Generative Agent-Based Models for Complex Systems Research: a review

Abstract: The advent of LLMs has significantly transformed the fields of natural and social sciences. Generative Agent-Based Models (GABMs), which utilize LLMs in place of real subjects, are gaining increasing public attention. Far from aiming for comprehensiveness, this paper aims to offer readers an opportunity to understand how LLMs are disrupting complex systems research and behavioral sciences. In particular, we evaluate recent advancements in various domains within complex systems, encompassing network science, evolutionary game theory, social dynamics, and epidemic propagation. Additionally, we propose possible directions for future research to further advance these fields.