Approximation models are frequently used in design optimization in order to reduce computational cost. Many new forms of approximation model have appeared recently, but it is difficult to know which one to choose for the problem at hand. This paper investigates a model selection strategy that estimates the accuracy of all the different approximation models under consideration, so that the most accurate one can be chosen
Published in:
Magnetics, IEEE Transactions on
(Volume:42
,
Issue:
4
)
Date of Publication: April 2006