数学误用四例

发布时间：2019-03-09 13:59

【摘要】：深浅程度不同地来源于实践过程的思考，本文就数学方法的误用提出不同形式的四例，四节各举一例。第一节指出：在细纱接头的工作管理中，为了求出生产率和机器利用率等指标，与其通过排队论模型间接推算，不如通过产量直接推算来得准确和省事。第二节指出：欧美一批试验设计书〔３～８，２０〕关于“当可能有非零的交互作用时，不许可安排部分实施”的观点，如果仅限于实施平方和分解的方差分析、把主效应和各级交互作用估准，，那还是正确的；但此观点不能引伸出去，因为正是低比例的正交表部分实施，不管有无交互作用为零，才是高效益地实现相对最优化的有效措施和根本手段。第三节指出了正交表田口式表头设计的不合理处，并说明了小表多排因素为什么能提高效益的理由。第三节后的附录指出两个设计例子的布点密集而不够分散。关于偏差的第四个例子已提前在〔２２〕中发表。
[Abstract]:In this paper, four cases of the misuse of mathematical methods are put forward, and one example is given in each of the four sections. In the first section, it is pointed out that in order to find out the indexes of productivity and machine utilization ratio in the work management of roving joint, it is not as accurate and easy to calculate directly through the output than to calculate indirectly by the queuing theory model. The second section points out that a batch of European and American experimental design books (3 ~ 8, 20) have a view that "when there may be non-zero interactions, no partial implementation can be arranged", if it is limited to the analysis of variance of the sum of squares decomposition. It is true to estimate the main effect and all levels of interaction; However, this point of view cannot be extended, because it is the partial implementation of the low proportion orthogonal table, no matter whether the interaction is zero or not, is the effective measure and the fundamental means to achieve the relative optimization effectively. The third section points out the irrationality of the design of the orthogonal table Taguchi head and explains the reason why the multi-row factor of the small table can improve the benefit. The appendix to section III indicates that the two design examples are densely distributed and not scattered enough. A fourth example of deviation has been published in advance in (22).
【作者单位】：中国科学技术大学研究生院
【分类号】：C931.1

【相似文献】