Studies in Efficient Design of Experiments

首席研究员计划研究实验设计中的几个问题。我们将研究正交表和其他一些设计的投影性质。最近的研究表明，某些正交表的有趣的投影性质具有重要的统计学意义。还将研究新引入的估计能力的概念，即衡量设计处理和估计涉及相互作用的不同潜在模型的能力。除了析因设计，首席调查员还将研究随机区组效应模型下的最优和有效回归设计。其他拟议的研究包括首席研究员与李克洲教授合作研究降维技术，以分析设计实验的数据。这种技术与动态图形相结合，可以帮助提取传统分析可能无法提供的其他有用信息。由此产生的设计问题也将得到研究。实验设计在广泛的科学和工业研究中被广泛使用。在工业实验中，往往要研究大量的因素，但进行实验的费用很高。部分析因设计对于处理这个问题特别有用。近年来，析因设计受到了相当大的关注，这主要是因为日本成功地将其应用于提高工业制造的质量。这种兴趣的重现为重新审视一些传统观点提供了机会。在未来几年，这将是首席调查员研究活动的最高优先事项。除了理论研究外，还将构建良好的设计供实验者使用。其他拟议的研究包括最近的验光实验引起的问题，以及对一种新的数据分析工具在实验数据上的性能的研究。

The Principal Investigator plans to study several problems in design of experiments. The projection properties of orthogonal arrays and some other designs will be investigated. Recent research has shown that interesting projection properties of certain orthogonal arrays have important statistical implications. The newly introduced concept of estimation capacity, a measure of the capability of a design to handle and estimate different potential models involving interactions, will also be investigated. In addition to factorial designs, the Principal Investigator will study optimal and efficient regression designs under random block-effects models. Other proposed research includes the Principal Investigator's collaboration with Professor Ker-Chau Li on the study of dimension reduction techniques for the analysis of data from designed experiments. Such techniques, in conjunction with dynamic graphics, can help extract additional useful information that may not be provided by traditional analyses. The resulting design issues will also be studied. Experimental design is used extensively in a wide range of scientific and industrial investigations. In industrial experiments, often a large number of factors have to be studied, but the experiments are expensive to conduct. Fractional factorial designs are particularly useful to handle this problem. In recent years, factorial designs have received considerable attention, mainly due to the Japanese success in applying them to improve quality in industrial manufacturing. This re-surgence of interest provides an opportunity to re-examine some conventional wisdom. This will be given the highest priority in the Principal Investigator's research activities during the coming years. In addition to theoretical studies, good designs will also be constructed for experimenters' use. Other proposed research includes a problem arising from a recent optometry experiment, and the study of the performance of a new data-analyt ic tool on experimental data.