Optimization of Multi-Fidelity Computer Experiments via the EQIE Criterion

Journal Article · Mon Jan 30 23:00:00 EST 2017 · Technometrics

DOI: https://doi.org/10.1080/00401706.2016.1142902 · OSTI ID:1405181

He, Xu ^[1]; Tuo, Rui ^[2]; Jeff Wu, C. F. ^[3]

---

1. Chinese Academy of Sciences (CAS), Beijing (China); Georgia Tech
2. Chinese Academy of Sciences (CAS), Beijing (China)
3. Georgia Inst. of Technology, Atlanta, GA (United States)

Computer experiments based on mathematical models are powerful tools for understanding physical processes. This article addresses the problem of kriging-based optimization for deterministic computer experiments with tunable accuracy. Our approach is to use multi- delity computer experiments with increasing accuracy levels and a nonstationary Gaussian process model. We propose an optimization scheme that sequentially adds new computer runs by following two criteria. The first criterion, called EQI, scores candidate inputs with given level of accuracy, and the second criterion, called EQIE, scores candidate combinations of inputs and accuracy. Here, from simulation results and a real example using finite element analysis, our method out-performs the expected improvement (EI) criterion which works for single-accuracy experiments.

• Research Organization: Georgia Tech Research Corp., Atlanta, GA (United States)
• Sponsoring Organization: USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR) (SC-21)
• Grant/Contract Number: SC0010548
• OSTI ID: 1405181
• Report Number(s): DOE-GT--0010548-7; FG02-13ER26159
• Journal Information: Technometrics, Journal Name: Technometrics Journal Issue: 1 Vol. 59; ISSN 0040-1706
• Publisher: Taylor & FrancisCopyright Statement
• Country of Publication: United States
• Language: English