DOI QR코드

DOI QR Code

Weight Function-based Sequential Maximin Distance Design to Enhance Accuracy and Robustness of Surrogate Model

대체모델의 정확성 및 강건성 향상을 위한 가중함수 기반 순차 최소거리최대화계획

  • Jang, Junyong (Dept. of Automotive Engineering, College of Engineering, Hanyang Univ.) ;
  • Cho, Su-Gil (Dept. of Automotive Engineering, College of Engineering, Hanyang Univ.) ;
  • Lee, Tae Hee (Dept. of Automotive Engineering, College of Engineering, Hanyang Univ.)
  • 장준용 (한양대학교 공과대학 미래자동차공학과) ;
  • 조수길 (한양대학교 공과대학 미래자동차공학과) ;
  • 이태희 (한양대학교 공과대학 미래자동차공학과)
  • Received : 2014.12.31
  • Accepted : 2015.02.21
  • Published : 2015.04.01

Abstract

In order to efficiently optimize the problem involving complex computer codes or computationally expensive simulation, surrogate models are widely used. Because their accuracy significantly depends on sample points, many experimental designs have been proposed. One approach is the sequential design of experiments that consider existing information of responses. In earlier research, the correlation coefficients of the kriging surrogate model are introduced as weight parameters to define the scaled distance between sample points. However, if existing information is incorrect or lacking, new sample points can be misleading. Thus, our goal in this paper is to propose a weight function derived from correlation coefficients to generate new points robustly. To verify the performance of the proposed method, several existing sequential design methods are compared for use as mathematical examples.

효율적인 최적설계를 위해 공학분야에 도입된 대체모델의 정확성은 표본점에 큰 영향을 받는다. 대체모델의 정확성을 높이는 방법으로 기 추출한 응답을 이용하는 순차실험계획이 제안되었다. 크리깅 대체모델의 상관계수를 가중치로 적용하여 대체모델의 정확성을 향상시킨 연구가 있었으나, 주어진 정보가 부족하거나 상관계수가 잘못 추정된 경우 표본점이 잘못 추출되어 대체모델의 강건성이 저하된다. 본 논문에서는 기존 순차실험계획의 여러 문제점을 제시하고, 이를 해결하기 위한 가중함수 기반 순차 최소거리최대화계획을 제안한다. 제안하는 순차실험계획의 효용성을 수학 함수에 적용하여 기존 순차실험계획들과 비교하여 정확성과 강건성이 향상됨을 예시한다.

Keywords

References

  1. Johnson, M., Moore, L. and Ylvisaker, D., 1990, "Minimax and Maximin Distance Designs," Journal of Statistical Planning and Inference, Vol. 26, pp. 131-148. https://doi.org/10.1016/0378-3758(90)90122-B
  2. Shewry, M.C. and Wynn, H.P, 1987, "Maximum Entropy Sampling," Journal of Applied Statistics, Vol. 40, No. 4, pp. 165-170.
  3. Lee, T.H., and Jung, J.J., 2004, "Maximum Eigenvalue Sampling of Kriging Model," 10th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference, Albany, New York, August, AIAA 2004-4488.
  4. Jin, R., Chen, W. and Sudjianto, A., 2002, "On Sequential Sampling for Global Metamodeling for in Engineering Design," Proceedings ASME 2002 Design Engineering Technical Conferences and Computer and Information in Engineering Conference, Montreal, Canada, September, DETC2002/DAC-34092.
  5. Crombecq, K., Laermans, E. and Dhaene, T., 2011, "Efficient Space-Filling and Non-Collapsing Sequential Design Strategies for Simulation-Based Modeling," European Journal of Operation Research, Vol. 214, No. 3, pp. 683-696. https://doi.org/10.1016/j.ejor.2011.05.032
  6. Cho, S.K, Byun, H. and Lee, T.H., 2009, "Selection Method of Global Model and Correlation Coefficients for Kriging Metamodel," Trans. Korean Soc. Mech. Eng. A, Vol. 33, No. 8, pp. 813-818. https://doi.org/10.3795/KSME-A.2009.33.8.813
  7. Jang, J., Lee, J.M. and Lee, T.H., 2011, "A Study of C1-Continuity of Split Region Kriging model According to Correlation Functions," Proceedings of the KSME(A ), pp. 174-179.
  8. Park, J.S., 1994, "Optimal Latin-Hypercube Designs for Computer Experiments," Journal of Statistical Planning and Inference. Vol. 39, No. 1, pp. 95-111. https://doi.org/10.1016/0378-3758(94)90115-5