Research and Applications of Multi-Objective Optimization Methods

• Main Authors: Kuang-Hung Hsien, 薛光宏
• Other Authors: Shyh-Chour Huang
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/93839414374237922132

博士 === 國立高雄應用科技大學 === 機械與精密工程研究所 === 101 === In recent years, the optimization method is applied in a variety of fields to solve a variety of complex practical problems. The Taguchi method is one of the most used method and save the cost of experiments. The optimal problems usually have multiple objectives, so that users can not easily set the weight of the objective, and the optimal solution obtained does not meet the requirements. The objective of this dissertation is to study multi-objective optimization problem based on the Taguchi method, such as Taguchi, Utility and Fuzzy multi-objective optimization methods. By application of statistical of the contribution of ANOVA determining the weights of the objective functions, the multi-objective problems can be able to obtain more reasonable optimal solution satisfying the real requirement. Through redefined and improved the multi-objective optimization problems, several new methods, TMOO, W-Utility, HTB-Utility, W-Fuzzy and L-Fuzzy, were developed. The results shown these methods enable more efficient processing of multi-objective problems. By mixing up above methods, some hybrid multi-objective optimization methods were developed as follows: HTU (Hybrid Taguchi-utility), HUF (Hybrid utility-fuzzy), HTF (Hybrid Taguchi-fuzzy) and HTUF (Hybrid Taguchi-utility-fuzzy). The results shown these hybrid multi-objective optimization methods can define the weights for every objective functions clearly, and can obtain the more reasonable optimal solution.