WO2022126705A1 - 一种电信核心网故障干预时机决策方法 - Google Patents

一种电信核心网故障干预时机决策方法 Download PDF

Info

Publication number
WO2022126705A1
WO2022126705A1 PCT/CN2020/139251 CN2020139251W WO2022126705A1 WO 2022126705 A1 WO2022126705 A1 WO 2022126705A1 CN 2020139251 W CN2020139251 W CN 2020139251W WO 2022126705 A1 WO2022126705 A1 WO 2022126705A1
Authority
WO
WIPO (PCT)
Prior art keywords
network
fault
core network
intervention
timing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2020/139251
Other languages
English (en)
French (fr)
Inventor
叶可江
林彦颖
须成忠
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Publication of WO2022126705A1 publication Critical patent/WO2022126705A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/04—Arrangements for maintaining operational condition
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Definitions

  • the invention belongs to the technical field of telecommunications, and in particular relates to a decision method for fault intervention timing of a telecommunications core network based on an uncertainty deep neural network.
  • the core telecom network is an important telecom infrastructure for data exchange. It consists of a large number of important equipment such as routers, switches and firewalls, interconnecting various networks and providing a path for exchanging data between different LANs or subnets , whose main function is routing and switching.
  • the fault propagation range will also be from a local diffusion effect (a certain area or a certain network type) to a very extensive diffusion;
  • the network itself has certain adjustment capabilities, such as automatically modifying the topology of the network to allow some traffic to pass through the normal link, thereby avoiding network disconnection.
  • This adjustment is very efficient, but on the one hand, the recovery period of the faulty link is long, and when the overall load of the core network is high, this adjustment method is very dangerous. It may drastically increase the load of other links and increase the failure probability of other links.
  • the embodiments of the present invention provide a telecommunication core network fault intervention timing decision method, so as to solve the problem in the prior art that the prior art does not take the network as a whole as a consideration, but focuses on the failure of a certain device or a certain protocol. Effective modeling of complex and uncertain behaviors in the network cannot determine the appropriate timing of operation and maintenance intervention, which may lead to unnecessary overhead or problems of fault propagation.
  • a telecommunication core network fault intervention timing decision-making method characterized by comprising the following steps:
  • S3 Use the Bayesian inference process of deep Bayesian neural networks to make decisions on the timing of network interventions.
  • the fault information reported by the core network includes the fault unit, the fault cause, and the fault level.
  • the network state is modeled by using the fault information reported by the core network within a period of time, and the modeling formula is as follows:
  • x ⁇ [1, 4] represents the fault level of the equipment
  • w represents the amount of fault alarms per unit time
  • si and k are a set of constant coefficients used to adjust the weights.
  • the weight parameter of the deep Bayesian neural network in the step S2 is a random variable; the output of the deep Bayesian neural network is used to describe the likelihood of a specific probability distribution.
  • the Bayesian inference process of the deep Bayesian neural network adopts the following formula:
  • Di represents the ith time period
  • ⁇ is 0 ⁇ 1
  • KL is the KL divergence. It is defined that the same random variable x has two separate probability distributions p(x) and q(x), using the KL divergence To measure the difference between p(x) and q(x), the calculation formula of KL divergence is:
  • the present invention has the following beneficial effects:
  • Fig. 1 is a flow chart of a method for determining the timing of fault intervention in a telecom core network provided by the present invention
  • Fig. 2 is a network state quantization diagram provided by the present invention.
  • Fig. 3 is a linear Bayesian neural network structure diagram provided by the present invention.
  • Fig. 4 is the weight distribution diagram of Bayesian deep neural network provided by the present invention.
  • FIG. 5 is a comparison diagram of a Bayesian deep neural network provided by the present invention and a simple neural network method
  • FIG. 6 is a result diagram of a single network operation and maintenance intervention decision-making experiment of the telecommunication core network fault intervention timing decision method provided by the present invention
  • FIG. 7 is a result diagram of an experiment of repeating network operation and maintenance intervention decision-making for the method for determining the timing of fault intervention in a telecom core network provided by the present invention.
  • the telecommunication core network fault intervention timing decision method according to the embodiment of the present invention is implemented based on the uncertainty deep neural network.
  • Figure 5 shows the comparison between the regression of the Bayesian neural network and the regression of the simple neural network method. And the importance of measurement uncertainty is explained. While both methods perform well over the range of training data, in cases where extrapolation is required, the probabilistic method provides the full distribution of the function's output rather than the point estimates provided by the neural network. The output distributions provided by probabilistic methods allow the development of credible models because they can identify uncertainties in predictions.
  • the method for deciding the timing of fault intervention in a telecom core network includes the following steps:
  • the fault information reported by the core network includes information such as the fault unit, fault cause, and fault level.
  • the modeling formula is as follows:
  • x ⁇ [1, 4] represents the fault level of the device.
  • level 1 indicates that the network condition is good, the higher the level, the worse the network status is, and the level 4 indicates that the network status has been serious
  • w means per unit time
  • si and k are a set of constant coefficients used to adjust the weights, and the values of si and k depend on the divergence of the specific data.
  • FIG. 2 visually shows the network status.
  • S2 Model the randomness and uncertainty of the overall state change of the core network using a deep Bayesian neural network with uncertain feature modeling.
  • the deep Bayesian neural network differs from traditional deep neural networks such as convolutional neural networks as follows:
  • the weight parameter of the deep Bayesian neural network is a random variable, not a definite value. As shown in Figure 3, it combines probability modeling and neural network, and can give the confidence of the prediction result. Its priors are used to describe key parameters and serve as the input to the neural network.
  • the output of the neural network is used to describe the likelihood of a particular probability distribution.
  • the posterior distribution is computed by sampling or variational inference. This is critical for many problems, and Bayesian Neural Networks (BNNs) are very robust due to their uncertainty quantification capabilities.
  • Di represents the ith time period
  • It refers to the cross entropy function (Cross Entropy)
  • KL refers to the KL divergence
  • ⁇ is the calculation constraint
  • the value range of ⁇ is 0 ⁇ 1. Therefore, the prior distribution is a proportional mixture of two Gaussian distributions.
  • KL divergence also known as relative entropy, defines that the same random variable x has two separate probability distributions p(x) and q(x).
  • KL divergence Kullback-Leibler(KL) divergence

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Telephonic Communication Services (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明适用于电信技术领域，提供了电信核心网故障干预时机决策方法，包括以下步骤：利用一段时间内的核心网络上报的故障信息对网络状态进行建模；利用具有不确定性特征建模的深度贝叶斯神经网络对核心网络整体状态变化的随机性和不确定性进行建模：利用深度贝叶斯神经网络的贝叶斯推断过程对网络干预时机进行决策。本发明将电信网络整体作为考量，利用具有不确定性特征建模的深度贝叶斯神经网络对核心网络整体状态变化的随机性和不确定性进行建模的方法，利用不确定性深度神经网络对电信核心网络中复杂的不确定性行为进行了有效的建模，并利用深度贝叶斯神经网络的贝叶斯推断过程对网络干预时机进行决策的方法进行了有效的决策。

Description

一种电信核心网故障干预时机决策方法 技术领域
本发明属于电信技术领域，尤其涉及一种基于不确定性深度神经网络的电信核心网故障干预时机决策方法。
背景技术
核心电信网络是用于数据交换的重要电信基础架构，它由大量的路由器，交换机和防火墙等重要设备组成，互连各种网络，为在不同的LAN或子网之间交换数据提供了一条路径，其主要功能是路由和交换。
随着云计算、5G/6G、物联网等业务的迅猛发展，目前电信网络的核心网络面临着以下几个趋势：
1)、核心网内的设备种类从较为单一迅速增长为多样化，体现为多厂家、多协议、多设备制式等，由此带来对电信核心网的整体状态较为难以把握；
2)、完成一个业务所需要的设备规模从较少变为需要大规模设备协同，体现为跨区域、跨网络类型、跨通信协议；
3)、设备分布从较为集中式的核心网络机房到分布式的网络机房，较为显著的是5G的随着从“通信站”走向“社会站”，大量小微站点将与广泛的社会资源融为一体；
4)、故障传播范围也将从局部的扩散影响(某个地区或者某种网络类型)到非常广泛的扩散；
5)、网络整体的变化将呈现不确定性的趋势。
因此，这种核心网络规模和复杂性的扩张带来一个显著的问题是网络稳定性难以控制，运维成本显著上升。
通常情况下，网络自身有一定的调节能力，比如自动修改网络的拓扑结构，使部分流量往正常的链路通过，从而避免网络断路。这种调节是十分高效的，但一方面故障链路的恢复周期较长，并且在核心网络整体负载较高的情况下，这种调节方式十分危险。其可能剧烈增大其他链路的负载，提高其他链路的故障概率。
故在高负载的情况下通常由经验丰富的运维工程师来处理网络故障，这带来两个当前非常受到关注的问题：
第一，电信核心网络的整体状态如何量化。目前网络由于上述分析的几个趋势可以得出，网络状态难以量化。
第二，什么时候可以对网络进行运维干预，如果过早地对网络进行人工干预，此时网络性能瓶颈还远没有产生，此时运维将会导致十分显著的额外开销，并且对接下来的整体状态继续变化的趋势并无明显帮助；而如果干预不及时，网络性能瓶颈产生，将引发网络故障大规模扩散，所以核心问题是干预的时机。
发明内容
有鉴于此，本发明实施例提供了电信核心网故障干预时机决策方法，以解决现有技术中现有技术没有将网络整体作为考量，而注重某一种设备或者某一种协议故障情况，无法对网络中复杂的不确定性行为进行有效建模，无法确定较为妥当的运维干预时机，可能导致不必要的开销或者故障扩散的问题。
为达到上述目的，本发明提供的方案是：
一种电信核心网故障干预时机决策方法，其特征在于，包括以下步骤：
S1：利用一段时间内的核心网络上报的故障信息对网络状态进行建模；
S2：利用具有不确定性特征建模的深度贝叶斯神经网络对核心网络整体状态变化的随机性和不确定性进行建模；
S3：利用深度贝叶斯神经网络的贝叶斯推断过程对网络干预时机进行决策。
作为一种改进方式，所述步骤S1中，所述核心网络上报的故障信息包括故 障单元、故障原因、故障等级。
作为一种改进方式，所述利用一段时间内的核心网络上报的故障信息对网络状态进行建模，建模公式如下：
Figure PCTCN2020139251-appb-000001
其中，x∈[1，4]表示设备的故障等级，w表示每单位时间的故障警报量，si和k是用于调整权重的一组常数系数。
作为一种改进方式，所述步骤S2中的深度贝叶斯神经网络的权重参数是随机变量；所述深度贝叶斯神经网络的输出用来描述特定的概率分布的似然。
作为一种改进方式，所述深度贝叶斯神经网络的贝叶斯推断过程采用如下公式：
Figure PCTCN2020139251-appb-000002
其中，Di代表第i个时间段，
Figure PCTCN2020139251-appb-000003
为交叉熵函数，β的取值范围为0<β<1，KL为KL散度，定义同一个随机变量x有两个单独的概率分布p(x)和q(x)，使用KL散度来衡量p(x)和q(x)的差异，则KL散度的计算公式为：
Figure PCTCN2020139251-appb-000004
对KL散度的计算公式进行变形可得：
Figure PCTCN2020139251-appb-000005
KL的值越小，表示q(x)分布和p(x)分布越接近。
本发明与现有技术相比存在的有益效果是：
(1)将电信网络整体作为考量，利用具有不确定性特征建模的深度贝叶斯神经网络对核心网络整体状态变化的随机性和不确定性进行建模的方法，从整体性出发对网络的状态进行可靠的建模；
(2)利用深度贝叶斯神经网络的贝叶斯推断过程对网络干预时机进行决策的方法进行了有效的决策。
附图说明
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1是本发明提供的电信核心网故障干预时机决策方法的流程框图；
图2是本发明提供的网络状态量化图；
图3是本发明提供的线性贝叶斯神经网络结构图；
图4是本发明提供的贝叶斯深度神经网络的权重分布图；
图5是本发明提供的贝叶斯深度神经网络与简单的神经网络方法对比图；
图6是本发明提供的电信核心网故障干预时机决策方法的单次网络运维干预决策实验结果图；
图7是本发明提供的电信核心网故障干预时机决策方法的重复多次网络运维干预决策实验结果图。
具体实施方式
以下描述中，为了说明而不是为了限定，提出了诸如特定系统结构、技术之类的具体细节，以便透彻理解本发明实施例。然而，本领域的技术人员应当 清楚，在没有这些具体细节的其它实施例中也可以实现本发明。在其它情况中，省略对众所周知的系统、装置、电路以及方法的详细说明，以免不必要的细节妨碍本发明的描述。
为了说明本发明所述的技术方案，下面通过具体来进行说明。
本发明实施例的电信核心网故障干预时机决策方法是基于不确定性深度神经网络实现的，下图5示出了贝叶斯神经网络的回归与简单的神经网络方法的回归的进行了比较，并说明了测量不确定度的重要性。虽然这两种方法在训练数据的范围内都表现良好，在需要外推法的情况下，概率方法提供了函数输出的完整分布，而不是由神经网络提供的点估计。概率方法提供的输出分布允许开发可信的模型，因为它们可以识别预测中的不确定性。
如图1所示，本发明实施例的电信核心网故障干预时机决策方法包括如下步骤：
S1:利用一段时间内的核心网络上报的故障信息对网络状态进行建模，通过故障信息来量化网络整体状态。具体地，核心网络上报的故障信息包括故障单元、故障原因、故障等级等信息，建模公式如下：
Figure PCTCN2020139251-appb-000006
其中，x∈[1，4]表示设备的故障等级，设备的故障等级中，级别1表示网络状况良好，级别越高表示网络状态越差，级别4表示网络状态已经严重；w表示每单位时间的故障警报量；si和k是用于调整权重的一组常数系数，si和k的取值取决于具体数据的散度情况。附图2直观地示出了网络状态情况。
S2：利用具有不确定性特征建模的深度贝叶斯神经网络对核心网络整体状态变化的随机性和不确定性进行建模。
需要说明的是，深度贝叶斯神经网络与传统的深度神经网络如卷积神经网络有如下区别：
深度贝叶斯神经网络其权重参数是随机变量，而非确定的值，如图3所示，它是通过概率建模和神经网络结合起来，并能够给出预测结果的置信度。其先验用来描述关键参数，并作为神经网络的输入。
神经网络的输出用来描述特定的概率分布的似然。通过采样或者变分推断来计算后验分布。这对于很多问题来说非常关键，由于贝叶斯神经网络(BNN)具有不确定性量化能力，所以具有非常强的鲁棒性。
因此，在建模过程中，使用贝叶斯神经网络为基础，构建线性贝叶斯神经网络(如图3所示)。
S3：利用深度贝叶斯神经网络的贝叶斯推断过程对网络干预时机进行决策，贝叶斯推断过程采用如下公式：
Figure PCTCN2020139251-appb-000007
其中，Di代表第i个时间段，
Figure PCTCN2020139251-appb-000008
是指交叉熵函数(Cross Entropy)，KL是指KL散度，β为计算约束，β的取值范围为0<β<1。因此，先验分布为两个高斯分布组成的比例混合。
KL散度又称相对熵,定义同一个随机变量x有两个单独的概率分布p(x)和q(x)，我们可以使用KL散度(Kullback-Leibler(KL)divergence)来衡量这两个分布的差异，KL散度的计算公式：
Figure PCTCN2020139251-appb-000009
KL的值越小，表示q分布和p分布越接近。对KL散度的计算的公式进行变形可得：
Figure PCTCN2020139251-appb-000010
等式的后一部分即是前文所述交叉熵函数，即：
Figure PCTCN2020139251-appb-000011
本发明实施例的电信核心网故障干预时机决策方法有以下优点：
(1)将电信网络整体作为考量，利用具有不确定性特征建模的深度贝叶斯神经网络对核心网络整体状态变化的随机性和不确定性进行建模的方法，从整体性出发对网络的状态进行可靠的建模；
(2)利用深度贝叶斯神经网络的贝叶斯推断过程对网络干预时机进行决策的方法进行了有效的决策。
本实施例的电信核心网故障干预时机决策方法进行了充分的实验，效果显著且被证明可行。实验结果如下图所示：
(1)单次“网络运维干预决策”实验准确率达到95.21％，实验结果如图6所示，图中横纵坐标表示实验中“一个决策事件”，颜色中深色代表“不进行运维干预”，浅色代表“进行运维干预”。
(2)重复100次实验，均值在92.271％以上，实验结果如图7所示，图中横坐标表示“实验的轮数”，纵坐标表示“每轮实验的准确率”。
以上所述实施例仅用以说明本发明的技术方案，而非对其限制；尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本发明各实施例技术方案的精神和范围，均应包含在本发明的保护范围之内。

Claims (5)

  1. 一种电信核心网故障干预时机决策方法，其特征在于，包括以下步骤：
    S1：利用一段时间内的核心网络上报的故障信息对网络状态进行建模；
    S2：利用具有不确定性特征建模的深度贝叶斯神经网络对核心网络整体状态变化的随机性和不确定性进行建模；
    S3：利用深度贝叶斯神经网络的贝叶斯推断过程对网络干预时机进行决策。
  2. 如权利要求1所述的电信核心网故障干预时机决策方法，其特征在于：所述步骤S1中，所述核心网络上报的故障信息包括故障单元、故障原因和故障等级。
  3. 如权利要求2所述的电信核心网故障干预时机决策方法，其特征在于：所述利用一段时间内的核心网络上报的故障信息对网络状态进行建模，建模公式如下：
    Figure PCTCN2020139251-appb-100001
    其中，x∈[1，4]表示设备的故障等级，w表示每单位时间的故障警报量，si和k是用于调整权重的一组常数系数。
  4. 如权利要求1所述的电信核心网故障干预时机决策方法，其特征在于，所述步骤S2中，所述深度贝叶斯神经网络的权重参数是随机变量；所述深度贝叶斯神经网络的输出用来描述特定的概率分布的似然。
  5. 如权利要求1所述的电信核心网故障干预时机决策方法，其特征在于，所述步骤S3中，所述深度贝叶斯神经网络的贝叶斯推断过程采用如下公式：
    Figure PCTCN2020139251-appb-100002
    其中，Di代表第i个时间段，
    Figure PCTCN2020139251-appb-100003
    为交叉熵函数，β的取值范围为0<β<1，KL为KL散度，定义同一个随机变量x有两个单独的概率分布p(x)和q(x)，使用KL散度来衡量p(x)和q(x)的差异，则KL散度的计算公式为：
    Figure PCTCN2020139251-appb-100004
    对KL散度的计算公式进行变形可得：
    Figure PCTCN2020139251-appb-100005
    KL的值越小，表示q(x)分布和p(x)分布越接近。
PCT/CN2020/139251 2020-12-18 2020-12-25 一种电信核心网故障干预时机决策方法 Ceased WO2022126705A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011510414.0 2020-12-18
CN202011510414.0A CN112637879A (zh) 2020-12-18 2020-12-18 一种电信核心网故障干预时机决策方法

Publications (1)

Publication Number Publication Date
WO2022126705A1 true WO2022126705A1 (zh) 2022-06-23

Family

ID=75318134

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/139251 Ceased WO2022126705A1 (zh) 2020-12-18 2020-12-25 一种电信核心网故障干预时机决策方法

Country Status (2)

Country Link
CN (1) CN112637879A (zh)
WO (1) WO2022126705A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101917297A (zh) * 2010-08-30 2010-12-15 烽火通信科技股份有限公司 基于贝叶斯网络的核心网故障诊断方法及系统
US20190215230A1 (en) * 2018-01-08 2019-07-11 Cisco Technology, Inc. Analyzing common traits in a network assurance system
CN110738242A (zh) * 2019-09-25 2020-01-31 清华大学 一种深度神经网络的贝叶斯结构学习方法及装置
CN111368888A (zh) * 2020-02-25 2020-07-03 重庆邮电大学 基于深度动态贝叶斯网络的服务功能链故障诊断方法

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101567814B (zh) * 2009-05-22 2011-03-30 清华大学 基于SNMP和随机Petri网的自动化网络管理方法
WO2019076463A1 (en) * 2017-10-19 2019-04-25 Nokia Technologies Oy EXECUTING ANALYSIS AND MANAGEMENT OF A MOBILE COMMUNICATION NETWORK BASED ON PERFORMANCE INFORMATION, CONFIGURATION INFORMATION, AND ENVIRONMENTAL INFORMATION

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101917297A (zh) * 2010-08-30 2010-12-15 烽火通信科技股份有限公司 基于贝叶斯网络的核心网故障诊断方法及系统
US20190215230A1 (en) * 2018-01-08 2019-07-11 Cisco Technology, Inc. Analyzing common traits in a network assurance system
CN110738242A (zh) * 2019-09-25 2020-01-31 清华大学 一种深度神经网络的贝叶斯结构学习方法及装置
CN111368888A (zh) * 2020-02-25 2020-07-03 重庆邮电大学 基于深度动态贝叶斯网络的服务功能链故障诊断方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
JIA, JINLIANG ET AL.: "Research on fault diagnosis and testing method of VoLTE core network", COMMUNICATIONS MANAGEMENT AND TECHNOLOGY, no. 4, 15 August 2017 (2017-08-15), pages 49 - 52, XP055942421 *

Also Published As

Publication number Publication date
CN112637879A (zh) 2021-04-09

Similar Documents

Publication Publication Date Title
US11669751B2 (en) Prediction of network events via rule set representations of machine learning models
US10536344B2 (en) Privacy-aware model generation for hybrid machine learning systems
Guo et al. Capturing spatial–temporal correlations with Attention based Graph Convolutional Network for network traffic prediction
Chen et al. Building autonomic elastic optical networks with deep reinforcement learning
CN120321136B (zh) 基于图神经网络的智能优化与资源动态重分配方法和系统
CN117896163B (zh) 一种网络流量监控系统
CN115423041A (zh) 一种基于深度学习的边缘云故障预测方法及系统
CN120614616A (zh) 一种电力5g网络服务质量监测方法、系统及存储介质
Islam et al. Software-defined networking-based resilient proactive routing in smart grids using graph neural networks and deep Q-networks
CN116562631B (zh) 一种电力光纤通信网络风险量化评价方法、系统和设备
Ghaderi et al. Learning traffic encoding matrices for delay-aware traffic engineering in SD-WANs
Memon et al. Design and Implementation of a Robust Convolutional Neural Network‐Based Traffic Matrix Estimator for Cloud Networks
Fang et al. A novel network delay prediction model with mixed multi-layer perceptron architecture for edge computing
US7363386B1 (en) Method for estimating a traffic matrix in a communication network using non-stationary data
WO2022126705A1 (zh) 一种电信核心网故障干预时机决策方法
Zhang et al. A novel virtual network fault diagnosis method based on long short-term memory neural networks
CN111901134A (zh) 一种基于循环神经网络模型rnn的预测网络质量的方法和装置
US10812206B2 (en) Method and apparatus for predicting the bitrate of a repaired communication channel
CN119939378A (zh) 基于数字孪生和复杂网络理论的电力通信网风险评估方法
Barrera et al. Statistical detection of congestion in routers
Mesquita et al. Traffic matrix prediction for optical networks
CN108418724A (zh) 基于云计算的下一代关键信息基础设施网络智能管理系统
CN118200223B (zh) 基于损失数据分类分级的网络链路选择方法、系统及装置
CN115473797B (zh) 一种基于端到端业务性能指标的智能网络故障定位方法
CN118337694B (zh) 一种基于服务功能链的路由优化方法和系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20965718

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20965718

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 20965718

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 20965718

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC ( EPO FORM 1205A DATED 28/05/2024 )