Batched bandit problems

Authors:
  • VPVianney Perchet
    Vianney Perchet
  • PRPhilippe Rigollet
    Philippe Rigollet
  • SCSylvain Chassang
    Sylvain Chassang
  • ESErik Snowberg
    Erik Snowberg

Top Cited Papers
Open Access
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.

This publication has 32 references indexed in Scilit: