CN-121984830-A - Network planning method, apparatus, device, storage medium, and computer program product

CN121984830ACN 121984830 ACN121984830 ACN 121984830ACN-121984830-A

Abstract

The application discloses a network planning method, a device, equipment, a storage medium and a computer program product, which are used for solving the problem that the existing network planning method lacks a targeted fusion mechanism for the multi-system data characteristics of the communication industry, so that a model is difficult to meet the operation requirements of differentiation, localization and security in actual deployment. The method comprises the steps of obtaining network planning data of at least two data sources and knowledge data uploaded by users, carrying out data integration processing on the network planning data and the knowledge data to obtain a training data set, training a pre-built original network planning model according to the training data set to obtain a network planning model, carrying out segmentation processing on the network planning model to obtain at least two network planning sub-models, and respectively deploying the network planning sub-models to edge nodes so as to enable network planning tasks to be executed through the network planning sub-models and output planning results.

Inventors

CAO HUAN
ZHOU HUIFANG
CHENG JIN
TUO JIAN

Assignees

中国移动通信集团设计院有限公司
中国移动通信集团有限公司

Dates

Publication Date: 20260505
Application Date: 20251230

Claims (10)

1. A method of network planning, comprising: Acquiring network planning data of at least two data sources and knowledge data uploaded by a user; performing data integration processing on the network planning data and the knowledge data to obtain a training data set; training a pre-constructed original network planning model according to the training data set to obtain a network planning model; Dividing the network planning model to obtain at least two network planning sub-models; and deploying each network planning sub-model to an edge node respectively, so that a network planning task is executed through the network planning sub-model, and a planning result is output.
2. The method according to claim 1, wherein the data integration processing is performed on the network planning data and the knowledge data, and specifically includes: performing association and fusion processing on cross-system data from different service systems; carrying out priority scheduling update processing on dynamic data from an external data source; And receiving and structuring the knowledge data uploaded by the user.
3. The method according to claim 2, wherein the performing a priority scheduling update process on the dynamic data from the external data source specifically comprises: Determining data updating priority of each external data source according to the attribute information of the external data source, wherein the attribute information comprises at least one of administrative region grades corresponding to the external data source, service growth indexes corresponding to the administrative region of the external data source and network planning related data; and according to the update priority, scheduling and updating the data processing sequence of the external data source.
4. The method according to claim 2, wherein the associating and fusing of cross-system data from different service systems specifically comprises: The method comprises the steps of respectively obtaining current situation data related to network planning from at least two independent service systems, wherein the independent service systems comprise a service supporting system, an operation supporting system, a network management system and a plan construction management system; And carrying out weighted fusion calculation on the current situation data according to a preset feature mapping rule corresponding to the communication network planning scene to obtain fusion features.
5. The method according to claim 2, wherein the receiving and structuring the knowledge data uploaded by the user specifically comprises: Receiving multi-mode knowledge data uploaded by a user through an interactive interface; Identifying the knowledge type of the knowledge data and extracting the structural information of the knowledge data; and according to the structured information, converting the knowledge data into vector features, and storing the vector features into a knowledge database.
6. The method of claim 1, wherein the partitioning the network planning model comprises: dividing the network planning model into a plurality of model fragments according to the functional modules corresponding to the network planning model; generating a corresponding encryption key for each model fragment, and encrypting each model fragment by using the corresponding encryption key; And respectively transmitting the encrypted model fragments and the corresponding encryption keys to the target edge node through separate transmission channels.
7. A network planning apparatus, comprising: the planning data acquisition unit is used for acquiring network planning data of at least two data sources and knowledge data uploaded by a user; the data integration unit is used for carrying out data integration processing on the network planning data and the knowledge data to obtain a training data set; the model training unit is used for training the pre-constructed original network planning model according to the training data set to obtain a network planning model; The segmentation unit is used for carrying out segmentation processing on the network planning model to obtain at least two network planning sub-models; and the planning unit is used for respectively deploying the network planning sub-models to the edge nodes so as to enable the network planning sub-models to execute network planning tasks and output planning results.
8. A network planning apparatus comprising: processor, and A memory arranged to store computer executable instructions that, when executed, cause the processor to: Acquiring network planning data of at least two data sources and knowledge data uploaded by a user; performing data integration processing on the network planning data and the knowledge data to obtain a training data set; training a pre-constructed original network planning model according to the training data set to obtain a network planning model; Dividing the network planning model to obtain at least two network planning sub-models; and deploying each network planning sub-model to an edge node respectively, so that a network planning task is executed through the network planning sub-model, and a planning result is output.
9. A computer readable storage medium storing one or more programs, which when executed by an electronic device comprising a plurality of application programs, cause the electronic device to perform the network planning method of any of claims 1-6.
10. A computer program product comprising a computer program which, when executed by a processor, implements a network planning method according to any one of claims 1-6.

Description

Network planning method, apparatus, device, storage medium, and computer program product Technical Field The present application relates to the field of wireless communications technologies, and in particular, to a network planning method, apparatus, device, storage medium, and computer program product. Background Communication network planning is a key link for ensuring efficient and reliable operation of a network, and relates to fusion analysis and intelligent decision of multidimensional heterogeneous data. Conventional network planning systems typically rely on manually integrating data from multiple independent systems, including user traffic, network load, asset status, engineering resources, etc., of a business support system, an operational support system, a management support system, a planning construction management system, etc. However, the data sources are isolated from each other, have different formats and different update periods, so that a data island phenomenon exists in the planning process, and real-time and dynamic knowledge update and decision support are difficult to realize. With the development of artificial intelligence technology, particularly a large language model, an industry large model provides a new technical path for communication network planning. Industry knowledge can be embedded into the model through knowledge injection and knowledge distillation technology, and capability migration from a large-scale teacher model to a lightweight chemical model can be realized, so that the model is suitable for scenes with limited edge equipment resources. However, the existing large-model deployment scheme still faces a plurality of challenges in the application of the communication industry, on one hand, communication data has the characteristics of wide distribution, strong sensitivity, high compliance requirement and the like, and is difficult to directly adopt a general large-model training and deployment flow, and on the other hand, the existing scheme lacks a targeted fusion mechanism for the multi-system data characteristics of the communication industry and dynamic adaptation capability for province, city and county multilevel management architecture, so that the model is difficult to meet the operation requirements of differentiation, localization and security in actual deployment. Therefore, how to construct a communication network planning system which can realize the dynamic fusion of multi-source data, support the lightweight safe deployment and has industry suitability becomes a key problem to be solved in the current technical field. Disclosure of Invention The embodiment of the application provides a network planning method which is used for solving the problems that the existing PDP filtering method lacks a targeted fusion mechanism for multi-system data characteristics in the communication industry and is oriented to dynamic adaptation capability of a province, city and county multi-level management architecture, so that the model is difficult to meet the operation requirements of differentiation, localization and security in actual deployment. The embodiment of the application also provides a network planning device which is used for solving the problems that the existing network planning lacks a targeted fusion mechanism for the multi-system data characteristics of the communication industry and the dynamic adaptation capability of a multi-level management architecture for provinces, cities and counties causes that the model is difficult to meet the operation requirements of differentiation, localization and security in actual deployment. The embodiment of the application also provides network planning equipment which is used for solving the problems that the existing network planning method lacks a targeted fusion mechanism for the multi-system data characteristics of the communication industry and is oriented to the dynamic adaptation capability of a province, city and county multi-level management architecture, so that the model is difficult to meet the operation requirements of differentiation, localization and security in actual deployment. The embodiment of the application also provides a computer readable storage medium which is used for solving the problems that the existing network planning method lacks a targeted fusion mechanism for the multi-system data characteristics of the communication industry and the dynamic adaptation capability of a multi-level management architecture for provinces, cities and counties causes that the model is difficult to meet the operation requirements of differentiation, localization and security in actual deployment. A computer program product is used for solving the problems that the existing network planning method lacks a targeted fusion mechanism for multi-system data characteristics of the communication industry and dynamic adaptation capability of a multi-level management architecture for provinces, cities and count