Abstract:The massive population election simulation aims to model the preferences of specific groups in particular election scenarios. It has garnered significant attention for its potential to forecast real-world social trends. Traditional agent-based modeling (ABM) methods are constrained by their ability to incorporate complex individual background information and provide interactive prediction results. In this paper, we introduce ElectionSim, an innovative election simulation framework based on large language models, designed to support accurate voter simulations and customized distributions, together with an interactive platform to dialogue with simulated voters. We present a million-level voter pool sampled from social media platforms to support accurate individual simulation. We also introduce PPE, a poll-based presidential election benchmark to assess the performance of our framework under the U.S. presidential election scenario. Through extensive experiments and analyses, we demonstrate the effectiveness and robustness of our framework in U.S. presidential election simulations.

What problem does this paper attempt to address?

This paper attempts to solve the following three main problems: 1. **How to achieve high - precision election simulation at the individual level?** - Traditional agent - based modeling (ABM) methods are difficult to integrate complex individual background information, resulting in insufficiently accurate simulations of individual behaviors. Although large language models (LLMs) can generate human - like behaviors, they lack sufficient personalized input data in large - scale election simulations and cannot fully capture the diversity of voter behaviors, motivations, and decision - making processes. This directly affects the accuracy of macro - level results. 2. **How to generate customized distributions consistent with real - world statistical data?** - Accurate election simulations require ensuring that simulated individuals can represent the diversity and distribution of the real - world population. Although random sampling can capture diversity, it is difficult to align with the real - world demographic distribution and is prone to bias. Therefore, a carefully planned sampling strategy needs to be designed to ensure that the distribution of simulated individuals is consistent with that of actual voters. 3. **How to systematically evaluate the performance of election simulations?** - Existing election simulation evaluation methods mainly focus on prediction accuracy, which is too single - minded to comprehensively evaluate all aspects of simulation results. In order to more comprehensively evaluate the effectiveness of election simulations, a multi - faceted evaluation method needs to be designed to provide a more comprehensive analysis. To solve these problems, the paper proposes the **ElectionSim** framework, which uses large - language - model - driven agents for large - scale election simulations. Specifically: - **ElectionSim**: By collecting a large amount of user data from social media platforms, a diverse voter pool of millions of users is constructed, and methods such as iterative proportional fitting (IPF) are used to ensure that the sample distribution is consistent with the real - world voter distribution. - **PPE Benchmark**: A poll - based presidential election (PPE) benchmark is introduced to evaluate the accuracy of simulation results. Experiments show that the performance of this framework in the simulation of the US presidential election is very close to the actual election results, accurately matching the actual results in the predictions of 46/51 states and 12/15 swing states. Through these methods, the ElectionSim framework not only improves the simulation accuracy at the individual level but also ensures a high degree of consistency between the simulation results and the real world and provides a systematic evaluation method.

ElectionSim: Massive Population Election Simulation Powered by Large Language Model Driven Agents

Electionsim: Massive population election simulation powered by large language model driven agents

A Large-scale Empirical Study on Large Language Models for Election Prediction

Forecasting Elections with Agent-Based Modeling: Two Live Experiments.

Representation Bias in Political Sample Simulations with Large Language Models

Donald Trumps in the Virtual Polls: Simulating and Predicting Public Opinions in Surveys Using Large Language Models

Towards More Accurate US Presidential Election via Multi-step Reasoning with Large Language Models

Build An Influential Bot In Social Media Simulations With Large Language Models

GenSim: A General Social Simulation Platform with Large Language Model based Agents

Casevo: A Cognitive Agents and Social Evolution Simulator

Political Actor Agent: Simulating Legislative System for Roll Call Votes Prediction with Large Language Models

From Individual to Society: A Survey on Social Simulation Driven by Large Language Model-based Agents

S3: Social-network Simulation System with Large Language Model-Empowered Agents

Large Language Models Empowered Agent-based Modeling and Simulation: A Survey and Perspectives

Unveiling the Truth and Facilitating Change: Towards Agent-based Large-scale Social Movement Simulation

Intelligent Computing Social Modeling and Methodological Innovations in Political Science in the Era of Large Language Models

Fusing Dynamics Equation: A Social Opinions Prediction Algorithm with LLM-based Agents

Simulating Social Media Using Large Language Models to Evaluate Alternative News Feed Algorithms

Agentic Society: Merging skeleton from real world and texture from Large Language Model

LMAgent: A Large-scale Multimodal Agents Society for Multi-user Simulation