This paper examines the combinations of two decision-making models. The first, district-voting, is a commonly used method both for political elections (in the US, UK, Canada etc,), as well as in any setting with multiple groups: each district or group have an election, and the election results are aggregated in a second stage. The second model is that of iterative voting. In iterative voting, participants are assumed to adapt their vote to the current situation as they see it (e.g., via polls). Voters attempt to increase the chance the winner will be a candidate they prefer, and in order to do so, they vote differently than they believe. We show that while some voting rules are known to converge when using iterative voting, the process no longer converges when using iterative voting in district-based elections. However, we use simulations to compare the election outcomes using multiple voting rules and multiple distributions, showing which voting rules result in higher quality winners, and which voting rules seem to be less suited for district-base iterative voting.
|Original language||English GB|
|Title of host publication||Games and Information Workshop (GAIW), Montreal, Canada|
|Number of pages||8|
|State||Published - 2019|