AI-guided design optimization of core–shell BCC lattices to achieve ultrahigh stiffness via multi-material 3D printing

Lattice structures have gained prominence as mechanical metamaterials capable of being lightweight, achieving high specific stiffness, and exhibiting specific energy-absorbing characteristics through geometry-driven design. Furthermore, the advent of additive manufacturing has enabled the realization of complex geometries and multi-material integration. Specifically, the mechanical performance of core–shell lattice architectures can be enhanced by independently tuning the inner core and outer shell components. However, with studies focusing on single-material systems or simple beam geometries, comprehensive frameworks that couple beam shape and material arrangement in an integrated design process remain limited. This study introduces an inverse design methodology wherein the shapes of the core and shell beams are independently parameterized via Bézier curves and mapped onto body-centered cubic lattice structures. Finite-element analysis and homogenization are performed to evaluate the stress distributions and effective stiffnesses of the lattice structures. To explore the high-dimensional design space efficiently, a neural network–boosted genetic algorithm–based optimization framework is employed. Experimental validation is conducted by fabricating the optimized lattice specimens using PolyJet multi-material 3D printing and conduct uniaxial compression tests. The experimental results closely match the simulated data, and performance comparisons are made under identical volume-fraction constraints. Under the controlled condition, the optimized design exhibits substantial improvements in stiffness, yield strength, and energy absorption relative to a baseline cylindrical core–shell design, confirming the effectiveness of the proposed optimization strategy. The proposed framework thus offers an integrated pipeline for the design, simulation, and realization of high-performance multi-material lattice structures.