Exclusive Content & Downloads from ASQ

Robust Design Optimization With Quadratic Loss Derived From Gaussian Process Models

Summary: Gaussian process models are often used to model responses as a function of control or noise factors. Robust design optimization is often performed using an expected quadratic loss based on the assumption that the posterior mean of the Gaussian process model were the actual response function. The authors propose instead estimating expected quadratic loss criteria using numerical inversion of the Lugannani-Rice sadlepoint approximation. The accuracy of this method is compared with the use of moment-matching techniques on real data.

Anyone with a subscription, including Site and Enterprise members, can access this article.

Other Ways to Access content:

Join ASQ

Join ASQ as a Full member. Enjoy all the ASQ member benefits including access to many online articles.

  • Topics: Software and Technology (for statistics, measurement, analysis)
  • Keywords: Computer experiments, Computer models, Control factors, Loss functions, Noise, Response, Robust design, Lugannani–Rice saddlepoint approximation
  • Author: Tan, Matthias H. Y.; Wu, C. F. Jeff;
  • Journal: Technometrics