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

数学科学：关于最优和鲁棒实验设计的确定和构建

• 批准号: 8700945
• 负责人: Michael Jacroux
• 金额: --
• 依托单位: Washington State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1987
• 资助国家: 美国
• 起止时间: 1987-06-01 至 1989-11-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8700945&HistoricalAwards=false
• 关键词: Mathematical; Sciences; Determination; Constructionof; Optimal

The investigator will study optimal and robust experimental designs. The first set of problems deal with optimally estimating the effects of treatments and comparing test treatments with a standard treatment. The second problem of robust designs will involve nearest neighbour analysis. The main benefit of the nearest neighbour analysis is to obtain unbiased estimates of treatment effects even when unknown trends are present. This research will construct designs which are efficient under nearest neighbour analysis. "Clustering" as an alternative to nearest neighbour methods will be explored also.

研究者将研究最佳和稳健的实验设计。第一组问题处理最佳估计的治疗效果和比较测试治疗与标准治疗。稳健设计的第二个问题将涉及最近邻分析。最近邻分析的主要好处是即使在存在未知趋势的情况下也能获得治疗效果的无偏估计。本研究将建构在最近邻分析下有效的设计。还将探讨“聚类”作为最近邻方法的替代办法。