The historic center of Maastricht with the planned Sphinx Quarter in the foreground: Image: woneninhetsphinxkwartier.nl

REPORT

REPORT
“Artificial Societies” – An intensive academic course on agent-based models

REPORT

“Artificial Societies” – An Inspiring Intensive Course on Agent-Based Models

Updated:
08 Oct 2025, 13:49 | Reading time: 5 min

By Phillipp Krüger

In brief

KIWIT member Phillipp Krüger participates in an interdisciplinary ABM course at the United Nations University Maastricht
He studied physics, mathematics, and educational sciences in Göttingen and Oldenburg
Doctoral project on paradigm development in AI research

In Maastricht’s Sphinxkwartier—the repurposed site of the Sphinx ceramics factory founded in 1830—industrial history meets the present. What was once the city’s largest factory has been transformed into a cultural and innovation district, including the Sphinx Passage, whose thousands of tiles narrate the local ceramic tradition. Against this backdrop, the United Nations University pursues a guiding theme well aligned with the site: innovation as a social process.

UNU-MERIT—an institute of the United Nations University closely affiliated with Maastricht University—advances policy and innovation research, provides education, and mobilizes knowledge for inclusive and sustainable development. Innovation is not a force of nature, but a social practice. This perspective also underpinned my intensive course Agent-Based Modelling for Social Sciences: over five days, we modeled socio-economic phenomena through extensive lectures, hands-on exercises, and discussions.

Agent-based models (ABM) are rooted in cybernetics: feedback loops, control mechanisms, and simple local rules can give rise to self-organized order. The artificial-life tradition rendered this bottom-up principle tangible—from cellular automata to swarm behavior—demonstrating how complex, non-linear dynamics emerge from numerous micro-level decisions. Within this framework, scholars speak of “artificial societies”: deliberately simplified, computational models of societies in which large numbers of agents interact and social processes are investigated through simulation. A central publication venue in this field is the Journal of Artificial Societies and Social Simulation (JASSS), an interdisciplinary open-access journal that has brought together research on social processes via computer simulation since 1998.

How do agent-based models work?
(Explanatory excerpt from a video lecture by RWTH Aachen University, Human-Computer Interaction Center)

Imagine simulating the segregation behavior of a heterogeneous neighborhood: Every person follows the norm, "I am open to strangers—as long as no more than half of my immediate neighbors are completely different from me." This small, local rule unintentionally leads to the emergence of visible clusters after many moves: The neighborhood becomes more segregated, even though no one wanted segregation. This unexpected overall pattern is called emergence . ABM (Action-Based Management) shows step by step how such patterns develop—and which levers (information flows, housing options, neighborhood programs) can be used to change the dynamics.

ABMs are not numerical tricks, but rather open thought experiments : They generate knowledge about mechanisms ("If X, then Y → pattern Z"), are suitable for "what-if" analyses, and test causal plausibility beyond correlations. At the same time, they neither replace dense meaning analyses nor do they provide reliable effect sizes without data coupling. Their sweet spot lies where processes are the focus, mechanisms are disputed, and interventions are to be pre-tested.

The PhD students came predominantly from the Netherlands, Belgium, and the DACH region (Germany, Austria, and Switzerland); their academic backgrounds were equally diverse, ranging from economics and urban planning to climate modeling, innovation research, and sociology. This very mix made the exchange so valuable: we examined assumptions from different perspectives, refined models, and explored new questions. It was enriching both professionally and personally.