网络运行态势感知技术及其模型

发布时间：2018-05-31 06:33

本文选题：网络态势感知 + 运行态势感知　；参考：《解放军理工大学学报(自然科学版)》2015年01期

【摘要】：针对现有网络态势感知技术无法反映网络整体运行质量的问题,提出了网络运行态势的概念,并给出其模型。首先,将马尔科夫链的转移概率引入网络链路权重评价中以衡量相邻节点的可达性,进而求得网络链路权重;随后,为一致、稳定地评估网络运行质量并克服基于支持向量机评估方法在训练过程中存在的参数难以确定、过拟合等问题,提出了基于相关向量机的网络运行质量评估方法;最后,将各链路的运行质量评估值按照权值加权得到网络运行态势,实现网络运行质量局部评估向网络运行态势整体呈现的提升。实验结果表明,该模型获得的网络态势值与通过灰色模糊综合评估得到的态势值的平均绝对误差仅为1.61%,优于基于支持向量机的网络评估方法;通过对态势均值及方差的比较,可对不同网络建设方案提供必要的技术指导。
[Abstract]:In order to solve the problem that the existing network situation awareness technology can not reflect the overall operation quality of the network, the concept of the network operation situation is proposed and its model is given. Firstly, the transfer probability of Markov chain is introduced into the evaluation of network link weight to measure the reachability of adjacent nodes, and then the network link weight is obtained. To evaluate the network running quality stably and overcome the difficulties of parameter determination and over-fitting in the training process based on support vector machine (SVM), a network quality evaluation method based on correlation vector machine (RVM) is proposed. The network running situation is obtained by weighting the operation quality evaluation value of each link according to the weight value, and the overall performance of the network running situation can be improved by the local evaluation of the network operation quality. The experimental results show that the average absolute error between the network situation value obtained by this model and the situation value obtained by grey fuzzy comprehensive evaluation is only 1.61a, which is superior to the network evaluation method based on support vector machine, and compared with the situation mean value and variance. It can provide necessary technical guidance for different network construction schemes.
【作者单位】：空军工程大学信息与导航学院;
【基金】：陕西省自然科学基础研究计划重点资助项目(2012JZ8005)
【分类号】：TP393.08

【参考文献】