【摘要】 针对汽车前端结构优化设计中存在代理模型拟合精度低、多目标优化性能不理想的问题，本文提出了一种基于动态集成模型和新型多目标浣熊优化算法的汽车前端结构优化方法。首先，构建汽车前端结构参数化模型，定义结构性能指标和设计变量，通过有限元仿真分析获取数据。然后，提出自适应过滤机制和多维加权组合策略，构建动态集成模型。接着，基于仿真数据，建立结构性能与设计变量之间的高精度代理模型。最后，通过重构种群初始化策略、优化过程和更新机制，提出一种全新的多目标浣熊优化算法，并对汽车前端结构进行多目标优化设计。优化结果表明，优化后的结构在各性能上显著优于基准模型，且各性能指标的平均提升率超10%。本文所提出的优化方法能有效提高汽车前端结构的耐撞性，并为汽车结构领域的智能设计与优化提供理论参考。更多还原

【Abstract】 For the problems of low fitting accuracy in surrogate models and suboptimal performance in multi-objective optimization for automotive front-end structure design, an optimization approach based on a dynamic ensemble model and a novel multi-objective coati optimization algorithm is proposed in this paper. Firstly, a parameterized model of the automotive front-end structure is established, defining structural performance metrics and design variables, with data obtained through finite element simulation analysis. Next, a dynamic ensemble model is constructed based on an adaptive filtering mechanism and a multidimensional weighted combination strategy. Subsequently, a high-accuracy surrogate model is developed based on simulation data to map the relationship between structural performance and design variables. Finally, a novel multi-objective coati optimization algorithm is introduced by reconstructing population initialization strategies, optimization processes, and update mechanism, enabling multi-objective optimization of the front-end structure. The optimization results show that the optimized structure significantly outperforms the baseline model, with an average improvement rate exceeding 10% across all performance metrics. The proposed method effectively enhances the crashworthiness of automotive front-end structures and provides a theoretical foundation for intelligent design and optimization in the field of automotive structural engineering.更多还原