Comparative Study on Kriging Surrogate Models for Metaheuristic Optimization of Multidimensional Electromagnetic Problems

Author

Bin Xia ; Ziyan Ren ; Chang-Seop Koh

Author_Institution

Coll. of Electr. & Comput. Eng., Chungbuk Nat. Univ., Cheongju, South Korea

Volume

51

Issue

3

fYear

2015

fDate

Mar-15

Firstpage

1

Lastpage

4

Abstract

This paper presents a comparative study on different surrogate models for an application to multidimensional optimal design of electromagnetic devices. The surrogate model is combined with a heuristic optimization algorithm, which usually requires a lot of objective function calls. For this, various Kriging surrogate models, including dynamic Kriging (DK) and multiquadric radial basis function with optimal shape parameter, are investigated and discussed. In the DK surrogate model, the optimal combination of the basis functions is obtained using an improved binary particle swarm optimization algorithm.

Keywords

computational electromagnetics; electromagnetic devices; electromagnetism; optimisation; particle swarm optimisation; radial basis function networks; statistical analysis; DK surrogate model; dynamic Kriging; electromagnetic devices; heuristic optimization algorithm; improved binary particle swarm optimization algorithm; kriging surrogate models; metaheuristic optimization; multidimensional electromagnetic problems; multidimensional optimal design; multiquadric radial basis function; optimal shape parameter; Accuracy; Adaptation models; Algorithm design and analysis; Biological system modeling; Computational modeling; Optimization; Shape; Binary particle swarm optimization (BPSO); Kriging surrogate models; dynamic Kriging (DK); multiquadric radial basis function (MQ-RBF);

fLanguage

English

Journal_Title

Magnetics, IEEE Transactions on

Publisher

ieee

ISSN

0018-9464

Type

jour

DOI

10.1109/TMAG.2014.2361956

Filename

7093488