Skip to main content Skip to main navigation



Social Simulation for Analysis of Infectious Disease Control

  • Duration:

The current COVID 19 pandemic is a major challenge for society and the healthcare system. The behavior of people is very much dependent on their environment (emergence effect), as shown, for example, by the purchases of toilet paper at the beginning of the pandemics in early 2020. Here, social contagion has taken place that can be explained by means of social mechanisms. However, means of statistics are limited for analysis of emergent effects. Unlike other modeling and analysis approaches, AI as well as agent-based modeling and simulation generates macro-level behavior through the local interaction of micro-level actors. The Cognitive Social Simulation group at DFKI has extensive experience in agent-based modeling and social simulation, which are also suitable for simulations of escalating infectious diseases. Based on an influenza model (Timm and Lasner, 2013), we have developed a new simulation model for COVID-19.

As a first comparison, the simulation model was used to compare spread and impact of COVID-19 with those of common influenza. As an initial result, influenca without any measures taken shows comparable virus spread as CODIV-19 with the measures taken in early 2020: contact restriction, home-office, and school-closing. Our simulation model allows for activation and deactivation of private and political measures for containment of the disease. Doing so, "social distancing"-measures and their effectiveness can be tested in the simulation (strictly speaking, however, this is rather "physical distancing", since the measures are intended to achieve a spatial separation of people, but not a complete break-off of social contacts). With SoSAD, we are contributing to the evaluation of measures, especially on regional level. The structure of the model is formed by agents representing e.g. people, households, schools, hospitals, leisure and work places. Connections between the agents are possible points of contact. Aspects / parameters considered so far were all researched from available sources or derived from published data. In particular, information and data from Imperial College in London, the European Centre for Disease Control (ECDC) and the Robert Koch Institute were used as sources.


Seit dem Sommer besteht eine Kooperation mit dem Fraunhofer-Institut für Techno- und Wirtschaftsmathematik in Kaiserslautern (ITWM) und deren System EpideMSE Die Stadt Kaiserslautern hat am 11.9.2020 die Zusammenarbeit im COVID-19-Pandemiemanagement mit dem DFKI und dem ITWM bekannt gegeben (SWR Beitrag)

Since summer there is a cooperation with the Fraunhofer Institute for Industrial Mathematics in Kaiserslautern (ITWM) and their system EpideMSE The City of Kaiserslautern announced the cooperation in COVID-19 pandemic management with DFKI and ITWM on 9/11/2020 (SWR Beitrag)


DFKI Eigenprojekt


Publications about the project

Lukas Tapp; Veronika Kurchyna; Falco Nogatz; Jan Ole Berndt; Ingo Timm

In: Multi-Agent-Based Simulation XXIII. International Workshop on Multi-Agent Systems and Agent-Based Simulation (MABS-2022), located at AAMAS 2022, May 8-9, Auckland, New Zealand, Pages 95-106, Lecture Notes in Artificial Intelligence (LNAI), Springer, 2023.

To the publication

Ingo J. Timm; Jan Ole Berndt

Whitepaper, Deutsches Forschungszentrum für Künstliche Intelligenz (DFKI), Whitepaper, Vol. 1, 8/2020.

To the publication