Stochastic and Adversarial Combinatorial Bandits - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2015

Stochastic and Adversarial Combinatorial Bandits

Résumé

This paper investigates stochastic and adversarial combinatorial multi-armed bandit problems. In the stochastic setting, we first derive problem-specific regret lower bounds, and analyze how these bounds scale with the dimension of the decision space. We then propose COMBUCB, algorithms that efficiently exploit the combinatorial structure of the problem, and derive finite-time upper bound on their regrets. These bounds improve over regret upper bounds of existing algorithms, and we show numerically thatCOMBUCB significantly outperforms any other algorithm. In the adversarial setting, we propose two simple algorithms, namely COMBEXP-1 and COMBEXP-2 for semi-bandit and bandit feedback, respectively. Their regrets have similar scaling as state-of-the-art algorithms, in spite of the simplicity of their implementation
Fichier non déposé

Dates et versions

hal-01137956 , version 1 (31-03-2015)

Identifiants

  • HAL Id : hal-01137956 , version 1

Citer

Richard Combes, Marc Lelarge, Alexandre Proutière, Sadegh Talebi. Stochastic and Adversarial Combinatorial Bandits. 2015. ⟨hal-01137956⟩
363 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More