Affiliations: [a] School of Electrical Engineering and Computer Science,
Seoul National University, Shillim-dong, Kwanak-gu, Seoul, Korea
Correspondence:
[*]
Corresponding author: Dong-Kyun Woo, School of Electrical
Engineering, Seoul National University, Shillim-dong, Kwanak-gu, Seoul,
151-742, Korea. Tel.: +82 2 880 7262; Fax: +82 2 878 1452; E-mail:
[email protected]
Abstract: This paper presents evolution strategy (ES) assisted by Kriging
metamodel to preselect the most promising solutions and reduce the number of
fitness function evaluations. The role of Kriging metamodel is to predict an
objective function value for a new candidate solution by exploiting information
recorded during previous generations. The use of promising solutions for
screening the candidate solutions makes it possible to significantly reduce the
computational cost of ES. The usefulness of the proposed method is verified by
means of mathematical test cases and the optimal rotor structure design of
interior permanent magnet synchronous machine.
