Response surface approximations: Noise, error repair, and modeling errors

In the past few Sears there has been interest in using response surface tec hniques to create surrogates to computer simulations. Response surface tech niques allow detection and correction of errors as well as filtering out nu merical noise, but these techniques introduce additional modeling errors. M ethods of reducing both noise and modeling errors are explored. It is also demonstrated that repairing designs with large errors is preferable to elim inating these designs from consideration. Response surface approximations f or a high-speed civil transport wing weight equation created from the resul ts of a large number of structural optimizations are used for demonstration . It is shown that the statistical tools available for response surface tec hniques are effective for error detection and for filtering out noise. Once the noise is reduced, modeling errors can also be reduced by increasing th e order of the approximation.