Large Language Models for Designing Participatory Budgeting Rules
Nguyen Thach, Xingchen Sha, Hau Chan
TL;DR
This work introduces LLMRule, a first framework that uses large language models within an evolutionary search to automate the design of participatory budgeting (PB) rules, bridging utilitarian and fairness objectives. It defines Strong-EJR approximation as a computable fairness proxy and develops efficient verification and acceleration techniques to enable rapid evaluation during search. Across 617 real PB instances with approval and cardinal ballots, LLMRule produces greedy PB rules that achieve superior utilitarian welfare while maintaining fairness close to baseline proportional rules, and it demonstrates robust performance as completion methods for non-exhaustive MES rules. The approach reduces manual domain knowledge requirements and offers a scalable pathway to generate PB rules with provable fairness properties, with potential impact on the practical deployment of participatory budgeting in cities. The study also situates its contributions within the broader literature on PB rule design and LLM-assisted algorithmic design, highlighting both methodological advances and remaining challenges for fairness verification and scalability.
Abstract
Participatory budgeting (PB) is a democratic paradigm for deciding the funding of public projects given the residents' preferences, which has been adopted in numerous cities across the world. The main focus of PB is designing rules, functions that return feasible budget allocations for a set of projects subject to some budget constraint. Designing PB rules that optimize both utility and fairness objectives based on agent preferences had been challenging due to the extensive domain knowledge required and the proven trade-off between the two notions. Recently, large language models (LLMs) have been increasingly employed for automated algorithmic design. Given the resemblance of PB rules to algorithms for classical knapsack problems, in this paper, we introduce a novel framework, named LLMRule, that addresses the limitations of existing works by incorporating LLMs into an evolutionary search procedure for automating the design of PB rules. Our experimental results, evaluated on more than 600 real-world PB instances obtained from the U.S., Canada, Poland, and the Netherlands with different representations of agent preferences, demonstrate that the LLM-generated rules generally outperform existing handcrafted rules in terms of overall utility while still maintaining a similar degree of fairness.