistics Theory
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. We propose a simple policy, and 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.
Keywords
Cite
@article{arxiv.1505.00369,
title = {Batched bandit problems},
author = {Vianney Perchet and Philippe Rigollet and Sylvain Chassang and Erik Snowberg},
journal= {arXiv preprint arXiv:1505.00369},
year = {2016}
}
Comments
Published at http://dx.doi.org/10.1214/15-AOS1381 in the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)