Social Cognition Simulation with Large Language Model-driven Agents

Authors

Mingxin Zhang, College of Politics, National Defense University, Shanghai 200433, China
Jinxuan Wu, College of Computer Science and Artificial Intelligence, Fudan University, Shanghai 200438, China
Rui Zhu, College of Computer Science and Artificial Intelligence, Fudan University, Shanghai 200438, China
Yunlong Wang, College of Politics, National Defense University, Shanghai 200433, China
Wenjuan Meng, College of Computer Science and Artificial Intelligence, Fudan University, Shanghai 200438, China
Zhe Liu, College of Computer Science and Artificial Intelligence, Fudan University, Shanghai 200438, China
Xu Li, College of Computer Science and Artificial Intelligence, Fudan University, Shanghai 200438, China
Xiaolei Chen, College of Computer Science and Artificial Intelligence, Fudan University, Shanghai 200438, China
Yuxuan Liang, College of Computer Science and Artificial Intelligence, Fudan University, Shanghai 200438, China
Yi Zheng, College of Computer Science and Artificial Intelligence, Fudan University, Shanghai 200438, China
Xiangyang Xue, College of Computer Science and Artificial Intelligence, Fudan University, Shanghai 200438, China

Abstract

Abstract: With the continuous evolution of the capabilities of generative LLMs, their application in social cognition simulation is demonstrating paradigm-shifting potential. Traditional social simulation methods predominantly rely on static rules and simplified behavioral models, making it difficult to capture the dynamic evolution and cultural complexity of human social behavior. LLM-driven agents, equipped with contextual understanding and natural language generation capabilities, are emerging as novel tools for modeling social cognitive mechanisms, enabling the simulation of complex sociopsychological processes such as identity construction, value judgment, and intentional reasoning. This paper briefly introduced the technical foundations of LLMs and highlighted their suitability for social cognition simulation. It constructed a framework for agent-based social cognition modeling, encompassing attribute modeling, memory management, planning, and action. At the simulation process level, the paper proposed a technical pipeline consisting of “data collection, agent collaboration, and multidimensional evaluation,” while delving into challenges such as cognitive interpretability and simulation-reality alignment. It summarized the current application progress in fields such as sociology, economics, and military science and discussed emerging trends and future directions for LLM-based social cognition simulation.

Recommended Citation

Zhang, Mingxin; Wu, Jinxuan; Zhu, Rui; Wang, Yunlong; Meng, Wenjuan; Liu, Zhe; Li, Xu; Chen, Xiaolei; Liang, Yuxuan; Zheng, Yi; and Xue, Xiangyang (2026) "Social Cognition Simulation with Large Language Model-driven Agents," Journal of System Simulation: Vol. 38: Iss. 2, Article 2.
DOI: 10.16182/j.issn1004731x.joss.25-0612
Available at: https://dc-china-simulation.researchcommons.org/journal/vol38/iss2/2

First Page

261

Last Page

277

CLC

TP391.9

Recommended Citation

Zhang Mingxin, Wu Jinxuan, Zhu Rui, et al. Social Cognition Simulation with Large Language Modeldriven Agents[J]. Journal of System Simulation, 2026, 38(2): 261-277.

Corresponding Author

Wu Jinxuan

DOI

10.16182/j.issn1004731x.joss.25-0612