Dokument: Voting and Judgment Aggregation: An Axiomatic and Algorithmic Analysis
| Titel: | Voting and Judgment Aggregation: An Axiomatic and Algorithmic Analysis | |||||||
| URL für Lesezeichen: | https://docserv.uni-duesseldorf.de/servlets/DocumentServlet?id=58445 | |||||||
| URN (NBN): | urn:nbn:de:hbz:061-20220104-104736-0 | |||||||
| Kollektion: | Dissertationen | |||||||
| Sprache: | Englisch | |||||||
| Dokumententyp: | Wissenschaftliche Abschlussarbeiten » Dissertation | |||||||
| Medientyp: | Text | |||||||
| Autor: | Selker, Ann-Kathrin [Autor] | |||||||
| Dateien: |
| |||||||
| Beitragende: | Jun.-Prof. Dr. Baumeister, Dorothea [Gutachter] [im Online-Personal- und -Vorlesungsverzeichnis LSF anzeigen] Prof. Dr. Rothe, Jörg [Gutachter] [im Online-Personal- und -Vorlesungsverzeichnis LSF anzeigen] | |||||||
| Dewey Dezimal-Klassifikation: | 000 Informatik, Informationswissenschaft, allgemeine Werke » 004 Datenverarbeitung; Informatik | |||||||
| Beschreibung: | This thesis deals with preference and judgment aggregation in the context of computational social choice, a subfield of multiagent systems and artificial intelligence. In preference aggregation, agents (typically called voters) have preferences over a given set of candidates and the goal is to aggregate ballots derived from these preferences to determine an output (e.g., a single candidate, a set of candidates, or a linear order over the candidates). In judgment aggregation, agents (typically called judges) have judgments over a given set of issues and the goal is to aggregate these judgments into a collective set of judgments.
In the context of multiwinner elections that elect a committee of candidates, this thesis studies a new type of ballot called l-ballots. In contrast to existing ballot types, l-ballots are a compromise between ordinal and cardinal ballots. This thesis focuses on the axiomatic properties of two newly defined types of multiwinner voting rules using these l-ballots as input, and proposes a generalization of l-ballots to fully cardinal ballots. Furthermore, this thesis explores the computational complexity of strategic attacks on voting rules. For several prominent iterative voting rules, i.e., voting rules that proceed in rounds, it is shown that shift bribery is NP-complete in all considered cases. In the context of iterative voting, voters are encouraged to repeatedly update their ballots as a reaction to the current state of the election. This thesis uses a model where voters are connected via an underlying social network and compute their information about the current state of the election both by observing their neighbors' ballots and an opinion poll announced by a polling agency. This thesis explores the manipulation power of the polling agency for the voting rules plurality and veto and shows that manipulation is para-NP-hard even for very restricted social networks and permitted ballot deviations. In particular, the thesis focuses on distance restrictions for the polling agency in regard to the manipulated opinion poll and for the voters in regard to their deviations. Finally, this thesis deals with control in judgment aggregation under various preference types of the chair where a chair tries to achieve a better outcome by changing the structure of the aggregation process, e.g., by adding or deleting judges. This thesis shows NP-completeness for most considered problems for the judgment aggregation procedures uniform (constant) premise-based quota rules. | |||||||
| Lizenz: |  Urheberrechtsschutz | |||||||
| Fachbereich / Einrichtung: | Mathematisch- Naturwissenschaftliche Fakultät » WE Informatik | |||||||
| Dokument erstellt am: | 04.01.2022 | |||||||
| Dateien geändert am: | 04.01.2022 | |||||||
| Promotionsantrag am: | 29.07.2021 | |||||||
| Datum der Promotion: | 16.12.2021 |
