The design domain has significant influence on the result of topology optimization. When several components are to be topology optimized separately, the allocation of the available space to each of them will be crucial for the resulting mechanical performance of the system [1, 2, 3]. Unfortunately, the optimal allocation depends on the results of component optimizations which are not known initially. This chicken-or-egg problem was solved in [2] by conducting several component topology optimization problems for varying design domains as preparation and making the relationship between resulting component mass and design space available as sample data. Based on this sample data, good design domains could be chosen manually. Unfortunately, this technique is limited to very simple design problems. To extend it to more complex design problems and to automate the approach, this paper proposes the following alternative: For a sample of varying design domains, topology optimizations are performed for each component. Then, based on these data, meta models are trained to estimate (1) the feasibility and (2) the mass of individual components, as a function of the dimensions of the allocated design domain. The actual design domains are then allocated by numerical optimization using these meta models. Final geometries are determined by detailed topology optimization on these design domains. The method is applied to the steering mechanism of a glider plane. A reduction in mass by 17%, compared to a manual distribution of design domains could be achieved. REFERENCES [1] E. Tyflopoulos and M. Steinert, Messing with boundaries - quantifying the potential loss by pre-set parameters in topology optimization. Procedia CIRP , Vol. 84, 979- 985, 2019. https://doi.org/10.1016/j.procir.2019.04.307 [2] F. Endress, T. Kipouros, and M. Zimmermann, Distributing Design Domains for Topology Optimization in Systems Design, Proceedings of the ASME 2023 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference. Volume 2: 43rd Computers and Information in Engineering Conference (CIE). Boston, Massachusetts, USA. August 20–23, 2023. V002T02A079. ASME. https://doi.org/10.1115/DETC2023-1148832 [3] L. Krischer, M. Zimmermann, Decomposition and optimization of linear structures using meta models. Struct Multidisc Optim, Vol. 64, 2393–2407, 2021. https://doi.org/10.1007/s00158-021-02993-1