Computational strategies for the design of high-strength and high-thermal-conductivity casting Mg/Al alloys

High-strength and high-thermal-conductivity Mg/Al alloys are pivotal for lightweight thermal management applications in aerospace and electric vehicles, yet their development is hindered by the intrinsic trade-off between solute strengthening and electron scattering. This review summarizes the evolution of alloy design approaches from empirical trial-and-error methods to advanced artificial intelligence (AI)-driven strategies. First, the physical mechanisms governing strength and thermal conductivity (TC) are examined, identifying “matrix purification” and specific precipitation architectures as key microstructural design goals to minimize solute scattering. Then, the review critically evaluates traditional computational tools, including CALPHAD (CALculation of PHAse Diagrams) for phase equilibrium prediction and density functional theory for intrinsic transport predictions. Crucially, we demonstrate that integrating these physics-based descriptors as input features into machine learning models significantly enhances prediction accuracy for complex multicomponent systems. Subsequently, the integration of these approaches for the concurrent optimization of strength and TC is discussed, highlighting the role of multi-objective optimization algorithms in mapping the Pareto frontier. Finally, the review discusses the future potential of emerging frontiers in Generative AI, such as generative adversarial networks (GANs), variational autoencoders (VAEs), and inverse design, highlighting the critical role of expert knowledge-guided constraints and physics-informed priors in steering model training and generation. Such hybrid frameworks are envisioned to autonomously navigate high-dimensional compositional spaces while maintaining physical interpretability and thermodynamic consistency, thereby accelerating the discovery of next-generation multifunctional alloys.