期刊
IEEE TRANSACTIONS ON MAGNETICS
卷 46, 期 8, 页码 2868-2871出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TMAG.2010.2043654
关键词
Kriging surrogate model; multiobjective optimization; optimal design; parallel computing; surrogate-assisted optimization; TEAM workshop problem 22
Conceiving electromagnetic devices using finite element modeling tools is a complex task and also time-costly. The Efficient Global Optimization method, based on the progressive construction of surrogate models, is studied. The method uses Kriging models and allows for multiobjective optimization. An original infill criterion, combining the surrogate models and an estimate of their error is proposed. Moreover, two techniques for the calculi distribution, adapted to the algorithm, are tested on an eight-core machine. An advantage of the method consists in its capability of providing sufficiently accurate models for each objective and constraint function around the obtained Pareto front. The SMES device of the TEAM problem 22 is used as benchmark.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据