- Stochastic Modeling and Node Influence Analysis in Liquid Democracy Systems
- Project Year:
2025
- REU Student (s):
Xiangzhe Xu | Tsinghua University (China)
- Student 1 Institution:
Tsinghua University (China)
- Project Mentor:
Lirong Xia
- Project Mentor Area:
DIMACS
- Project Abstract:
This paper introduces a stochastic observation model for liquid democracy that accounts for uncertainty in delegation decisions, addressing limitations of existing deterministic models. By analyzing key properties such as Positive Proxy Gain (PPG) and Delegated Nash Harmony (DNH), alongside node influence metrics, we evaluate system performance. Theoretical analysis and simulations validate the model's generalizability and PPG property, though DNH may not hold in certain configurations. We incorporate machine learning and heuristic methods to predict node influence, providing robust analytical tools for liquid democracy systems.
