基于SA-BP神经网络算法的光接入网络通道质量评估方法  

A SA-BP neural network algorithm based channel quality evaluation method in optical access network

作　　者：蔡冰清 徐思雅[1] 亓峰[1] 葛维春 周桂平 于波涛 李悦悦 CAI Bingqing;XU Siya;QI Feng;GE Weichun;ZHOU Guiping;YU Botao;LI Yueyue(State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications, Beijing 100876, China;State Grid Liaoning Electric Power Co., Ltd., Shenyang 110006, China;Beijing GuoDianTong Network Technology Co., Ltd., Beijing 100070, China)

机构地区：[1]北京邮电大学网络与交换技术国家重点实验室,北京100876 [2]国网辽宁省电力有限公司,辽宁沈阳110006 [3]北京国电通网络技术有限公司,北京100070

出　　处：《电信科学》2018年第4期162-172,共11页Telecommunications Science

基　　金：国家重点研发计划基金资助项目(No.2016YFB0901200)~~

摘　　要：目前,针对光接入网通道质量评估的研究主要集中在设备层和网络层,缺乏涵盖物理层、网络层和业务层的综合评估方法,致使运维人员难以通过网络监测等方式全面、准确地判断网络通道的实际质量。为解决以上问题,首先对影响光接入网通道质量的关键因素进行了深入分析,提出了面向多层次、多指标的光接入网通道质量综合评估模型,然后设计了SA-BP神经网络算法对多指标参数进行训练,从而实现通道质量的准确评估。通过仿真表明,提出的评估方法具有较高的评估精度和稳定度,便于提升网络运维的质量和效率。So far, existing works related to quality evaluation of channel mainly focused on equipment layer and network layer in optical access network. Most of researches lack synthesis evaluation methods covering physical layer, network layer and business layer. As a result, it is hard for administrators to judge the actual quality of channel by network monitoring completely and accurately. To solve this problem, key influence factors of channel quality in optical access network was analyzed firstly, and a comprehensive multi-layer-multi-index evaluation model of channel quality was proposed. Then a SA-BP neural network algorithm was designed to train the parameters of the proposed evaluation model to make the result more accurate. The simulation results show that the proposed method has higher accuracy and stability, and can improve the quality and efficiency of network operation and maintenance.

关 键 词：光接入网络 通道质量 评估模型 SA-BP神经网络算法 

分 类 号：G304[文化科学]