Artificial intelligence assisted multiobjective design of functionally graded materials using hybrid optimization and experimental validation

Designing functionally graded materials (FGMs) through conventional trial-and-error approaches is time-intensive and difficult to scale for large multi-material design spaces. This study presents an AI-assisted, computational-intelligence-based framework for compatibility-aware material selection and gradient-profile optimization of FGM structures. The framework combines a curated database of 180 candidate materials with a mixed-variable hybrid genetic algorithm–particle swarm optimization approach to balance weight, cost, stiffness, strength, and thermal resistance. Rapid objective-function screening was followed by finite element verification of shortlisted designs and proof-of-concept experimental validation using powder-metallurgy-fabricated Al₂O₃/Ti-6Al-4V FGM beam specimens. For the representative beam case considered, the optimized FGM showed approximately 25% lower weight, 30% lower deflection, corresponding to approximately 43% higher effective stiffness under identical loading, and 60% lower peak thermal stress compared with the corresponding homogeneous reference under the same modeling assumptions. Experimental measurements from three specimens showed differences of approximately 3% for elastic modulus and 4% for ultimate strength relative to numerical predictions, supporting the feasibility of the workflow for the present case while indicating the need for broader statistical validation. The estimated per-design screening and verification time decreased from approximately 480 h for a conventional trial-and-error route to approximately 18 h after database preparation, although this comparison is case-specific and depends on model fidelity, hardware, and available material data. Overall, the results suggest that AI-assisted multi-objective optimization can serve as a useful decision-support tool for narrowing FGM design spaces before detailed numerical and experimental qualification.