Fig. 1. A pyramid framework for visualizing multi-scale simulation strategies.

Fig. 2. (a) The generalized fault energy and surface energy of the Mg, Mg-Zn and Mg-Zn-R models were calculated in the DFT, and the formability and surface stability of magnesium alloys were evaluated; (b) Details of the compression twin and slip trace by MD at 300 K [52], [76].

Fig. 3. (a) Two-dimensional slices of the pure Mg NC model, the Mg-0.3Y NC model containing randomly distributed Y atoms and the Y-segregated Mg-0.3Y NC model; (b) Molecular dynamics analysis of plastic deformation of Mg-14Li-7Al alloy; (c) The atomic climb processes of <a> and <c+a> dislocations in Mg lattice and the corresponding energy evolution during <a> and <c+a> dislocation climb [81], [84], [86].

Fig. 4. (a) The corrected BKS model describes the influence of precipitated phase size on strengthening and effectively predicts the variation trend of CRSS with precipitated phase size and distribution; (b) Two-dimensional phase field model (PF-DDRX-SP) [21], [98].

Fig. 5. A linked single crystal to polycrystal to FE macro-scale modeling framework for plastic deformation of polycrystals [110].

Fig. 6. (a) An integrated computing strategy based on the weakly coupled FEM-VPSC model; (b) Diffusion activation energy curve of Al in matrix and at grain boundary; (c) Dislocation configuration (Circles 1 and 2 show the 1/6<112> dislocation network and 1/3<100> dislocation network, respectively) [130], [136].

Fig. 7. A machine learning strategy integrating target performance determination, backward and forward modeling, and feature importance analysis [139].

Fig. 8. (a) Interpretable machine learning design strategy for high-performance low-alloyed Mg alloys, and excavating high-performance low-alloyed Mg alloys; (b) Schematic diagram of the active learning design strategy workflow [140], [142].

Fig. 9. By integrating DFT computational data points and machine learning methods, a widely applicable DLP was created to study the dislocation behavior of Mg [41].