“…The surrogate model-based optimization couples the optimization algorithms with surrogate models to ascertain optimal design parameters, offering the benefits of the simple model and high efficiency. Various surrogate models, such as response surface methodology (RSM) models [ 7 , 8 ] artificial neural network model (ANN) [ 9 ], and Kriging model [ 10 ], have been employed to depict the relationship between design parameters and objective function. On the other hand, the physical model-based optimization integrates the optimization algorithms with finite element models, eliminating the approximation errors inherent in surrogate models.…”