The field of artificial intelligence (AI) offers transformative potential for the study and application of multibody system dynamics, a research area closely related to computational methods. This minisymposium aims to explore the synergistic intersection between machine learning (ML) techniques and multibody system dynamics to strengthen the development of predictive and prescriptive models, optimization techniques, and efficient computational strategies. Historically, the use of multibody systems in reinforcement learning serves as a practical framework for training AI agents, expanding the scope from robotics to complex mechanical systems. By using AI and ML, researchers can accurately predict the behavior of dynamical systems [1], enabling improved simulations and robust design methods. Generative AI is used to create multibody system models [2], and upcoming AI approaches start to be used to provide decisions and to prescribe tasks. The scope of the minisymposium also includes the development of reduced-order models that utilize data-driven techniques to simplify complex systems while retaining essential dynamics. The integration of AI into multibody systems promises not only to improve our understanding of kinematics or dynamic behavior, but also to develop new approaches to system design and control. Through extensive discussions and the exchange of novel research results, this minisymposium aims to enable collaboration and drive innovation for theoretical and applied aspects of multibody dynamics and machine learning.

[1] Slimak T., Zwölfer A., Todorov B., Rixen D.J., Overview of design considerations for data-driven time-stepping schemes applied to nonlinear mechanical systems, J. Computational and Nonlinear Dynamics, Vol. 19(7): 071012, 2024.
[2] Gerstmayr J., Manzl P., Pieber M., Multibody models generated from natural language, Multibody System Dynamics, 2024.