MoL-2018-24: van Oostveen, Lucy (2018) What You Know About People’s Preferences Matters: Investigating simpler notions of partial information in the context of strategic manipulation in voting. [Report]
Preview |
Text
MoL-2018-24.text.pdf Download (664kB) | Preview |
Abstract
The Gibbard-Satterthwaite Theorem tells us that most voting rules are susceptible to strategic manipulation, which means that a voter can benefit from voting something other than their true preference. In this theorem, however, it is assumed that voters have full information about the preferences of other voters, which is not always realistic in practice. While models of partial information of the preferences of other voters exist in the literature, the notion of partial information that is used is often very general and includes complicated instances that play an important role in the complexity analysis of voting. In this thesis, simpler and more plausible models of partial information in the context of voting are developed and investigated. We formalized four structures of partial information and looked at both the manipulability and the difficulty of manipulation of these structures in combination with the k-approval, Borda, and Copeland voting rules. We found that restricting partial information to simpler instances does not prevent manipulation and show there is a difference between knowing a little about the preferences of every voter and knowing everything about the preferences of some. Moreover, we show that for k-approval manipulation is computationally easy for these structures and argue why we believe this might also be the case for other voting rules. We believe that this thesis shows that the way in which partial information is modeled influences important properties of manipulability and we therefore believe that it is an important factor that should be taken into account.
| Item Type: | Report |
|---|---|
| Report Nr: | MoL-2018-24 |
| Series Name: | Master of Logic Thesis (MoL) Series |
| Year: | 2018 |
| Subjects: | Computation Logic |
| Depositing User: | Dr Marco Vervoort |
| Date Deposited: | 15 Oct 2018 12:09 |
| Last Modified: | 15 Oct 2018 12:09 |
| URI: | https://eprints.illc.uva.nl/id/eprint/1638 |
Actions (login required)
 |
View Item |
