摘要
稳健性参数设计一直是一个很重要的统计问题,目前通常使用的方法称为两步法.第一步求初始稳定中心,第二步线性调优.但是对初始稳定中心的求解方法却不同,而且不同的求解方法也导致最终求得的稳定中心也不同.电感电路稳健性参数设计是稳健性设计中的经典例子之一.但就是对这个经典例子,许多教科书给出的并非最优解.本文用我们创立的求解稳定中心的方法(称为全局-局部分析方法或者GL算法Global-Local Algo-rithm),得出了比以往更好的稳定中心.
Robust parameter design is an important statistical problem. The method ususally adopted is the following two-step procedure. The aim of a first step is to find out the former stable center. The second step is to optimize the former stable center linearly. But the way used to find out the former stable center is different among different statistical schools and that leads to obtain different stable centers. Identifying the stable center of an inductance circuit is a classical work of robust parameter design. However, the stable center found in a lot of books is not optimal. In this paper, a new stable center has been found by using our methods for solving stable centers, called GL Algorithm （Global-Local Algorithm）.
出处
《河南师范大学学报（自然科学版）》
CAS
CSCD
北大核心
2006年第1期8-14,共7页
Journal of Henan Normal University(Natural Science Edition)
基金
国家自然科学基金资助项目(10571045)
关键词
稳定性
正交性
不变性
信息分解比
stability
orthogonality
invariance
information-decomposition-ratlo
