Hybrid Variational Quantum Algorithm Enhanced Subentire-Domain Basis Functions Method With High Learning Efficiency and Better Robustness
- 13 June 2025
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 73 (9) , 6707-6717
Abstract
The subentire-domain (SED) basis functions method is the most effective method for analyzing the electromagnetic properties of large-scale finite periodic structures (LFPSs). Recently, artificial neural networks have been employed to accelerate this method by rapidly predicting the expansion coefficients of SED basis functions without filling mutual coupling matrices. However, the training processes of prediction models can be further improved due to its classical computational paradigm. In this paper, a novel variational quantum algorithm enhanced SED basis functions method is proposed and the quantum computing paradigm is utilized to analyze LFPSs for the first time. In our algorithm, the array features are expanded and encoded onto few qubits, and the resulting quantum state is unitarily transformed into expansion coefficients by the parameterized quantum circuit. In addition, the algorithm is deployed on the quantum simulator for numerical experiments. The experimental results demonstrate that our method can accurately and quickly analyze LFPSs. Furthermore, the quantum-inspired models achieve 27%-62% improvements in learning efficiency for corner and edge cells, and 22%-59% improvements in robustness for all types of cells.Keywords
Funding Information
- Key Program of the National Natural Science Foundation of China (62231001)
- Special Fund for Key Basic Research in Jiangsu Province (BK20243015)
- Fundamental Research Funds for the Central Universities (2242022k60004)
- National Science Fund for Distinguished Young Scholars of China (61925103)
- National Natural Science Foundation of China (62101115)
This publication has 37 references indexed in Scilit:
- Variational quantum algorithmsNature Reviews Physics, 2021
- Training deep quantum neural networksNature Communications, 2020
- Supervised learning with quantum-enhanced feature spacesNature, 2019
- Quantum machine learningNature, 2017
- Deep learningNature, 2015
- Quantum Computation and Quantum Information2012
- A subdomain multilevel approach for the efficient MoM analysis of large planar antennasMicrowave and Optical Technology Letters, 2000
- Frequency Selective Surfaces2000
- Multilevel fast‐multipole algorithm for solving combined field integral equations of electromagnetic scatteringMicrowave and Optical Technology Letters, 1995
- Learning representations by back-propagating errorsNature, 1986