Abstract

Using results of the probability theory, we estimate a number of computational experiments required for choosing the most manufacturable design with a sufficient confidence level. Experiments with broadband optical monitoring are used as examples.

© 2010 The Optical Society

PDF Article