New paper in Structural and Multidisciplinary Optimization journal

© 2023 EPFL
The formulations we developed for our structural matchmaking algorithms also perform very well against classical optimization formulations. In this recent paper published in the journal of Structural and Multidisciplinary Optimization, we compare the formulations developed by Jan Brütting during his PhD at SXL, with two state-of-the-art formulations: Rasmussen and Stolpe (2008), and Ghattas and Grossmann (1991).
It's all about Mixed-Integer Linear Programming (MILP), the way the problem is written through (in-)equations, the number of variables in the problem, and the speed needed to converge to the optimal solution. Transforming a truss topology optimization problem (i.e. one that selects minimum-weight cross-sections from a finite set of predetermined ones) into a linear problem is not trivial. Large efforts have been made to make this paper as educational as possible. We highly recommend it to beginners in the field.
Jan Brütting, Gennaro Senatore, Corentin Fivet (2022) "MILP‑based discrete sizing and topology optimization of truss structures: new formulation and benchmarking", Structural and Multidisciplinary Optimization, Springer (2022) 65:277. DOI: 10.1007/s00158-022-03325-7
Link to Open Access paper: https://link.springer.com/article/10.1007/s00158-022-03325-7