Thompson Sampling for Linear-Quadratic Control Problems

Marc Abeille 1, 2 Alessandro Lazaric 1, 2
2 SEQUEL - Sequential Learning
Inria Lille - Nord Europe, CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille (CRIStAL) - UMR 9189
Abstract : We consider the exploration-exploitation tradeoff in linear quadratic (LQ) control problems, where the state dynamics is linear and the cost function is quadratic in states and controls. We analyze the regret of Thompson sampling (\ts) (a.k.a. posterior-sampling for reinforcement learning) in the frequentist setting, i.e., when the parameters characterizing the LQ dynamics are fixed. Despite the empirical and theoretical success in a wide range of problems from multi-armed bandit to linear bandit, we show that when studying the frequentist regret \ts in control problems, we need to trade-off the frequency of sampling optimistic parameters and the frequency of switches in the control policy. This results in an overall regret of $O(T^{2/3})$, which is significantly worse than the regret $O(\sqrt{T})$ achieved by the optimism-in-face-of-uncertainty algorithm in LQ control problems.
Type de document :
Communication dans un congrès
AISTATS 2017 - 20th International Conference on Artificial Intelligence and Statistics, Apr 2017, Fort Lauderdale, United States
Liste complète des métadonnées

https://hal.inria.fr/hal-01493564
Contributeur : Alessandro Lazaric <>
Soumis le : mardi 21 mars 2017 - 17:39:46
Dernière modification le : mardi 3 juillet 2018 - 11:37:07
Document(s) archivé(s) le : jeudi 22 juin 2017 - 14:00:20

Fichier

main.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01493564, version 1

Citation

Marc Abeille, Alessandro Lazaric. Thompson Sampling for Linear-Quadratic Control Problems. AISTATS 2017 - 20th International Conference on Artificial Intelligence and Statistics, Apr 2017, Fort Lauderdale, United States. 〈hal-01493564〉

Partager

Métriques

Consultations de la notice

437

Téléchargements de fichiers

124