一种基于模式识别的微波功分器性能优化方法
By combining the improved Hopfield neural network model with the lion herd optimization algorithm, the problems of insufficient efficiency and accuracy in traditional microwave power divider design are solved, achieving efficient and accurate microwave power divider performance optimization with the advantages of physical consistency and fast convergence.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HE FEI SHUN ZE TONG XIN KE JI YOU XIAN GONG SI
- Filing Date
- 2025-12-19
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional microwave power divider design methods suffer from reduced efficiency and accuracy in high-dimensional parameter spaces, making it difficult to meet the miniaturization, broadband, and intelligence requirements of modern RF systems. Furthermore, existing pattern recognition-based methods lack physical consistency and interpretability.
An improved Hopfield neural network model combined with a lion pack optimization algorithm is adopted. Through pattern recognition and physical constraint learning, the mapping relationship between structural parameters and performance indicators is realized. A dynamic hunting mechanism and adaptive step size adjustment are introduced. During the optimization process, energy conservation, impedance matching and signal symmetry are taken into account.
It improves the design efficiency and performance accuracy of microwave power dividers, realizes high-precision performance prediction and optimization, and features strong physical consistency, excellent interpretability and fast optimization convergence speed. It is suitable for intelligent design of microwave power dividers and other RF devices.
Smart Images
Figure CN121723955B_ABST