Coarse and Surrogate Model Assessment for Engineering Design Optimization with Space Mapping

Two important issues in space mapping optimization are: (i) the quality of the coarse model used in the optimization process, and (ii) the right choice of the space mapping surrogate model for a given optimization problem. Both issues are critical to the performance of the space mapping algorithm. In this paper we introduce methods of assessing the quality of coarse/surrogate models. The methods can be used to predict whether a given combination of a coarse model and a space mapping type can be successfully used in space mapping optimization. They also allow us to compare different surrogate models with respect to potential performance in optimization. Theoretical considerations are illustrated by examples.