File(s) under permanent embargo
Hurwiczα expected utility to rank p-approximated generalized lotteries of I Type with partially quantified uncertainty
conference contribution
posted on 2023-05-23, 13:43 authored by Nataliya NikolovaNataliya Nikolova, Mednikarov, B, Dimitrakiev, D, Kiril TenekedjievKiril TenekedjievWe discuss how to rank one-dimensional fuzzy-rational generalized lotteries of I type – lotteries with continuous real set of prizes, where uncertainty is partially quantifies by p-ribbon distribution functions (CDFs). The p-ribbon CDFs are interpolated on interval quantile estimates of the decision maker. We utilize the Hurwiczα criterion under strict uncertainty to approximate the p-ribbon distribution functions into classical ones. That allows us to transform p-fuzzy rational generalized lotteries of I type into classical ones and use Hurwiczα expected utility principle to rank those. Since the Hurwiczα method weighs the extreme pessimism and extreme optimism, we need to first approximate the p-ribbon distribution functions using the Wald and maximax criteria under strict uncertainty (representing the extreme pessimism and extreme optimism of the fuzzy rational decision maker) before we use the Hurwiczα criterion. A numerical example demonstrates the procedures.
History
Publication title
International Conference - Automatics and Informatics '18 - ProceedingsPagination
103-106ISSN
1313-1850Department/School
Australian Maritime CollegePublisher
John Atanasoff Society of Automatics and InformaticsPlace of publication
BulgariaEvent title
Automatics and Informatics '18Event Venue
Sofia, BulgariaDate of Event (Start Date)
2018-10-04Date of Event (End Date)
2018-10-06Rights statement
Copyright 2018 John Atanasoff Society of Automatics and InformaticsRepository Status
- Restricted