Space-mapping-based interpolation for engineering optimization

Abstract

We consider a simple and efficient space mapping (SM)-based interpolation scheme to work in conjunction with SM optimization algorithms. The technique is useful if the fine model (the one that is supposed to be optimized) is available only on a structured grid. It allows us to estimate the response of the fine model at off-grid points and, as a result, increases the effective resolution of the design variable domain search and improves the quality of the fine model solution found by the SM optimization algorithm. The proposed method requires little computational effort. In particular, no additional fine model evaluations are necessary. Several examples that verify the accuracy and robustness of our approach are provided