A Multi-Constrained Optimization Method for THz Backward Wave Oscillators

Rong Bao; Yongdong Li; Hongguang Wang; Chunliang Liu

doi:10.3390/app122010583

Applied Sciences (Oct 2022)

A Multi-Constrained Optimization Method for THz Backward Wave Oscillators

Rong Bao,
Yongdong Li,
Hongguang Wang,
Chunliang Liu

Affiliations

Rong Bao: Key Laboratory of Physical Electronics and Devices, Ministry of Education, School of Electronic Science and Engineering, Faculty of Electronic and Information Engineering, Xi’an Jiaotong University, Xi’an 710049, China
Yongdong Li: Key Laboratory of Physical Electronics and Devices, Ministry of Education, School of Electronic Science and Engineering, Faculty of Electronic and Information Engineering, Xi’an Jiaotong University, Xi’an 710049, China
Hongguang Wang: Key Laboratory of Physical Electronics and Devices, Ministry of Education, School of Electronic Science and Engineering, Faculty of Electronic and Information Engineering, Xi’an Jiaotong University, Xi’an 710049, China
Chunliang Liu: Key Laboratory of Physical Electronics and Devices, Ministry of Education, School of Electronic Science and Engineering, Faculty of Electronic and Information Engineering, Xi’an Jiaotong University, Xi’an 710049, China

DOI: https://doi.org/10.3390/app122010583
Journal volume & issue: Vol. 12, no. 20
p. 10583

Abstract

Read online

The current design period for various backward wave oscillators (BWOs) is still at least several months. How to find the best structure parameters with an efficient and stable optimization method is a problem facing researchers in both scientific research and engineering work. In this paper, a non-randomized iterative optimization method is proposed. It applies orthogonal design methods to find local solutions that can provide optimal ‘gradient direction’ for several successive next iteration steps. An evaluation function is designed to distinguish the better ones from the local solutions in the multi-constrained optimization of such BWOs. Optimizations from different starting points are performed separately for a global optimal solution. Two BWOs at different frequency ranges are optimized using the proposed method. The validity and stability of the method are verified. It is believed that the method can provide the global optimum and shorten the design period of THz BWOs.

Published in Applied Sciences

ISSN: 2076-3417 (Online)
Publisher: MDPI AG
Country of publisher: Switzerland
LCC subjects: Technology: Engineering (General). Civil engineering (General); Science: Biology (General); Science: Physics; Science: Chemistry
Website: http://www.mdpi.com/journal/applsci

About the journal

Abstract

Keywords