[TuC1] Optical Amplifiers

9:45 AM - 10:00 AM JST(12:45 AM - 1:00 AM UTC)

[TuC1-2] NN-IGBM: A Full-physics Raman Numerical Solution Solving Framework Initialized by Neural Network

○Zihang Wang¹, Yihao Zhang¹, Xiaomin Liu¹, Yicheng Xu¹, Lilin Yi¹, Weisheng Hu¹, Qunbi Zhuge¹ (1. Shanghai Jiao Tong University (China))

Keywords:

Artificial intelligence and machine learning for optical transmission systems and subsystems,Transmission impairments and their mitigations

We propose a full-physics solver framework for Raman amplifiers with third-order pumps. Simulations demonstrate that it improves the convergence rate to 98.37% and simulation speed by 19.85 times on average over 10,000 simulation cases.