Lightweight structures are essential for reducing carbon dioxide emissions because lower structural mass directly decreases fuel consumption. Recent advancements in additive manufacturing have enabled the production of components with complex geometries, including Voronoi-type structures, which are commonly used to reduce weight in automotive and aerospace applications. Although often described as bioinspired, such structures are frequently generated using the Poisson point process, which involves a high level of randomness. Since tree leaves have evolved over time and possess important mechanical properties that enable them to withstand environmental conditions, this paper proposes a new heuristic, called BOTH, based on Bauhinia forficata . This heuristic initializes Voronoi structures with an emphasis on enhancing structural stiffness. To analyze leaf morphology and evaluate the mechanical behavior of the resulting structures, computer vision, machine learning algorithms, and finite element simulations were utilized. Validation of the heuristic showed that structures initialized with BOTH exhibit higher stiffness and less variability than those generated using the Poisson point process, with the minimum stiffness increasing from 14.85 to 176.12 kN/mm, corresponding to more than an elevenfold improvement in allowable design stress.
Almeida et al. (2026) studied this question.