Analysis of Thompson Sampling for Stochastic Sleeping Bandits

Aritra Chatterjee, Ganesh Ghalme, Shweta Jain, Rohit Vaish, Y. Narahari
Proceedings of the 33rd Conference on Uncertainty in Artificial Intelligence, PMLR R15:91-100, 2017.

Abstract

We study a variant of the stochastic multi- armed bandit problem where the set of avail- able arms varies arbitrarily with time (also known as the sleeping bandit problem). We focus on the Thompson Sampling algorithm and consider a regret notion defined with re- spect to the best available arm. Our main re- sult is an O(log T) regret bound for Thompson Sampling, which generalizes a similar bound known for this algorithm from the classical bandit setting. Our bound also matches (up to constants) the best-known lower bound for the sleeping bandit problem. We show via simu- lations that Thompson Sampling outperforms the UCB-style AUER algorithm for the sleep- ing bandit problem.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR15-chatterjee17a, title = {Analysis of {T}hompson Sampling for Stochastic Sleeping Bandits}, author = {Chatterjee, Aritra and Ghalme, Ganesh and Jain, Shweta and Vaish, Rohit and Narahari, Y.}, booktitle = {Proceedings of the 33rd Conference on Uncertainty in Artificial Intelligence}, pages = {91--100}, year = {2017}, editor = {Elidan, Gal and Kersting, Kristian}, volume = {R15}, series = {Proceedings of Machine Learning Research}, month = {11--15 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r15/main/assets/chatterjee17a/chatterjee17a.pdf}, url = {https://proceedings.mlr.press/r15/chatterjee17a.html}, abstract = {We study a variant of the stochastic multi- armed bandit problem where the set of avail- able arms varies arbitrarily with time (also known as the sleeping bandit problem). We focus on the Thompson Sampling algorithm and consider a regret notion defined with re- spect to the best available arm. Our main re- sult is an O(log T) regret bound for Thompson Sampling, which generalizes a similar bound known for this algorithm from the classical bandit setting. Our bound also matches (up to constants) the best-known lower bound for the sleeping bandit problem. We show via simu- lations that Thompson Sampling outperforms the UCB-style AUER algorithm for the sleep- ing bandit problem.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Analysis of Thompson Sampling for Stochastic Sleeping Bandits %A Aritra Chatterjee %A Ganesh Ghalme %A Shweta Jain %A Rohit Vaish %A Y. Narahari %B Proceedings of the 33rd Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2017 %E Gal Elidan %E Kristian Kersting %F pmlr-vR15-chatterjee17a %I PMLR %P 91--100 %U https://proceedings.mlr.press/r15/chatterjee17a.html %V R15 %X We study a variant of the stochastic multi- armed bandit problem where the set of avail- able arms varies arbitrarily with time (also known as the sleeping bandit problem). We focus on the Thompson Sampling algorithm and consider a regret notion defined with re- spect to the best available arm. Our main re- sult is an O(log T) regret bound for Thompson Sampling, which generalizes a similar bound known for this algorithm from the classical bandit setting. Our bound also matches (up to constants) the best-known lower bound for the sleeping bandit problem. We show via simu- lations that Thompson Sampling outperforms the UCB-style AUER algorithm for the sleep- ing bandit problem. %Z Reissued by PMLR on 04 October 2026.
APA
Chatterjee, A., Ghalme, G., Jain, S., Vaish, R. & Narahari, Y.. (2017). Analysis of Thompson Sampling for Stochastic Sleeping Bandits. Proceedings of the 33rd Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R15:91-100 Available from https://proceedings.mlr.press/r15/chatterjee17a.html. Reissued by PMLR on 04 October 2026.

Related Material