Hostname: page-component-7b7567c8c-tnkh4 Total loading time: 0 Render date: 2026-11-16T23:16:22.656Z Has data issue: false hasContentIssue false

Worth of perfect information in bernoulli bandits

Published online by Cambridge University Press:  01 July 2016

Donald A. Berry*
Affiliation:
University of Minnesota
Robert P. Kertz*
Affiliation:
Georgia Institute of Technology
*
∗ Postal address: School of Statistics, University of Minnesota, Minneapolis, MN 55455, USA. Supported in part by NSF Grants DMS 85-05023 and DMS 88-03087.
∗∗ Postal address: School of Mathematics, Georgia Institute of Technology, Atlanta, GA 30332, USA.

Abstract

For k-armed Bernoulli bandits with discounting, sharp comparisons are given between average optimal rewards for a gambler and for a ‘perfectly informed' gambler, over natural collections of prior distributions. Some of these comparisons are proved under general discounting, and others under non-increasing discount sequences. Connections are made between these comparisons and the concept of ‘regret' in the minimax approach to bandit processes. Identification of extremal cases in the sharp comparisons is emphasized.

Information

Type
Research Article
Copyright
Copyright © Applied Probability Trust 1991 

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Article purchase

Temporarily unavailable