A method for statistical design centering

Author

Wojciechowski, Jacek ; Vlach, Jiri

Author_Institution

Inst. of Electron. Fundamentals, Warsaw Univ. of Technol., Poland

Volume

1

fYear

1992

fDate

10-13 May 1992

Firstpage

33

Abstract

A method for geometrical centering is extended by including statistical noncorrelated properties of the design parameters. A metric is introduced that replaces evaluation of a multidimensional integral by evaluation of a number of simple integrals. Yield maximization is formulated and solved as a geometrical problem in n-dimensional metric space; yield derivatives are not required. Optimization techniques are used in solving the problem, simultaneously also providing means for the constraint region approximation and yield estimation. The optimization procedure requires that parameter probability density functions be sufficiently regular. An approximation meeting this condition is proposed

Keywords

constraint handling; network synthesis; probability; statistical analysis; constraint region approximation; geometrical centering; n-dimensional metric space; parameter probability density functions; statistical design centering; statistical noncorrelated properties; yield derivatives; Circuit synthesis; Constraint optimization; Design methodology; Extraterrestrial measurements; Gravity; Multidimensional systems; Optimization methods; Probability; Sampling methods; Statistical distributions;

fLanguage

English

Publisher

ieee

Conference_Titel

Circuits and Systems, 1992. ISCAS '92. Proceedings., 1992 IEEE International Symposium on

Conference_Location

San Diego, CA

Print_ISBN

0-7803-0593-0

Type

conf

DOI

10.1109/ISCAS.1992.230021

Filename

230021