Exclusive Content & Downloads from ASQ

Multivariate Gaussian Process Emulators With Nonseparable Covariance Structures

Summary: The Gaussian process regression model is a popular type of “emulator” used as a fast surrogate for computationally expensive simulators (deterministic computer models). For simulators with multivariate output, common practice is to specify a separable covariance structure for the Gaussian process. Though computationally convenient, this can be too restrictive, leading to poor performance of the emulator, particularly when the different simulator outputs represent different physical quantities. Also, treating the simulator outputs as independent can lead to inappropriate representations of joint uncertainty. We develop nonseparable covariance structures for Gaussian process emulators, based on the linear model of coregionalization and convolution methods. Using two case studies, we compare the performance of these covariance structures both with standard separable covariance structures and with emulators that assume independence between the outputs. In each case study, it was discovered that only emulators with non-separable covariances structures have sufficient flexibility both to give good predictions and to represent joint uncertainty about the simulator outputs appropriately. This article has supplementary material online.

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: Engineering
  • Keywords: Case study, Covariance, Uncertainty, Regression analysis, Computer models
  • Author: Fricker, Thomas E.; Oakley, Jeremy E. ; Urban, Nathan M.
  • Journal: Technometrics