EF45 – Multi-Agent Social Systems



Agent-Based Models of SARS-CoV2 Transmission: Multilevel Identification and Network-Based Reduction

Project Heads

Nataša Djurdjevac Conrad, Martin Weiser, Sarah Wolf, Edda Klipp

Project Members

Jan-Hendrik Niemann (FU Berlin), Björn Goldenbogen (HU Berlin)

Project Duration

01.01.2022 – 31.12.2023

Located at

FU Berlin, HU Berlin


Epidemiological agent-based models (ABMs) such as the GEospatially Referenced Demographic ABM (GERDA) [1] can simulate and forecast infection spreading in much detail and, if parametrized correctly, with high accuracy. They incur, however, a huge computational effort and a large set of uncertain parameters. We develop multilevel representations of ABMs using network-based coarsening and model identification techniques in order to improve the understanding of pattern formation and to accelerate parameter identification and policy optimization.


Efficient multilevel optimization requires suitable coarse models. Obvious candidates such as the established SIR models are often not rich enough to reflect spatial or subpopulation effects and related parameters such as in GERDA. We find coarse models providing a good trade-off between complexity, approximation quality, and extrapolation power via network-based coarsening and model identification techniques, see, e.g., [3, 4]. Their effectivity is evaluated in multilevel optimization and compared with available coarse-grained and SIR-type models.


Another essential step towards nonlinear multilevel parameter identification and policy optimization is the efficient computation of derivatives with respect to model parameters. Since straightforward numerical differentiation of (stochastic and discontinuous) ABMs is infeasibly expensive due to their computational complexity and variance, we study the combined use of hierarchical Monte Carlo methods and adjoint concepts for efficient gradient computation. Figure 1 A shows the trajectories for the fraction of susceptible, infected and recovered agents in the GERDA model, divided between adults and children. Figure 1 B shows the optimal policy computed using the multilevel optimization algorithm in [1]. The computational cost of this algorithm is compared to the cost of a state-of-the-art inexact gradient descent algorithm in Figure 1 C. The multilevel approach is initially superior as it approaches the optimum faster with less ABM simulations needed.


In a complementary approach we are aiming for a coarse grained ABM that reflects the key characteristics of the original model but is reduced with respect to time steps or agent number.  Based on the temporal interaction and infection networks spanned by the original ABM, we analyze the involvement and degrees of connection of population subgroups and search for dominant infection patterns. Further we develop methods for compression of the undirected interaction graphs and the directed infection graphs  retaining the relevant information, to eventually establish algorithms to coarse grain GERDA .

Project Webpages

Selected Publications

  1. B. Goldenbogen, S.O. Adler, O. Bodeit, J. AH Wodke, X. Escalera-Fanjul, A. Korman, M. Krantz, L. Bonn, R. Morán-Torres, J.E.L. Haffner, M. Karnetzki, I. Maintz, L. Mallis, H. Prawitz, P.S. Segelitz, M. Seeger, R. Linding and E. Klipp. Adaptive combination of interventions required to reach population immunity due to stochastic community dynamics and limited vaccination
  2. J.-H. Niemann, S. Uram, S. Wolf, N. Djurdjevac Conrad, M. Weiser. Multilevel Optimization for Policy Design with Agent-Based Epidemic Models. J. Comput. Sci., 102242, 2024. (https://doi.org/10.1016/j.jocs.2024.102242)
  3. S. Klus, F. Nüske, S. Peitz, J.-H. Niemann, C. Clementi, and C. Schütte. Data-driven approximation of the Koopman generator: Model reduction, system identification, and control. Physica D: Nonlinear Phenomena, 406:132416, 2020. (Physica D)
  4. J.-H. Niemann, S. Klus, and C. Schütte. Data-driven model reduction of agent-based systems using the Koopman generator. Plos One, 16 (5), 2020. (Plos One)
  5. J.-H. Niemann, S. Klus, N. Djurdjevac Conrad, and C. Schütte. Koopman-Based Surrogate Models for Multi-Objective Optimization of Agent-Based Systems, 2023. (under review)

Selected Pictures

Figure 1

A: Trajectories of susceptible (blue), infected (red) and recovered (green) adults (solid) and children (dashed) computed using the GERDA model with the optimal control/policy shown in B.

B: Optimal control/policy for the GERDA model computed using the multilevel algorithm. A value of 0 corresponds to no home schooling resp. home office, while a value of 1 means the opposite, i.e., neither presence at work nor at school.

C: Computational cost in terms of ABM simulations of the multilevel algorithm compared to a state-of-the-art inexact gradient descent algorithm. The multilevel approach is superior in the initial phase as it (i) captures the nonlinear structure better the a first- order second-order Taylor model, and (ii) serves as a preconditioner and thus reduces “zig-zagging”.

Please insert any kind of pictures (photos, diagramms, simulations, graphics) related to the project in the above right field (Image with Text), by choosing the green plus image on top of the text editor. (You will be directed to the media library where you can add new files.)
(We need pictures for a lot of purposes in different contexts, like posters, scientific reports, flyers, website,…
Please upload pictures that might be just nice to look at, illustrate, explain or summarize your work.)

As Title in the above form please add a copyright.

And please give a short description of the picture and the context in the above textbox.

Don’t forget to press the “Save changes” button at the bottom of the box.

If you want to add more pictures, please use the “clone”-button at the right top of the above grey box.