传统船舶复杂曲面设计依赖经验迭代，存在参数化程度低、多目标优化冲突难平衡的问题，本文首先分析了传统DE算法的原理，提出融合主成分分析（PCA）降维与分层搜索的策略，构建嵌入NSGA-Ⅱ非支配排序机制及AHP-熵权法自适应权重的多目标DE算法，以船艏NURBS曲面为优化对象，将改进DE算法与传统DE算法、BP神经网络优化算法进行仿真对比，研究结果表明，改进DE算法的兴波阻力较传统DE算法降低11.5 %；初稳性高度较传统DE算法提升9.8 %，制造复杂度降至4.8 ，且收敛速度与多目标分布均匀性均优于对比算法，有效提升船舶曲面优化精度与效率。
The design of complex surfaces for traditional ships relies on empirical iteration, which has problems such as low parameterization degree and difficulty in balancing multi-objective optimization conflicts. This paper first analyzes the principle of the traditional DE algorithm and proposes a strategy that integrates principal component analysis (PCA) dimensionality reduction and hierarchical search. A multi-objective DE algorithm embedded with the NSGA-II non-dominated sorting mechanism and the adaptive weight of the AHP-entropy weight method was constructed. Taking the bow NURBS surface as the optimization object, the improved DE algorithm was simulated and compared with the traditional DE algorithm and the BP neural network optimization algorithm. The research results show that the ripple resistance of the improved DE algorithm is reduced by 11.5% compared with the traditional DE algorithm. The initial stability height is increased by 9.8% compared with the traditional DE algorithm, the manufacturing complexity is reduced to 4.8, and both the convergence speed and the uniformity of multi-objective distribution are superior to the comparison algorithms, effectively improving the accuracy and efficiency of ship surface optimization.
2025,47(21): 48-52 收稿日期：2025-5-24
DOI：10.3404/j.issn.1672-7649.2025.21.009
分类号：U622
作者简介：吴铭川(2005-),男,研究方向为数学建模及微分方程、船舶曲面设计及应用
参考文献：
[1] 杨维. 基于计算机辅助的船舶复杂曲面结构设计[J]. 舰船科学技术, 2024, 46(13): 67-70
YANG W. Computer-aided structural design of marine complex surfaces[J]. Ship Science and Technology, 2024, 46(13): 67-70
[2] 钱强. 基于运动学约束的5轴加工复杂曲面刀矢优化探究[J]. 模具制造, 2024, 24(12): 152-154
QIAN Q. Research on tool vector optimization for 5-axis machining of complex surfaces based on kinematic constraints[J]. Die & Mould Manufacturing, 2024, 24(12): 152-154
[3] 童疏影. 基于3D打印技术的复杂曲面结构设计与优化方法[J]. 中国机械, 2023(25): 7-10
TONG S Y. Structural design and optimization method of complex surfaces based on 3d printing technology[J]. China Machinery, 2023(25): 7-10
[4] 喻宏伟, 关新, 阮鹏, 等. 基于改进微分进化算法的微网多目标优化的研究[J]. 电气时代, 2019(9): 51-53
YU H W, GUAN X, RUAN P, et al. Research on multi-objective optimization of microgrid based on improved differential evolution algorithm[J]. Electric Age, 2019(9): 51-53
[5] 姚玮, 罗建军, 马卫华, 等. 一种改进的微分进化算法及其在多目标访问任务中的应用[J]. 宇航学报, 2017, 38(9): 919-926
YAO W, LUO J J, MA W H, et al. An improved differential evolution algorithm and its application in multi-objective visiting tasks[J]. Journal of Astronautics, 2017, 38(9): 919-926
[6] 何小龙, 白俊强, 李宇飞. 基于全局灵敏度分析的改进微分进化算法[J]. 西北工业大学学报, 2016, 34(3): 411-417
HE X L, BAI J Q, LI Y F. Improved differential evolution algorithm based on global sensitivity analysis[J]. Journal of Northwestern Polytechnical University, 2016, 34(3): 411-417
