Thesis (M.Sc.) - Cairo University - Faculty of Engineering - Department of Mathematics and Physics

Optimization is a crucial step for optimal system design. The optimization process requires many function evaluations to obtain the optimal design of computationally expensive engineering systems (CEES). Nowadays, the increasing complexity of engineering systems in different areas (such as structural mechanics, fluid dynamics, electrical engineering and electromagnetics) implies a significant increase of the CPU time consumption in the design and optimization stage. The computational overhead makes the optimization of such CEES practically prohibitive. To overcome this problem, computationally cheap surrogate models (Response Surfaces, Space Mapping, Kriging Models and Neural Networks) are used. Throughout the optimization process, iteratively updated surrogates are employed to replace the computationally expensive models.

Issued also as CD