If you need an accessible version of this item please contact JSTOR User Support

Bayesian Design and Analysis of Computer Experiments: Use of Derivatives in Surface Prediction

Max D. Morris, Toby J. Mitchell and Donald Ylvisaker
Technometrics
Vol. 35, No. 3 (Aug., 1993), pp. 243-255
DOI: 10.2307/1269517
Stable URL: http://www.jstor.org/stable/1269517
Page Count: 13
  • Download PDF
  • Cite this Item

You are not currently logged in.

Access your personal account or get JSTOR access through your library or other institution:

login

Log in to your personal account or through your institution.

If you need an accessible version of this item please contact JSTOR User Support
Bayesian Design and Analysis of Computer Experiments: Use of Derivatives in Surface Prediction
Preview not available

Abstract

This article is concerned with the problem of predicting a deterministic response function y0 over a multidimensional domain T, given values of y0 and all of its first derivatives at a set of design sites (points) in T. The intended application is to computer experiments in which y0 is an output from a computer model of a physical system and each point in T represents a particular configuration of the input parameters. It is assumed that the first derivatives are already available (e. g., from a sensitivity analysis) or can be produced by the code that implements the model. A Bayesian approach in which the random function that represents prior uncertainty about y0 is taken to be a stationary Gaussian stochastic process is used. The calculations needed to update the prior given observations of y0 and its first derivatives at the design sites are given and are illustrated in a small example. The issue of experimental design is also discussed, in particular the criterion of maximizing the reduction in entropy, which leads to a kind of D optimality. It is shown that, for certain classes of correlation functions in which the intersite correlations are very weak, D-optimal designs necessarily maximize the minimim distance between design sites. A simulated annealing algorithm is described for constructing such maximin distance designs. An example is given based on a demonstration model of eight inputs and one output, in which predictions based on a maximin design, a Latin hypercube design, and two compromise designs are evaluated and compared.

Page Thumbnails

  • Thumbnail: Page 
243
    243
  • Thumbnail: Page 
244
    244
  • Thumbnail: Page 
245
    245
  • Thumbnail: Page 
246
    246
  • Thumbnail: Page 
247
    247
  • Thumbnail: Page 
248
    248
  • Thumbnail: Page 
249
    249
  • Thumbnail: Page 
250
    250
  • Thumbnail: Page 
251
    251
  • Thumbnail: Page 
252
    252
  • Thumbnail: Page 
253
    253
  • Thumbnail: Page 
254
    254
  • Thumbnail: Page 
255
    255