Chapter 6: Infill Criteria with Noisy Data

The infill criteria presented in chapter 3 were formulated on the premise that the true engineering function to be approximated is smooth and continuous. Thus we can use a smooth, continuous interpolating surrogate in lieu of calculations of the true function. Most functions that are encountered do indeed have smooth and continuous trends, but function evaluations are often scattered about this trend due to errors in the physical experiment or computer simulation used to calculate the function. One of the key advantages of using a surrogate is that the scatter, or `noise’, can be filtered out, leaving a smooth trend to be searched by the optimization algorithm.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s