Artificial intelligence-optimized gypsum-based composites for stealth technology: Integrating high strength performance and radar absorption

The demand for lightweight, high-performance materials with electromagnetic shielding capability is increasing, particularly in aerospace, defense, and infrastructure applications. This study introduces a novel gypsum-based composite system enhanced with carbon black (CB), magnetite (Fe3O4), boron nitride (BN), and hybrid glass-basalt fibers to simultaneously improve mechanical strength and electromagnetic wave absorption. A limited experimental dataset of 13 formulations was extended to 260 using physics-consistent data augmentation. Multi-objective optimization was then performed using the Artificial Bee Colony (ABC) algorithm to maximize compressive, flexural, and splitting tensile strengths while minimizing return loss in the 8–12 GHz (X-band) range. The optimized formulation (85% gypsum, 15% resin, 6.8% CB, 18.5% Fe3O4, 2% BN, 1% glass fiber, 0.5% basalt fiber) achieved 34.41 MPa compressive strength (+6.1%), 7.71 MPa flexural strength (+5.3%), and 4.45 MPa splitting tensile strength (+5.7%). Simultaneously, the minimum return loss improved from −22.5 to −24.8 dB, resulting in an enhancement of electromagnetic absorption by 10.2% at 9.72 GHz. Validation tests confirmed the model's accuracy within a 1.35% margin of error. This hybrid methodology, which combines experimental science, AI-driven data expansion, and bio-inspired optimization, offers a scalable and cost-effective route to multifunctional composite design. The findings are directly relevant for radar-absorbing structures, EMI shielding in electronics, and smart civil infrastructure.