CN114189475B

CN114189475B - Network load optimization method and device, computer storage medium and electronic equipment

Info

Publication number: CN114189475B
Application number: CN202111537857.3A
Authority: CN
Inventors: 周笑笑; 谢卓罡; 卢哲钊; 姚莉; 贝旭峰
Original assignee: China Telecom Corp Ltd
Current assignee: China Telecom Corp Ltd
Priority date: 2021-12-15
Filing date: 2021-12-15
Publication date: 2024-03-19
Anticipated expiration: 2041-12-15
Also published as: CN114189475A

Abstract

The disclosure relates to a network load optimization method and device, a computer storage medium and electronic equipment, and relates to the technical field of big data processing, wherein the method comprises the following steps: acquiring high load monitoring points in a target range, and acquiring a high load monitoring model through first network key data of the high load monitoring points; verifying the high-load monitoring model through the second network key data to obtain target monitoring points in the high-load monitoring points, and obtaining the target monitoring model through the network key data of the target monitoring points; acquiring neighbor monitoring points of the target monitoring points, and acquiring a neighbor monitoring model through third network key data of the neighbor monitoring points; obtaining a first predicted value and a second predicted value through a target monitoring model and a neighbor monitoring model, and obtaining a load balancing suggestion list of high-load monitoring points according to the first predicted value and the second predicted value; and optimizing the network load of the high-load monitoring points according to the load balancing suggestion list. The present disclosure improves the efficiency of network load optimization.

Description

Network load optimization method and device, computer storage medium and electronic equipment

Technical Field

The embodiment of the disclosure relates to the technical field of big data processing, in particular to a network load optimization method and device, a computer readable storage medium and electronic equipment.

Background

In existing networks, particularly in larger cities, the use of the network has significant tidal and periodic effects, as well as network loading over time during the day.

In the existing wireless optimization system, the monitoring of network key data and the adjustment of network parameters are relatively separated, when the condition of unbalanced load occurs in the network, a network optimization engineer usually acquires the network key data in a network key data monitoring system to manually analyze the network key data, and then manually adjusts the static parameters in the network management, so that the network optimization system has larger time delay, has poorer portability of the adjustment of the static parameters of different monitoring points, consumes a large amount of resources, and has low network optimization efficiency.

Accordingly, there is a need to provide a new network load optimization method.

It should be noted that the information disclosed in the above background section is only for enhancing understanding of the background of the present disclosure and thus may include information that does not constitute prior art known to those of ordinary skill in the art.

Disclosure of Invention

It is an object of the present disclosure to provide a network load optimizing method, a network load optimizing apparatus, a computer-readable storage medium, and an electronic device, which further overcome, at least to some extent, the problem of low network optimization efficiency due to limitations and drawbacks of the related art.

According to one aspect of the present disclosure, there is provided a network load optimization method including:

acquiring high load monitoring points in a target range and first network key data of the high load monitoring points, and acquiring a high load monitoring model through the first network key data;

verifying the high-load monitoring model through second network key data to obtain target monitoring points in the high-load monitoring points, and obtaining a target monitoring model through the network key data of the target monitoring points;

acquiring neighbor cell monitoring points of the target monitoring points and third network key data of the neighbor cell monitoring points, and acquiring a neighbor cell monitoring model through the third network key data;

obtaining a first predicted value of the high-load monitoring point and a second predicted value of a neighboring monitoring point of the high-load monitoring point through the target monitoring model and the neighboring monitoring model, and obtaining a load balancing suggestion list of the high-load monitoring point according to the first predicted value and the second predicted value;

And optimizing the network load of the high-load monitoring points according to the load balancing suggestion list.

In one exemplary embodiment of the present disclosure, obtaining high load monitoring points within a target range, first network key data of the high load monitoring points, includes:

acquiring network key data of monitoring points in a target range;

filtering the network key data of the monitoring points according to a first missing threshold value to obtain filtered network key data;

processing the filtering network key data by an interpolation method to obtain target network key data in the target range;

and distinguishing the monitoring points according to preset high-load monitoring point judging conditions to obtain the high-load monitoring points in the target range and first network key data corresponding to the high-load monitoring points.

In an exemplary embodiment of the present disclosure, distinguishing the monitoring points according to a preset high-load monitoring point judgment condition includes:

when the frequency band of the monitoring point is 800M and the bandwidth is 5M, acquiring the utilization rate of the physical resource block, the total flow, the number of users and the maximum wireless resource control connection number of the monitoring point in the self-busy time;

In a preset time period, when any monitoring point is busy, the utilization rate of the physical resource blocks is not less than a first preset utilization rate, and the times of the total flow not less than the first preset flow are greater than preset times; or (b)

When the utilization rate of the physical resource block in the self busy time is not less than a first preset utilization rate, the number of users is not less than a first preset use number, and the number of times that the maximum wireless resource control connection number is not less than the first preset connection number is greater than a preset number of times, determining the monitoring point as a high-load monitoring point;

wherein the self-busy hour is the hour with the largest total flow of the packet data convergence protocol layer in 24 hours.

when the frequency band of the monitoring point is 1.8G/2.1G bandwidth is 20M, acquiring the utilization rate of the physical resource block, the total flow, the number of users and the maximum wireless resource control connection number of the monitoring point in the self-busy time;

in a preset time period, when any monitoring point is busy, the physical resource block utilization rate is not less than a second preset utilization rate, and the times that the total flow is not less than the second preset flow are greater than preset times; or (b)

When the utilization rate of the physical resource block in the self busy time is not less than a second preset utilization rate, the number of users is not less than a second preset use number, and the number of times that the maximum wireless resource control connection number is not less than the second preset connection number is greater than a preset number of times, determining the monitoring point as a high-load monitoring point;

In an exemplary embodiment of the present disclosure, the network load optimization method further includes:

acquiring the downlink physical resource block utilization rate, the maximum radio resource control connection number and the total flow of a packet data convergence protocol layer included in the first network key data;

training the preset neural network model by utilizing the downlink physical resource utilization rate of the high-load monitoring points in the target range to obtain a resource utilization rate prediction model;

training the preset neural network model by utilizing the maximum wireless resource control connection number of the high-load monitoring points in the target range to obtain a resource control connection number prediction model;

and training the preset neural network model by utilizing the total flow of the packet data convergence protocol layer of the high-load monitoring points in the target range to obtain a flow prediction model.

In an exemplary embodiment of the present disclosure, verifying the high load monitoring model by the second network key data, to obtain a target monitoring point of the high load monitoring points includes:

acquiring second network key data of the high-load monitoring point in a second preset time period, and inputting the second network key data into the high-load monitoring model to obtain a high-load prediction result;

the downlink physical resource block utilization rate, the maximum wireless resource control connection number and the total flow of the packet data convergence protocol layer of the high-load monitoring point are respectively input into a resource utilization rate prediction model, a resource control connection number prediction model and a flow prediction model to obtain a resource block utilization rate prediction value, a connection number prediction value and a flow prediction value;

verifying the high load prediction result, the resource block utilization ratio prediction value, the connection number prediction value and the flow prediction value to obtain prediction accuracy;

and determining any monitoring point of the high load monitoring points as a target monitoring point according to the accuracy rate of the high load occurrence time point in the high load prediction result of any monitoring point, the accuracy rate and recall rate of the high load occurrence time point prediction, the occurrence frequency of the high load occurrence time point and the prediction accuracy rate.

In an exemplary embodiment of the present disclosure, obtaining a neighbor cell monitoring point of the target monitoring point includes:

acquiring the coverage range of the target monitoring points and the coverage range of the monitoring points included in the target range;

determining a coverage polygon of the target monitoring point according to the coverage of the target monitoring point and determining the coverage polygon of the monitoring point according to the coverage of the detection point included in the target range;

when the overlapping of the coverage polygon of the target monitoring point and the coverage polygon of the monitoring point is determined, determining the monitoring point as a neighboring monitoring point of the target monitoring point, and generating a neighboring table according to the neighboring monitoring point;

and acquiring neighbor monitoring points with the times of reaching high load in a third preset time period included in the neighbor table being greater than the third preset times, and deleting the neighbor monitoring points from the neighbor table.

In an exemplary embodiment of the present disclosure, obtaining a load balancing suggestion list of the high load monitoring point according to the first predicted value and the second predicted value includes:

acquiring a high load occurrence time point included in the first predicted value and a predicted value of a neighbor cell monitoring point included in the neighbor cell table included in the second predicted value, and deleting the neighbor cell monitoring point in the neighbor cell table according to the high load occurrence time point and the predicted value of the neighbor cell monitoring point to obtain a neighbor cell monitoring point of a target monitoring point;

Sorting the neighbor monitoring points according to the frequency bands and the priorities;

and acquiring ordered neighbor monitoring points, and taking the ordered neighbor monitoring points as a load balancing suggestion list of the high-load monitoring points.

According to one aspect of the present disclosure, there is provided a network load optimizing apparatus including:

the high load monitoring model training module is used for acquiring high load monitoring points in a target range and first network key data of the high load monitoring points, and training a preset neural network model through the first network key data to obtain a high load monitoring model;

the target monitoring model training module is used for verifying the high-load monitoring model through second network key data to obtain target monitoring points in the high-load monitoring points, and training the preset neural network model through the network key data of the target monitoring points to obtain a target monitoring model;

the neighbor cell monitoring model training module is used for acquiring neighbor cell monitoring points of the target monitoring points and third network key data of the neighbor cell monitoring points, and training the preset neural network model through the third network key data to obtain a neighbor cell monitoring model;

And the suggestion list generation module is used for obtaining a first predicted value of the high-load monitoring point and a second predicted value of a neighboring cell monitoring point of the high-load monitoring point through the target monitoring model and the neighboring cell monitoring model, and obtaining a load balancing suggestion list of the high-load monitoring point according to the first predicted value and the second predicted value.

And the network load optimization module is used for optimizing the network load of the high-load monitoring points according to the load balancing suggestion list.

According to one aspect of the present disclosure, there is provided a readable storage medium having stored thereon a computer program which, when executed by a processor, implements the network load optimization method according to any of the above-described exemplary embodiments.

According to one aspect of the present disclosure, there is provided an electronic device including:

a processor; and

a memory for storing executable instructions of the processor;

wherein the processor is configured to perform the network load optimization method of any of the above-described exemplary embodiments via execution of the executable instructions.

According to the network load optimization method provided by the embodiment of the disclosure, on one hand, after network key data of the monitoring points are obtained, a target monitoring model and a neighbor monitoring model are obtained through the network key data, a first predicted value of the high-load monitoring points and a second predicted value of the neighbor monitoring points of the high-load monitoring points are obtained through the target monitoring model and the neighbor monitoring model, a load balancing suggestion list of the high-load monitoring points is obtained according to the first predicted value and the second predicted value, and network load of the high-load monitoring points is optimized according to the load balancing suggestion list, so that the problem that in the prior art, when network load is unbalanced, a network optimization engineer is required to manually analyze the network key data first, and then manually adjust parameters in a network management, so that network optimization efficiency is low is solved; on the other hand, after the high-load monitoring model is obtained, the high-load monitoring model is verified through the second network key data, the target monitoring points included in the high-load monitoring points are obtained, the preset neural network model is trained through the network key data of the target monitoring points, the target monitoring model is obtained, the network key data of the high-load monitoring points are predicted through the target monitoring model, and the accuracy of prediction is improved.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the disclosure and together with the description, serve to explain the principles of the disclosure. It will be apparent to those of ordinary skill in the art that the drawings in the following description are merely examples of the disclosure and that other drawings may be derived from them without undue effort.

Fig. 1 schematically illustrates a flow chart of a network load optimization method according to an example embodiment of the present disclosure.

Fig. 2 schematically illustrates a block diagram of a network load optimization system according to an example embodiment of the present disclosure.

Fig. 3 schematically illustrates a method flow diagram for acquiring high load monitoring points within a target range and first network key data for the high load monitoring points, according to an example embodiment of the present disclosure.

Fig. 4 schematically illustrates a flow chart of a method of distinguishing monitoring points according to preset high load monitoring point judgment conditions according to an example embodiment of the present disclosure.

Fig. 5 schematically illustrates a method flowchart for differentiating monitoring points according to preset high load monitoring point judgment conditions according to an example embodiment of the present disclosure.

Fig. 6 schematically illustrates a flow chart of a network load optimization method according to an example embodiment of the present disclosure.

Fig. 7 schematically illustrates a flowchart of a method for validating a high load monitoring model for target monitoring points in a high load monitoring point by second network critical data according to an example embodiment of the disclosure.

Fig. 8 schematically illustrates a flowchart of a method of obtaining neighbor monitoring points for a target monitoring point according to an example embodiment of the present disclosure.

Fig. 9 schematically illustrates a method flowchart for deriving a load balancing inventory proposal for high load monitoring points from a first predicted value, a second predicted value, according to an example embodiment of the disclosure.

Fig. 10 schematically illustrates a method flow diagram for network load optimization in accordance with an example embodiment of the present disclosure.

Fig. 11 schematically illustrates an apparatus block diagram for network load optimization in accordance with an example embodiment of the present disclosure.

Fig. 12 schematically illustrates an electronic device for implementing the above-described network load optimization method according to an example embodiment of the present disclosure.

Detailed Description

Example embodiments will now be described more fully with reference to the accompanying drawings. However, the exemplary embodiments may be embodied in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the present disclosure. One skilled in the relevant art will recognize, however, that the aspects of the disclosure may be practiced without one or more of the specific details, or with other methods, components, devices, steps, etc. In other instances, well-known technical solutions have not been shown or described in detail to avoid obscuring aspects of the present disclosure.

Furthermore, the drawings are merely schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus a repetitive description thereof will be omitted. Some of the block diagrams shown in the figures are functional entities and do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software or in one or more hardware modules or integrated circuits or in different networks and/or processor devices and/or microcontroller devices.

In existing networks, especially in larger cities, there are often significant tidal effects and periodic effects (such as weekend effects), for example, during working hours, users are mostly concentrated in office areas, while during working hours, users are mostly concentrated in residences, and during weekends or holidays, the network load in places such as malls, scenic spots, stations and the like can also vary significantly.

In the existing wireless optimization system, the monitoring of network key data and the adjustment of network parameters are relatively separated, network optimizers often manually acquire network monitoring data from a monitoring system of the network key data and analyze the network monitoring data, and then manually acquire static parameters in a network management system and adjust the static parameters, so that the network optimization system often has larger time delay. In addition, one of the most common means for optimizing the high-load cell in the related art is a large-scale scene, non-differential configuration reselection and switching parameters, and the manufacturer has the characteristic function of load balancing, but can only be applied to the same station and the same manufacturer. The scheme cannot match the business change of the tidal effect, when the number of high-load cells is obviously increased, the manual adjustment cannot be completed in time, and the manual adjustment range is extremely limited.

Based on one or more of the above problems, in this exemplary embodiment, there is provided a network load optimization method, which may be executed on a server, a server cluster, or a cloud server, etc.; of course, those skilled in the art may also operate the method of the present invention on other platforms as required, and this is not a particular limitation in the present exemplary embodiment. Referring to fig. 1, the network load optimization method may include the steps of:

s110, acquiring high-load monitoring points in a target range and first network key data of the high-load monitoring points, and acquiring a high-load monitoring model through the first network key data;

s120, verifying the high-load monitoring model through second network key data to obtain target monitoring points in the high-load monitoring points, and obtaining a target monitoring model through the network key data of the target monitoring points;

s130, acquiring neighbor cell monitoring points of the target monitoring points and third network key data of the neighbor cell monitoring points, and obtaining a neighbor cell monitoring model through the third network key data;

s140, obtaining a first predicted value of the high-load monitoring point and a second predicted value of a neighboring cell monitoring point of the high-load monitoring point through the target monitoring model and the neighboring cell monitoring model, and obtaining a load balancing suggestion list of the high-load monitoring point according to the first predicted value and the second predicted value;

And S150, optimizing the network load of the high-load monitoring point according to the load balancing suggestion list.

According to the network load optimization method, on one hand, after network key data of the monitoring points are obtained, a target monitoring model and a neighbor monitoring model are obtained through the network key data, a first predicted value of the high-load monitoring points and a second predicted value of neighbor monitoring points of the high-load monitoring points are obtained through the target monitoring model and the neighbor monitoring model, a load balancing suggestion list of the high-load monitoring points is obtained according to the first predicted value and the second predicted value, and network loads of the high-load monitoring points are optimized according to the load balancing suggestion list, so that the problem that in the prior art, when the network loads are unbalanced, a network optimization engineer is required to manually analyze the network key data, and then manually adjust parameters in a network management system, so that network optimization efficiency is low is solved; on the other hand, after the high-load monitoring model is obtained, the high-load monitoring model is verified through the second network key data, the target monitoring points included in the high-load monitoring points are obtained, the preset neural network model is trained through the network key data of the target monitoring points, the target monitoring model is obtained, the network key data of the high-load monitoring points are predicted through the target monitoring model, and the accuracy of prediction is improved.

The steps involved in the network load optimization method according to the exemplary embodiment of the present disclosure are explained and described in detail below.

First, application scenarios and purposes of the exemplary embodiments of the present disclosure are explained and explained. In particular, the exemplary embodiments of the present disclosure may be applied to network load optimization, mainly to study how to improve the efficiency of network load optimization.

In the method, network key data of monitoring points in a received target range are used as a basis, firstly, the obtained network key data are preprocessed, network key data with the missing proportion larger than a preset missing proportion threshold value are removed, the monitoring points are distinguished, and high-load monitoring points and network key data of the high-load monitoring points are obtained; then training a preset neural network model through network key data of the high-load monitoring point to obtain a high-load monitoring model, and verifying the high-load monitoring model through network key data of another time period of the high-load monitoring point to obtain a target monitoring point included in the high-load monitoring point; then training a preset neural network model through network key data of the target monitoring point to obtain the target monitoring model, and simultaneously acquiring neighbor monitoring points of the target monitoring point and the network key data of the neighbor monitoring point, and training the preset neural network model through the network key data of the neighbor monitoring point to obtain the neighbor monitoring model; and finally, obtaining a first predicted value and a second predicted value of the high-load monitoring point and the neighbor monitoring point of the high-load monitoring point through the target monitoring model and the neighbor monitoring model, obtaining a load balancing suggestion list of the high-load monitoring point through the first predicted value and the second predicted value, and carrying out load balancing on the high-load monitoring point according to the load balancing suggestion list, thereby improving the efficiency of network load optimization of the high-load monitoring point.

Next, explanation and explanation will be made of a network load optimizing system involved in an exemplary embodiment of the present disclosure. Referring to fig. 2, the network load optimization system may include a watchpoint classification module 210, a network critical data preprocessing module 220, a model training module 230, and a load balancing module 240. The monitoring point classification module 210 is configured to obtain a frequency band, a bandwidth, a PRB (Physical Resource Block ) utilization rate, a maximum RRC (Radio Resource Control ) connection number, a PDCP (Packet Data Convergence Protocol ) layer total flow, a total flow, and a number of users of the monitoring point, and classify the monitoring point according to the obtained data, so as to obtain a high-load monitoring point; the network key data preprocessing module 220 is connected with the monitoring point classification module 210 in a network manner, and is used for acquiring network key data of the monitoring points, filtering the network key data with the missing proportion higher than a first missing threshold value in the acquired network key data, filtering the network key data of an input model through a second missing threshold value when predicting, and filling the rest network key data through an interpolation method; the model training module 230 is in network connection with the network key data preprocessing module 220 and is used for acquiring network key data of the high-load monitoring points, training a preset neural network model through the network key data of the high-load monitoring points to obtain a high-load monitoring model, verifying the high-load monitoring model to obtain target monitoring points included in the high-load monitoring points, acquiring the network key data of the target monitoring points, training the preset neural network model through the network key data of the target monitoring points to obtain a target monitoring model, and acquiring neighbor monitoring points of the target monitoring points, training the preset neural network model through the network key data of the neighbor monitoring points to obtain a neighbor monitoring model; the load balancing module 240 is in network connection with the model training module 230, and is configured to obtain network key data of the high load monitoring point and network key data of a neighboring cell monitoring point of the high load monitoring point, respectively input the network key data of the high load monitoring point and the network key data of the neighboring cell monitoring point into the neighboring cell monitoring model of the target monitoring model to obtain a first predicted value and a second predicted value, obtain a load balancing suggestion list of the high load monitoring point through the first predicted value and the second predicted value, and optimize a network load of the high load monitoring point through the load balancing suggestion list.

Hereinafter, the steps S110 to S150 will be explained and described in detail with reference to fig. 2.

In step S110, a high load monitoring point in a target range and first network key data of the high load monitoring point are obtained, and a high load monitoring model is obtained through the first network key data.

The target range may be any city, any community, or any public building, and in this example embodiment, the target range is not specifically limited. All monitoring points in the target range can be obtained, and the monitoring points in the target range are judged according to preset high-load monitoring point judgment conditions, so that the high-load monitoring points in the target range are obtained. The first network key data may be network key data of a self-busy hour within a preset time, for example, the first network key data may be network key data of a self-busy hour of 60 days; the network key data of the monitoring points comprise: downlink PRB utilization, maximum RRC connection number, PDCP layer total traffic. Training a preset neural network model through first network key data to obtain a high-load monitoring model, wherein the preset neural network model is an Amazon probabilistic neural network deep AR model, neurons of the model adopt a Memory cycle neural network model type, namely LSTM (Long Short-Term Memory) or GRU (Gate Recurrent Unit, gate control cycle unit), and an reasoning process is divided into two stages of a training process and a prediction process.

In this example embodiment, referring to fig. 3, the obtaining the high load monitoring point in the target range and the first network key data of the high load monitoring point may include step S310 to step S340:

s310, acquiring network key data of monitoring points in a target range;

s320, filtering the network key data of the monitoring points according to a first missing threshold value to obtain filtered network key data;

s330, processing the filter network key data through an interpolation method to obtain target network key data in the target range;

and S340, distinguishing the monitoring points according to preset high-load monitoring point judging conditions to obtain the high-load monitoring points in the target range and first network key data corresponding to the high-load monitoring points.

Hereinafter, step S310 to step S340 will be explained and explained. Specifically, firstly, monitoring points in a target range and network key data of the monitoring points are obtained, and instability is brought to prediction of a model due to the fact that high missing proportion often occurs in the network data, however, the model can process the situation that the data is really not high, so that the obtained network key data of the monitoring points can be filtered through a first missing threshold value during model training to obtain filtered network key data; the first missing threshold may be 0.3 or 0.4, and in this exemplary embodiment, the first missing threshold is not specifically limited; the filtering network key data can be filled through an interpolation method, the interpolation method is to estimate an approximate value of a function at a missing point by using a value of a known time point through a one-dimensional function, a fitting curve is required to be as possible to be a known data point, and the fitting variance is minimum, wherein the piecewise interpolation method is to fit different functions to different parts of a sequence by considering the periodicity of the time sequence, and the curves among the functions are smoothly butted, so that the accuracy of local fitting is improved; finally, the monitoring points can be distinguished according to preset high-load monitoring point judging conditions, and the high-load monitoring points included in the monitoring points and first network key data corresponding to the high-load monitoring points are obtained.

In addition to filtering the network key data of the monitoring points in the training stage, in the prediction stage, the network key data of the monitoring points can be filtered through the second missing threshold value, so that the accuracy of prediction is improved.

Further, in this exemplary embodiment, referring to fig. 4, distinguishing the monitoring points according to the preset high load monitoring point judgment condition may include steps S410 to S430:

s410, when the frequency band of the monitoring point is 800M and the bandwidth is 5M, acquiring the utilization rate of the physical resource block, the total flow, the number of users and the maximum wireless resource control connection number of the monitoring point in the self-busy time;

s420, in a preset time period, when any monitoring point is busy, the utilization rate of the physical resource block is not less than a first preset utilization rate, and the frequency of the total flow not less than the first preset flow is greater than a preset frequency; or (b)

S430, determining the monitoring point as a high-load monitoring point when the utilization rate of the physical resource block in the self busy hour is not less than a first preset utilization rate, the number of users is not less than a first preset use number, and the number of times that the maximum wireless resource control connection number is not less than the first preset connection number is greater than a preset number of times;

Hereinafter, step S410 to step S430 will be explained and explained. Specifically, firstly, judging the frequency band and the bandwidth of the monitoring point, when the frequency band of the monitoring point is 800M and the bandwidth is 5M, acquiring the utilization rate of the PRB of the monitoring point in the busy hour, the total flow of the monitoring point, the number of users used by the monitoring point and the maximum RRC connection number of the monitoring point, wherein the judging of the monitoring point can comprise the following steps: the PRB utilization rate of the monitoring point in the self busy hour is more than or equal to a first preset utilization rate, and the total flow of the monitoring point is more than or equal to a first preset flow; condition II: the PRB utilization rate of the monitoring point in the self busy hour is more than or equal to the first preset utilization rate, the number of users of the monitoring point is more than or equal to the first preset use number, and the maximum RRC connection number is more than or equal to the first preset connection number; when the number of times that the first condition is met is larger than the preset number of times in the preset time or the number of times that the second condition is met is larger than the preset number of times in the preset time, the monitoring point can be determined to be a high-load monitoring point. Wherein, the self busy time is the hour of maximum total flow of the PDCP layer in 24 hours. The first preset usage rate may be 50% or 55%, which is not specifically limited in this example embodiment, and the first preset flow rate may be 1.5GB or 2GB, which is not specifically limited in this example embodiment; when the scene of the monitoring point is a cell, the first preset usage number may be a number of users using a network in the cell, and the value of the first preset usage number may be 75 or 80, which is not specifically limited in this example embodiment; the first preset number of connections may be 50 or 55, which is not particularly limited in this exemplary embodiment. The number of times that the first condition or the second condition is met in the preset time meets the preset number of times, the whole day data of 30 days of the monitoring points in the target range can be obtained, when the first condition or the second condition is met in at least 4 days of self-busy time in any one continuous 7 days of 30 days, the monitoring points can be determined to be high-load monitoring points.

In addition, when the frequency band of the monitoring point is 1.8G/2.1G with 20M, referring to fig. 5, the step S510-step S530 may be included to distinguish the monitoring points according to a preset high load monitoring point judgment condition:

s510, when the frequency band of the monitoring point is 1.8G/2.1G bandwidth is 20M, acquiring the utilization rate of the physical resource block, the total flow, the number of users and the maximum wireless resource control connection number of the monitoring point in the busy hour;

s520, in a preset time period, when any monitoring point is busy, the utilization rate of the physical resource block is not less than a second preset utilization rate, and the frequency of the total flow not less than the second preset flow is greater than a preset frequency; or (b)

S530, determining the monitoring point as a high-load monitoring point when the utilization rate of the physical resource block in the self busy hour is not less than a second preset utilization rate, the number of users is not less than a second preset use number, and the number of times that the maximum wireless resource control connection number is not less than the second preset connection number is greater than the preset number of times;

Hereinafter, step S510 to step S530 will be explained and explained. Specifically, when the frequency band of the monitoring point is 1.8G/2.1G M bandwidth is 20M, the self-busy PRB utilization rate of the monitoring point, the total flow of the monitoring point, the number of users used by the monitoring point, and the maximum RRC connection number of the monitoring point can be obtained, and the judging of the monitoring point can include the condition one: the PRB utilization rate of the monitoring point in the self busy hour is more than or equal to a second preset utilization rate, and the total flow of the monitoring point is more than or equal to a second preset flow; condition II: the PRB utilization rate of the monitoring point in the self busy hour is more than or equal to the second preset utilization rate, the number of users of the monitoring point is more than or equal to the second preset use number, and the maximum RRC connection number is more than or equal to the second preset connection number; when the number of times that the first condition is met is larger than the preset number of times in the preset time or the number of times that the second condition is met is larger than the preset number of times in the preset time, the monitoring point can be determined to be a high-load monitoring point. Wherein, the self busy time is the hour of maximum total flow of the PDCP layer in 24 hours. The second preset usage rate may be 50% or 55%, which is not specifically limited in this exemplary embodiment, and the second preset flow rate may be 6GB or 6.5GB, which is not specifically limited in this exemplary embodiment; when the scene of the monitoring point is a cell, the first preset usage number may be a number of users using a network in the cell, and the value of the first preset usage number may be 300 or 320, which is not specifically limited in this example embodiment; the first preset number of connections may be 200 or 250, which is not specifically limited in this exemplary embodiment. The number of times that the first condition or the second condition is met in the preset time meets the preset number of times, the whole day data of 30 days of the monitoring points in the target range can be obtained, when the first condition or the second condition is met in at least 4 days of self-busy time in any one continuous 7 days of 30 days, the monitoring points can be determined to be high-load monitoring points.

After the high-load monitoring points are obtained, network key data of the high-load monitoring points can be obtained, and a preset neural network model is trained through the obtained network key data of the high-load monitoring points to obtain a high-load monitoring model, wherein the preset neural network model can be an Amazon probabilistic neural network deep AR model. In the related art, ARIMA (Autoregressive Integrated Moving Average model, differential integration moving average autoregressive model), holt-windows (holter-temperature) method, exponential smooth moving average line, LSTM (Long Short-Term Memory network) neural network, and fbarophet (time series prediction) algorithm have the following limitations: (1) The prediction of a single variable is limited, and the prediction of various network key data of a large number of monitoring points cannot be adapted; (2) The requirement on the regularity of the time sequence is high, a plurality of assumption conditions need to be met, but network parameters of different monitoring points are complex and the situation is changeable; (3) Models in the related art generally cannot predict a long time window (for example, 24 hours a day), and errors of prediction are accumulated continuously along with the extension of time; however, the deep ar model can solve the above-mentioned drawbacks, and the deep ar model can learn the potential probability distribution parameters of the time series through the memory neural network structure, and sample the prediction result from the probability distribution model in the output stage, so as to improve the stability of the prediction result in a long time window and reduce the deviation of the prediction value.

In the example, the high-load monitoring points and the network key data of the high-load monitoring points are acquired by filtering, filling and judging the network key data of the monitoring points, so that the accuracy of model training is improved.

In step S120, the high load monitoring model is verified by the second network key data, so as to obtain a target monitoring point of the high load monitoring points, and the target monitoring model is obtained by the network key data of the target monitoring point.

The second network key data may be network key data of the high load monitoring point in the time of 60 days of self-busy, or may be network key data of the high load monitoring point in the time of 65 days of self-busy, which is not specifically limited in this example embodiment. Training a preset neural network model through network key data of target monitoring points to obtain a target monitoring model, wherein the preset neural network model is a deep AR model.

In this example embodiment, multiple data included in the network key data may also be obtained, and the preset neural network model may be trained by the multiple data to obtain a corresponding prediction model, and referring to fig. 6, the network load optimization method may further include steps S610 to S640:

S610, acquiring the downlink physical resource block utilization rate, the maximum radio resource control connection number and the total flow of a packet data convergence protocol layer included in the first network key data;

s620, training the preset neural network model by utilizing the downlink physical resource utilization rate of the high-load monitoring points in the target range to obtain a resource utilization rate prediction model;

s630, training the preset neural network model by using the maximum wireless resource control connection number of the high-load monitoring points in the target range to obtain a resource control connection number prediction model;

and S640, training the preset neural network model by using the total flow of the packet data convergence protocol layer of the high-load monitoring points in the target range to obtain a flow prediction model.

Hereinafter, step S610 to step S640 will be explained and explained. Specifically, the network key data includes: the downlink PRB utilization rate, the maximum RRC connection number and the total flow of the PDCP layer can be obtained, the downlink PRB utilization rate in the first network key data of the high-load monitoring point can be obtained, and a preset neural network model is trained through the downlink PRB utilization rate, so that a resource utilization rate prediction model is obtained; the method comprises the steps that the maximum RRC connection number in first network key data can be obtained, and a preset neural network model is trained through the maximum RRC connection number, so that a resource control connection number prediction model is obtained; and acquiring the total flow of the PDCP layer in the first network key data, and training a preset neural network model through the total flow of the PDCP layer to obtain a flow prediction model. The deep AR model can also model and predict a large number of network parameter time sequences with similar characteristics, so that the prediction efficiency of network key data is improved.

After the prediction models corresponding to the various data included in the network key data are obtained, the target monitoring points included in the high-load monitoring points can be obtained through the various prediction models and the high-load monitoring model. Referring to fig. 7, the verification of the high load monitoring model by the second network key data to obtain the target monitoring point of the high load monitoring points may include steps S710 to S740:

s710, acquiring second network key data of the high-load monitoring point in a second preset time period, and inputting the second network key data into the high-load monitoring model to obtain a high-load prediction result;

s720, respectively inputting the utilization rate of the downlink physical resource block, the maximum wireless resource control connection number and the total flow of the packet data convergence protocol layer of the high-load monitoring point into a resource utilization rate prediction model, a resource control connection number prediction model and a flow prediction model to obtain a resource block utilization rate prediction value, a connection number prediction value and a flow prediction value;

s730, verifying the high load prediction result, the resource block utilization ratio prediction value, the connection number prediction value and the flow prediction value to obtain prediction accuracy;

And S740, determining any monitoring point of the high load monitoring points as a target monitoring point according to the accuracy rate of the high load occurrence time point in the high load prediction result of any monitoring point, the accuracy rate and recall rate of the high load occurrence time point prediction, the occurrence frequency of the high load occurrence time point and the prediction accuracy rate.

Hereinafter, step S710 to step S740 will be explained and explained. Specifically, first, second network key data of the high-load monitoring point may be acquired, where the second network key data may be self-busy time network key data of the high-load monitoring point for 10 days or 15 days, and this is not specifically limited in this example embodiment; after the second network data of the high-load monitoring point in the second preset time period is obtained, the second network data can be input into a high-load monitoring model to obtain a high-load prediction result, wherein the high-load prediction result is the high-load occurrence time point of the predicted monitoring point; after the prediction result of the high-load monitoring point is obtained, the downlink PRB utilization rate, the maximum RRC connection number and the total flow of the PDCP layer which are included in the network key data can be respectively input into a corresponding prediction model to respectively obtain a resource block utilization rate prediction value, a connection number prediction value and a flow prediction value; then, verifying the obtained resource block utilization rate predicted value, the connection number predicted value and the flow predicted value to obtain the accuracy of the predicted value; finally, determining the accuracy of the high load occurrence time point included in the high load prediction result of any monitoring point, the accuracy and recall rate of the high load occurrence time point prediction, the occurrence frequency of the high load occurrence time point and the prediction accuracy; and obtaining target monitoring points included in the high-load monitoring points according to the accuracy of the high-load occurrence time points of the high-load monitoring points, the accuracy and recall rate of the prediction of the high-load occurrence time points, the occurrence frequency of the high-load occurrence time points and the prediction accuracy.

In this example embodiment, after the target monitoring points included in the high-load monitoring points are obtained, network key data within a preset time period of the target monitoring points may be obtained, and a preset neural network model is trained by the network key data, so as to obtain the target monitoring model. The preset neural network model is a deep AR model.

In step S130, neighboring cell monitoring points of the target monitoring points and third network key data of the neighboring cell monitoring points are obtained, and a neighboring cell monitoring model is obtained through the third network key data.

The neighbor cell monitoring points of the target monitoring points can be obtained through covering the polygons, and the neighbor cell monitoring points and the target monitoring points are overlapped through grids; training a preset neural network model through third network key data to obtain a neighbor cell monitoring model, wherein the preset neural network model is a deep AR model.

In this example embodiment, referring to fig. 8, acquiring the neighbor cell monitoring point of the target monitoring point may include steps S810-S840:

step S810, acquiring the coverage range of the target monitoring points and the coverage range of the monitoring points included in the target range;

S820, determining a coverage polygon of the target monitoring point according to the coverage of the target monitoring point and determining the coverage polygon of the monitoring point according to the coverage of the detection point included in the target range;

step S830, when the overlapping of the covering polygon of the target monitoring point and the covering polygon of the monitoring point is determined, determining the monitoring point as a neighboring monitoring point of the target monitoring point, and generating a neighboring table according to the neighboring monitoring point;

and S840, acquiring neighbor monitoring points with the times of reaching high load in a third preset time period included in the neighbor table being greater than the third preset times, and deleting the neighbor monitoring points from the neighbor table.

Hereinafter, the steps S810 to S840 will be explained and explained. Specifically, first, the coverage of the target monitoring points and the coverage of the monitoring points included in the target range can be obtained; then, covering the target monitoring points and monitoring points, except the target monitoring points, included in the target range by using a covering polygon, wherein the covering polygon may be a 20m×20m grid or a 30m×30m grid, and in this example embodiment, the covering polygon is not specifically limited; when grids are overlapped between the covering polygon of any monitoring point included in the target range and the covering polygon of the target monitoring point after covering the target monitoring point and the monitoring point in the target range, determining any monitoring point included in the target range as a neighboring area monitoring point of the target monitoring point, and generating a neighboring area table of the target monitoring point, wherein the neighboring area table comprises all neighboring area monitoring points with grids overlapped with the target monitoring point. And after acquiring the neighbor cell table of the target monitoring point, acquiring neighbor cell monitoring points with the times of reaching high load in a third preset time period included in the neighbor cell table being larger than the third preset times, and deleting the monitoring points from the neighbor cell table. The third preset time period may be 7 days in the past, the third preset times may be 4 times, and the third preset time period and the third preset times are not specifically limited in this example embodiment.

In this example embodiment, after the neighbor cell monitoring point of the target monitoring point is obtained, third network key data of the neighbor cell monitoring point may be obtained, and the preset neural network model is trained by the third network key data, so as to obtain the neighbor cell monitoring model.

In step S140, a first predicted value of the high load monitoring point and a second predicted value of the neighboring cell monitoring point of the high load monitoring point are obtained through the target monitoring model and the neighboring cell monitoring model, and a load balancing suggestion list of the high load monitoring point is obtained according to the first predicted value and the second predicted value.

Before the first predicted value and the second predicted value are obtained through the target monitoring model and the neighbor monitoring model, network key data input to the target monitoring model and network key data input to the neighbor monitoring model can be filtered through a second missing threshold.

In this example embodiment, referring to fig. 9, the obtaining, according to the first predicted value and the second predicted value, a load balancing suggestion list of the high load monitoring point may include steps S910 to S930:

s910, obtaining a high load occurrence time point included in the first predicted value and a predicted value of a neighbor cell monitoring point included in the neighbor cell table included in the second predicted value, and deleting the neighbor cell monitoring point in the neighbor cell table according to the high load occurrence time point and the predicted value of the neighbor cell monitoring point to obtain a neighbor cell monitoring point of a target monitoring point;

S920, sorting the neighbor monitoring points according to the frequency bands and the priorities;

and S930, acquiring ordered neighbor monitoring points, and taking the ordered neighbor monitoring points as a load balancing suggestion list of the high-load monitoring points.

Hereinafter, step S910 to step S930 will be explained and explained. Specifically, first, a high load occurrence time point of a high load monitoring point included in a first predicted value and a high load occurrence time point of a neighboring cell monitoring point included in a second predicted value are obtained, and the neighboring cell monitoring point included in a neighboring cell table is deleted according to the high load occurrence time point in the first predicted value and the high load occurrence time point in the second predicted value, specifically: and determining an adjustment time interval of the high load monitoring point according to the high load occurrence time point of the high load monitoring point, and deleting the neighbor monitoring point from the neighbor table when the high load occurrence time point of the neighbor monitoring point is within a preset time period of the adjustment time interval. The preset period of the adjustment time interval may be 2 hours before and after the adjustment time interval, or may be 1 hour before and after the adjustment time interval, which is not specifically limited in this example embodiment. After neighbor monitoring points of the target monitoring points are obtained, sorting can be performed according to the frequency bands and the priorities of the neighbor monitoring points, and specifically, when the frequency band of the neighbor monitoring points is higher, the ranking of the neighbor monitoring points is higher; when the frequency bands of the neighbor cell monitoring points are the same, the neighbor cell monitoring points with high priority (1.8G/2.1G) can be ranked according to the priority, and the ranking of the neighbor cell monitoring points with high priority (800M) is positioned before the neighbor cell monitoring points with low priority. When adjacent cell monitoring points are in the same frequency band and the same priority, the overlapping area occupation ratio of the adjacent cell monitoring points and the target monitoring points can be obtained, and the adjacent cell monitoring points are ordered in a reverse order according to the overlapping area occupation ratio, so that ordered adjacent cell monitoring points are obtained; after the ordered neighbor monitoring points are obtained, neighbor monitoring points positioned in front N in the ordering process can be obtained, and the front N neighbor monitoring points are used as a load balancing suggestion list of the high-load monitoring points. Wherein N is a positive integer.

In the embodiment of the present example, on one hand, by predicting the neighboring cell monitoring points and making screening of the neighboring cell monitoring points, the neighboring cell monitoring points which will generate high load in the future are effectively filtered, so as to avoid invalid or negative adjustment of the high load monitoring points; on the other hand, corresponding load balancing suggestion lists are generated aiming at different high-load monitoring points, so that the network quality is ensured while the load balancing efficiency of the high-load monitoring points is improved; on the other hand, comprehensive optional neighbor monitoring points are provided, so that network optimization personnel can select according to actual production environments, and the actual effectiveness of load balancing is ensured.

Further, the load balancing suggestion list includes: the method comprises the steps of monitoring point information of high load monitoring points, adjustment time intervals of the high load monitoring points, N adjacent area monitoring points before sequencing, high load occurrence time points of the adjacent area monitoring points, basic information of the adjacent area monitoring points, priority of the adjacent area monitoring points, load balancing time periods of the adjacent area monitoring points and coverage relation between the high load monitoring points and the adjacent area monitoring points.

In step S150, optimizing the network load of the high load monitoring point according to the load balancing suggestion list.

In this example embodiment, after the load balancing suggestion list of the high load monitoring point is obtained, the network load of the high load monitoring point may be optimized according to the load balancing suggestion list.

The network load optimization method provided by the example embodiment of the present disclosure has at least the following advantages: on one hand, before model training, data input into the model are screened and filled, so that the complexity of the data and the number of model training are reduced, and the accuracy of model training is improved; on the other hand, by training the deep AR model to obtain different monitoring models, the potential probability distribution function of the time sequence can be learned through the memory neural network structure, and the prediction result is sampled from the probability distribution model in the output stage, so that the stability of the prediction result in a long-time window is improved, and the deviation of the prediction value is reduced; by simultaneously modeling and predicting the time series of a large number of network key data with similar characteristics, the general rule of the overall data can be learned, and the network key data of different monitoring points can be adjusted, so that the training time and the resource cost are reduced; on the other hand, when the load balancing suggestion list is determined, neighbor monitoring points with high load in the future can be filtered, and invalid or negative adjustment of the high load monitoring points is avoided; corresponding load balancing suggestion lists are generated aiming at different high-load monitoring points, so that the network quality is ensured while the load balancing efficiency of the high-load monitoring points is improved; and the comprehensive optional neighbor monitoring points are provided, so that network optimization personnel can select according to actual production environments, and the actual effectiveness of load balancing is ensured.

The network load optimization method according to the exemplary embodiment of the present disclosure is further explained and illustrated below with reference to fig. 10. Wherein, fig. 10 is a logic diagram for implementing a network load optimization method, the network load optimization method may include:

s1010, inputting network key data of the high-load monitoring points into a high-load monitoring model as input data, and verifying the high-load monitoring model to obtain target monitoring points;

s1020, network key data of the target monitoring points are used as input data and are input into a preset neural network model to obtain a target monitoring model;

s1030, network key data of the high-load monitoring points are used as input data and are input into a target monitoring model, and predicted values of the high-load monitoring points are obtained;

s1040, acquiring neighbor monitoring points of the high-load monitoring points, taking network key data of the neighbor monitoring points as input, and inputting the network key data into a neighbor monitoring model to obtain predicted values of the neighbor monitoring points;

s1050, obtaining a load balancing list of the high-load monitoring points according to the predicted value of the high-load monitoring points and the predicted value of the neighbor monitoring points;

and S1060, carrying out load balancing adjustment on the high-load monitoring points according to the load balancing list.

The exemplary embodiment of the present disclosure further provides a network load optimization apparatus, which may include: a high load monitoring model training module 1110, a target monitoring model training module 1120, a neighbor monitoring model training module 1130, a suggestion list generation module 1140, and a network load optimization module 1150. Wherein:

the high load monitoring model training module is used for acquiring high load monitoring points in a target range and first network key data of the high load monitoring points, and acquiring a high load monitoring model through the first network key data;

the target monitoring model training module is used for verifying the high-load monitoring model through second network key data to obtain target monitoring points in the high-load monitoring points, and obtaining a target monitoring model through the network key data of the target monitoring points;

the neighbor cell monitoring model training module is used for acquiring neighbor cell monitoring points of the target monitoring points and third network key data of the neighbor cell monitoring points, and acquiring a neighbor cell monitoring model through the third network key data;

The specific details of each module in the above network load optimization device are described in detail in the corresponding network load optimization method, so that the details are not repeated here.

acquiring network key data of monitoring points in a target range;

It should be noted that although in the above detailed description several modules or units of a device for action execution are mentioned, such a division is not mandatory. Indeed, the features and functionality of two or more modules or units described above may be embodied in one module or unit in accordance with embodiments of the present disclosure. Conversely, the features and functions of one module or unit described above may be further divided into a plurality of modules or units to be embodied.

Furthermore, although the steps of the methods in the present disclosure are depicted in a particular order in the drawings, this does not require or imply that the steps must be performed in that particular order or that all illustrated steps be performed in order to achieve desirable results. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step to perform, and/or one step decomposed into multiple steps to perform, etc.

In an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided.

Those skilled in the art will appreciate that the various aspects of the present disclosure may be implemented as a system, method, or program product. Accordingly, various aspects of the disclosure may be embodied in the following forms, namely: an entirely hardware embodiment, an entirely software embodiment (including firmware, micro-code, etc.) or an embodiment combining hardware and software aspects may be referred to herein as a "circuit," module "or" system.

An electronic device 1200 according to such an embodiment of the present disclosure is described below with reference to fig. 12. The electronic device 1200 shown in fig. 12 is merely an example, and should not be construed as limiting the functionality and scope of use of the disclosed embodiments.

As shown in fig. 12, the electronic device 1200 is in the form of a general purpose computing device. Components of electronic device 1200 may include, but are not limited to: the at least one processing unit 1210, the at least one memory unit 1220, a bus 1230 connecting the different system components (including the memory unit 1220 and the processing unit 1210), and a display unit 1240.

Wherein the storage unit stores program code that is executable by the processing unit 1210 such that the processing unit 1210 performs steps according to various exemplary embodiments of the present disclosure described in the above-described "exemplary methods" section of the present specification. For example, the processing unit 1210 may perform step S110 shown in fig. 1: acquiring high-load monitoring points in a target range and first network key data of the high-load monitoring points, and training a preset neural network model through the first network key data to obtain a high-load monitoring model; s120: verifying the high-load monitoring model through second network key data to obtain target monitoring points in the high-load monitoring points, and training the preset neural network model through the network key data of the target monitoring points to obtain a target monitoring model; s130: acquiring neighbor cell monitoring points of the target monitoring points and third network key data of the neighbor cell monitoring points, and training the preset neural network model through the third network key data to obtain a neighbor cell monitoring model; s140: obtaining a first predicted value of the high-load monitoring point and a second predicted value of a neighboring monitoring point of the high-load monitoring point through the target monitoring model and the neighboring monitoring model, and obtaining a load balancing suggestion list of the high-load monitoring point according to the first predicted value and the second predicted value; s150: and optimizing the network load of the high-load monitoring points according to the load balancing suggestion list.

The storage unit 1220 may include a readable medium in the form of a volatile storage unit, such as a Random Access Memory (RAM) 12201 and/or a cache memory 12202, and may further include a Read Only Memory (ROM) 12203.

Storage unit 1220 may also include a program/utility 12204 having a set (at least one) of program modules 12205, such program modules 12205 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of which may include an implementation of a network environment.

Bus 1230 may be a local bus representing one or more of several types of bus structures including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or using any of a variety of bus architectures.

The electronic device 1200 may also communicate with one or more external devices 1300 (e.g., keyboard, pointing device, bluetooth device, etc.), one or more devices that enable a user to interact with the electronic device 1200, and/or any device (e.g., router, modem, etc.) that enables the electronic device 1200 to communicate with one or more other computing devices. Such communication may occur through an input/output (I/O) interface 1250. Also, the electronic device 1200 may communicate with one or more networks such as a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network such as the internet through the network adapter 1260. As shown, the network adapter 1260 communicates with other modules of the electronic device 1200 over bus 1230. It should be appreciated that although not shown, other hardware and/or software modules may be used in connection with electronic device 1200, including, but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, data backup storage systems, and the like.

From the above description of embodiments, those skilled in the art will readily appreciate that the example embodiments described herein may be implemented in software, or may be implemented in software in combination with the necessary hardware. Thus, the technical solution according to the embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (may be a CD-ROM, a U-disk, a mobile hard disk, etc.) or on a network, including several instructions to cause a computing device (may be a personal computer, a server, a terminal device, or a network device, etc.) to perform the method according to the embodiments of the present disclosure.

In an exemplary embodiment of the present disclosure, a computer-readable storage medium having stored thereon a program product capable of implementing the method described above in the present specification is also provided. In some possible implementations, various aspects of the disclosure may also be implemented in the form of a program product comprising program code for causing a terminal device to carry out the steps according to the various exemplary embodiments of the disclosure as described in the "exemplary methods" section of this specification, when the program product is run on the terminal device.

A program product for implementing the above-described method according to an embodiment of the present disclosure may employ a portable compact disc read-only memory (CD-ROM) and include program code, and may be run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto, and in this document, a readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium would include the following: an electrical connection having one or more wires, a portable disk, a hard disk, random Access Memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The computer readable signal medium may include a data signal propagated in baseband or as part of a carrier wave with readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A readable signal medium may also be any readable medium that is not a readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Program code for carrying out operations of the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device, partly on a remote computing device, or entirely on the remote computing device or server. In the case of remote computing devices, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., connected via the Internet using an Internet service provider).

Furthermore, the above-described figures are only schematic illustrations of processes included in the method according to the exemplary embodiments of the present disclosure, and are not intended to be limiting. It will be readily appreciated that the processes shown in the above figures do not indicate or limit the temporal order of these processes. In addition, it is also readily understood that these processes may be performed synchronously or asynchronously, for example, among a plurality of modules.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any adaptations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

Claims

1. A method for optimizing network load, comprising:

2. The network load optimization method according to claim 1, wherein obtaining high load monitoring points in a target range and first network key data of the high load monitoring points comprises:

acquiring network key data of monitoring points in a target range;

3. The network load optimizing method according to claim 2, wherein distinguishing the monitoring points according to a preset high load monitoring point judgment condition comprises:

4. The network load optimizing method according to claim 2, wherein distinguishing the monitoring points according to a preset high load monitoring point judgment condition comprises:

5. The network load optimization method according to claim 1, characterized in that the network load optimization method further comprises:

Training a preset neural network model by utilizing the downlink physical resource utilization rate of the high-load monitoring points in the target range to obtain a resource utilization rate prediction model;

6. The network load optimization method of claim 5, wherein verifying the high load monitoring model via second network key data to obtain target monitoring points of the high load monitoring points comprises:

7. The network load optimization method according to claim 1, wherein obtaining the neighbor monitoring point of the target monitoring point comprises:

8. The network load optimization method according to claim 7, wherein obtaining the load balancing suggestion list of the high load monitoring point according to the first predicted value and the second predicted value comprises:

9. A network load optimizing apparatus, comprising:

the suggestion list generation module is used for obtaining a first predicted value of the high-load monitoring point and a second predicted value of a neighboring cell monitoring point of the high-load monitoring point through the target monitoring model and the neighboring cell monitoring model, and obtaining a load balancing suggestion list of the high-load monitoring point according to the first predicted value and the second predicted value;

10. A readable storage medium having stored thereon a computer program, which when executed by a processor implements the network load optimization method of any of claims 1-8.

11. An electronic device, comprising:

a processor; and

a memory for storing executable instructions of the processor; wherein the processor is configured to perform the network load optimization method of any one of claims 1-8 via execution of the executable instructions.