Abstract:
|
Jerry Sacks was the central figure in the initial development of what is still often called “Design and Analysis of Computer Experiment,” abbreviated as DACE. While statisticians had certainly been involved in analyzing the output of deterministic computer models before the 1980’s, this was the period in which a general framework – largely based on Gaussian processes -- was laid out for experimental design, variable screening, and output prediction in settings where computer models of complex physical systems were the objects of study. Jerry’s leadership in this area included his programmatic emphasis at NSF, the University of Illinois, and NISS; his own substantial research contributions to important foundational efforts; and his insistence that the value of these developments be solidly demonstrated in a variety of real, difficult, and interesting applications of computer models. I will briefly review some of his accomplishments in this area, and look at how this effort set the stage for more recent DACE research and applications.
|