Abstract: To address the structural design optimization problem of a vehicle component, Hybrid Robust Arti?cial Bee Colony（HRABC） algorithm based on Taguchi method is proposed. Taguchi method is used to generate Analysis of Variance（ANOVA） table of objective function. Suitable region of design variables by which robust primary population of arti?cial bee colony algorithm is defined can be found by ANOVA table. Then, expression based on multiple design variables of objective function can be used to extract solution space of arti?cial bee colony algorithm and get optimization results. Finally, the effectiveness and robustness of proposed approach is verified on the structural design optimization of a vehicle component and a multi-tool milling optimization problem. Analysis results show that proposed HRABC method has the best optimization effect on rate of convergence and effectiveness comparing with several frequently-used optimization algorithms.

Key words: artificial bee colony, Taguchi method, vehicle component, optimal designing, analysis of variance

摘要： 针对车辆部件的结构设计制造优化问题，提出了基于Taguchi方法的混合鲁棒人工蜂群算法（HRABC）。采用Taguchi方法生成了目标函数的方差分析（ANOVA）表，根据ANOVA表寻找到设计变量的合理区间，并根据这些区间定义人工蜂群算法的鲁棒初始种群；利用基于目标函数的多个设计变量的影响提取人工蜂群算法的解空间，从而得到优化结果。在车辆部件结构设计优化及多刀具铣削优化问题上，验证了所提算法的有效性及鲁棒性。分析结果表明，与几种常用的优化算法相比，在收敛速度和有效性方面，该HRABC算法具有最好的优化效果。

关键词: 人工蜂群算法, Taguchi方法, 车辆部件, 优化设计, 方差分析