A Hybrid Algorithm for Electromagnetic Optimization Utilizing Neural Networks

被引：0

作者：

Liu, Yanan ^{[1
]}

Lu, Tianjian ^{[2
]}

Wu, Ken ^{[2
]}

Jin, Jian-Ming ^{[1
]}

机构：

[1] Univ Illinois, 306 N Wright St, Urbana, IL 61801 USA

[2] Google Inc, 1600 Amphitheatre Pkwy, Mountain View, CA 94043 USA

来源：

2018 IEEE 27TH CONFERENCE ON ELECTRICAL PERFORMANCE OF ELECTRONIC PACKAGING AND SYSTEMS (EPEPS) | 2018年

关键词：

neural networks; genetic algorithms; optimization;

D O I：

暂无

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

We propose a hybrid algorithm for electromagnetic optimization. The algorithm combines the genetic algorithm and gradient based optimization, leveraging the power of neural networks in both. The effectiveness and efficiency of the proposed algorithm are demonstrated using antenna and microwave filter design examples.

引用

页码：261 / 263

页数：3

共 8 条

[1]

[Anonymous], 1999, Microwave Journal

[2]

Bachiller C., 2006, 2006 IEEE Antennas and Propagation Society International Symposium (IEEE Cat. No. 06CH37758C), P1567, DOI 10.1109/APS.2006.1710855

[3] Neural network based design of metagratings [J].