A flexible approximation model for use with optimization
Rudi Balling, David Clark · 1992
Flexible models for a proximating objective and constraint Anctions in optimization problems are developed. The models are based on weighted sums of exact function values at sample points. Function gradients can be incorporated into the models. These gradients can be obtained through exact analysis or via an ap roximation based on the function vayues at sample oints. The models are tested on. wee-known optimization problems.