EUVIMEDEUVIMED entdecken
Crossref

Towards Differentiable Agent-Based Simulation

Philipp Andelfinger

ACM Transactions on Modeling and Computer Simulation · 31.10.2022 · 32 (4) · 1-26

DOI: 10.1145/3565810 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ACM Transactions on Modeling and Computer Simulation, online erschienen am 11.01.2023. Verfasst von Philipp Andelfinger. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 83 Literaturangaben. Gefördert von: Deutsche Forschungsgemeinschaft.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Simulation-based optimization using agent-based models is typically carried out under the assumption that the gradient describing the sensitivity of the simulation output to the input cannot be evaluated directly. To still apply gradient-based optimization methods, which efficiently steer the optimization towards a local optimum, gradient estimation methods can be employed. However, many simulation runs are needed to obtain accurate estimates if the input dimension is large. Automatic differentiation (AD) is a family of techniques to compute gradients of general programs directly. Here, we explore the use of AD in the context of time-driven agent-based simulations. By substituting common discrete model elements such as conditional branching with smooth approximations, we obtain gradient information across discontinuities in the model logic. On the examples of a synthetic grid-based model, an epidemics model, and a microscopic traffic model, we study the fidelity and overhead of the differentiable simulations as well as the convergence speed and solution quality achieved by gradient-based optimization compared with gradient-free methods. In traffic signal timing optimization problems with high input dimension, the gradient-based methods exhibit substantially superior performance. A further increase in optimization progress is achieved by combining gradient-free and gradient-based methods. We demonstrate that the approach enables gradient-based training of neural network-controlled simulation entities embedded in the model logic. Finally, we show that the performance overhead of differentiable agent-based simulations can be reduced substantially by exploiting sparsity in the model logic.

Förderung

  • Deutsche Forschungsgemeinschaft

Zitieren

APA

Andelfinger, P. (2022). Towards Differentiable Agent-Based Simulation. ACM Transactions on Modeling and Computer Simulation, 32(4), 1-26. https://doi.org/10.1145/3565810

Vancouver

Andelfinger P. Towards Differentiable Agent-Based Simulation. ACM Transactions on Modeling and Computer Simulation. 2022;32(4):1-26. doi:10.1145/3565810

BibTeX
@article{Andelfinger2022,
  author = {Andelfinger, Philipp},
  title = {Towards Differentiable Agent-Based Simulation},
  journal = {ACM Transactions on Modeling and Computer Simulation},
  year = {2022},
  volume = {32},
  number = {4},
  pages = {1--26},
  doi = {10.1145/3565810}
}
Kostenlos starten →