Evaluating sets of multi-attribute alternatives with uncertain preferences

Simple item page
dc.availability.bitstreamopenaccess
dc.contributor.advisorWilson, Nicen
dc.contributor.authorToffano, Federico
dc.contributor.funderScience Foundation Irelanden
dc.contributor.funderUnited Technologies Research Centeren
dc.date.accessioned2021-05-11T11:09:17Z
dc.date.available2021-05-11T11:09:17Z
dc.date.issued2020-09
dc.date.submitted2020-09
dc.description.abstractIn a decision-making problem, there can be uncertainty regarding the user preferences concerning the available alternatives. Thus, for a decision support system, it is essential to analyse the user preferences to make personalised recommendations. In this thesis we focus on Multiattribute Utility Theory (MAUT) which aims to define user preference models and elicitation procedures for alternatives evaluated with a vector of a fixed number of conflicting criteria. In this context, a preference model is usually represented with a real value function over the criteria used to evaluate alternatives, and an elicitation procedure is a process of defining such value function. The most preferred alternative will then be the one that maximises the value function. With MAUT models, it is common to represent the uncertainty of the user preferences with a parameterised value function. Each instantiation of this parameterisation then represents a user preference model compatible with the preference information collected so far. For example, a common linear value function is the weighted sum of the criteria evaluating an alternative, which is parameterised with respect to the set of weights. We focus on this type of preference models and in particular on value functions evaluating sets of alternatives rather single alternatives. These value functions can be used for example to define if a set of alternatives is preferred to another one, or which is the worst-case loss in terms of utility units of recommending a set of alternatives. We define the concept of setwise minimal equivalent subset (SME) and algorithms for its computation. Briefly, SME is the subset of an input set of alternatives with equivalent value function and minimum cardinality. We generalise standard preference relations used to compare single alternatives with the purpose of comparing sets of alternatives. We provide computational procedures to compute SME and evaluate preference relations with particular focus on linear value functions. We make extensive use of the Minimax Regret criterion, which is a common method to evaluate alternatives for potential questions and recommendations with uncertain value functions. It prescribes an outcome that minimises the worst-case loss with respect to all the possible parameterisation of the value function. In particular, we focus on its setwise generalisation, namely \textit{Setwise Minimax Regret} (SMR), which is the worst-case loss of recommending a set of alternatives. We provide a novel and efficient procedure for the computation of the SMR when supposing a linear value function. We also present a novel incremental preference elicitation framework for a supplier selection process, where a realistic medium-size factory inspires constraints and objectives of the underlying optimization problem. This preference elicitation framework applies for generic multiattribute combinatorial problems based on a linear preference model, and it is particularly useful when the computation of the set of Pareto optimal alternatives is practically unfeasible.en
dc.description.statusNot peer revieweden
dc.description.versionAccepted Versionen
dc.format.mimetypeapplication/pdfen
dc.identifier.citationToffano, F. 2020. Evaluating sets of multi-attribute alternatives with uncertain preferences. PhD Thesis, University College Cork.en
dc.identifier.endpage141en
dc.identifier.urihttps://hdl.handle.net/10468/11277
dc.language.isoenen
dc.publisherUniversity College Corken
dc.relation.projectinfo:eu-repo/grantAgreement/SFI/SFI Research Centres/12/RC/2289/IE/INSIGHT - Irelands Big Data and Analytics Research Centre/en
dc.rights© 2020 Federico Toffano.en
dc.rights.urihttps://creativecommons.org/licenses/by-nc-nd/4.0/en
dc.subjectMultiattribute utility theoryen
dc.subjectUncertain preferencesen
dc.subjectMultiobjective optimisationen
dc.subjectInteractive preference elicitationen
dc.subjectPreference relationsen
dc.subjectMinimax regreten
dc.titleEvaluating sets of multi-attribute alternatives with uncertain preferencesen
dc.typeDoctoral thesisen
dc.type.qualificationlevelDoctoralen
dc.type.qualificationnamePhD - Doctor of Philosophyen
Files
Original bundle
Now showing 1 - 3 of 3
Thumbnail Image
Name:
ToffanoF_PhD2020.pdf
Size:
1.77 MB
Format:
Adobe Portable Document Format
Description:
Full Text E-thesis
Download
Thumbnail Image
Name:
ToffanoF_PhD2020.zip
Size:
7.39 MB
Format:
http://www.iana.org/assignments/media-types/application/zip
Description:
Supplementary Data
Thumbnail Image
Name:
3. _SubmissionforExaminationForm2- - Federico Toffano (1).pdf
Size:
1.23 MB
Format:
Adobe Portable Document Format
Description:
Submission for Examination Form
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
5.2 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
Research Theses Submission
College of Science, Engineering and Food Science - Doctoral Theses
Computer Science - Doctoral Theses
Insight Centre for Data Analytics - Doctoral Theses