基于神经网络模型的光纤链路传输质量预测模型研究

打开文本图片集

中图分类号：TN913：TP181

文献标志码：A 文章编码：1672-7274（2025）09-0034-03

Research on Fiber Optic Link Transmission Quality Prediction Model Based on NeuralNetwork Model

PAN Zhibin

（CNOOCInformation TechnologyCo.，Ltd.，Shenzhen518o52，China）

Abstract： Fiber optic linksare subject to various factors in actual operation，suchas nonlinear effects，fiber attenuation，environmentalconditions，and equipment performance，which can cause distortion or interruption of transmissionsignals.Accurately predicting the transmisionqualityoffiberoptic links can identify potential problems inadvance，optimizeresource alocation，and effectivelyreduceoperationand maintenance costs.Thisarticle improves the BP neural network by using dynamic learning rate and introducing momentum term，effectively enhancing the prediction performance of fiber optic link transmission quality.

Keywords： fiber optic link; transmission quality; neural network

1 研究背景

光纤通信广泛应用于城域网、骨干网以及接入网，其传输质量直接影响数据通信的可靠性和服务质量。（剩余3120字）