Robust global optimal method for electromagnetic device design using approximation models

This paper presents a fast global optimal method for the optimization design of electromagnetic device, which based on the approximation models and the global optimization algorithms. Two approximate models, radial basis functions model and compactly supported radial basis functions model, are fully discussed in this paper. In order to illustrate the efficiency and robustness of the proposed method, an analytical test function and the TEAM Workshop problem 22 are investigated. Analysis and comparison show that the presented method is promising. The finite element sampling points can be obviously reduced by the proposed method.