Mathematical Sciences: On the Determination and Constructionof Efficient and Robust Experimental Designs

使用实验条件的最佳选择或实验单元或材料的安排，可以最大限度地提高从实验中得出的推论的精度。本研究将构建在相邻实验单元之间存在相关性的情况下有效且稳健的数学设计，基于有效且允许检测偏离模型假设的聚类技术的设计，以及可同时或随后用于分离因素或因素组合的设计，这些因素或因素组合在初始分数因子筛选实验中混淆。

Using the optimal choice of experimental conditions or of arrangements of experimental units or material maximizes the precision of the inferences drawn from the experiment. This research will construct mathematically designs which are efficient and robust in the presence of correlations among adjacent experimental units, designs based on clustering techniques which are efficient and also permit detection of departures from model assumptions, and designs which can be used simultaneously or subsequently to separate effects of factors or combinations of factors which are confounded in an initial fractional factorial screening experiment.