WO2023236548A1 - 网络配置参数的调整方法、装置及网管系统 - Google Patents

网络配置参数的调整方法、装置及网管系统 Download PDF

Info

Publication number
WO2023236548A1
WO2023236548A1 PCT/CN2023/073441 CN2023073441W WO2023236548A1 WO 2023236548 A1 WO2023236548 A1 WO 2023236548A1 CN 2023073441 W CN2023073441 W CN 2023073441W WO 2023236548 A1 WO2023236548 A1 WO 2023236548A1
Authority
WO
WIPO (PCT)
Prior art keywords
cells
network configuration
configuration parameters
feature vectors
target
Prior art date
Application number
PCT/CN2023/073441
Other languages
English (en)
French (fr)
Inventor
杨磊
饶慧珍
Original Assignee
中兴通讯股份有限公司
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 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2023236548A1 publication Critical patent/WO2023236548A1/zh

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06—Management of faults, events, alarms or notifications
    • H04L41/0654—Management of faults, events, alarms or notifications using network fault recovery
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0803—Configuration setting
    • H04L41/0813—Configuration setting characterised by the conditions triggering a change of settings
    • H04L41/0816—Configuration setting characterised by the conditions triggering a change of settings the condition being an adaptation, e.g. in response to network events
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/04—Arrangements for maintaining operational condition
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L61/00—Network arrangements, protocols or services for addressing or naming

Definitions

  • the present disclosure relates to the field of wireless communication technology, and in particular, to a method, device and network management system for adjusting network configuration parameters.
  • Network reliability is one of the important indicators for evaluating network performance.
  • classification and hierarchical decision trees are usually used to detect faults in the network.
  • the present disclosure provides a method, device and network management system for adjusting network configuration parameters.
  • the present disclosure provides a method for adjusting network configuration parameters.
  • the method includes: obtaining feature vectors corresponding to N cells with a target fault type, wherein the feature vectors corresponding to the N cells are represented by
  • the target fault type is any fault type among multiple fault types, and N is an integer greater than or equal to 1; according to the fault type of each of the N cells
  • the geographical location is clustered, and the N cells are divided into M areas, where M is an integer greater than or equal to 1 and less than or equal to N; regression analysis is performed on the network configuration parameters of each area in the M areas, Obtain the adjustment value of the network configuration parameters of each area, the network configuration parameters are parameters associated with the target fault type, and the target fault type is determined based on the feature vectors corresponding to the N cells; and
  • the network configuration parameters of each area are adjusted based on the adjustment values of the network configuration parameters of each area.
  • the present disclosure also provides a device for adjusting network configuration parameters.
  • the device includes: an acquisition module configured to acquire feature vectors corresponding to N cells with target fault types, wherein the N The feature vector corresponding to the cell is used to characterize the fault types existing in the N cells.
  • the target fault type is any fault type among multiple fault types, and N is an integer greater than or equal to 1;
  • the clustering module configured to perform clustering according to the geographical location of each of the N cells, and divide the N cells into M areas, where M is an integer greater than or equal to 1 and less than or equal to N;
  • the analysis module is configured to perform network configuration parameters of each of the M areas. Perform regression analysis to obtain the adjustment values of network configuration parameters in each area.
  • the network configuration parameters are parameters associated with the target fault type.
  • the target fault type is based on the feature vectors corresponding to the N cells. and an adjustment module configured to adjust the network configuration parameters of each area based on the adjustment value of the network configuration parameters of each area.
  • the present disclosure also provides a network management system, including a processor, a communication interface, a memory, and a communication bus.
  • the processor, the communication interface, and the memory complete communication with each other through the communication bus; the memory is used to store the computer.
  • the present disclosure also provides a computer-readable storage medium on which a computer program is stored.
  • the computer program is executed by a processor, the method for adjusting network configuration parameters as described in the first aspect is implemented.
  • Figure 1 is a schematic flowchart of a method for adjusting network configuration parameters provided by an embodiment of the present disclosure
  • Figure 2 is a schematic structural diagram of an apparatus for adjusting network configuration parameters provided by an embodiment of the present disclosure.
  • Figure 3 is a schematic structural diagram of a network management system provided by an embodiment of the present disclosure.
  • the classification and hierarchical decision tree method can effectively diagnose network faults, the diagnostic conclusions are relatively simple and cannot form regular subsequent adjustments to network configuration parameters. Therefore, after a network fault is diagnosed based on a classified and hierarchical decision tree, the network configuration parameters still need to be manually adjusted based on human experience, making the adjustment of network configuration parameters less timely and accurate.
  • Figure 1 is a schematic flowchart of a method for adjusting network configuration parameters provided by an embodiment of the present disclosure. As shown in Figure 1, the method for adjusting network configuration parameters may include the following steps 101 to 104.
  • Step 101 Obtain the feature vectors corresponding to the N cells with the target fault type, where the feature vectors corresponding to the N cells are used to characterize the fault types existing in the N cells, and the target fault type is one of multiple fault types.
  • N is an integer greater than or equal to 1.
  • the method for adjusting network configuration parameters can be applied to network management systems or other electronic devices installed with network management type software, and is not specifically limited in this disclosure.
  • network management system For convenience of explanation, subsequent embodiments will take the network management system as an example for explanation.
  • the above target fault type is any fault type among multiple fault types existing in the wireless network.
  • the target fault type can be a first-level fault type, such as access fault type, call drop fault type, intra-system switching fault type, etc.; it can also be a second-level fault type, such as transmission fault type under access fault type. , coverage fault type, capacity fault type, etc.; it can also be a three-level fault type, such as weak coverage fault type, uplink interference fault type, downlink interference fault type, etc. under the coverage fault type.
  • the target fault type can be flexibly set according to the fault type that actually needs to be detected, and is not specifically limited in this disclosure.
  • the above feature vectors can be classified according to the first-level fault types.
  • the access fault type corresponds to the feature vector of the access fault type
  • the call drop fault type corresponds to the feature vector of the call drop fault type
  • the intra-system switching fault type corresponds to the system Characteristic vectors of internal switching fault types, etc.
  • the length of each feature vector can be determined based on the number of all sub-fault types (including second-level fault types, third-level fault types,%) included in the first-level fault type.
  • obtaining the feature vectors corresponding to the N cells with the target fault type is to obtain the feature vectors of the N cells with the first-level fault type; when the target fault type is a certain sub-unit
  • selecting a fault type it is necessary to first obtain the feature vector corresponding to the first-level fault type to which the sub-fault type exists in each cell, and then select the feature vector of the sub-fault type based on the feature vector corresponding to the first-level fault type.
  • Step 102 Cluster according to the geographical location of each of the N cells, and divide the N cells into M areas, where M is an integer greater than or equal to 1 and less than or equal to N.
  • the distance between any two cells can be calculated based on the geographical location of each cell in the N cells, and then the cells with closer distances can be clustered in Together, in this way, N cells can be divided into M areas, where each area includes at least one cell. Since the types of network faults in cells that are geographically close are similar and the network configuration parameters that need to be adjusted are also similar, the network configuration parameters of cells belonging to the same area can be adjusted together, thereby improving the efficiency of network configuration parameter adjustment of the cells.
  • the distance between any two cells such as Euclidean distance, Bachchan distance, Mahalanobis distance, etc.
  • Step 103 Perform regression analysis on the network configuration parameters of each area in the M areas to obtain the adjustment values of the network configuration parameters of each area.
  • the network configuration parameters are parameters associated with the target fault type.
  • the target fault type is based on N
  • the characteristic vector corresponding to the cell is determined.
  • the target fault type can be determined based on the feature vectors corresponding to the N cells, and then the network configuration parameters associated with the target fault type can be determined based on the target fault type. For example, suppose the goal If the fault type is an access fault type, then the network configuration parameters can be the antenna transmit power and antenna pitch angle that are related to the access fault type. Therefore, the antenna transmit power and antenna pitch angle of each area can be Perform regression analysis on network configuration parameters to obtain the adjustment values of network configuration parameters such as antenna transmit power and antenna pitch angle in each area.
  • Step 104 Adjust the network configuration parameters of each area based on the adjustment values of the network configuration parameters of each area.
  • the network configuration parameters of each area can be adjusted directly based on the adjustment values of the network configuration parameters of each area, so that the network configuration parameters are optimized.
  • the fault type existing in each cell can be identified based on the feature vector of each cell.
  • clustering can also be performed based on the geographical location of the N cells, and regression analysis can be performed on the network configuration parameters of each clustered area to obtain the adjustment values of the network configuration parameters, and Network configuration parameters are automatically adjusted based on the adjustment value without relying on personal experience, thereby improving the timeliness and accuracy of network configuration parameter adjustment.
  • the above-mentioned step 101 obtaining the feature vectors corresponding to the N cells with the target fault type, includes: acquiring the feature vectors of multiple fault types of the L cells, where L is an integer greater than N, and the N cells are L N cells in the community; delete the feature vectors corresponding to normal cells among the obtained feature vectors of L cells, and obtain the feature vectors of K cells with faults.
  • K is greater than or equal to N, and less than or An integer equal to L; and filtering the feature vectors of K cells to obtain feature vectors corresponding to N faulty cells.
  • the above L cells can be understood as all cells within the entire wireless network coverage area that need to be analyzed for network faults.
  • the L cells include the N cells in which the target fault type exists.
  • the above-mentioned K cells refer to the cells with faults except normal cells among the L cells.
  • the K cells include the above-mentioned N cells with the target fault type and K-N cells with other fault types.
  • feature vectors of multiple different fault types of L cells can be first obtained, and then based on the obtained feature vectors, the feature vectors of normal cells in the L cells are screened out and deleted, and the feature vectors with faults are retained.
  • Feature vectors of K cells and then filter the feature vectors of K cells with faults based on fault types to obtain feature vectors corresponding to N faulty cells with target fault types.
  • the feature vectors corresponding to fault-free cells can be screened out, and the feature vectors of cells with the target fault type can be further accurately located, which greatly improves the positioning efficiency of the target cells and reflects the performance data vectorized processing.
  • the above steps of obtaining feature vectors of multiple fault types of L cells include: obtaining input data corresponding to multiple fault types of L cells, where the input data includes multiple fields and values corresponding to each field. , different fields represent different network performance parameters; determine whether the value corresponding to the field meets the preset conditions corresponding to the fields matching the fields in the preset fault type classification table, where the preset fault type classification table includes multiple fault types and the preset conditions required to meet multiple fault types.
  • the preset conditions are the value range of at least one field; in the field corresponding When the value satisfies the preset condition corresponding to the field, the flag position corresponding to the field is set to 1; when the value corresponding to the field does not meet the preset condition corresponding to the field, the flag position corresponding to the field is set to 0; and according to each field
  • the corresponding flag bits determine the feature vectors of multiple fault types of L cells.
  • the feature vector of each cell can be obtained according to the input data corresponding to multiple fault types of each cell and the preset fault type classification table.
  • the preset fault type classification table is preset data used to classify wireless network faults.
  • the preset fault type classification table may include multiple fault types and preset conditions required to satisfy the multiple fault types, where the preset condition is a value range of at least one field. For example, taking the access fault type as an example, the preset fault type classification table can be obtained as shown in the following table:
  • the AA, AB and other fields in the table respectively represent different network performance parameters.
  • a certain field meets the corresponding preset conditions, it means that the fault type corresponding to the field exists in the wireless network of the cell. Therefore, the input data corresponding to multiple different fault types of L cells can be obtained, the values corresponding to each field can be obtained, and then it can be determined whether the values corresponding to each field satisfy the requirements of the fields matching the fields in the preset fault type classification table. Preset conditions. If the value corresponding to the field meets the preset condition corresponding to the field, then the flag position corresponding to the field can be set to 1; if the value corresponding to the field does not meet the preset condition corresponding to the field, then the flag position corresponding to the field can be set to 1. Flag position 0.
  • the feature vector corresponding to the access fault type of a certain cell is 1011100010010000.
  • the characteristic vectors of call drop fault types and the characteristic vectors of intra-system handover faults can be obtained in the same manner as above.
  • the feature vectors of different fault types of a certain cell within a specific time period can be obtained, and then the feature vectors of multiple fault types of L cells can be obtained.
  • the input data can be obtained according to different time granularities, for example, 15 minutes as the granularity, 1 hour as the granularity, or 1 day as the granularity, etc., which is not specifically limited in this disclosure.
  • the eigenvector includes: performing an AND operation on the preset eigenvector and the eigenvectors of K cells to obtain the operation result.
  • the eigenvectors of K cells and the preset eigenvectors are expressed in binary, and the length of the preset eigenvector is equal to
  • the length of the feature vectors of K cells is consistent, and the value of the field corresponding to the target fault type in the preset feature vector is 1, and the value of the other fields except the field corresponding to the target fault type is 0; and according to the operation result, Characteristic vectors corresponding to N faulty cells are determined from the characteristic vectors of K cells.
  • the above-mentioned preset feature vector is a feature vector customized according to actual screening needs.
  • the preset feature vector is used to screen feature vectors of cells with faults to obtain feature vectors of cells with target fault types.
  • completely fault-free feature vectors may be eliminated first.
  • the feature vectors of the K cells with faults can be screened based on the preset feature vectors. For example, assume that the feature vector of the access fault type of a certain cell at points 0 to 5 is obtained as follows:
  • Each row represents a feature vector, and different feature vectors correspond to different times, representing from 0 o'clock to 5 o'clock respectively.
  • Each feature vector includes 16 flag bits, and different flag bits represent different sub-fault types in the access fault type.
  • the preset feature vector can be defined as:
  • the flag bits of the fields related to the coverage type fault type from 0 to 4 points are all set to 1, and the flag bits from 0 to 4 points are related to the coverage type fault type.
  • the flag bits of fields unrelated to the fault type are all set to 0, and all flag bits at point 5 are set to 0.
  • the preset feature vector can be defined as:
  • the feature vectors of the access fault types at 2 points of the cell can be obtained, as shown in the following figure:
  • the preset feature vector can be customized to obtain the feature vector of the required fault type. Different preset feature vectors can be defined according to different requirements to achieve parallel analysis and processing of different fault types and improve the efficiency of analysis and processing.
  • the original analysis process and analysis conclusions are completely digitized, realizing the separation of operations and data.
  • the formed data is also easy to store, making it convenient to further extract the characteristic data of the community, and match the configuration data to form an operation and maintenance library.
  • the data filtering method turns the original table lookup operation into a binary AND operation, which greatly saves computing overhead and significantly improves both efficiency and convenience.
  • the above-mentioned step 103 is to perform regression analysis on the network configuration parameters of each area in the M areas to obtain the adjustment value of the network configuration parameters of each area, including: based on the feature vectors corresponding to the N cells, determining the network configuration parameters in the M areas.
  • Network configuration parameters of each area perform regression analysis on the network configuration parameters to obtain the objective function.
  • the objective function includes network configuration parameters and target parameters.
  • the objective parameters are any other parameters in the objective function except the network configuration parameters.
  • the objective function is used Calculate the optimal solution of network configuration parameters; and use the objective function to calculate the adjustment of network configuration parameters in each region. Integer value.
  • the network configuration parameters associated with all fault types can be set in advance, and then the network configuration parameters that need to be adjusted can be determined according to the target fault type, and then regression analysis is performed on the network configuration parameters of each area to obtain the user
  • the objective function is used to calculate the optimal solution of the network configuration parameters, and finally based on the objective function, the adjustment values of the network configuration parameters in each region are determined.
  • the adjustment values of the network configuration parameters can be calculated by themselves, and the network configuration parameters can be adjusted. That is, an adaptive adjustment mode is formed through machine learning. Once the network data deteriorates, the configuration adjustment can be automatically made, and the adjustment process Completely machine learning and does not rely on personal experience.
  • the above steps perform regression analysis on the network configuration parameters to obtain the target function, including: using the least squares estimation algorithm to calculate the partial derivative of the target parameter; when the partial derivative of the target parameter is 0, calculate the target parameter estimated values; and determine the objective function based on the estimated values of the objective parameters.
  • the regression equations associated with the feature vectors of all fault types can also be set in advance, and then the selected regression equation can be determined based on the feature vectors of the target fault type. Then, regression analysis is performed on the network configuration parameters through the regression equation to obtain the objective function.
  • the network configuration parameters may include antenna transmit power and antenna pitch angle. The antenna transmit power is recorded as X 1 and the antenna pitch angle is recorded as The index value that needs to be adjusted for the problem, Channel Quality Indication (CQI for short), is recorded as Y. It is assumed that the following regression equation can be used for regression analysis:
  • ⁇ 1 is the coefficient of X 1
  • ⁇ 2 is the coefficient of X 2
  • ⁇ 0 is a constant.
  • the estimated values of the three target parameters ⁇ 0 , ⁇ 1 , and ⁇ 2 can be obtained, thereby obtaining the objective function. Then, with the objective function known, the adjustment values of the network configuration parameters X 1 and X 2 can be calculated through multiple Y values.
  • the relevant data can be entered into the objective function to obtain the values of the antenna transmit power and antenna pitch angle that need to be adjusted, so that the network configuration parameters can be directly configured. Optimize on.
  • the above-mentioned step 102 is to perform clustering according to the geographical location of each of the N cells and divide the N cells into M areas, including: obtaining the geographical location corresponding to the N cells; based on the geographical location corresponding to the N cells , calculate the Euclidean distance between any two cells in N cells; and cluster the cells whose Euclidean distance is less than the preset threshold to form M regions.
  • the Euclidean distance between any two cells can be calculated based on the geographical location of each of the N cells, and then the cells whose Euclidean distance is less than a preset threshold can be clustered together.
  • the N cells are divided into M areas, where each area includes at least one cell. Since the types of network faults in cells that are geographically close are similar and the network configuration parameters that need to be adjusted are also similar, the network configuration parameters of cells belonging to the same area can be adjusted together, thereby improving the efficiency of network configuration parameter adjustment of the cells.
  • Figure 2 is a schematic structural diagram of an apparatus for adjusting network configuration parameters provided by an embodiment of the present disclosure.
  • the device 200 for adjusting network configuration parameters includes an acquisition module 201 , a clustering module 202 , an analysis module 203 and an adjustment module 204 .
  • the acquisition module 201 is configured to obtain feature vectors corresponding to N cells with target fault types, where the feature vectors corresponding to the N cells are used to characterize the fault types existing in the N cells, and the target fault types are multiple Any fault type among the fault types, N is an integer greater than or equal to 1.
  • the clustering module 202 is configured to perform clustering according to the geographical location of each of the N cells, and divide the N cells into M areas, where M is an integer greater than or equal to 1 and less than or equal to N.
  • the analysis module 203 is configured to perform regression analysis on the network configuration parameters of each area in the M areas, and obtain the adjustment value of the network configuration parameters of each area.
  • the network configuration parameters are parameters associated with the target fault type, and the target fault type It is determined based on the feature vectors corresponding to N cells.
  • the adjustment module 204 is configured to adjust the network configuration parameters of each area based on the adjustment values of the network configuration parameters of each area.
  • the acquisition module 201 includes a first acquisition sub-module, a deletion sub-module and a filtering sub-module.
  • the first acquisition sub-module is configured to acquire feature vectors of multiple fault types of L cells, where L is an integer greater than N, and the N cells are N cells among the L cells.
  • the deletion submodule is configured to delete the feature vectors corresponding to the normal cells among the obtained feature vectors of the L cells, and obtain the feature vectors of the K cells with faults, where K is greater than or equal to N, and less than or equal to L integer.
  • the screening submodule is configured to screen the feature vectors of K cells to obtain feature vectors corresponding to N faulty cells.
  • the first acquisition sub-module includes an acquisition unit, a judgment unit, a first processing unit, a second processing unit and a first determination unit.
  • the acquisition unit is configured to acquire input data corresponding to multiple fault types of L cells, where the input data includes multiple fields and values corresponding to each field, and different fields represent different network performance parameters.
  • the judgment unit is configured to judge whether the value corresponding to the field satisfies the preset condition corresponding to the field matching the field in the preset fault type classification table, wherein the preset fault type classification table includes a plurality of fault types and those satisfying a plurality of types of faults.
  • the preset conditions are the value range of at least one field.
  • the first processing unit is configured to set the flag position corresponding to the field to 1 when the value corresponding to the field satisfies the preset condition corresponding to the field.
  • the second processing unit is configured to set the flag bit of the field to 0 when the value corresponding to the field does not meet the preset condition corresponding to the field.
  • the first determination unit is configured to determine the feature vectors of multiple fault types of the L cells according to the flag bits corresponding to each field.
  • the screening sub-module includes an operation unit and a second determination unit.
  • the operation unit is configured to perform an AND operation on the preset feature vector and the feature vectors of the K cells to obtain the operation result, in which the feature vectors of the K cells and the preset feature vector are both expressed in binary, and the length of the preset feature vector is It is consistent with the length of the feature vectors of K cells, and the value of the field corresponding to the target fault type in the preset feature vector is 1, and the value of the other fields except the field corresponding to the target fault type is 0.
  • the second determination unit is configured to determine the feature vectors corresponding to the N faulty cells from the feature vectors of the K cells based on the operation results.
  • the analysis module 203 includes a determination sub-module, an analysis sub-module and a first calculation sub-module.
  • the determination sub-module is configured to determine the network configuration parameters of each of the M areas based on the feature vectors corresponding to the N cells.
  • the analysis submodule is configured to perform regression analysis on the network configuration parameters to obtain the objective function.
  • the objective function includes the network configuration parameters and the target parameters.
  • the objective parameters are any other parameters in the objective function except the network configuration parameters.
  • the objective function is used for Calculate the optimal solution for network configuration parameters.
  • the first calculation sub-module is configured to use the objective function to calculate the adjustment value of the network configuration parameters of each area.
  • the analysis sub-module includes a first calculation unit, a second calculation unit and a third determination unit.
  • the first calculation unit is configured to calculate the partial derivative of the target parameter using a least squares estimation algorithm.
  • the second calculation unit is configured to calculate the estimated value of the target parameter when the partial derivative of the target parameter is 0.
  • the third determination unit is configured to determine the objective function according to the estimated value of the objective parameter.
  • the clustering module 202 includes a second acquisition sub-module, a second calculation sub-module and a clustering sub-module.
  • the second acquisition sub-module is configured to acquire the geographical locations corresponding to N cells.
  • the second calculation submodule is configured to calculate the Euclidean distance between any two cells among the N cells based on the geographical locations corresponding to the N cells.
  • the clustering submodule is configured to cluster cells whose Euclidean distance is smaller than a preset threshold to form M regions.
  • the device 200 for adjusting network configuration parameters can implement the steps of the method for adjusting network configuration parameters as provided in any of the foregoing method embodiments, and achieve the same technical effect, which will not be described again here.
  • the embodiment of the present disclosure provides a network management system, including a processor 311, a communication interface 312, a memory 313, and a communication bus 314.
  • the processor 311, the communication interface 312, and the memory 313 are completed through the communication bus 314. communication between each other.
  • Memory 313 is used to store computer programs.
  • the processor 311 is used to implement the network configuration parameter adjustment method provided in any of the foregoing method embodiments when executing the program stored on the memory 313, including: obtaining N data with the target fault type. Characteristic vectors corresponding to cells, where the characteristic vectors corresponding to N cells are used to characterize the fault types existing in N cells.
  • the target fault type is any fault type among multiple fault types, and N is greater than or equal to An integer of 1; cluster according to the geographical location of each of the N communities, and divide the N communities into M regions, where M is an integer greater than or equal to 1 and less than or equal to N; for each of the M regions Perform regression analysis on the network configuration parameters to obtain the adjustment values of the network configuration parameters in each area.
  • the network configuration parameters are parameters that are associated with the target fault type.
  • the target fault type is determined based on the feature vectors corresponding to N cells; and Based on the adjustment value of the network configuration parameters of each area, the network configuration parameters of each area are adjusted.
  • Embodiments of the present disclosure also provide a computer-readable storage medium on which a computer program is stored.
  • the computer program is executed by a processor, the method for adjusting network configuration parameters as provided in any of the foregoing method embodiments is implemented.
  • feature vectors corresponding to N cells with target fault types are obtained, where the feature vectors corresponding to N cells are used to characterize the fault types existing in N cells, and the target fault types are multiple Any fault type among the fault types; cluster the N cells according to the geographical location of each cell, and divide the N cells into M areas, where M is an integer greater than or equal to 1 and less than or equal to N; Perform regression analysis on the network configuration parameters of each area in the M areas to obtain the adjustment values of the network configuration parameters in each area.
  • the network configuration parameters are parameters that are associated with the target fault type.
  • the target fault type is based on the N cells.
  • the characteristic vector of is determined; based on the adjustment value of the network configuration parameters of each area, the network configuration parameters of each area are adjusted.
  • the fault type existing in each cell can be identified based on the feature vector of each cell.
  • the N cells of the target fault type can also be clustered based on the geographical location of the N cells, and regression analysis is performed on the network configuration parameters of each clustered area to obtain the adjustment value of the network configuration parameter, and based on the adjustment value Automatically adjust network configuration parameters without relying on personal experience, thus improving the timeliness and accuracy of network configuration parameter adjustment.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本公开涉及一种网络配置参数的调整方法、装置及网管系统，该方法包括：获取存在目标故障类型的N个小区所对应的特征向量，其中，N个小区所对应的特征向量用于表征N个小区所存在的故障类型，目标故障类型为多种故障类型中的任意一种故障类型，N为大于或等于1的整数；根据N个小区中各小区的地理位置进行聚类，将N个小区划分至M个区域，M为大于或等于1，且小于或等于N的整数；对M个区域中各区域的网络配置参数进行回归分析，得到各区域的网络配置参数的调整值；基于各区域的网络配置参数的调整值，对各区域的网络配置参数进行调整。

Description

网络配置参数的调整方法、装置及网管系统
相关申请的交叉引用
本公开要求享有2022年06月06日提交的名称为“网络配置参数的调整方法、装置及网管系统”的中国专利申请CN202210634449.8的优先权，其全部内容通过引用并入本公开中。
技术领域
本公开涉及无线通信技术领域，尤其涉及一种网络配置参数的调整方法、装置及网管系统。
背景技术
网络的可靠性是评价网络性能的重要指标之一，针对网络中出现的故障，目前通常是采用分类分层的决策树来进行检测的。
发明内容
本公开提供了一种网络配置参数的调整方法、装置及网管系统。
第一方面，本公开提供了一种网络配置参数的调整方法，所述方法包括：获取存在目标故障类型的N个小区所对应的特征向量，其中，所述N个小区所对应的特征向量用于表征所述N个小区所存在的故障类型，所述目标故障类型为多种故障类型中的任意一种故障类型，N为大于或等于1的整数；根据所述N个小区中各小区的地理位置进行聚类，将所述N个小区划分至M个区域，M为大于或等于1，且小于或等于N的整数；对所述M个区域中各区域的网络配置参数进行回归分析，得到各区域的网络配置参数的调整值，所述网络配置参数为与所述目标故障类型存在关联关系的参数，所述目标故障类型是基于所述N个小区所对应的特征向量确定得到；以及基于各区域的所述网络配置参数的调整值，对各区域的所述网络配置参数进行调整。
第二方面，本公开还提供了一种网络配置参数的调整装置，所述装置包括：获取模块，被配置为获取存在目标故障类型的N个小区所对应的特征向量，其中，所述N个小区所对应的特征向量用于表征所述N个小区所存在的故障类型，所述目标故障类型为多种故障类型中的任意一种故障类型，N为大于或等于1的整数；聚类模块，被配置为根据所述N个小区中各小区的地理位置进行聚类，将所述N个小区划分至M个区域，M为大于或等于1，且小于或等于N的整数；分析模块，被配置为对所述M个区域中各区域的网络配置参数进 行回归分析，得到各区域的网络配置参数的调整值，所述网络配置参数为与所述目标故障类型存在关联关系的参数，所述目标故障类型是基于所述N个小区所对应的特征向量确定得到；以及调整模块，被配置为基于各区域的所述网络配置参数的调整值，对各区域的所述网络配置参数进行调整。
第三方面，本公开还提供了一种网管系统，包括处理器、通信接口、存储器和通信总线，其中，处理器、通信接口、存储器通过通信总线完成相互间的通信；存储器，用于存放计算机程序；处理器，用于执行存储器上所存放的程序时，实现第一方面所述的网络配置参数的调整方法。
第四方面，本公开还提供了一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现如第一方面所述的网络配置参数的调整方法。
附图说明
此处的附图被并入说明书中并构成本说明书的一部分，示出了符合本公开的实施例，并与说明书一起用于解释本公开的原理。
为了更清楚地说明本公开实施例或相关技术中的技术方案，下面将对实施例或相关技术描述中所需要使用的附图作简单地介绍，显而易见地，对于本领域普通技术人员而言，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为本公开实施例提供的网络配置参数的调整方法的流程示意图；
图2为本公开实施例提供的网络配置参数的调整装置的结构示意图；以及
图3为本公开实施例提供的网管系统的结构示意图。
具体实施方式
为使本公开实施例的目的、技术方案和优点更加清楚，下面将结合本公开实施例中的附图，对本公开实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本公开的一部分实施例，而不是全部的实施例。基于本公开中的实施例，本领域普通技术人员在没有做出创造性劳动的前提下所获得的所有其他实施例，都属于本公开保护的范围。
虽然采用分类分层的决策树方式可以有效诊断网络故障，但诊断结论较为单一，无法对网络配置参数形成有规律性的后续调整。因此，在根据分类分层的决策树诊断出网络故障后，还需要根据人为经验手动调整网络配置参数，使得网络配置参数的调整的及时性和准确性较低。
参见图1，图1为本公开实施例提供的网络配置参数的调整方法的流程示意图之一。如图1所示，该网络配置参数的调整方法可以包括以下步骤101至步骤104。
步骤101、获取存在目标故障类型的N个小区所对应的特征向量，其中，N个小区所对应的特征向量用于表征N个小区所存在的故障类型，目标故障类型为多种故障类型中的任意一种故障类型，N为大于或等于1的整数。
需要说明的是，该网络配置参数的调整方法可以应用于网管系统或者其他安装有网管类型软件的电子设备，本公开不做具体限定。为方便说明，后续实施例均以网管系统为例进行解释说明。
上述目标故障类型为无线网络中存在的多种故障类型中的任意一种故障类型。该目标故障类型可以为一级故障类型，如接入类故障类型、掉话类故障类型、系统内切换故障类型等；还可以为二级故障类型，如接入类故障类型下的传输故障类型、覆盖故障类型、容量故障类型等；亦可以为三级故障类型，如覆盖故障类型下的弱覆盖故障类型、上行干扰故障类型、下行干扰故障类型等。该目标故障类型可以根据实际需要检测的故障类型进行灵活设置，本公开不做具体限定。上述特征向量可以根据一级故障类型进行分类，如接入类故障类型对应接入类故障类型的特征向量，掉话类故障类型对应掉话类故障类型的特征向量，系统内切换故障类型对应系统内切换故障类型的特征向量等。每个特征向量的长度可以根据一级故障类型内所包含的所有子故障类型(包括二级故障类型、三级故障类型、……)的数量进行确定。当目标故障类型为一级故障类型时，获取存在目标故障类型的N个小区所对应的特征向量即为获取存在该一级故障类型的N个小区的特征向量；当目标故障类型为某一子故障类型时，需要先获取各小区中存在该子故障类型所属的一级故障类型所对应的特征向量，再基于该一级故障类型所对应的特征向量选择出存在该子故障类型的特征向量。
步骤102、根据N个小区中各小区的地理位置进行聚类，将N个小区划分至M个区域，M为大于或等于1，且小于或等于N的整数。
在该步骤中，可以根据N个小区中各小区的地理位置，计算任意两个小区之间的距离如欧氏距离、巴氏距离、马氏距离等，进而将距离较近的小区聚类在一起，这样，就可以将N个小区划分至M个区域，其中，每个区域包括至少一个小区。由于地理位置靠近的小区存在的网络故障类型类似，需要调整的网络配置参数也类似，因而可以将属于同一区域的小区的网络配置参数一并进行调整，从而提高小区的网络配置参数调整效率。
步骤103、对M个区域中各区域的网络配置参数进行回归分析，得到各区域的网络配置参数的调整值，网络配置参数为与目标故障类型存在关联关系的参数，目标故障类型是基于N个小区所对应的特征向量确定得到。
由于特征向量可以识别出故障类型，因而可以基于N个小区所对应的特征向量确定得到目标故障类型，进而根据目标故障类型确定与之关联的网络配置参数。例如，假设目标 故障类型为接入类故障类型，那么该网络配置参数可以为与该接入类故障类型存在关联关系的天线发射功率和天线俯仰角等，因而可以对每个区域的天线发射功率和天线俯仰角等网络配置参数进行回归分析，得到各区域的天线发射功率和天线俯仰角等网络配置参数的调整值。
步骤104、基于各区域的网络配置参数的调整值，对各区域的网络配置参数进行调整。
在得到各区域的网络配置参数的调整值后，可以直接基于各区域的网络配置参数的调整值对各区域的网络配置参数进行调整，使得网络配置参数最优。
在本实施例中，可以基于每个小区的特征向量，识别出每个小区存在的故障类型。同时，对于存在目标故障类型的N个小区，还可以基于N个小区的地理位置进行聚类，对聚类后的每个区域的网络配置参数进行回归分析，得到网络配置参数的调整值，并基于该调整值自动调整网络配置参数，而无需依赖个人经验，从而提高了网络配置参数调整的及时性和准确性。
进一步地，上述步骤101、获取存在目标故障类型的N个小区所对应的特征向量，包括：获取L个小区的多种故障类型的特征向量，L为大于N的整数，N个小区为L个小区中的N个小区；将获取到的L个小区的特征向量中属于正常小区所对应的特征向量进行删除，得到存在故障的K个小区的特征向量，K为大于或等于N，且小于或等于L的整数；以及对K个小区的特征向量进行筛选，得到N个故障小区所对应的特征向量。
上述L个小区可以理解为需要进行网络故障分析的整个无线网络覆盖区域内的所有小区。该L个小区包括上述存在目标故障类型的N个小区。上述K个小区是指该L个小区中除去正常小区外的存在故障的小区，该K个小区包括上述存在目标故障类型的N个小区和存在其他故障类型的K-N个小区。
在一些实施例中，可以先获取L个小区的多种不同故障类型的特征向量，再基于获取到的特征向量筛选出L个小区中正常小区的特征向量，并将其删除，保留存在故障的K个小区的特征向量，然后基于故障类型对存在故障的K个小区的特征向量进行筛选，得到存在目标故障类型的N个故障小区所对应的特征向量。通过这种方式，可以筛选掉无故障小区对应的特征向量，并进一步精确定位到存在目标故障类型的小区的特征向量，大大提高了目标小区的定位效率，体现出了性能数据向量化处理后的优势。
进一步地，上述步骤、获取L个小区的多种故障类型的特征向量，包括：获取L个小区的多种故障类型对应的输入数据，其中，输入数据包括多个字段和每个字段对应的数值，不同字段代表不同的网络性能参数；判断字段对应的数值是否满足预设故障类型分类表中与字段相匹配的字段所对应的预设条件，其中，预设故障类型分类表包括多种故障类型和满足多种故障类型所需的预设条件，预设条件为至少一个字段的取值范围；在字段对应的 数值满足字段所对应的预设条件的情况下，将字段对应的标志位置1；在字段对应的数值不满足字段所对应的预设条件的情况下，将字段的标志位置0；以及根据各字段对应的标志位，确定L个小区的多种故障类型的特征向量。
在一些实施例中，可以根据各小区的多种故障类型对应的输入数据和预设故障类型分类表获取各小区的特征向量。预设故障类型分类表为预设设置的用于对无线网络故障进行划分的数据。该预设故障类型分类表可以包括多种故障类型和满足多种故障类型所需的预设条件，预设条件为至少一个字段的取值范围。例如，以接入类故障类型为例，可以得到预设故障类型分类表如下表所示：

表中的AA、AB等字段分别表示不同的网络性能参数，当某一字段满足对应的预设条件，则表示小区的无线网络中存在该字段对应的故障类型。因而，可以获取L个小区的多种不同故障类型对应的输入数据，得到各字段对应的数值，再判断各字段对应的数值是否满足预设故障类型分类表中与字段相匹配的字段所对应的预设条件，如果字段对应的数值满足字段所对应的预设条件，那么可以将字段对应的标志位置1；如果字段对应的数值不满足字段所对应的预设条件的情况下，那么将字段的标志位置0。继续基于上表进行解释说明，假设各字段对应的数值如下：
这样，就可以得到某小区的接入类故障类型对应的特征向量为1011100010010000。依次类推，可以根据上述相同的方式得到掉话类故障类型的特征向量和系统内切换故障的特征向量等等。这样，就可以得到某小区在特定时间段内的不同故障类型的特征向量，进而得到L个小区的多种故障类型的特征向量。需要说明的是，输入数据可以按照不同的时间粒度进行获取，例如，以15分钟为粒度、以1小时为粒度或者以1天为粒度等，本公开不做具体限定。
进一步地，上述步骤、对K个小区的特征向量进行筛选，得到N个故障小区所对应的特 征向量，包括：将预设特征向量与K个小区的特征向量进行与运算，得到运算结果，其中，K个小区的特征向量和预设特征向量均采用二进制表示，预设特征向量的长度与K个小区的特征向量的长度一致，且预设特征向量中与目标故障类型对应的字段的数值为1，除目标故障类型对应的字段之外的其余字段的数值为0；以及根据运算结果，从K个小区的特征向量中确定出N个故障小区所对应的特征向量。
上述预设特征向量为根据实际筛选需要自定义的特征向量，该预设特征向量用于对存在故障的小区的特征向量进行筛选，得到存在目标故障类型的小区的特征向量。
在一些实施例中，可以先排除掉完全无故障的特征向量。例如，在接入类故障类型的特征向量中，可以按照上述方法判断AA、AB字段的数值是否满足对应的预设条件，来得到特征向量的第一位，如果这个标志位为“0”，那么表示接入类是无故障的，这样的特征向量就可以直接排除掉。然后，针对不同一级故障类型的特征向量需要单独处理，如果同一个小区，存在多个不同一级故障类型的特征向量需要分别进行单独处理。
在筛选完无故障的特征向量后，可以再基于预设特征向量对存在故障的K个小区的特征向量进行筛选。例如，假设获取到某小区在0～5点的接入类故障类型的特征向量如下：
每一行代表一个特征向量，不同的特征向量对应的时间不同，分别表示从0点至5点。每个特征向量包括16个标志位，不同标志位表示的接入类故障类型中不同的子故障类型。
假设需要筛选出上表中0～4点的覆盖类故障类型(即接入类故障类型中的子故障类型)的数据，那么预设特征向量可以定义为：
该预设特征向量中0～4点的与覆盖类故障类型相关的字段的标志位(即第3列标志位至第6列标志位)均置1，并将0～4点的与覆盖类故障类型无关的字段的标志位均置0，且5点的所有标志位置0。这样，通过将预设特征向量与该小区的特征向量进行与运算，可以得到该 小区的覆盖类故障类型的特征向量，如下图所示：
又假设需要筛选出上表中2点的接入类故障类型的数据，那么预设特征向量可以定义为：
将预设特征向量与该小区的特征向量进行与运算，可以得到该小区的2点的接入类故障类型的特征向量，如下图所示：
在本实施例中，可以自定义预设特征向量获取所需的故障类型的特征向量，针对不同的需求，定义不同的预设特征向量，实现不同故障类型的并行分析处理，提高分析处理效率。同时，将原有的分析过程和分析结论完全数据化，实现了运算和数据的分离，形成的数据也方便存储，方便进一步提取小区的特征数据，匹配配置数据形成运维库。数据筛选的方式将原来的查表运算变成了二进制与运算，大大节省了计算开销，无论从效率上，还是便利程度上都显著进步。
进一步地，上述步骤103、对M个区域中各区域的网络配置参数进行回归分析，得到各区域的网络配置参数的调整值，包括：基于N个小区所对应的特征向量，确定M个区域中各区域的网络配置参数；对网络配置参数进行回归分析，得到目标函数，目标函数包括网络配置参数和目标参数，目标参数为目标函数中除网络配置参数之外的其他任何参数，目标函数用于计算网络配置参数的最优解；以及利用目标函数计算各区域的网络配置参数的调 整值。
在一些实施例中，可以预先对所有故障类型所关联的网络配置参数进行设置，然后可以根据目标故障类型，确定需要调整的网络配置参数，再对各区域的网络配置参数进行回归分析，得到用于计算网络配置参数的最优解的目标函数，最后基于目标函数，确定各区域的网络配置参数的调整值。这样，可以自行计算出网络配置参数的调整值，并对网络配置参数进行调整，即通过机器学习后形成了自适应的调整模式，一旦出现网络数据的恶化即可自动进行配置调整，而且调整过程完全由机器学习不依赖个人经验。
进一步地，上述步骤、对网络配置参数进行回归分析，得到目标函数，包括：利用最小二乘估计算法计算目标参数的偏导数；在目标参数的偏导数为0的情况下，计算得到目标参数的估计值；以及根据目标参数的估计值，确定目标函数。
在一些实施例中，还可以预先对所有故障类型的特征向量所关联的回归方程进行设置，然后可以根据目标故障类型的特征向量，确定选用的回归方程。进而通过回归方程对网络配置参数进行回归分析，得到目标函数。例如，假设目标故障类型为接入类故障类型，网络配置参数可以包括天线发射功率和天线俯仰角，分别将天线发射功率记为X1,将天线俯仰角记为X2，将调整接入类问题需要调整的指标值信道质量指示(Channel Quality Indication,简称为CQI)优良率记为Y，假设可以使用如下回归方程做回归分析：
Y＝β0+β2X1+β2X2
其中，β1为X1的系数、β2为X2的系数，β0为常量。
对于多回归方程，在模型和数据满足前文的基本假定的前提下，参数估计可以通过最小二乘估计来得到，对其求偏导数，并令其结果等于0，得到如下回归方程：
通过上述公式可以得到β0、β1、β2这三个目标参数的估计值，从而得到目标函数。然后，在已知目标函数的情况下，通过多个Y值，即可计算出网络配置参数X1和X2的调整值。
以下列举一个数据样例进行解释说明，实际数据情况不限于该样例所列举的数据：

通过上表进行计算，最后可以得到的目标函数为：Y＝75+0.033X1+22.333X2。
利用这个目标函数，可以在需要对接入类故障进行优化时，将相关的数据进入该目标函数内，即可得到需要调整的天线发射功率和天线俯仰角的值，从而可以直接对网络配置参数上进行优化。
进一步地，上述步骤102、根据N个小区中各小区的地理位置进行聚类，将N个小区划分至M个区域，包括：获取N个小区对应的地理位置；基于N个小区对应的地理位置，计算N个小区中任意两个小区之间的欧氏距离；以及将欧氏距离小于预设阈值的小区进行聚类，形成M个区域。
在一些实施例中，可以根据N个小区中各小区的地理位置，计算任意两个小区之间的欧氏距离，进而将欧氏距离小于预设阈值的小区聚类在一起，这样，就可以将N个小区划分至M个区域，其中，每个区域包括至少一个小区。由于地理位置靠近的小区存在的网络故障类型类似，需要调整的网络配置参数也类似，因而可以将属于同一区域的小区的网络配置参数一并进行调整，从而提高小区的网络配置参数调整效率。
参见图2，图2为本公开实施例提供的网络配置参数的调整装置的结构示意图。如图2所示，该网络配置参数的调整装置200包括获取模块201、聚类模块202、分析模块203以及调整模块204。
获取模块201，被配置为获取存在目标故障类型的N个小区所对应的特征向量，其中，N个小区所对应的特征向量用于表征N个小区所存在的故障类型，目标故障类型为多种故障类型中的任意一种故障类型，N为大于或等于1的整数。
聚类模块202，被配置为根据N个小区中各小区的地理位置进行聚类，将N个小区划分至M个区域，M为大于或等于1，且小于或等于N的整数。
分析模块203，被配置为对M个区域中各区域的网络配置参数进行回归分析，得到各区域的网络配置参数的调整值，网络配置参数为与目标故障类型存在关联关系的参数，目标故障类型是基于N个小区所对应的特征向量确定得到。
调整模块204，被配置为基于各区域的网络配置参数的调整值，对各区域的网络配置参数进行调整。
进一步地，获取模块201包括第一获取子模块、删除子模块以及筛选子模块。
第一获取子模块，被配置为获取L个小区的多种故障类型的特征向量，L为大于N的整数，N个小区为L个小区中的N个小区。
删除子模块，被配置为将获取到的L个小区的特征向量中属于正常小区所对应的特征向量进行删除，得到存在故障的K个小区的特征向量，K为大于或等于N，且小于或等于L的 整数。
筛选子模块，被配置为对K个小区的特征向量进行筛选，得到N个故障小区所对应的特征向量。
进一步地，第一获取子模块包括获取单元、判断单元、第一处理单元、第二处理单元以及第一确定单元。
获取单元，被配置为获取L个小区的多种故障类型对应的输入数据，其中，输入数据包括多个字段和每个字段对应的数值，不同字段代表不同的网络性能参数。
判断单元，被配置为判断字段对应的数值是否满足预设故障类型分类表中与字段相匹配的字段所对应的预设条件，其中，预设故障类型分类表包括多种故障类型和满足多种故障类型所需的预设条件，预设条件为至少一个字段的取值范围。
第一处理单元，被配置为在字段对应的数值满足字段所对应的预设条件的情况下，将字段对应的标志位置1。
第二处理单元，被配置为在字段对应的数值不满足字段所对应的预设条件的情况下，将字段的标志位置0。
第一确定单元，被配置为根据各字段对应的标志位，确定L个小区的多种故障类型的特征向量。
进一步地，筛选子模块包括运算单元以及第二确定单元。
运算单元，被配置为将预设特征向量与K个小区的特征向量进行与运算，得到运算结果，其中，K个小区的特征向量和预设特征向量均采用二进制表示，预设特征向量的长度与K个小区的特征向量的长度一致，且预设特征向量中与目标故障类型对应的字段的数值为1，除目标故障类型对应的字段之外的其余字段的数值为0。
第二确定单元，被配置为根据运算结果，从K个小区的特征向量中确定出N个故障小区所对应的特征向量。
进一步地，分析模块203包括确定子模块、分析子模块以及第一计算子模块。
确定子模块，被配置为基于N个小区所对应的特征向量，确定M个区域中各区域的网络配置参数。
分析子模块，被配置为对网络配置参数进行回归分析，得到目标函数，目标函数包括网络配置参数和目标参数，目标参数为目标函数中除网络配置参数之外的其他任何参数，目标函数用于计算网络配置参数的最优解。
第一计算子模块，被配置为利用目标函数计算各区域的网络配置参数的调整值。
进一步地，分析子模块包括第一计算单元、第二计算单元以及第三确定单元。
第一计算单元，被配置为利用最小二乘估计算法计算目标参数的偏导数。
第二计算单元，被配置为在目标参数的偏导数为0的情况下，计算得到目标参数的估计值。
第三确定单元，被配置为根据目标参数的估计值，确定目标函数。
进一步地，聚类模块202包括第二获取子模块、第二计算子模块以及聚类子模块。
第二获取子模块，被配置为获取N个小区对应的地理位置。
第二计算子模块，被配置为基于N个小区对应的地理位置，计算N个小区中任意两个小区之间的欧氏距离。
聚类子模块，被配置为将欧氏距离小于预设阈值的小区进行聚类，形成M个区域。
需要说明的是，该网络配置参数的调整装置200可以实现如前述任意一个方法实施例提供的网络配置参数的调整方法的步骤，且达到相同的技术效果，在此不再一一赘述。
如图3所示，本公开实施例提供了一种网管系统，包括处理器311、通信接口312、存储器313和通信总线314，其中，处理器311、通信接口312、存储器313通过通信总线314完成相互间的通信。存储器313，用于存放计算机程序。在本公开一些实施例中，处理器311，用于执行存储器313上所存放的程序时，实现前述任意一个方法实施例提供的网络配置参数的调整方法，包括：获取存在目标故障类型的N个小区所对应的特征向量，其中，N个小区所对应的特征向量用于表征N个小区所存在的故障类型，目标故障类型为多种故障类型中的任意一种故障类型，N为大于或等于1的整数；根据N个小区中各小区的地理位置进行聚类，将N个小区划分至M个区域，M为大于或等于1，且小于或等于N的整数；对M个区域中各区域的网络配置参数进行回归分析，得到各区域的网络配置参数的调整值，网络配置参数为与目标故障类型存在关联关系的参数，目标故障类型是基于N个小区所对应的特征向量确定得到；以及基于各区域的网络配置参数的调整值，对各区域的网络配置参数进行调整。
本公开实施例还提供了一种计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现如前述任意一个方法实施例提供的网络配置参数的调整方法。
在本公开实施例中，通过获取存在目标故障类型的N个小区所对应的特征向量，其中，N个小区所对应的特征向量用于表征N个小区所存在的故障类型，目标故障类型为多种故障类型中的任意一种故障类型；根据N个小区中各小区的地理位置进行聚类，将N个小区划分至M个区域，M为大于或等于1，且小于或等于N的整数；对M个区域中各区域的网络配置参数进行回归分析，得到各区域的网络配置参数的调整值，网络配置参数为与目标故障类型存在关联关系的参数，目标故障类型是基于N个小区所对应的特征向量确定得到；基于各区域的网络配置参数的调整值，对各区域的网络配置参数进行调整。通过这种方式，可以基于每个小区的特征向量，识别出每个小区存在的故障类型。同时，对于存在 目标故障类型的N个小区，还可以基于N个小区的地理位置进行聚类，对聚类后的每个区域的网络配置参数进行回归分析，得到网络配置参数的调整值，并基于该调整值自动调整网络配置参数，而无需依赖个人经验，从而提高了网络配置参数调整的及时性和准确性。
需要说明的是，在本文中，诸如“第一”和“第二”等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来，而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的过程、方法、物品或者设备中还存在另外的相同要素。
以上所述仅是本公开的具体实施方式，使本领域技术人员能够理解或实现本公开。对这些实施例的多种修改对本领域的技术人员来说将是显而易见的，本文中所定义的一般原理可以在不脱离本公开的精神或范围的情况下，在其它实施例中实现。因此，本公开将不会被限制于本文所示的这些实施例，而是要符合与本文所公开的原理和新颖特点相一致的最宽的范围。

Claims (10)

  1. 一种网络配置参数的调整方法，包括：
    获取存在目标故障类型的N个小区所对应的特征向量，其中，所述N个小区所对应的特征向量用于表征所述N个小区所存在的故障类型，所述目标故障类型为多种故障类型中的任意一种故障类型，N为大于或等于1的整数；
    根据所述N个小区中各小区的地理位置进行聚类，将所述N个小区划分至M个区域，M为大于或等于1，且小于或等于N的整数；
    对所述M个区域中各区域的网络配置参数进行回归分析，得到各区域的网络配置参数的调整值，所述网络配置参数为与所述目标故障类型存在关联关系的参数，所述目标故障类型是基于所述N个小区所对应的特征向量确定得到；以及
    基于各区域的所述网络配置参数的调整值，对各区域的所述网络配置参数进行调整。
  2. 根据权利要求1所述的方法，其中，所述获取存在目标故障类型的N个小区所对应的特征向量，包括：
    获取L个小区的多种故障类型的特征向量，L为大于N的整数，所述N个小区为所述L个小区中的N个小区；
    将获取到的所述L个小区的特征向量中属于正常小区所对应的特征向量进行删除，得到存在故障的K个小区的特征向量，K为大于或等于N，且小于或等于L的整数；以及
    对所述K个小区的特征向量进行筛选，得到所述N个故障小区所对应的特征向量。
  3. 根据权利要求2所述的方法，其中，所述获取L个小区的多种故障类型的特征向量，包括：
    获取L个小区的多种故障类型对应的输入数据，其中，所述输入数据包括多个字段和每个所述字段对应的数值，不同所述字段代表不同的网络性能参数；
    判断所述字段对应的数值是否满足预设故障类型分类表中与所述字段相匹配的字段所对应的预设条件，其中，所述预设故障类型分类表包括多种故障类型和满足所述多种故障类型所需的预设条件，所述预设条件为至少一个所述字段的取值范围；
    在所述字段对应的数值满足所述字段所对应的预设条件的情况下，将所述字段对应的标志位置1；
    在所述字段对应的数值不满足所述字段所对应的预设条件的情况下，将所述字段的标志位置0；以及
    根据各字段对应的标志位，确定所述L个小区的多种故障类型的特征向量。
  4. 根据权利要求2所述的方法，其中，所述对所述K个小区的特征向量进行筛选，得到所述N个故障小区所对应的特征向量，包括：
    将预设特征向量与所述K个小区的特征向量进行与运算，得到运算结果，其中，所述K个小区的特征向量和所述预设特征向量均采用二进制表示，所述预设特征向量的长度与所述K个小区的特征向量的长度一致，且所述预设特征向量中与所述目标故障类型对应的字段的数值为1，除所述目标故障类型对应的字段之外的其余字段的数值为0；以及
    根据所述运算结果，从所述K个小区的特征向量中确定出所述N个故障小区所对应的特征向量。
  5. 根据权利要求1所述的方法，其中，所述对所述M个区域中各区域的网络配置参数进行回归分析，得到各区域的网络配置参数的调整值，包括：
    基于所述N个小区所对应的特征向量，确定所述M个区域中各区域的网络配置参数；
    对所述网络配置参数进行回归分析，得到目标函数，所述目标函数包括所述网络配置参数和目标参数，所述目标参数为所述目标函数中除所述网络配置参数之外的其他任何参数，所述目标函数用于计算所述网络配置参数的最优解；以及
    利用所述目标函数计算各区域的所述网络配置参数的调整值。
  6. 根据权利要求5所述的方法，其中，所述对所述网络配置参数进行回归分析，得到目标函数，包括：
    利用最小二乘估计算法计算所述目标参数的偏导数；
    在所述目标参数的偏导数为0的情况下，计算得到所述目标参数的估计值；以及
    根据所述目标参数的估计值，确定所述目标函数。
  7. 根据权利要求1所述的方法，其中，所述根据所述N个小区中各小区的地理位置进行聚类，将所述N个小区划分至M个区域，包括：
    获取所述N个小区对应的地理位置；
    基于所述N个小区对应的地理位置，计算所述N个小区中任意两个小区之间的欧氏距离；以及
    将所述欧氏距离小于预设阈值的小区进行聚类，形成所述M个区域。
  8. 一种网络配置参数的调整装置，包括：
    获取模块，被配置为获取存在目标故障类型的N个小区所对应的特征向量，其中，所述N个小区所对应的特征向量用于表征所述N个小区所存在的故障类型，所述目标故障类型为多种故障类型中的任意一种故障类型，N为大于或等于1的整数；
    聚类模块，被配置为根据所述N个小区中各小区的地理位置进行聚类，将所述N个小区划分至M个区域，M为大于或等于1，且小于或等于N的整数；
    分析模块，被配置为对所述M个区域中各区域的网络配置参数进行回归分析，得到各区域的网络配置参数的调整值，所述网络配置参数为与所述目标故障类型存在关联关系的参数，所述目标故障类型是基于所述N个小区所对应的特征向量确定得到；以及
    调整模块，被配置为基于各区域的所述网络配置参数的调整值，对各区域的所述网络配置参数进行调整。
  9. 一种网管系统，包括处理器、通信接口、存储器和通信总线，其中，处理器、通信接口、存储器通过通信总线完成相互间的通信；
    存储器，用于存放计算机程序；以及
    处理器，用于执行存储器上所存放的程序时，实现权利要求1-7任一项所述的网络配置参数的调整方法。
  10. 一种计算机可读存储介质，其上存储有计算机程序，其中，所述计算机程序被处理器执行时实现如权利要求1-7任一项所述的网络配置参数的调整方法。
PCT/CN2023/073441 2022-06-06 2023-01-20 网络配置参数的调整方法、装置及网管系统 WO2023236548A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210634449.8 2022-06-06
CN202210634449.8A CN117241301A (zh) 2022-06-06 2022-06-06 网络配置参数的调整方法、装置及网管系统

Publications (1)

Publication Number Publication Date
WO2023236548A1 true WO2023236548A1 (zh) 2023-12-14

Family

ID=89088549

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/073441 WO2023236548A1 (zh) 2022-06-06 2023-01-20 网络配置参数的调整方法、装置及网管系统

Country Status (2)

Country Link
CN (1) CN117241301A (zh)
WO (1) WO2023236548A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016112670A1 (zh) * 2015-01-16 2016-07-21 中兴通讯股份有限公司 小区切换优化评估方法及装置
CN110875825A (zh) * 2018-08-30 2020-03-10 中国移动通信集团广东有限公司 故障判决方法及装置
WO2021151503A1 (en) * 2020-01-31 2021-08-05 Telefonaktiebolaget Lm Ericsson (Publ) Analytics node and method thereof
WO2022017012A1 (zh) * 2020-07-21 2022-01-27 华为技术有限公司 一种网络配置方法及装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016112670A1 (zh) * 2015-01-16 2016-07-21 中兴通讯股份有限公司 小区切换优化评估方法及装置
CN110875825A (zh) * 2018-08-30 2020-03-10 中国移动通信集团广东有限公司 故障判决方法及装置
WO2021151503A1 (en) * 2020-01-31 2021-08-05 Telefonaktiebolaget Lm Ericsson (Publ) Analytics node and method thereof
WO2022017012A1 (zh) * 2020-07-21 2022-01-27 华为技术有限公司 一种网络配置方法及装置

Also Published As

Publication number Publication date
CN117241301A (zh) 2023-12-15

Similar Documents

Publication Publication Date Title
CN111356148B (zh) 一种实现网络优化的方法及相关设备
WO2020155755A1 (zh) 基于谱聚类的异常点比例优化方法、装置及计算机设备
WO2020155752A1 (zh) 异常点检测模型验证方法、装置、计算机设备及存储介质
WO2020147317A1 (zh) 一种网络异常行为确定方法、装置、设备及可读存储介质
US20220124517A1 (en) Anomaly detection method and device, terminal and storage medium
CN111459135B (zh) 基于物联网的智能家居故障状态追溯方法及中控中心
CN112243249B (zh) 5g nsa组网下lte新入网锚点小区参数配置方法和装置
CN108768695B (zh) Kqi的问题定位方法及装置
CN106027317B (zh) 信任感知的Web服务质量预测系统及方法
US20210026341A1 (en) Network analysis program, network analysis device, and network analysis method
CN104283737A (zh) 数据流的处理方法和装置
CN114610559A (zh) 设备运行环境评估方法、判决模型训练方法、电子设备
CN112020098A (zh) 负荷均衡方法、装置、计算设备及计算机存储介质
CN115309605A (zh) 一种基于大数据的异常监测的方法及装置
CN109783553A (zh) 一种配电网海量数据质量提升系统
US11425635B2 (en) Small cell identification using machine learning
WO2023236548A1 (zh) 网络配置参数的调整方法、装置及网管系统
CN117081909B (zh) 异常宽带修正方法、装置、电子设备及存储介质
JP7434690B2 (ja) ユーザ機器ｕｅのユーザデータを検出するための方法、装置、システム、デバイスおよびコンピュータプログラム
CN109526027A (zh) 一种小区容量优化方法、装置、设备及计算机存储介质
CN109976974B (zh) 一种针对运行状态判断的云计算环境下系统监测方法
CN114205247B (zh) 配电物联网的接入方法、装置、计算机设备及存储介质
CN117376084A (zh) 故障检测方法、电子设备及其介质
CN114968992A (zh) 数据识别清洗及补偿方法、装置、电子设备及存储介质
CN112417446A (zh) 一种软件定义网络异常检测架构

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: 23818716

Country of ref document: EP

Kind code of ref document: A1