In this study, a novel multi-objective reliability-based design optimization (MORBDO) method considering the maximum allowable deviation range of design variables is proposed for the reducer housing of electric vehicles. First, the numerical model of the reducer housing is established by ABAQUS and verified by experiments. A radial basis function (RBF) neural network model is used to construct the approximate finite element model. The structural parameters of the RBF are optimized using the heuristic global optimization ability of the particle swarm optimization (PSO) algorithm. Sequential quadratic programming (SQP) and non-dominated sorting genetic algorithm II (NSGA II) are used to perform the MORBDO. Finally, the technique for order preference by similarity to ideal solution, a multi-criteria decision-making (MCDM) method, is used to select the ideal design in multi-objective Pareto points. The optimization method generated a set of Pareto non-dominated solutions with three objectives, which can be selected for a more feasible scheme using MCDM. The proposed method comprehensively measures the requirements of manufacturing and performance criteria, and the optimization results provide a variety of optimization design schemes for the reducer housing of electric vehicles.

在这项研究中，提出了一种新的基于多目标可靠性的设计优化（MORBDO）方法，该方法考虑了设计变量的最大允许偏差范围，用于电动汽车减速器外壳。首先，通过ABAQUS建立了减速器壳体的数值模型，并进行了实验验证。径向基函数（RBF）神经网络模型用于构建近似有限元模型。RBF的结构参数使用粒子群优化（PSO）算法的启发式全局优化能力进行优化。顺序二次规划 (SQP) 和非支配排序遗传算法 II (NSGA II) 用于执行 MORBDO。最后，通过与理想解决方案相似度的顺序偏好技术，一种多标准决策（MCDM）方法，用于在多目标 Pareto 点中选择理想设计。该优化方法生成了一组具有三个目标的 Pareto 非支配解，可以使用 MCDM 选择更可行的方案。该方法综合衡量了制造要求和性能标准，优化结果为电动汽车减速器外壳提供了多种优化设计方案。