한국산업융합학회 논문집 (Journal of the Korean Society of Industry Convergence)
- 제12권1호
- /
- Pages.35-42
- /
- 2009
- /
- 1226-833X(pISSN)
- /
- 2765-5415(eISSN)
평균과 분산에 관한 설계자 선호에 기초한 설계 최적화
Design Optimization Based on Designer's Preferences for the Mean and Variance
- Park, Jong-Cheon (School of Mechanical Engineering, KIT) ;
- Kim, Kyung-Mo (School of Mechanical Engineering, KIT) ;
- Kim, Kwang-Ho (School of Mechanical Engineering, KIT)
- 투고 : 2008.11.28
- 심사 : 2009.02.13
- 발행 : 2009.02.28
초록
In Taguchi's quadratic expected loss function used as robustness metric of performance characteristics, the mean and variance contributions are confounded. The consolidation of the mean and variance in the expected loss function may not always be the ideal approach. This paper presents a procedure for multi-attributes design optimization, where the mean and variance of performance characteristics are considered as separate attributes having designer's relative preferences for them and Technique for Order Preference by Similarity to Ideal Solution(TOPSIS) is introduced to attain robust optimal design. The effectiveness of proposed approach is shown with an example of a weld line minimization problem in the injection molding process.