A novel multi-objective shielding optimization method: DNN-PCA-NSGA-Ⅱ

•We have proposed a novel multi-objective shielding optimization method which couples deep neural network (DNN), PCA and NSGA-Ⅱ.•The proposed method can effectively remove redundant objectives in optimization calculation, so as to reduce the bad influence of the redundant objective on the optimization results.•The proposed method has shown its feasibility in solving many-objectives shielding optimization problems with redundant objectives.•The proposed method has shown its applicability in engineering optimization problems (i.e. Savannah marine nuclear power reactor shielding model).