A Gradient Taguchi Method for Engineering Optimization

5th Asia Conference on Mechanical and Materials Engineering

机械制造;材料科学

Hwang, Shun-Fa^1 ; Wu, Jen-Chih^1 ; He, Rong-Song^2

Department of Mechanical Engineering, National Yunlin University of Science and Technology, 123 University Road, Touliu

640, Taiwan^1

Department of Mechanical Engineering, WuFeng University, 117 Chian-Kuo Road, Ming-Hsiung, Chia-Yi

621, Taiwan^2

关键词: Computation costs;    Convergence speed;    Engineering optimization;    Fast convergence speed;    Hybrid algorithms;    Hybrid genetic algorithms;    Optimum combination;    Vibration testing;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/241/1/012022/pdf
DOI  :  10.1088/1757-899X/241/1/012022

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

To balance the robustness and the convergence speed of optimization, a novel hybrid algorithm consisting of Taguchi method and the steepest descent method is proposed in this work. Taguchi method using orthogonal arrays could quickly find the optimum combination of the levels of various factors, even when the number of level and/or factor is quite large. This algorithm is applied to the inverse determination of elastic constants of three composite plates by combining numerical method and vibration testing. For these problems, the proposed algorithm could find better elastic constants in less computation cost. Therefore, the proposed algorithm has nice robustness and fast convergence speed as compared to some hybrid genetic algorithms.