Abstract:We consider committee election of $k \geq 3$ (out of $m \geq k+1$) candidates, where the voters and the candidates are associated with locations on the real line. Each voter's cardinal preferences over candidates correspond to her distance to the candidate locations, and each voter's cardinal preferences over committees is defined as her distance to the nearest candidate elected in the committee. We consider a setting where the true distances and the locations are unknown. We can nevertheless have access to degraded information which consists of an order of candidates for each voter. We investigate the best possible distortion (a worst-case performance criterion) wrt. the social cost achieved by deterministic committee election rules based on ordinal preferences submitted by $n$ voters and few additional distance queries. We show that for any $k \geq 3$, the best possible distortion of any deterministic algorithm that uses at most $k-3$ distance queries cannot be bounded by any function of $n$, $m$ and $k$. We present deterministic algorithms for $k$-committee election with distortion of $O(n)$ with $O(k)$ distance queries and $O(1)$ with $O(k \log n)$ distance queries.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: in committee elections with 1 - Euclidean preferences, how to use a small number of distance queries to achieve a bounded (or even constant) distortion. Specifically, the paper focuses on how to design deterministic committee election rules to minimize the social cost (i.e., the sum of the distances from all voters to the nearest elected candidate) based on the ordinal preferences submitted by voters and a small number of additional distance queries when the locations of voters and candidates are unknown. The paper specifically studies how many distance queries are required to achieve a bounded distortion under 1 - Euclidean preferences, that is, when voters and candidates are located on the real number line.
### Problem Background
In multi - winner voting, voters usually rank candidates according to their ordinal preferences rather than providing specific cardinal utilities. Although this ordinal preference information is easy to obtain, it may lead to a loss of social welfare in the election results, and this loss is called "distortion". Distortion is defined as the worst - case approximation ratio between the social cost generated by the election rule and the optimal social cost.
### Research Motivation
The research motivations of the paper are as follows:
1. **1 - Euclidean Preferences**: In this setting, both voters and candidates are located on the real number line, and voters' preferences for candidates are determined by the distances between them. This preference structure allows for more efficient use of ordinal preference information to achieve low distortion.
2. **Use of Distance Queries**: By introducing a small number of distance queries, the deficiency of ordinal preference information can be compensated, thereby improving the quality of election results.
3. **Deterministic Rules**: The paper focuses on deterministic election rules rather than random rules, which makes the problem more challenging but also more in line with practical applications.
### Main Contributions
The main contributions of the paper include:
1. **Lower Bound Analysis**: It is proved that for any \(k\geq3\), if at most \(k - 3\) distance queries are used, then a bounded distortion cannot be guaranteed. Specifically, the paper constructs a set of instances, showing that in this case the distortion may be infinite.
2. **Algorithm Design**: A greedy algorithm is proposed, which uses at most \(O(k)\) distance queries and can achieve an \(O(n)\) distortion. In addition, an algorithm using \(O(k\log n)\) distance queries is also proposed, which can achieve a constant - level distortion.
3. **Representative Candidate Set**: By selecting a small representative candidate set, low distortion can be achieved with a small number of distance queries. Specifically, the paper shows how to use \(O(k\log n)\) distance queries to calculate a representative set containing \(O(k\log n)\) candidates, thereby achieving a distortion of 5.
### Conclusion
In general, through theoretical analysis and algorithm design, this paper explores how to use a small number of distance queries to achieve a bounded distortion under 1 - Euclidean preferences. These results not only enrich the theoretical basis of multi - winner elections but also provide valuable guidance for practical applications.