Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


osu1110297243.pdf (1.03 MB)

ETD Abstract Container

Abstract Header

Experimental planning and sequential kriging optimization using variable fidelity data

Huang, Deng
http://rave.ohiolink.edu/etdc/view?acc_num=osu1110297243

Abstract Details

2005, Doctor of Philosophy, Ohio State University, Industrial and Systems Engineering.
Engineers in many industries routinely need to improve the product or process designs using data from the field, lab, and computer experiments. This research seeks to develop experimental planning and optimization schemes using data form multiple experimental sources. We use the term "fidelity" to refer to the extent to which a surrogate experimental system can reproduce results of the system of interest. For experimental planning, we present perhaps the first optimal designs for variable fidelity experimentation, using an extension of the Expected Integrated Mean Squared Error (EIMSE) criterion, where the Generalized Least Squares (GLS) method was used to generate the predictions. Numerical tests are used to compare the method performance with alternatives and to investigate the robustness of incorporated assumptions. The method is applied to automotive engine valve heat treatment process design in which real world data were mixed with data from two types of computer simulations. Sequential Kriging Optimization (SKO) is a method developed in recent years for solving expensive black-box problems. We propose an extension of the SKO method, named Multiple Fidelity Sequential Kriging Optimization (MFSKO), where surrogate systems are exploited to reduce the total evaluation cost. As a pre-step to MFSKO, we extended SKO to address stochastic black-box systems. Empirical studies showed that SKO compared favorably with alternatives in terms of consistency in finding global optima and efficiency as measured by number of evaluations. Also, in the presence of noise, the new expected improvement criterion achieves desired balance between the need for global and local searches. In the proposed MFSKO method, data on all experimental systems are integrated to build a kriging meta-model that provides a global prediction of the system of interest and a measure of prediction uncertainty. The location and fidelity level of the next evaluation are selected by maximizing an augmented expected improvement function, which is connected with the evaluation costs. The proposed method was applied to test functions from the literature and metal-forming process design problems via Finite Element simulations. The method manifests sensible search patterns, robust performance, and appreciable reduction in total evaluation cost as compared to the original method.
Richard Miller (Advisor)
120 p.
Multiple Fidelity; Optimal Experimental Design; Surrogate Systems; Expected Integrated Mean Squared Error; Efficient Global Optimization; Stochastic Black-box Systems; Kriging

Recommended Citations

Citations

  • Huang, D. (2005). Experimental planning and sequential kriging optimization using variable fidelity data [Doctoral dissertation, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1110297243

    APA Style (7th edition)

  • Huang, Deng. Experimental planning and sequential kriging optimization using variable fidelity data. 2005. Ohio State University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1110297243.

    MLA Style (8th edition)

  • Huang, Deng. "Experimental planning and sequential kriging optimization using variable fidelity data." Doctoral dissertation, Ohio State University, 2005. http://rave.ohiolink.edu/etdc/view?acc_num=osu1110297243

    Chicago Manual of Style (17th edition)

osu1110297243
2,603
© 2005, all rights reserved.
This open access ETD is published by The Ohio State University and OhioLINK.