Home > Research > Publications & Outputs > On the identification and mitigation of weaknes...

Electronic data

  • 1607.05970v1

    Rights statement: https://www.cambridge.org/core/journals/probability-in-the-engineering-and-informational-sciences The final, definitive version of this article has been published in the Journal,Probability in the Engineering and Informational Sciences, 31 (2), pp 239-263 2017, © 2016 Cambridge University Press.

    Accepted author manuscript, 457 KB, PDF document

    Available under license: CC BY-NC: Creative Commons Attribution-NonCommercial 4.0 International License

Links

Text available via DOI:

View graph of relations

On the identification and mitigation of weaknesses in the Knowledge Gradient policy for multi-armed bandits

Research output: Contribution to journalJournal article

Published
<mark>Journal publication date</mark>04/2017
<mark>Journal</mark>Probability in the Engineering and Informational Sciences
Issue number2
Volume31
Number of pages25
Pages (from-to)239-263
Publication statusPublished
Early online date13/09/16
Original languageEnglish

Abstract

The Knowledge Gradient (KG) policy was originally proposed for online ranking and selection problems but has recently been adapted for use in online decision making in general and multi-armed bandit problems (MABs) in particular. We study its use in a class of exponential family MABs and identify weaknesses, including a propensity to take actions which are dominated with respect to both exploitation and exploration. We propose variants of KG which avoid such errors. These new policies include an index heuristic which deploys a KG approach to develop an approximation to the Gittins index. A numerical study shows this policy to perform well over a range of MABs including those for which index policies are not optimal. While KG does not make dominated actions when bandits are Gaussian, it fails to be index consistent and appears not to enjoy a performance advantage over competitor policies when arms are correlated to compensate for its greater computational demands.

Bibliographic note

https://www.cambridge.org/core/journals/probability-in-the-engineering-and-informational-sciences The final, definitive version of this article has been published in the Journal,Probability in the Engineering and Informational Sciences, 31 (2), pp 239-263 2017, © 2016 Cambridge University Press.