Perpetual Voting: Fairness in Long-Term Decision Making
M. Lackner
AAAI 2020
Abstract
In this paper we introduce a new voting formalism to support long-term collective decision making: perpetual voting rules. These are voting rules that take the history of previous decisions into account. Due to this additional information, perpetual voting rules may offer temporal fairness guarantees that cannot be achieved in singular decisions. In particular, such rules may enable minorities to have a fair (proportional) influence on the decision process and thus foster long-term participation of minorities. This paper explores the proposed voting rules via an axiomatic analysis as well as a quantitative evaluation by computer simulations. We identify two perpetual voting rules as particularly recommendable in long-term collective decision making.
Experiments:
| Election type |
Culture |
Candidates |
Voters |
Instances |
Parameters |
| Approval |
Euclidean 2D |
{20} |
{5} |
200000 |
Voters are split in two groups and are placed on the 2d plane by a bivariate normal distribution. For the first group (6 voters) both x- and y-coordinates are independently drawn from N (-0.5, 0.2); for the second group (14 voters) x- and y-coordinates are from N (0.5, 0.2). That is, the first, smaller group is centered around (-0.5, -0.5), the second, larger group around (0.5, 0.5). Alternatives are distributed uniformly in the rectangle [-1, 1] × [-1, 1]. Voters approve all alternatives that have a distance of at most 1.5 times the distance to the closest alternative. This yields approval sets of size 1.8 on average. It is important to note that alternatives change in every round and thus even voters that are close to each other do not necessarily have the same approval sets each round. |