Compromise and elicitation in social choice : a study of egalitarianism and incomplete information in voting
Compromis et élicitation en choix social : une étude sur l'égalitarisme et l'information incomplète dans le vote
Napolitano, Beatrice (2022), Compromise and elicitation in social choice : a study of egalitarianism and incomplete information in voting, doctoral thesis prepared under the supervision of Sanver, M. Remzi, Université Paris sciences et lettres
Author(s)
Napolitano, BeatriceUnder the direction of
Sanver, M. RemziAbstract (FR)
Les règles de vote représentent des moyens formels d'agréger les préférences d'un groupe d'électeurs en une décision collective. Mais quelle règle choisir dépend du contexte et de la conception de la justice de la société concerné. Bien que toutes les règles de vote puissent être considérées comme un compromis entre individus, certaines le sont plus que d'autres. Dans ce manuscrit, nous proposons deux définition de compromis basées sur une notion de <Abstract (EN)
Voting rules are formal means of aggregating preferences of a group of voters over a set of alternatives into a collective decision. But which rule to use is a difficult question and the answer depends on the conception of justice of the concerned society or committee. Although all voting rules can be seen as a compromise between individuals, some are more so than others. In this manuscript we provide two notions of compromise based on an equal-loss perspective, favoring an outcome where every voter concedes as equally as possible. We study several rules that reflect some idea of compromise and show how, although imposing a willingness to compromise on individuals, they fail to ensure an outcome where everyone has effectively compromised. Moreover, we question the assumption, often made in the classical setting, that the voting rule is fixed beforehand and full information on the preferences of the individuals is provided. Focusing on positional scoring rules, we propose elicitation procedures based on minimax regret, able to interleave questions to the voters and to the entity responsible for designing the voting rule in order to quickly obtain the most relevant information. Considering Majority Judgment rule, we analyze a particular preference elicitation procedure used in real voting scenarios. Given a partial information on individuals preferences, we analyze the probability of selecting a different winner from the one obtained by considering the full profile and study the minimum number of questions to ask voters in order for this probability to be low.Subjects / Keywords
Choix social; Élicitation de préférence; Prise de décision; Intelligence artificielle; Théorie de la décision; Modélisation des préférences; Computational Social Choice; Preference Modeling; Artificial Intelligence; Decision Theory; Preference ElicitationRelated items
Showing items related by title and author.
-
Lang, Jérôme; Pini, Maria Silvia; Rossi, Francesca; Salvagnin, Domenico; Venable, Kristen Brent; Walsh, Toby (2012) Article accepté pour publication ou publié
-
Napolitano, Beatrice; Cailloux, Olivier; Viappiani, Paolo (2019) Communication / Conférence
-
Wilczynski, Anaëlle (2018-12-13) Thèse
-
Oduro, Franklin; Oforiwaa Amanfo-Tetteh, Regina (2016) Rapport
-
Sali, Mustapha; Giard, Vincent (2013) Article accepté pour publication ou publié