论文部分内容阅读
本文阐述解决连续或离散最坏情况容差分配问题,以及在面向用户的计算机程序中的实现方法,这是应当和选择最适宜的标称设计同时解决的问题。容差问题实质上是在满足性能或别的指标要求的前提下指定误差的问题。我们的方法被认为是微波设计方面的新途径,能够解决各种容差和有关的问题。采用了达金(Dakin)的树形搜索法,新的准牛顿极小化法,以及最小的P次近似法。程序本身是妥善编成的,性能指标和约束条件的附加与删除,成本函数和最优化方法的代换等都容易实现。使用选择方式和拖欠值来扩大灵活性。完整的Fortran程序清单和有关文件将是有用的。
This article addresses solving continuous or discrete worst-case tolerance assignments and implementing them in user-oriented computer programs, a problem that should be addressed at the same time as choosing the most appropriate nominal design. Tolerance problem is essentially the specified error in the premise of meeting the performance or other indicators. Our method is considered as a new approach to microwave design that solves various tolerances and related problems. Dakin tree search, new quasi-Newton minimization, and the smallest P-approximation are used. The program itself is well organized, with additions and deletions of performance indicators and constraints, cost functions, and substitution of optimization methods. Use options and defaults to increase flexibility. A complete list of Fortran programs and related files will be useful.