光通信技术2025，Vol.49Issue(5)：89-93,5.DOI:10.13921/j.cnki.issn1002-5561.2025.05.016
基于特征提取的KNN路由优化算法
KNN routing optimization algorithm based on feature extraction
摘要
Abstract
To improve the routing efficiency and communication performance of large-scale Benes optical network,a K-nearest neighbor(KNN)routing optimization algorithm is proposed based on feature extraction.Features such as the waveguide cross-ing position and number are extracted to construct a routing table.In doing so,the traditional KNN algorithm is preprocessed and optimized.An optical network simulation platform is built based on the four-level pulse amplitude modulation(PAM4)modulation system to test and analyze the extinction ratio,bandwidth,and symbol error rate of different routing paths.Experi-mental results show that the proposed method improves the routing screening accuracy from 34.48%to 71.85%.At a transmis-sion rate of 30 Gbit/s,the superior path selected by the improved KNN routing algorithm requires 0.8 dBm less minimum re-ceived power than the inferior path.关键词
Benes光网络/K近邻算法/消光比/带宽/误符号率Key words
Benes optical network/K-nearest neighbor algorithm/extinction ratio/bandwidth/symbol error rate分类
信息技术与安全科学引用本文复制引用
赵莉,石昕宇,孙宗伟..基于特征提取的KNN路由优化算法[J].光通信技术,2025,49(5):89-93,5.基金项目
国家自然科学基金项目(62201338)资助. （62201338）
