“…Surrogate-assisted evolutionary algorithms (SAEAs) have been proposed to solve this issue and use a computationally inexpensive surrogate model to replace expensive EC 39,7 high-fidelity simulations for fitness evaluations in evolutionary algorithms (Wang et al, 2021;Alizadeh et al, 2020;Koziel and Bekasiewicz, 2020). To this end, various surrogate models are used in SAEAs, including the response surface method (RSM) (Zhang et al, 2017), neural networks (NNs) (Carneiro et al, 2019), the radial basis function (RBF) (Jing et al, 2019), support vector machines (SVMs) (Tao et al, 2018) and the Kriging model (You et al, 2021;Du and Leifsson, 2020), among others. Ensemble surrogates, multiple surrogates and adaptive surrogates have also been implemented to enhance the accuracy of fitness approximation (Garbo and German, 2019).…”