最优试验次序的几类构造方法

Several Construction Methods for Optimal Run Order

许多领域都离不开试验,例如在新产品的研发以及测试中都需要进行精心设计的科学试验.当在试验中因子的水平改变非常困难时,如何合理安排试验次序是一个非常重要的问题.本文研究了具有最小和最大水平变化次数的试验次序的一些基本理论,并针对完全因析设计、非正规部分因析设计和均匀设计等设计讨论了最优试验次序构造方法.作为实际中广泛应用的一些设计,利用本文的结果给出了其具有最小和最大水平变化次数的试验次序及相应的水平变化次数.

Experiments are indispensable in many fields,for example,well-designed scientific experiments are needed in the research,development and testing of new products.When it is very difficult to change the level of factors in the experiment,how to arrange the run order reasonably is a very important problem.In this paper,some basic theories of run order with the minimum and maximum level changes are studied,and the optimal run order construction methods are discussed for full factorial designs,nonregular fractional factorial designs and uniform designs.As for some designs are widely used in practice,the run order with the minimum and maximum level changes and the corresponding of level changes are given by using the results of this paper.