Key points are not available for this paper at this time.
Biologically inspired staggered composites are renowned for their exceptional mechanical properties, making them highly promising for applications in aerospace and rail transit. This creates a critical need to address the inverse design problem—that is the challenge of determining the optimal microstructure to achieve a desired macroscopic performance. This study develops an inverse design framework by integrating machine learning with genetic algorithms. First, through finite element method simulations, we construct a comprehensive database, which maps the microstructural parameters to the macroscopic elastoplastic responses. Leveraging this dataset, a high-fidelity machine learning surrogate model is then developed to rapidly capture the complex structure–property relationships. This model is subsequently coupled with genetic algorithms to efficiently navigate the vast design space and identify optimal configurations that match a target stress–strain curve. The mechanical responses of the newly designed composites exhibit excellent agreement with the target curves, validating the framework’s efficacy. Our work not only provides an accurate and efficient tool for the inverse design of staggered composites, but also establishes a generalizable strategy for the tailored design of a broad class of advanced functional materials.
Wu et al. (Tue,) studied this question.