Economic Simulations in Swarm: 20 Years On

At the start of this century, Francesco Luna and Benedikt Stefansson published the first edited volume on agent-based computational economics. By addressing three critical aspects of agent-based modeling (ABM) — “Esperanto”, computability, and object-oriented programming — this volume established a landmark vision for the future development of agent-based modeling in the 21st century. In this paper, we trace the evolution of these three foundational concepts over the past two decades and present an overview of three key developments in ABM since then. First, NetLogo has emerged as the “English” of ABM, replacing Swarm, while the pursuit of an “Esperanto” for ABM remains elusive. Second, agent-based computational economics has largely avoided concerns of uncomputability, a critical issue that continues to challenge mainstream analytical economics. Third, the rise of object-oriented programming as an alternative to procedurals and declarative programming, alongside ABM’s emergence as an alternative to equation-based modeling, reflects a deeper shared pursuit: the conceptualization of autonomous agents. This focus on autonomous agents serves as a bridge of consilience between C.P. Snow’s “two cultures”, uniting the sciences and the humanities.