Gradient‐based criteria for sequential experiment design

Abstract

Computer experiments are often used as inexpensive alternatives to real-world experiments. Statistical metamodels of the computer model's input-output behavior can be constructed to serve as approximations of the response surface of the real-world system. The suitability of a metamodel depends in part on its intended use. While decision makers may want to understand the entire response surface, they may be particularly keen on finding interesting regions of the design space, such as where the gradient is steep. We present an adaptive, value-enhanced batch sequential algorithm that samples more heavily in such areas while still providing an understanding of the entire surface. The design points within each batch can be run in parallel to leverage modern multi-core computing assets. We illustrate our approach for deterministic computer models, but it has potential for stochastic simulation models as well.

Research on evaluation indices and calculation method of experimental design

“…At present, many evaluation indices have been well applied to evaluate and select designs, such as the resolution and minimum aberration, optimal design criterion, analysis of variance, correlation between factors, spatial uniformity, and so on. [1][2][3][4][5][6][7][8] But it remains unclear which indices are appropriate for evaluating and comparing different designs. Additionally, it is crucial to understand the characteristics of the designs measured by these indices, how to calculate them, and how to integrate multiple indices to evaluate designs.…”

“…Therefore, the comparison and evaluation of different experimental designs have become a problem of great concern. At present, many evaluation indices have been well applied to evaluate and select designs, such as the resolution and minimum aberration, optimal design criterion, analysis of variance, correlation between factors, spatial uniformity, and so on 1–8 . But it remains unclear which indices are appropriate for evaluating and comparing different designs.…”