When designing a sport utility vehicle (SUV), designers strive to improve the vehicle’s rollover crashworthiness while avoiding a significant increase in its weight. To aid in optimizing such a trade-off, this paper proposes a multi-disciplinary and multi-objective hybrid optimization algorithm that combines particle swarm optimization and the artificial immune method. First, the SUV structure’s influence on body mass and rollover crashworthiness is studied using contribution analysis, and structural improvements are discussed according to Federal Motor Vehicle Safety Standard 216. Building on the analysis results, the SUV’s rollover crashworthiness and weight optimization model are proposed. Radial basis function neural network and a genetic algorithm are used to build and optimize surrogate models of total weight, maximum contact force, and torsion frequency. The proposed algorithm then utilizes particle swarm and artificial immune to seek Pareto solutions that optimize SUV structure. Finally, the technique for order preference by similarity to ideal solution method determines a final solution from Pareto-optimal solutions. Compared to previous studies, the results show that the proposed hybrid optimization algorithm improves the Pareto solution sets’ diversity and distribution uniformity, enhances SUV rollover crashworthiness, and reduces SUV structure components’ weight.

在设计运动型多用途车（SUV）时，设计人员努力提高车辆的侧翻防撞性，同时避免其重量显着增加。为了帮助优化这种折衷，本文提出了一种结合粒子群优化和人工免疫方法的多学科，多目标混合优化算法。首先，通过贡献分析研究了SUV结构对车身质量和侧翻防撞性的影响，并根据联邦机动车安全标准216讨论了结构改进。在分析结果的基础上，提出了SUV侧翻防撞性和重量优化模型。使用径向基函数神经网络和遗传算法来构建和优化总重量，最大接触力，和扭转频率。然后，提出的算法利用粒子群算法和人工免疫算法来寻找优化SUV结构的Pareto解决方案。最后，通过类似于理想解法的顺序偏好技术从帕累托最优解中确定最终解。与以往的研究相比，结果表明，提出的混合优化算法提高了帕累托解集的多样性和分布均匀性，提高了SUV侧翻的耐撞性，并减轻了SUV结构部件的重量。