摘要

When investigating many continuous factors under experi-mental cost constraints, an experimenter often selects a screening design from among a variety of published options. This article introduces a novel design selection methodology based on decision trees that identifies which screening design option minimizes the total expected cost of a multistage experiment under model uncertainty. The methodology is illustrated by an example in which a set of traditional and recently published experimental designs are considered for the screening of six potential factors.

  • 出版日期2014-4-3