5. Design Optimization Using Response Correction Techniques

The material presented in Chap. 4 was intended to provide the reader with a general introduction to surrogate modeling and surrogate-based optimization. Among the methods presented in there, optimization using physics-based models belongs to the most promising approaches in terms of computational efficiency. In Sects. 4.​3 and 4.​5, we outlined basic methods of constructing the physics-based surrogates by enhancing the underlying low-fidelity models, as well as utilizing such models in the design optimization process. At that point, it was mentioned that response correction is perhaps the simplest way of implementing such a model enhancement. In this chapter, we formulate the response correction task in a more rigorous manner, and then broadly discuss available response correction techniques while categorizing them into parametric and non-parametric ones. We also discuss in more depth the issue of low-fidelity modeling, including the low-fidelity model setup and finding a proper balance between the model speed and accuracy. More detailed exposition of various response correction methods and their applications for solving computationally expensive design problems in various engineering fields are given in the remaining chapters of this book.