基于二叉树算法的5G承载网故障定位方法及实践

Method and Practice of 5G Bearer Network Fault Location Based on Binary Tree Algorithm

2019年运营商提出移动及光网业务的“双提升”专项行动要求,围绕高质量发展主线,贯彻落实“事先化、主动化、自动化”工作要求,持续提升移动网质量和端到端运营能力。采用机器学习中二叉决策树的方法,根据5G承载网的层级关系,自动建立链路、收集告警,梳理告警关系建立决策树模型,按不同的告警设计了对应的推理树,并以此开发推理机,最终实现对故障的快速诊断定位与归并,并在湖北现场予以了正式使用,应用效果良好。

In 2019,operators propose a special action requirement for the"dual promotion"of mobile and optical network services and focus on the main line of developing in the high-quality way,implementing the requirements of''predictable,proactive,and automated",and continuously improving the quality of mobile networks and the end-to-end operational capabilities.According to the hierarchical relationship of 5G bearer network,the method of binary decision tree in machine learning is adopted to automatically establish links,collect alarms,sort out alarm relations.Then the decision tree model is established and the corresponding inference tree is designed according to different alarms,and thus the inference machine is developed.Finally the rapid diagnosis,location and merge of faults are realized,which was formally used in the Hubei field testing with good application effect.