Open Access Repository

Expected utility analysis of infinite compound lotteries

Cullum, JE ORCID: 0000-0001-6763-7821, Nikolova, N ORCID: 0000-0001-6160-6282 and Tenekedjiev, K ORCID: 0000-0003-3549-0671 2018 , 'Expected utility analysis of infinite compound lotteries' , International Journal of General Systems, vol. 48, no. 2 , pp. 112-138 , doi: 10.1080/03081079.2018.1548443.

Full text not available from this repository.

Abstract

Lotteries can be used to model alternatives with uncertain outcomes. Decision theory uses compound ordinary lotteries to represent a structure of lotteries within lotteries, but can only rank the finite compound lottery structure. We expand upon this approach to introduce solutions for infinite compound ordinary lotteries (ICOL). We describe a novel procedure to simplify any ICOL as much as possible to a maximum reduced ICOL, which is not a unique representation. We limit our discussion to ICOLs of first order, which are defined as maximum reduced ICOLs with a single maximum reduced ICOL in their direct outcome. Two special cases of ICOLs of first order are discussed. These are recursive and semi-recursive ICOLs. We provide an analytical approach to find the expected utility of recursive ICOLs, and a numerical algorithm for semi-recursive ICOLs. We demonstrate our solution methods by evaluating example decision problems involving: a randomizing device with unsuccessful trials, the St. Petersburg paradox, and training with virtual reality.

Item Type: Article
Authors/Creators:Cullum, JE and Nikolova, N and Tenekedjiev, K
Keywords: ICOLs of first order, recursive ICOL, semi-recursive ICOL, expected utility, generalized St. Petersburg paradox
Journal or Publication Title: International Journal of General Systems
Publisher: Taylor & Francis Ltd
ISSN: 0308-1079
DOI / ID Number: 10.1080/03081079.2018.1548443
Copyright Information:

© 2018 Informa UK Limited, trading as Taylor & Francis Group

Related URLs:
Item Statistics: View statistics for this item

Actions (login required)

Item Control Page Item Control Page
TOP