A load balancing migration method and system based on machine learning prediction

Abstract

The application discloses a load balancing migration method and system based on machine learning prediction, which comprises the following steps: acquiring and analyzing historical traffic load data of each service cell, determining key data features affecting traffic load, and constructing a prediction model based on the key data features to predict traffic load prediction data of each service cell; determining a high-load service cell and a target neighboring cell based on the traffic load prediction data, determining the load rate difference value of the two to determine whether load migration is needed; after determining that load migration is needed, determining user information of the high-load service cell and the target neighboring cell, and determining a migration strategy based on the user information and the load rate difference value. The application improves the accuracy and effectiveness of load balancing through an automatic and intelligent load balancing migration method, reduces the need for manual intervention, improves the operation and maintenance efficiency and operation stability of the communication network, can effectively optimize the utilization of communication network resources, and improves network performance and user experience.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a load balancing migration method and system based on machine learning prediction. Background Technology

[0002] Traditional mobile communication system load balancing methods typically perform balancing operations only when the load of the serving cell reaches a certain preset limit. However, in some specific application scenarios, user load may experience short-term peaks, requiring timely migration of high loads to low-load locations to maintain load balance. If the migration is not timely, congestion problems may occur in a short period of time, affecting the stability and normal use of the system. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention provides a load balancing migration method and system based on machine learning prediction, comprising: Obtain historical service traffic load data for each serving cell, analyze the historical service traffic load data, and determine the key data characteristics that affect service traffic load. A service traffic load prediction model is constructed based on key data features and a pre-set neural network model to predict the service traffic load of each serving cell. High-load service cells are identified based on business traffic load prediction data, and target neighbor cells that meet the requirements are identified around the high-load service cells. Determine the load rate difference between the predicted service traffic load of the high-load serving cell and the target neighboring cell, and determine whether load migration is necessary based on the load rate difference. After determining that load migration is necessary, user information for the high-load serving cell and target neighboring cells is identified, and a load balancing migration strategy is determined based on the user information and the load rate difference.

[0004] Furthermore, the acquisition of historical service traffic load data for each serving cell, and the analysis of this historical service traffic load data to determine key data characteristics affecting service traffic load, includes: Obtain historical service traffic load data for each serving cell and extract data features from the historical service traffic load data; Determine the correlation between each data feature and the business traffic load, and identify data features with a correlation greater than a preset threshold as key data features affecting the business traffic load.

[0005] Furthermore, the process of constructing a service traffic load prediction model based on key data features and a preset neural network model to predict service traffic load for each serving cell includes: A dataset is constructed based on historical business traffic load data and corresponding key data features, and the dataset is input into a preset neural network model to build an initial model for business traffic load prediction. The dataset is divided into training and test sets according to a preset ratio, and the training and test sets are input into the initial model for business traffic load prediction. The initial model for predicting business traffic load is trained and tested until it meets the preset convergence conditions, thus obtaining the business traffic load prediction model. The system acquires real-time traffic load data for each serving cell, inputs this data into the traffic load prediction model, and obtains the predicted traffic load data for each serving cell over a future period.

[0006] Furthermore, the step of determining high-load serving cells based on service traffic load prediction data and determining target neighboring cells that meet the requirements around the high-load serving cells includes: Based on the service traffic load prediction data of each serving cell, a first load sequence is constructed in descending order of load rate, and the serving cell that ranks first in the first load sequence is identified as a high-load serving cell. Identify the neighboring cells in the list of high-load serving cells, construct a second load sequence based on the service traffic load prediction data of each neighboring cell in descending order of load rate, and identify the last neighboring cell in the second load sequence as the target neighboring cell that meets the requirements around the high-load serving cell.

[0007] Furthermore, determining the load rate difference between the predicted service traffic load values ​​of the high-load serving cell and the target neighboring cell, and judging whether load migration is needed based on the load rate difference anomaly, includes: Determine the predicted traffic load of the high-load serving cell and the predicted traffic load of the target neighboring cell, and calculate the difference between the predicted traffic load of the high-load serving cell and the predicted traffic load of the target neighboring cell to obtain the load rate difference value. Determine a pre-set load rate difference threshold, and based on the relationship between the load rate difference value and the load rate difference threshold, determine whether it is necessary to migrate the service traffic load of the high-load serving cell to the target neighboring cell; If the load rate difference is less than the load rate difference threshold, it is determined that there is no need to migrate the service traffic load of the high-load serving cell to the target neighboring cell. If the load rate difference value is greater than or equal to the load rate difference threshold, it is determined that the service traffic load of the high-load serving cell needs to be migrated to the target neighboring cell.

[0008] Furthermore, after determining that load migration is necessary, the process of identifying user information for the high-load serving cell and the target neighboring cell, and determining a load balancing migration strategy based on the user information and the load rate difference, includes: After determining that load migration is necessary, the user information of the high-load serving cell and the target neighboring cell is determined. The user information includes the number of users and the historical service traffic load data of each user, and the load rate difference value is determined. The expected number of users to migrate from the high-load serving cell to the target neighboring cell is determined based on the load rate difference value and the number of users in the target neighboring cell. The migration order of users is determined based on the historical reference signal received power (RSRP) data of each user in the high-load serving cell to the target neighboring cell. The load balancing migration strategy of users is constructed based on the expected number of users and the migration order.

[0009] Furthermore, determining the expected number of users migrating from the high-load serving cell to the target neighboring cell based on the load rate difference value and the number of users in the target neighboring cell includes: Determine the predicted service traffic load of the high-load serving cell, and calculate the ratio of the load rate difference to the predicted service traffic load of the target neighboring cell to obtain the ratio coefficient; The estimated number of users to be migrated from the high-load serving cell to the target neighboring cell is calculated based on the ratio coefficient and the number of users in the target neighboring cell.

[0010] Furthermore, the formula for calculating the expected number of users is as follows: N = int((k*M) / 2), Where N is the expected number of users, int() is the floor function, k is the ratio coefficient, and M is the number of users in the target neighboring cell.

[0011] Furthermore, the determination of user migration order based on the historical reference signal received power (RSRP) data of each user in the high-load serving cell to the target neighbor cell includes: Based on the historical reference signal received power (RSRP) data of each user in the high-load serving cell to the target neighbor cell, the average historical reference signal received power (RSRP) of each user to the target neighbor cell is determined, and the first migration sequence is constructed in descending order based on the average historical reference signal received power (RSRP) of each user to the target neighbor cell. In the first migration sequence, the user arrangement order corresponding to the average received power (RSRP) of each historical reference signal is determined, and the migration order of each user is determined in order from front to back according to the user arrangement order.

[0012] The present invention also provides a load balancing migration system based on machine learning prediction, comprising: The acquisition module is used to acquire historical service traffic load data of each serving cell, and analyze the historical service traffic load data to determine the key data features that affect service traffic load. The prediction module is used to build a service traffic load prediction model based on key data features and a preset neural network model to make predictions and obtain service traffic load prediction data for each serving cell. The analysis module is used to identify high-load serving cells based on service traffic load prediction data, and to identify target neighboring cells that meet the requirements around the high-load serving cells. The judgment module is used to determine the load rate difference between the predicted service traffic load values ​​of the high-load serving cell and the target neighboring cell, and to determine whether load migration is needed based on the load rate difference value. The determination module is used to determine the user information of the high-load serving cell and the target neighboring cell after determining that load migration is required, and to determine the load balancing migration strategy based on the user information and the load rate difference value.

[0013] Compared with existing technologies, the load balancing migration method and system based on machine learning prediction proposed in this invention have the following advantages: This invention analyzes historical service traffic load data to identify key data features affecting service traffic load, and builds a neural network model based on these features for prediction. This can accurately predict the service traffic load of each serving cell and help to identify potential load peaks in advance. Based on service traffic load prediction data, this invention can quickly identify high-load service cells and further identify target neighbor cells that meet the requirements, thereby enabling targeted load balancing management. By comparing the load rate difference between the predicted service traffic load rate of the high-load serving cell and the target neighboring cell, the system can intelligently determine whether load migration is needed, avoiding unnecessary migration operations and improving system efficiency. When this invention determines that load migration is necessary, it formulates a personalized load balancing migration strategy based on user information and load rate differences between the high-load serving cell and the target neighboring cell, thereby achieving rapid load migration and optimizing network performance. In summary, this invention improves the accuracy and effectiveness of load balancing through an automated and intelligent load balancing migration method, reduces the need for manual intervention, enhances the operation and maintenance efficiency and stability of communication networks, effectively optimizes the utilization of communication network resources, improves network performance and user experience, and reduces operation and maintenance costs. It has high practical value and technical effectiveness. Attached Figure Description

[0014] Figure 1This is a schematic diagram of the process structure of the load balancing migration method based on machine learning prediction in an embodiment of the present invention; Figure 2 This is a schematic diagram of the composition of a load balancing migration system based on machine learning prediction in an embodiment of the present invention. Detailed Implementation

[0015] The specific embodiments of this application will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0016] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the platform or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0017] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying a coefficient of relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "multiple" means two or more.

[0018] like Figure 1 As shown in the embodiments of this application, a load balancing migration method based on machine learning prediction is provided, including: S100: acquiring historical service traffic load data of each serving cell, and analyzing the historical service traffic load data to determine key data features affecting service traffic load; S200: constructing a service traffic load prediction model based on key data features and a preset neural network model to predict the service traffic load of each serving cell; S300: determining high-load serving cells based on the service traffic load prediction data, and determining target neighboring cells around the high-load serving cells that meet the requirements; S400: determining the load rate difference between the predicted service traffic load rates of the high-load serving cells and the target neighboring cells, and determining whether load migration is required based on the load rate difference; S500: after determining that load migration is required, determining the user information of the high-load serving cells and the target neighboring cells, and determining a load balancing migration strategy based on the user information and the load rate difference.

[0019] Furthermore, this invention analyzes historical service traffic load data to identify key data features affecting service traffic load, and constructs a neural network model based on these features for prediction. This accurately predicts the service traffic load of each serving cell, helping to identify potential load peaks in advance. Based on the predicted service traffic load data, this invention can quickly identify high-load serving cells and further identify target neighboring cells that meet the requirements, thereby enabling targeted load balancing management. By comparing the load rate difference between the predicted service traffic load rates of high-load serving cells and target neighboring cells, the system can intelligently determine whether load migration is necessary, avoiding unnecessary migration operations and improving system efficiency. When load migration is determined to be necessary, this invention formulates a personalized load balancing migration strategy based on the user information and load rate difference between the high-load serving cell and the target neighboring cell, achieving rapid load migration and optimizing network performance. In summary, this invention, through an automated and intelligent load balancing migration method, improves the accuracy and effectiveness of load balancing, reduces the need for manual intervention, improves the operation and maintenance efficiency and operational stability of communication networks, effectively optimizes the utilization of communication network resources, improves network performance and user experience, and reduces operation and maintenance costs, demonstrating high practical value and technical effectiveness.

[0020] In the embodiments of this application, a load balancing migration method based on machine learning prediction is provided. The step of obtaining historical service traffic load data of each serving cell and analyzing the historical service traffic load data to determine key data features affecting service traffic load includes: obtaining historical service traffic load data of each serving cell and extracting data features of the historical service traffic load data; determining the correlation between each data feature and service traffic load, and determining data features with a correlation greater than a preset threshold as key data features affecting service traffic load.

[0021] In embodiments of this application, a load balancing migration method based on machine learning prediction is provided. The method involves constructing a service traffic load prediction model based on key data features and a preset neural network model to obtain service traffic load prediction data for each serving cell. This includes: constructing a dataset based on historical service traffic load data and corresponding key data features, and inputting the dataset into a preset neural network model to construct an initial service traffic load prediction model; dividing the dataset into training and testing sets according to a preset ratio, and inputting the training and testing sets into the initial service traffic load prediction model; training and testing the initial service traffic load prediction model until it meets a preset convergence condition to obtain a service traffic load prediction model; acquiring real-time service traffic load data for each serving cell, and inputting the real-time service traffic load data into the service traffic load prediction model for prediction to obtain service traffic load prediction data for each serving cell over a future period.

[0022] In an embodiment of this application, a load balancing migration method based on machine learning prediction is provided. The method for determining high-load serving cells based on service traffic load prediction data and determining target neighboring cells around the high-load serving cells that meet the requirements includes: constructing a first load sequence based on the service traffic load prediction data of each serving cell in descending order of load rate, and determining the serving cell at the first position in the first load sequence as the high-load serving cell; determining neighboring cells in the neighboring cell list of the high-load serving cell; constructing a second load sequence based on the service traffic load prediction data of each neighboring cell in descending order of load rate, and determining the neighboring cell at the last position in the second load sequence as the target neighboring cell around the high-load serving cell that meets the requirements.

[0023] In embodiments of this application, a load balancing migration method based on machine learning prediction is provided. The method involves determining the load rate difference between the predicted service traffic load values ​​of a high-load serving cell and a target neighboring cell, and determining whether load migration is necessary based on the load rate difference value. This includes: determining the predicted service traffic load values ​​of the high-load serving cell and the target neighboring cell, and calculating the difference between the predicted service traffic load values ​​of the high-load serving cell and the target neighboring cell to obtain the load rate difference value; determining a pre-set load rate difference threshold, and determining whether to migrate the service traffic load of the high-load serving cell to the target neighboring cell based on the relationship between the load rate difference value and the load rate difference threshold; if the load rate difference value is less than the load rate difference threshold, it is determined that there is no need to migrate the service traffic load of the high-load serving cell to the target neighboring cell; if the load rate difference value is greater than or equal to the load rate difference threshold, it is determined that the service traffic load of the high-load serving cell needs to be migrated to the target neighboring cell.

[0024] In embodiments of this application, a load balancing migration method based on machine learning prediction is provided. After determining that load migration is necessary, the method involves identifying user information for the high-load serving cell and the target neighboring cell, and determining a load balancing migration strategy based on the user information and the load rate difference value. This includes: after determining that load migration is necessary, identifying user information for the high-load serving cell and the target neighboring cell, where the user information includes the number of users and historical service traffic load data for each user, and determining the load rate difference value; determining the expected number of users to be migrated from the high-load serving cell to the target neighboring cell based on the load rate difference value and the number of users in the target neighboring cell; determining the migration order of users based on the historical reference signal received power (RSRP) data of each user in the high-load serving cell to the target neighboring cell; and constructing a load balancing migration strategy for users based on the expected number of users and the migration order.

[0025] In an embodiment of this application, a load balancing migration method based on machine learning prediction is provided. The method for determining the expected number of users to be migrated from a high-load serving cell to a target neighboring cell based on the load rate difference value and the number of users in the target neighboring cell includes: determining the predicted value of the service traffic load of the high-load serving cell, and calculating the ratio of the load rate difference value to the predicted value of the service traffic load rate of the target neighboring cell to obtain a ratio coefficient; and calculating the expected number of users to be migrated from the high-load serving cell to the target neighboring cell based on the ratio coefficient and the number of users in the target neighboring cell.

[0026] In an embodiment of this application, a load balancing migration method based on machine learning prediction is provided, wherein the formula for calculating the expected number of users is: N = int((k*M) / 2), Where N is the expected number of users, int() is the floor function, k is the ratio coefficient, and M is the number of users in the target neighboring cell.

[0027] In an embodiment of this application, a load balancing migration method based on machine learning prediction is provided. The method for determining the migration order of users based on the historical reference signal received power (RSRP) data of each user to the target neighboring cell includes: determining the average historical reference signal received power (RSRP) of each user to the target neighboring cell based on the historical reference signal received power (RSRP) data of each user to the target neighboring cell, and constructing a first migration sequence based on the average historical reference signal received power (RSRP) of each user to the target neighboring cell in descending order; determining the user arrangement order corresponding to each historical reference signal received power (RSRP) average value in the first migration sequence, and determining the migration order of each user in descending order according to the user arrangement order.

[0028] like Figure 2As shown in the embodiments of this application, a load balancing migration system based on machine learning prediction is provided, comprising: an acquisition module, used to acquire historical service traffic load data of each serving cell, and analyze the historical service traffic load data to determine key data features affecting service traffic load; a prediction module, used to construct a service traffic load prediction model based on the key data features and a preset neural network model to predict the service traffic load of each serving cell; an analysis module, used to determine high-load serving cells based on the service traffic load prediction data, and determine target neighboring cells around the high-load serving cells that meet the requirements; a judgment module, used to determine the load rate difference value between the predicted service traffic load values ​​of the high-load serving cell and the target neighboring cells, and determine whether load migration is required based on the load rate difference value; and a determination module, used to determine the user information of the high-load serving cell and the target neighboring cells after determining that load migration is required, and determine the load balancing migration strategy based on the user information and the load rate difference value.

[0029] In summary, this invention provides a load balancing migration method and system based on machine learning prediction, comprising: acquiring and analyzing historical service traffic load data of each serving cell to determine key data features affecting service traffic load; constructing a prediction model based on the key data features and a preset model to predict the service traffic load of each serving cell; determining high-load serving cells and their surrounding target neighboring cells based on the service traffic load prediction data, determining the load rate difference between the two, and determining whether load migration is required based on the difference; after determining that load migration is required, determining the user information of the high-load serving cell and the target neighboring cells, and determining a load balancing migration strategy based on the user information and the load rate difference. This invention, through an automated and intelligent load balancing migration method, improves the accuracy and effectiveness of load balancing, reduces the need for manual intervention, improves the operation and maintenance efficiency and operational stability of communication networks, effectively optimizes the utilization of communication network resources, and improves network performance and user experience.

[0030] Finally, it should be noted that those skilled in the art can obviously make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims and their equivalents, this invention also intends to include these modifications and variations.

[0031] The above description is merely one embodiment of the present invention, and should not be construed as limiting the scope of the invention. Any structural changes made based on the present invention, as long as they do not depart from the essence of the invention, should be considered as falling within the protection scope of the present invention and subject to its restrictions. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process and related descriptions of the platform described above can be referred to the corresponding processes in the foregoing platform embodiments, and will not be repeated here.

[0032] The term "comprising" or any other similar term is intended to cover non-exclusive inclusion, such that a process, platform, article, or device / platform that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to those processes, platforms, articles, or devices / platforms.

[0033] The technical solutions of the present invention have been described in conjunction with the accompanying drawings and further embodiments. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to closely related technical features, and the technical solutions resulting from such changes or substitutions will all fall within the scope of protection of the present invention.

[0034] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention.

Claims

1. A load balancing migration method based on machine learning prediction, characterized in that, include: Obtain historical service traffic load data for each serving cell, analyze the historical service traffic load data, and determine the key data characteristics that affect service traffic load. A service traffic load prediction model is constructed based on key data features and a pre-set neural network model to predict the service traffic load of each serving cell. High-load service cells are identified based on business traffic load prediction data, and target neighbor cells that meet the requirements are identified around the high-load service cells. Determine the load rate difference between the predicted service traffic load of the high-load serving cell and the target neighboring cell, and determine whether load migration is necessary based on the load rate difference. After determining that load migration is necessary, the user information of the high-load serving cell and the target neighboring cell is identified, and a load balancing migration strategy is determined based on the user information and the load rate difference. After determining that load migration is necessary, the process involves identifying user information for the high-load serving cell and the target neighboring cell, and determining a load balancing migration strategy based on the user information and the load rate difference. This includes: After determining that load migration is necessary, the user information of the high-load serving cell and the target neighboring cell is determined. The user information includes the number of users and the historical service traffic load data of each user, and the load rate difference value is determined. The expected number of users to migrate from the high-load serving cell to the target neighbor cell is determined based on the load rate difference value and the number of users in the target neighbor cell. The migration order of users is determined based on the historical reference signal received power (RSRP) data of each user in the high-load serving cell to the target neighbor cell. The load balancing migration strategy of users is constructed based on the expected number of users and the migration order. The process of determining the expected number of users to migrate from the high-load serving cell to the target neighboring cell based on the load rate difference value and the number of users in the target neighboring cell includes: Determine the predicted service traffic load of the high-load serving cell, and calculate the ratio of the load rate difference to the predicted service traffic load of the target neighboring cell to obtain the ratio coefficient. The estimated number of users to be migrated from the high-load serving cell to the target neighboring cell is calculated based on the ratio coefficient and the number of users in the target neighboring cell. The formula for calculating the expected number of users is as follows: N=int((k M) / 2), Where N is the expected number of users, int() is the floor function, k is the ratio coefficient, and M is the number of users in the target neighboring cell.

2. The load balancing migration method based on machine learning prediction according to claim 1, characterized in that, The process of acquiring historical service traffic load data for each serving cell and analyzing this historical service traffic load data to determine key data characteristics affecting service traffic load includes: Obtain historical service traffic load data for each serving cell and extract data features from the historical service traffic load data; Determine the correlation between each data feature and the business traffic load, and identify data features with a correlation greater than a preset threshold as key data features affecting the business traffic load.

3. The load balancing migration method based on machine learning prediction according to claim 2, characterized in that, The service traffic load prediction model, constructed based on key data features and a pre-set neural network model, is used to predict the service traffic load of each serving cell, including: A dataset is constructed based on historical business traffic load data and corresponding key data features, and the dataset is input into a preset neural network model to build an initial model for business traffic load prediction. The dataset is divided into training and test sets according to a preset ratio, and the training and test sets are input into the initial model for business traffic load prediction. The initial model for predicting business traffic load is trained and tested until it meets the preset convergence conditions, thus obtaining the business traffic load prediction model. The system acquires real-time traffic load data for each serving cell, inputs this data into the traffic load prediction model, and obtains the predicted traffic load data for each serving cell over a future period.

4. The load balancing migration method based on machine learning prediction according to claim 3, characterized in that, The process of identifying high-load serving cells based on service traffic load prediction data and identifying target neighboring cells that meet the requirements around high-load serving cells includes: Based on the service traffic load prediction data of each serving cell, a first load sequence is constructed in descending order of load rate, and the serving cell that ranks first in the first load sequence is identified as a high-load serving cell. Identify the neighboring cells in the list of high-load serving cells, construct a second load sequence based on the service traffic load prediction data of each neighboring cell in descending order of load rate, and identify the last neighboring cell in the second load sequence as the target neighboring cell that meets the requirements around the high-load serving cell.

5. The load balancing migration method based on machine learning prediction according to claim 4, characterized in that, The step of determining the load rate difference between the predicted service traffic load values ​​of the high-load serving cell and the target neighboring cell, and determining whether load migration is needed based on the load rate difference value, includes: Determine the predicted traffic load of the high-load serving cell and the predicted traffic load of the target neighboring cell, and calculate the difference between the predicted traffic load of the high-load serving cell and the predicted traffic load of the target neighboring cell to obtain the load rate difference value. Determine a pre-set load rate difference threshold, and based on the relationship between the load rate difference value and the load rate difference threshold, determine whether it is necessary to migrate the service traffic load of the high-load serving cell to the target neighboring cell; If the load rate difference is less than the load rate difference threshold, it is determined that there is no need to migrate the service traffic load of the high-load serving cell to the target neighboring cell. If the load rate difference value is greater than or equal to the load rate difference threshold, it is determined that the service traffic load of the high-load serving cell needs to be migrated to the target neighboring cell.

6. The load balancing migration method based on machine learning prediction according to claim 1, characterized in that, The process of determining the migration order of users based on the historical reference signal received power (RSRP) data of each user in a high-load serving cell to the target neighbor cell includes: Based on the historical reference signal received power (RSRP) data of each user in the high-load serving cell to the target neighbor cell, the average historical reference signal received power (RSRP) of each user to the target neighbor cell is determined, and the first migration sequence is constructed in descending order based on the average historical reference signal received power (RSRP) of each user to the target neighbor cell. In the first migration sequence, the user arrangement order corresponding to the average received power (RSRP) of each historical reference signal is determined, and the migration order of each user is determined in order from front to back according to the user arrangement order.

7. A load balancing migration system based on machine learning prediction, characterized in that, include: The acquisition module is used to acquire historical service traffic load data of each serving cell, and analyze the historical service traffic load data to determine the key data features that affect service traffic load. The prediction module is used to build a service traffic load prediction model based on key data features and a preset neural network model to make predictions and obtain service traffic load prediction data for each serving cell. The analysis module is used to identify high-load serving cells based on service traffic load prediction data, and to identify target neighboring cells that meet the requirements around the high-load serving cells. The judgment module is used to determine the load rate difference between the predicted service traffic load values ​​of the high-load serving cell and the target neighboring cell, and to determine whether load migration is needed based on the load rate difference value. The determination module is used to determine the user information of the high-load serving cell and the target neighboring cell after determining that load migration is required, and to determine the load balancing migration strategy based on the user information and the load rate difference value. After determining that load migration is necessary, the process involves identifying user information for the high-load serving cell and the target neighboring cell, and determining a load balancing migration strategy based on the user information and the load rate difference. This includes: After determining that load migration is necessary, the user information of the high-load serving cell and the target neighboring cell is determined. The user information includes the number of users and the historical service traffic load data of each user, and the load rate difference value is determined. The expected number of users to migrate from the high-load serving cell to the target neighbor cell is determined based on the load rate difference value and the number of users in the target neighbor cell. The migration order of users is determined based on the historical reference signal received power (RSRP) data of each user in the high-load serving cell to the target neighbor cell. The load balancing migration strategy of users is constructed based on the expected number of users and the migration order. The process of determining the expected number of users to migrate from the high-load serving cell to the target neighboring cell based on the load rate difference value and the number of users in the target neighboring cell includes: Determine the predicted service traffic load of the high-load serving cell, and calculate the ratio of the load rate difference to the predicted service traffic load of the target neighboring cell to obtain the ratio coefficient. The estimated number of users to be migrated from the high-load serving cell to the target neighboring cell is calculated based on the ratio coefficient and the number of users in the target neighboring cell. The formula for calculating the expected number of users is as follows: N=int((k M) / 2), Where N is the expected number of users, int() is the floor function, k is the ratio coefficient, and M is the number of users in the target neighboring cell.

Images