论文部分内容阅读
实验设计,尤其是混合水平实验设计,已经在科研领域和制造业中得到了广泛使用.虽然很多学者已经提出了许多实验设计理论以及实验设计用表,但是对于含有混合水平的大样本体系的实验设计方法仍然是相当困难的.本文根据Yong Guo提出的实验设计方法,并依据正交设计理论和遗传算法,把此方法改进为适用于大样本体系的实验设计方法.为了配合遗传算法,而改进了交叉与变异操作;为了从复杂体系的全实验设计中高效地获取某一次实验设计,引进了数学方法中的求余运算.最后,当把此实验设计方法应用于一个真实的、多因子多水平的体系时,获得了比较满意的实验方案.因此,对于大样本体系,此实验设计方法所得到的实验方案是可以被接受的.
Experimental design, especially mixed-level experimental design, has been widely used in the field of scientific research and manufacturing.Although many scholars have put forward many experimental design theory and experimental design of the table, but for a large sample system containing mixed levels of experiments The design method is still quite difficult.This paper is based on the experimental design method proposed by Yong Guo, and based on the orthogonal design theory and genetic algorithm, this method is improved to the experimental design method suitable for large sample system.In order to cooperate with the genetic algorithm, Crossover and mutation operation.In order to obtain a certain experimental design efficiently from the full experimental design of complex system, the remainder arithmetic in mathematics method was introduced.Finally, when applying this experimental design method to a real, multi-factorial Level system, a more satisfactory experimental scheme is obtained.Thus, for the large sample system, the experimental scheme obtained by this experimental design method is acceptable.