Experiments, Processes, and Selected Topics

摘要

The lead article in the May issue is "Analysis of Computer Experiments Using Penalized Likelihood in Gaussian Kriging Models" by Runze Li and Agus Sudjianto. Kriging is an analysis approach for computer experiments used to create a cheap-to-compute "metamodel" as a surrogate to a computationally expensive simulation model. In some situations, the likelihood function near the maximum may be flat, which leads to maximum likelihood estimates for covariance parameters with very large variance. To overcome this, the authors propose a penalized likelihood approach. In "Moment Aberration Projection for Nonregular Fractional Factorial Designs," Hongquan Xu and Lih-Yuan Deng propose a novel criterion they term moment aberration projection to rank and classify nonregular designs. The criterion measures the goodness of a design through moments of the number of coincidences between the rows of its projection designs. It is used to rank and classify designs of 16, 20, and 27 runs, and examples are used to illustrate that the subsequent ranking of designs is supported by other design criteria.