作者: Jack P. C. Kleijnen
DOI:
关键词:
摘要: Experiments with real systems, but especially simulated may involve hundreds of factors. However, only a few factors are really important. Detecting these important requires special designs such as random and group-screening designs. Random simple, they yield biased estimators. Group-screening is based on aggregation. The assumptions discussed in detail, seem not very restrictive. References to applications both design types given.