Objectives  To address the high computational cost and the difficulty of achieving global optimization using conventional methods in complex, high-dimensional full-hull design problems (e.g., a 30-dimensional design space), this study proposes a surrogate-model-based optimization method for accurate prediction and efficient optimization of ship resistance performance, providing a practical technical solution for high-dimensional hull form optimization.

Methods The Free-Form Deformation (FFD) technology was employed to parameterize the hull geometry with 30 design variables, enabling high-dimensional, high-degree-of-freedom shape representation. Latin Hypercube Sampling was used to generate the design sample set, with resistance data obtained through CFD simulation. A high-fidelity Kriging surrogate model was constructed to replace the high-cost CFD simulation. A hybrid strategy combining a Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) was developed to perform global optimization within the surrogate model space. In addition, sensitivity analysis was conducted to identify the key design variables affecting hull resistance.

Results The optimization results based on the KCS container ship model show that the prediction error of the constructed Kriging surrogate model was as low as 0.35%. The resistance coefficient of the optimized hull form was reduced by 8.52% compared with the baseline hull. Sensitivity analysis indicates that design variables associated with the bulbous bow region had a significant influence on hull resistance and constitute the primary influencing factors.

Conclusions This study achieves full-process, closed-loop verification of FFD parameterization, the Kriging surrogate model, and evolutionary optimization within a high-dimensional full-hull design framework. The proposed method greatly improves both computational efficiency and optimization effectiveness in high-dimensional hull form optimization. The demonstrated resistance reduction and the identification of key design variables provide clear engineering guidance for energy-efficient ship design, and offer a reproducible paradigm for applying mature methodologies to complex engineering optimization problems.