Title :
Parameter sensitivity of covariance-based response surfaces for modeling IC processes
Author :
Waring, Thomas G. ; Walton, Anthony J. ; Ferguson, R. Stuart ; Sprevak, Dan
Author_Institution :
Dept. of Electr. Eng., Edinburgh Univ., UK
fDate :
11/1/1999 12:00:00 AM
Abstract :
When modeling the results of computer experiments which do not suffer from random errors, it is desirable that the models used exactly reproduce all the simulated data points. This paper discusses a method that ensures this and examines how small variations of the estimated parameters of the model influence the accuracy of the prediction. These findings are used to justify a procedure whereby the conditioning difficulties inherent in the method can be overcome
Keywords :
covariance analysis; digital simulation; integrated circuit manufacture; semiconductor process modelling; surface fitting; IC process modelling; conditioning difficulties; covariance-based response surfaces; estimated model parameters; parameter sensitivity; simulated data points; Accuracy; Computational modeling; Computer errors; Computer simulation; Integrated circuit modeling; Parameter estimation; Polynomials; Predictive models; Response surface methodology; Surface fitting;
Journal_Title :
Semiconductor Manufacturing, IEEE Transactions on
