数据驱动型光纤多跨长距离传输预测模型研究数据

该数据来源于多跨长距离传输预测模型研究数据，展现了模型各个跨段的序贯训练方式及收敛态势。该数据由Python IDE、Optisystem和Matlab生成。基于级联网络的多跨长距离传输预测模型使用模块级联的方式，使其不仅能够有效预测链路整体的传输，还能预测每个跨段的传输效果。由于每个模块具有多头注意力机制作特征提取，因此模型的误差积累效应较弱。该成果发表于JLT上。支撑论文“Multi-Span Long-Haul Fiber Transmission Model Based on Cascaded Neural Networks With Multi-Head Attention Mechanism”和专利“一种多跨光纤传输信号预测系统的训练方法及装置”。该模型优化后可应对“千公里”级传输需求

This dataset is derived from research data for a multi-span long-haul transmission prediction model, showcasing the sequential training method and convergence trend of each span of the model. It was generated using Python IDE, Optisystem and Matlab. The multi-span long-haul transmission prediction model based on cascaded networks adopts a modular cascaded approach, enabling it to effectively predict not only the overall transmission performance of the optical fiber link but also the transmission effect of each individual span. Given that each module incorporates a multi-head attention mechanism for feature extraction, the model exhibits a relatively weak error accumulation effect. This work has been published in the Journal of Lightwave Technology (JLT). It supports the research paper titled "Multi-Span Long-Haul Fiber Transmission Model Based on Cascaded Neural Networks With Multi-Head Attention Mechanism" and the patent entitled "Training Method and Device for Multi-Span Optical Fiber Transmission Signal Prediction System". After optimization, this model is capable of handling transmission demands at the thousand-kilometer scale.

搜集汇总

背景与挑战

背景概述

该数据集聚焦于数据驱动的光纤多跨长距离传输预测模型研究，基于级联神经网络和多头注意力机制构建，能有效预测链路整体及每个跨段的传输效果，并减少误差积累。数据由Python IDE、Optisystem和Matlab生成，支撑了相关学术论文和专利，优化后适用于千公里级传输需求。