A method, apparatus, and electronic device for adjusting energy-saving strategies for network area clusters.

Abstract

This application discloses a method, apparatus, and electronic device for adjusting energy-saving strategies for network area clusters. The method includes: acquiring regional data collected from multiple network units; constructing a three-dimensional network coverage area model based on the regional data, and determining the regional network characteristics corresponding to each network unit based on the three-dimensional network coverage area model and the regional data; aggregating the actual network areas covered by network units with relevant regional network characteristics into corresponding target network area clusters based on the regional network characteristics of each network unit; and determining corresponding target energy-saving strategies for the network units contained in the target network area clusters. Implementing this application embodiment enables precise control over energy-saving strategies adopted by network units with different characteristics by dividing them into network area clusters, thereby effectively improving energy-saving effects and efficiency while ensuring network communication quality.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, specifically to a method, apparatus, and electronic device for adjusting energy-saving strategies for network area clusters. Background Technology

[0002] Currently, with the increasing demands for networks, the scale of network infrastructure deployment is constantly expanding, especially the continuously upgraded wireless communication infrastructure, which brings increasingly prominent energy consumption issues. However, aging and difficult-to-upgrade hardware, as well as slow response and poor adaptability in software, make it difficult for network units such as base stations to implement suitable energy-saving strategies, resulting in poor energy-saving effects for network units and potentially affecting network communication quality. Summary of the Invention

[0003] This application discloses a method, apparatus, and electronic device for adjusting energy-saving strategies for network area clusters. By dividing network area clusters, it can achieve precise control over the energy-saving strategies adopted by network units with different characteristics, thereby effectively improving energy-saving effect and efficiency while ensuring network communication quality.

[0004] The first aspect of this application discloses a method for adjusting a network area cluster energy-saving strategy, including: Acquire regional data collected separately for multiple network units; A three-dimensional network coverage area model is constructed based on the regional data, and the regional network characteristics corresponding to each network unit are determined based on the three-dimensional network coverage area model and the regional data. Based on the regional network characteristics corresponding to each network unit, the actual network regions covered by the network units with relevant regional network characteristics are aggregated into corresponding target network region clusters; For the network units contained in the target network region cluster, a corresponding target energy-saving strategy is determined.

[0005] The second aspect of this application discloses a network area cluster energy-saving strategy adjustment device, comprising: The data acquisition unit is used to acquire regional data collected from multiple network units respectively. The feature determination unit is used to construct a three-dimensional network coverage area model based on the regional data, and to determine the regional network features corresponding to each network unit based on the three-dimensional network coverage area model and the regional data. The region aggregation unit is used to aggregate the actual network regions covered by the network units with relevant regional network characteristics into corresponding target network region clusters based on the regional network characteristics corresponding to each network unit. The strategy adjustment unit is used to determine the corresponding target energy-saving strategy for the network units contained in the target network region cluster.

[0006] The third aspect of this application discloses an electronic device, including a memory and a processor. The memory stores a computer program, and when the computer program is executed by the processor, the processor enables the processor to implement any of the network area cluster energy-saving strategy adjustment methods disclosed in the first aspect of this application.

[0007] The fourth aspect of this application discloses a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any network area cluster energy-saving strategy adjustment method as disclosed in the first aspect of this application.

[0008] Compared with related technologies, the embodiments of this application have the following beneficial effects: In this embodiment, by acquiring regional data collected from multiple network units, a three-dimensional network coverage area model can be constructed based on this regional data. Based on this three-dimensional network coverage area model and the aforementioned regional data, the regional network characteristics corresponding to each network unit are determined. Furthermore, based on the regional network characteristics corresponding to each network unit, the actual network areas covered by network units with relevant regional network characteristics can be aggregated into corresponding target network area clusters. On this basis, corresponding target energy-saving strategies can be determined for the network units contained in the target network area clusters. Therefore, implementing this embodiment allows for the precise fusion of the areas actually covered by network units with different characteristics into corresponding network area clusters using three-dimensional network coverage area modeling. This enables the adoption of appropriate energy-saving strategies for different network area clusters, thereby achieving precise control over the energy-saving strategies of network area clusters. This division of network area clusters closely matches the actual wireless network environment, ensuring network communication quality while flexibly and dynamically determining and adjusting energy-saving strategies. It reduces mismatches in energy-saving strategies caused by base station response lag, thereby improving energy-saving effects and effectively enhancing overall energy efficiency. Attached Figure Description

[0009] Figure 1 This is a flowchart illustrating a method for adjusting a network area cluster energy-saving strategy according to an embodiment of this application; Figure 2 This is a flowchart illustrating another method for adjusting network area cluster energy-saving strategies disclosed in an embodiment of this application; Figure 3 This is a comparative schematic diagram of a network coverage cell segmentation method using two-dimensional and three-dimensional Thiessen polygons, as disclosed in an embodiment of this application. Figure 4This is a schematic diagram of a boundary revision of a network coverage cell based on a ray tracing model disclosed in an embodiment of this application; Figure 5 This is a flowchart illustrating another method for adjusting energy-saving strategies for network area clusters disclosed in an embodiment of this application; Figure 6 This is a schematic diagram illustrating an analytical process for adjusting the energy-saving strategy of a network area cluster disclosed in an embodiment of this application; Figure 7 This is a modular schematic diagram of a network area cluster energy-saving strategy adjustment device disclosed in an embodiment of this application; Figure 8 This is a modular schematic diagram of an electronic device disclosed in an embodiment of this application. Detailed Implementation

[0010] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0011] It should be noted that the terms "comprising" and "having" and any variations thereof in the embodiments of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0012] This application discloses a method, apparatus, and electronic device for adjusting energy-saving strategies for network area clusters. By dividing network area clusters, it can achieve precise control over the energy-saving strategies adopted by network units with different characteristics, thereby effectively improving energy-saving effect and efficiency while ensuring network communication quality.

[0013] The following will be described in detail with reference to the accompanying drawings.

[0014] Please see Figure 1 , Figure 1 This is a flowchart illustrating a method for adjusting a network area cluster energy-saving strategy according to an embodiment of this application. Figure 1 As shown, the method may include the following steps: S102. Obtain regional data collected separately for multiple network units.

[0015] In this embodiment, a network element may include network infrastructure such as a base station, and the wireless communication area covered by each base station may be considered as the network coverage cell (which may be divided into one or more sectors) corresponding to that base station. In some embodiments, the network coverage cell or the corresponding sector corresponding to the base station may also be regarded as a network element.

[0016] Based on this, regional data collected separately for multiple different network units may include regional network data and Digital Elevation Model (DEM) data, etc.

[0017] The aforementioned regional network data may include data collected through the network management system, such as base station basic information, base station configuration parameters, abnormal alarm information, and status monitoring data; data collected through the resource management system (i.e., asset management platform), such as the latitude and longitude of the base station location, its home area, antenna parameters, and equipment model; data collected through the performance management (PM) system, such as throughput, number of users, utilization rate, packet loss rate, and latency; and data collected through the signaling system (such as the X2 and S1 interfaces for communication between the base station and the core network and other base stations), such as handover relationships, handover success rate, RSRP (Reference Signal Receiving Power), and overlapping coverage information.

[0018] The aforementioned DEM data may include raster data (i.e., geographic information data) obtained based on the DEM to describe terrain elevation.

[0019] In some embodiments, the aforementioned regional data can also be divided into spatial characteristic data and service characteristic data based on its characteristics. Spatial characteristic data may include base station ID, cell ID, latitude and longitude and altitude of the base station location, azimuth and downtilt angle of the antenna, RSRP, SINR (Signal to Interference plus Noise Ratio), and other data related to the base station and its overall communication performance in space. Service characteristic data may include data related to specific communication services, such as the number of users, throughput, resource utilization, handover success rate, packet loss rate, and latency.

[0020] In some embodiments, the acquired regional data of different dimensions can be preprocessed to obtain preprocessed regional data corresponding to each network unit. In subsequent steps, a three-dimensional network coverage region model can be constructed based on the preprocessed regional data. Then, based on the three-dimensional network coverage region model and the preprocessed regional data corresponding to each network unit, the regional network features corresponding to each network unit can be determined, and the subsequent regional aggregation (i.e., fusion to obtain the corresponding network region clusters) steps can be performed accordingly.

[0021] For example, the above preprocessing process may include steps such as normalization, missing value handling, and outlier handling. In some embodiments, for regional data of different dimensions (e.g., different acquisition indicators), normalization can be performed on the corresponding data fields of each dimension first, and then missing value judgment processing and outlier detection processing can be performed on the normalized regional data to obtain the preprocessed regional data corresponding to each network unit.

[0022] The aforementioned normalization process can be used to unify the dimensions or orders of magnitude of different data, thereby improving the convergence speed and performance of subsequent machine learning algorithms and enhancing the accuracy of constructing a 3D network coverage area model. For example, for regional data with different collected indicators, linear normalization, zero-mean normalization, and other methods can be applied. For instance, linear normalization can be used for the number of users, packet loss rate, and RSRP, while zero-mean normalization can be used for throughput. This allows for appropriate normalization processing of each collected data field, ensuring it falls within a smaller data range.

[0023] Based on this, the missing value judgment process described above can be used to determine the missing data in each dimension region separately and perform targeted filling or deletion. For example, when the missing data rate of a certain collection indicator in a certain region is less than the first missing threshold (e.g., 3%, 5%), data filling can be performed using methods such as mean filling; when the missing data rate of a certain collection indicator in a certain region is greater than the second missing threshold (e.g., 20%, 30%), the relevant sample data can be directly deleted to reduce the analysis error caused by excessive data incompleteness.

[0024] The aforementioned outlier detection processing can be used to remove identified outliers, eliminating extreme value interference and improving algorithm robustness. This, in turn, helps improve the algorithm performance and accuracy of the network region cluster energy-saving strategy adjustment method. For example, outliers in the aforementioned region data can be identified using algorithms such as box plots (based on the IQR rule, i.e., Interquartile Range) and the 3σ principle. The intersection of the corresponding identification results is then taken, meaning only region data that is identified as outlier in all different outlier detection algorithms is removed.

[0025] S104. Construct a three-dimensional network coverage area model based on the above regional data, and determine the regional network characteristics corresponding to each network unit based on the three-dimensional network coverage area model and the above regional data.

[0026] In this embodiment of the application, in order to determine the corresponding regional network features for multiple different network units, an overall three-dimensional network coverage area model can be constructed based on the aforementioned regional data corresponding to each network unit.

[0027] For example, in constructing a three-dimensional network coverage area model, three-dimensional network coverage cells corresponding to each network unit can be constructed first based on the aforementioned regional data. It should be noted that the three-dimensional network coverage cells corresponding to different network units may or may not overlap. To assign each point in space to a unique corresponding network unit for accurate regional aggregation in subsequent steps, the boundaries of each of the aforementioned three-dimensional network coverage cells can be revised to determine the actual network area covered by each network unit (i.e., the area actually served by that network unit). Within the actual network area, the communication performance of each point relative to its corresponding network unit should be superior to that of other adjacent network units. Therefore, based on communication performance, the boundaries of each three-dimensional network coverage cell can be accurately delineated, resulting in a three-dimensional network coverage area model corresponding to each network unit.

[0028] Based on this, the three-dimensional network coverage area model constructed through the above process, combined with the regional data corresponding to each network unit, can determine the regional network characteristics corresponding to each network unit. For example, these regional network characteristics can be used to characterize the network traffic trends, fluctuations, and cycles of the actual network area covered by the corresponding network unit. In particular, they can include dynamic temporal characteristics (related to business characteristic data in the aforementioned regional data, such as the number of users and throughput at different times) and static spatial characteristics (related to spatial characteristic data in the aforementioned regional data), thereby enabling further clustering of each network unit based on characteristic patterns in subsequent steps.

[0029] S106. Based on the regional network characteristics corresponding to each network unit, the actual network regions covered by the network units with relevant regional network characteristics are aggregated into corresponding target network region clusters.

[0030] In this embodiment of the application, if the regional network features corresponding to certain network units are related, such as having the same or similar feature patterns, then these network units (which may include one or more) with related regional network features can be clustered to aggregate the actual network regions covered by these network units into corresponding target network region clusters.

[0031] In some embodiments, if the regional network characteristics corresponding to some network units indicate that they are symmetrical within a certain period (e.g., within 1 day, 1 week, etc.), such as the periodic increase and decrease of network traffic during commuting peak hours, weekday peak hours, etc., then the actual network areas covered by these network units can be aggregated into corresponding target network area clusters.

[0032] In other embodiments, if the regional network characteristics corresponding to some network units indicate that they are sudden within a certain period of time, such as a sudden surge in network traffic during large-scale offline events such as concerts or sports meets, then the actual network areas covered by these network units can be aggregated into corresponding target network area clusters.

[0033] In some embodiments, if the regional network characteristics corresponding to some network units indicate that they are both periodic and sudden, such as the opening of a commercial center (which also has periodic holiday peaks) or the start of a school year (which also has periodic holiday troughs), then the actual network areas covered by these network units can also be aggregated into the corresponding target network area clusters.

[0034] S108. For the network units contained in the target network region cluster, determine the corresponding target energy-saving strategy.

[0035] In the embodiments of this application, appropriate target energy-saving strategies can be formulated for different target network region clusters and their contained network units to adapt to the different characteristics and rules corresponding to each target network region cluster, so as to achieve precise control over the energy-saving strategies of network region clusters.

[0036] In some embodiments, the aforementioned target energy-saving strategy can be obtained from a pre-built energy-saving strategy library and determined as the target energy-saving strategy corresponding to the target network region cluster. For example, based on the historical performance data of each network unit and the current energy-saving strategy configuration, a corresponding energy-saving strategy library can be constructed through statistical induction, large-scale model learning, etc. This library may include energy-saving feature information (e.g., the network region cluster feature type to which the energy-saving strategy applies, including different trends, fluctuations, periods, etc.; and specific feature values, including time periods, fluctuation ranges, etc.), energy-saving strategies matching the energy-saving feature information, strategy execution parameters (e.g., operation type, operation parameters, priority, etc.), and strategy execution effects (e.g., the network status after the energy-saving strategy is executed, etc.).

[0037] By precisely controlling the energy-saving strategy of network area clusters, energy consumption can be reduced as much as possible on the basis of existing hardware, and the energy-saving strategy within the cluster can be adjusted in a timely manner to cope with different network conditions. This reduces the impact of problems such as slow response and poor adaptability of individual network units, thus helping to balance the overall network communication quality and system energy-saving requirements.

[0038] As can be seen, the network area cluster energy-saving strategy adjustment method described in the above embodiments can utilize three-dimensional network coverage area modeling to accurately merge the areas actually covered by different characteristic network units into corresponding network area clusters. This allows for the adoption of the most suitable energy-saving strategy for different network area clusters, thereby achieving precise control over the energy-saving strategy of network area clusters. By dividing network area clusters in this way, the actual wireless network environment can be closely matched. While ensuring network communication quality, it also allows for flexible and dynamic determination and adjustment of energy-saving strategies, reducing mismatches in energy-saving strategies caused by base station response lag, thus improving energy-saving effects and effectively enhancing overall energy efficiency.

[0039] Please see Figure 2 , Figure 2 This is a flowchart illustrating another method for adjusting network area cluster energy-saving strategies disclosed in an embodiment of this application. Figure 2 As shown, the method may include the following steps: S202. Obtain regional data collected separately for multiple network units.

[0040] Step S202 is similar to step S102 above, and will not be described again here.

[0041] S204. Preprocess the regional data for different dimensions to obtain the preprocessed regional data for each network unit.

[0042] Step S204 is similar in some aspects to step 102 described above. It should be noted that the regional data involved in subsequent steps can all be pre-processed regional data. For example, based on the pre-processed regional data, a three-dimensional network coverage area model can be constructed. Then, based on this three-dimensional network coverage area model and the pre-processed regional data corresponding to each network unit, the regional network characteristics corresponding to each network unit can be determined, and the subsequent steps of fusing network region clusters can be performed accordingly.

[0043] S206. Based on the preprocessed regional data, construct the three-dimensional network coverage cell corresponding to each network unit.

[0044] For example, the aforementioned regional data may include at least regional network data and DEM data. Therefore, when constructing a three-dimensional network coverage area model based on the regional data, the three-dimensional network coverage cell corresponding to each network cell can be constructed first according to the regional network data and DEM data corresponding to each network cell.

[0045] In some embodiments, taking the first network unit among the aforementioned multiple network units as an example, based on the regional network data and DEM data corresponding to the first network unit, a three-dimensional network coverage cell corresponding to the first network unit can be constructed using three-dimensional Thiessen polygons. The aforementioned regional network data may include the latitude and longitude of the base station (i.e., the first network unit), total elevation (which can be determined by combining the base station height with the terrain elevation in the DEM data), antenna height, antenna azimuth and downtilt angle, transmit power, coverage radius, etc. The aforementioned DEM data, in addition to being used to determine the total elevation in combination with the base station height, can also be used to comprehensively determine the actual network area covered by the base station based on terrain features such as mountains and buildings. Based on this, using three-dimensional Thiessen polygons, the communication coverage range of the first network unit in space can be comprehensively and three-dimensionally simulated, constructing a three-dimensional network coverage cell corresponding to the first network unit. That is, the boundary of the three-dimensional Thiessen polygon corresponding to the first network unit can correspond to the boundary of the three-dimensional network coverage cell corresponding to the first network unit, thereby more realistically reflecting the impact of terrain on signal propagation.

[0046] For example, please refer to Figure 3 , Figure 3 This is a comparative schematic diagram of a network coverage cell division method using two-dimensional and three-dimensional Thiessen polygons, as disclosed in an embodiment of this application. Figure 3As shown, compared to the two-dimensional Thiessen polygon on the left, the three-dimensional Thiessen polygon on the right ensures that each point within each polygon in the three-dimensional space (corresponding to different network units or their corresponding network coverage cells, which can be represented by base stations A, B, C, etc.) is closest to its respective network unit. This avoids the problem of significant errors between the theoretical and actual coverage areas of the modeled cells caused by the two-dimensional Thiessen polygon, which ignores terrain elevation and the built environment. This improves the accuracy of modeling the three-dimensional network coverage cells corresponding to network units, thereby helping to improve the precision of energy-saving strategy management.

[0047] S208. Revise the boundaries of each three-dimensional network coverage cell to determine the actual network area covered by each network unit, and obtain the three-dimensional network coverage area model corresponding to each network unit.

[0048] In this embodiment of the application, by revising the boundaries of each of the above-mentioned three-dimensional network coverage cells, the actual network area covered by each network unit (i.e., the area where the network unit actually provides services) can be determined respectively, so as to take into account the influence of terrain elevation, building environment, antenna direction and performance of base stations as much as possible, and further improve the modeling accuracy of each three-dimensional network coverage cell.

[0049] In some embodiments, the boundary revision process described above can be implemented based on a ray tracing model. For example, based on a ray tracing model, multiple sampling points can be set for the three-dimensional network coverage cell corresponding to the first network unit, and these sampling points can be spaced apart according to a preset spatial resolution (e.g., 5 meters, 6 meters, etc.). Furthermore, according to the network setting status of the first network unit, line-of-sight and beam determination can be performed on each sampling point to obtain the determination result corresponding to each sampling point.

[0050] The aforementioned network settings may include the antenna azimuth, downtilt (or elevation), beamwidth, etc. of the first network unit, so as to include sampling points in the area where the first network unit can actually provide services and make the aforementioned determination.

[0051] For example, the direction vector from the first network unit to the target sampling point is set as... If the target sampling point is one of the above multiple sampling points, then the direction vector It can be obtained using the following formula 1.

[0052] Formula 1:

[0053] in, Indicates the location of the target sampling point. This indicates the location of the first network element (or the corresponding network coverage cell / sector). The unit vector is for the main direction of the antenna beam. The calculation can be achieved by converting the antenna's azimuth and elevation angles into unit vectors in three-dimensional space, as shown in Formula 2 below.

[0054] Formula 2:

[0055] Based on this, the above direction vector can be calculated using the dot product method, as shown in Formula 3 below. Angle .

[0056] Formula 3:

[0057] Wherein, direction vector This can be a normalized direction vector. By determining the aforementioned angle... Is it less than half the antenna beamwidth, i.e. Whether the condition is met can determine whether the target sampling point is within the antenna beam range of the first network unit, and this can be used as the determination result of the line of sight and beam determination.

[0058] Furthermore, by performing path loss calculations based on the aforementioned determination results, the actual network coverage status corresponding to each sampling point can be determined according to the corresponding loss calculation results. The aforementioned path loss can include direct path loss, reflected path loss, and diffraction path loss, etc. For example, by superimposing the direct path loss, reflected path loss, and diffraction path loss corresponding to the target sampling point, the reachability and received power of the target sampling point can be calculated. Then, based on the reachability and received power, target sampling points that are blocked or strongly attenuated can be downweighted or eliminated.

[0059] Based on this, please refer to Figure 4 , Figure 4 This is a schematic diagram illustrating a method for revising the boundaries of network coverage cells based on a ray tracing model, as disclosed in an embodiment of this application. Figure 4 As shown, based on the actual network coverage status of each sampling point in the three-dimensional network coverage cell V corresponding to the first network unit, the physical serviceable subdomain S (i.e. the serviceable area corresponding to the first network unit) formed by the remaining sampling points that have not been completely downweighted or eliminated can be determined as the actual network area covered by the first network unit.

[0060] S210. Based on the above three-dimensional network coverage area model and the preprocessed area data, obtain the energy-saving characteristic service data corresponding to each network unit for the actual network area covered by each network unit.

[0061] S212. Periodically synchronize the energy-saving feature business data to obtain the spatiotemporal joint feature vector corresponding to each network unit, which serves as the regional network feature corresponding to each network unit.

[0062] In this embodiment, for the actual network area covered by each network unit (determined based on the boundary-revised three-dimensional network coverage area model), energy-saving characteristic service data corresponding to each network unit can be obtained from the aforementioned regional data based on the network traffic during the actual service provision process of each network unit. For example, the energy-saving characteristic service data may include various data related to system energy consumption from the aforementioned regional data, such as the number of users, peak load, total network traffic, average RSRP, average SINR, call drop rate (or packet loss rate), latency, number of complaints, etc., which are not specifically limited in this embodiment.

[0063] Based on this, by periodically synchronizing the aforementioned energy-saving feature business data using timestamps as a benchmark, the regional network features corresponding to each network unit can be divided both temporally (based on timestamps) and spatially (based on the 3D Thiessen polygon indices in the 3D network coverage area model). That is, through the aforementioned periodic synchronization process, the spatiotemporal joint feature vectors (including dynamic temporal features and static spatial features) corresponding to each network unit can be obtained, and these vectors can then be used as the regional network features corresponding to each network unit.

[0064] S214. Based on the regional network characteristics corresponding to each network unit, determine the corresponding time-series trend characteristics respectively.

[0065] For example, the aforementioned time-series trend characteristics may include at least periodic characteristics and / or bursty characteristics. By performing trend feature analysis and identification on the regional network characteristics corresponding to each network unit, the time-series trend characteristics corresponding to each regional network characteristic can be determined, and can also be used as the time-series trend characteristics corresponding to the corresponding network unit.

[0066] In this embodiment of the application, the time-series trend feature analysis is performed on the regional network features corresponding to each network unit based on the timestamp corresponding to the regional network feature. The corresponding periodic score and burst score can be calculated. Then, in subsequent steps, the periodic score and burst score can be used to determine whether the corresponding regional network feature is a periodic feature and / or a burst feature.

[0067] In some embodiments, the periodic score calculation for regional network features can be achieved by performing symmetry detection on the regional network features based on their time series within a certain periodic window. For example, if the regional network features corresponding to the first network unit can be represented as a time series... Based on the window length N (corresponding to energy-saving characteristic business data within a certain time period, such as 12 hours, 24 hours, etc.), it can be divided into multiple periodic windows. .

[0068] Furthermore, for each periodic window Its symmetry score can be calculated using the following formula 4. .

[0069] Formula 4:

[0070] in, Indicates the periodic window Reversal of time series, The correlation coefficient calculation function can be used, such as the Pearson correlation coefficient, to represent the periodicity of the calculated time series (the closer the correlation coefficient is to 1, the stronger the symmetry; the closer it is to 0, the weaker the symmetry; and a negative number indicates reverse symmetry).

[0071] Furthermore, based on the aforementioned symmetry score... The average of all periodic window symmetry scores can be calculated using the following formula 5, which serves as the corresponding periodic score. .

[0072] Formula 5:

[0073] Where M represents the total number of periodic windows.

[0074] Based on this, if the aforementioned periodic score meets the first judgment condition, it can be determined that the corresponding regional network feature corresponds to a periodic feature. For example, if the periodic score is greater than or equal to the first judgment threshold (e.g., 0.7, 0.8, etc.), it can be confirmed that the corresponding regional network feature has high periodicity (or strong periodicity), and thus it can be determined that it is a periodic feature.

[0075] In other embodiments, the burst score calculation for regional network features can be performed by analyzing and judging whether the regional network features have various types of burst features such as short duration, high intensity, non-periodicity, and structural mutation based on the time series of the regional network features within a certain periodic window. This allows for the calculation of the corresponding burst scores from multiple different dimensions, thereby minimizing misjudgments and providing comprehensive features that are easy to machine learn.

[0076] For example, by employing methods such as Zscore outlier detection, CPChou mutation detection, and IFScore isolated forest outlier detection, a weighted summation as shown in Formula 6 can be performed to comprehensively determine the burst score. .

[0077] Formula 6:

[0078] in, , , These are the adjustable weighting coefficients and the burst score, respectively. Can be normalized to an interval Within, and based on the specific interval to which it belongs, determine whether it is a high-suddenness event (e.g. ), and sudden events (e.g.) ), low-instability or non-instantaneous (e.g.) )wait.

[0079] Based on this, if the aforementioned burst score meets the second judgment condition, then the corresponding regional network feature can be determined to correspond to a burst feature. For example, if the burst score belongs to the score range corresponding to the aforementioned high or medium burst features, then the corresponding regional network feature can be confirmed as a burst feature.

[0080] S216. Based on the above time-series trend characteristics, cluster each network unit to obtain at least one type of network unit with relevant regional network characteristics, and aggregate the actual network regions covered by different types of network units into corresponding target network region clusters.

[0081] In this application embodiment, based on the above-mentioned time-series trend characteristics, network units corresponding to regional network characteristics (referred to as related regional network characteristics) with the same or similar characteristics can be aggregated into the same type of energy-saving region cluster. That is, the actual network areas covered by the same type of network units with related regional network characteristics are aggregated into the corresponding target network region cluster, and managed through a unified energy-saving strategy.

[0082] For example, based on the regional network characteristics and corresponding temporal trend characteristics corresponding to each of the above network units, the density-based clustering algorithm DBSCAN can be used to cluster each network unit based on feature patterns (without limiting the shape of the resulting region clusters), thereby obtaining at least one class of network units with relevant regional network characteristics. Based on this, the actual network regions covered by different classes of network units can be aggregated into corresponding target network region clusters, each of which can include spatially continuous or discontinuous network units of the same type.

[0083] S218. Based on the regional network characteristics corresponding to the network units contained in the target network region cluster, determine the target energy-saving strategy applicable to the characteristic rules corresponding to the regional network characteristics from the pre-built energy-saving strategy library.

[0084] In some embodiments, for different target network region clusters, target energy-saving strategies suitable for the corresponding feature patterns can be searched and matched in the energy-saving strategy library based on the regional network characteristics and corresponding time-series trend characteristics of the network units contained therein, so as to achieve precise control of the energy-saving strategy for the target network region cluster.

[0085] For example, for a target network area cluster with periodic characteristics, a target energy-saving strategy for periodic adjustment can be issued to the network units contained in the target network area cluster, so that the base station can be woken up in time during peak periods and put into sleep mode in time during off-peak periods.

[0086] For example, for a target network area cluster with sudden characteristics, a target energy-saving strategy for emergency scheduling can be issued to the network units contained in the target network area cluster, so that it can schedule appropriate base stations to cope with the surge in network traffic and restore the corresponding base station configuration after the network traffic subsides.

[0087] As can be seen, the network area cluster energy-saving strategy adjustment method described in the above embodiments can utilize three-dimensional network coverage area modeling to accurately merge the areas actually covered by different characteristic network units into corresponding network area clusters. This allows for the adoption of the most suitable energy-saving strategy for different network area clusters, thereby achieving precise control over the energy-saving strategy of network area clusters. This division of network area clusters closely matches the actual wireless network environment, ensuring network communication quality while flexibly and dynamically determining and adjusting energy-saving strategies. It reduces mismatches in energy-saving strategies caused by base station response lag, thus improving energy-saving effects and effectively enhancing overall energy efficiency. Furthermore, by using three-dimensional Thiessen polygons to model the network units in three-dimensional network coverage cells and further employing ray tracing models for boundary revision, the error between the constructed model and the actual network area covered by the network unit can be minimized, thereby improving the modeling accuracy for network coverage space and further enhancing the control precision of the energy-saving strategy.

[0088] Please see Figure 5 , Figure 5 This is a flowchart illustrating another method for adjusting network area cluster energy-saving strategies disclosed in an embodiment of this application. Figure 5 As shown, the method may include the following steps: S502. Acquire regional data collected separately for multiple network units.

[0089] S504. Preprocess the regional data for different dimensions to obtain the preprocessed regional data for each network unit.

[0090] The implementation of steps S502 and S504 is similar to some of the implementations of steps S102 and S104 described above, and will not be repeated here.

[0091] S506. Based on the preprocessed regional data, construct the three-dimensional network coverage cell corresponding to each network unit.

[0092] S508. Revise the boundaries of each three-dimensional network coverage cell to determine the actual network area covered by each network unit, and obtain the three-dimensional network coverage area model corresponding to each network unit.

[0093] Steps S506 and S508 are similar to steps S206 and S208 above, and will not be described again here.

[0094] S510. Based on the above three-dimensional network coverage area model and the preprocessed area data, obtain the energy-saving characteristic service data corresponding to each network unit for the actual network area covered by each network unit.

[0095] S512. Periodically synchronize the energy-saving feature business data to obtain the spatiotemporal joint feature vector corresponding to each network unit, which serves as the regional network feature corresponding to each network unit.

[0096] Steps S510 and S512 are similar to steps S210 and S212 above, and will not be described again here.

[0097] S514. Based on the timestamps corresponding to the regional network features, perform time-series trend feature analysis on the regional network features corresponding to each network unit, and calculate the corresponding periodic scores and burst scores.

[0098] Step S514 is similar in some implementations to step S214 described above. It should be noted that the periodicity score and burst score mentioned above can be used in the subsequent parallel steps S5161 and S5162 to determine the periodicity and burstiness of the regional network characteristics, respectively.

[0099] S5161. If the periodic score meets the first judgment condition, determine that the corresponding regional network feature corresponds to the periodic feature.

[0100] S5162. If the burst score meets the second judgment condition, determine that the corresponding regional network feature corresponds to the burst feature.

[0101] S518. Based on the regional network characteristics and corresponding temporal trend characteristics of each network unit, the density-based clustering algorithm DBSCAN is used to cluster each network unit according to the feature patterns, so as to obtain at least one class of network units with relevant regional network characteristics.

[0102] Step S518 is similar to some of the implementation methods of step S216 described above, and will not be repeated here.

[0103] S520. Based on the regional network characteristics and corresponding temporal trend characteristics of the network units contained in the target network region cluster, determine the target energy-saving strategy applicable to the corresponding characteristic patterns from the pre-constructed energy-saving strategy library.

[0104] Step S520 is similar in some implementations to step S218 described above. It should be noted that the aforementioned corresponding characteristic patterns may include characteristic patterns corresponding to the aforementioned regional network characteristics and corresponding time-series trend characteristics, such as the trend, fluctuation, and periodicity of regional network traffic. Based on this, load prediction can be performed on the aforementioned target network regional cluster to dynamically adjust the target energy-saving strategy applicable to the target network regional cluster according to the prediction results.

[0105] S522. Perform real-time network performance monitoring on the target network area cluster, and combine the real-time monitoring results to perform load prediction on the target network area cluster, and obtain short-term load prediction results.

[0106] S524. Based on the short-term load forecast results, the above-mentioned target energy-saving strategy shall be dynamically adjusted. The dynamic adjustment includes changing the target energy-saving strategy or suspending all energy-saving strategies.

[0107] In this embodiment, real-time monitoring of the network performance of the target network area cluster can be initiated after a target energy-saving strategy is determined for the target network area cluster, or it can be initiated continuously. Based on the real-time monitoring results, the energy-saving strategy (i.e., the aforementioned target energy-saving strategy) for the target network area cluster can be adjusted in a timely and dynamic manner. The indicators for the aforementioned real-time network performance monitoring may include throughput, latency, packet loss rate, handover success rate, network coverage quality, user satisfaction (or number of complaints), etc., and are not specifically limited in this embodiment.

[0108] For example, a machine learning model can be introduced to predict the load of the target network area cluster in the near future (e.g., several minutes or hours) based on the real-time monitoring results described above, thus obtaining the corresponding short-term load prediction results. These short-term load prediction results may include the number of users, network traffic, resource utilization, etc., which are not specifically limited in this embodiment.

[0109] Based on this, and according to the aforementioned short-term load forecast results, the target energy-saving strategies already applied to the target network area cluster (i.e., those distributed to its contained network units) can be dynamically adjusted, including changing the target energy-saving strategies or suspending all energy-saving strategies. Specifically, energy-saving strategy adaptation can be carried out in advance based on the aforementioned short-term load forecast results, including adaptation operations such as early wake-up of base stations and adjustment of network coverage, thereby effectively improving the response efficiency of network units and reducing the execution latency of energy-saving strategies.

[0110] Please refer to further information. Figure 6 , Figure 6 This is a schematic diagram illustrating the analytical process of the network area cluster energy-saving strategy adjustment method disclosed in an embodiment of this application. Figure 6As shown, after collecting regional data and performing necessary data preprocessing for each network unit, a three-dimensional network coverage area model corresponding to each network unit can be constructed based on the preprocessed regional data (including regional network data, DEM data, and three-dimensional coordinates of network units). Specifically, the aforementioned DEM data and three-dimensional coordinates of network units can first be used to construct the corresponding three-dimensional network coverage cells using three-dimensional Thiessen polygons. Then, based on a ray tracing model, boundary revisions can be performed according to the aforementioned regional network data and three-dimensional coordinates of network units to determine the actual network area covered by each network unit, thus obtaining the three-dimensional network coverage area model corresponding to each network unit.

[0111] Based on this, trend characteristic analysis and identification are performed on each network unit. Network areas corresponding to network units with periodic characteristics can be aggregated into high-period clusters, and network areas corresponding to network units with bursty characteristics can be aggregated into high-burst clusters. For different types of target network area clusters (including the aforementioned high-period clusters and high-burst clusters), target energy-saving strategies applicable to the corresponding characteristic patterns can be determined. Furthermore, based on real-time network performance monitoring and load prediction results, it can be determined whether adjustments to the target energy-saving strategy are needed. If no adjustment is needed, the target energy-saving strategy can be maintained; if adjustment is required, the target energy-saving strategy can be changed, or all energy-saving strategies can be suspended to adapt to changes in the actual network environment and communication service requirements.

[0112] As can be seen, the network area cluster energy-saving strategy adjustment method described in the above embodiments can utilize three-dimensional network coverage area modeling to accurately merge the areas actually covered by different characteristic network units into corresponding network area clusters. This allows for the adoption of the most suitable energy-saving strategy for different network area clusters, thereby achieving precise control over the energy-saving strategy of network area clusters. This division of network area clusters closely matches the actual wireless network environment, ensuring network communication quality while flexibly and dynamically determining and adjusting energy-saving strategies. It reduces mismatches in energy-saving strategies caused by base station response lag, thus improving energy-saving effects and effectively enhancing overall energy efficiency. Furthermore, by using three-dimensional Thiessen polygons to model the network units in three-dimensional network coverage cells and further employing ray tracing models for boundary revision, the error between the constructed model and the actual network area covered by the network unit can be minimized, thereby improving the modeling accuracy for network coverage space and further enhancing the control precision of the energy-saving strategy. Furthermore, by monitoring the network performance of network area clusters in real time, it is possible to make short-term predictions of network load within the area, which helps to adapt energy-saving strategies in advance, improve the response efficiency of network units, reduce the execution latency of energy-saving strategies, and further improve overall energy-saving efficiency.

[0113] Please see Figure 7 , Figure 7 This is a modular schematic diagram of a network area cluster energy-saving strategy adjustment device disclosed in an embodiment of this application. Figure 7 As shown, the network area cluster energy-saving strategy adjustment device may include a data acquisition unit 701, a feature determination unit 702, an area aggregation unit 703, and a strategy adjustment unit 704, wherein: The data acquisition unit 701 is used to acquire regional data collected from multiple network units respectively. The feature determination unit 702 is used to construct a three-dimensional network coverage area model based on the above-mentioned regional data, and to determine the regional network features corresponding to each network unit based on the three-dimensional network coverage area model and the above-mentioned regional data. The region aggregation unit 703 is used to aggregate the actual network regions covered by network units with relevant regional network characteristics into corresponding target network region clusters based on the regional network characteristics corresponding to each network unit. The strategy adjustment unit 704 is used to determine the corresponding target energy-saving strategy for the network elements contained in the target network region cluster.

[0114] As can be seen, the network area cluster energy-saving strategy adjustment device described in the above embodiments can accurately merge the areas actually covered by different characteristic network units into corresponding network area clusters by utilizing three-dimensional network coverage area modeling. This allows for the adoption of the most suitable energy-saving strategy for different network area clusters, thereby achieving precise control over the energy-saving strategy of network area clusters. By dividing network area clusters in this way, the actual wireless network environment can be closely matched. While ensuring network communication quality, it also allows for flexible and dynamic determination and adjustment of energy-saving strategies, reducing mismatches in energy-saving strategies caused by base station response lag, thus improving energy-saving effects and effectively enhancing overall energy efficiency.

[0115] In some embodiments, the feature determination unit 702 described above can be specifically used for: Based on regional data, construct three-dimensional network coverage cells corresponding to each network unit; The boundaries of each 3D network coverage cell are revised to determine the actual network area covered by each network cell, thus obtaining the 3D network coverage area model corresponding to each network cell.

[0116] In some embodiments, the aforementioned regional data may include at least regional network data and digital elevation model (DEM) data. Therefore, when the feature determination unit 702 constructs the three-dimensional network coverage cell corresponding to each network cell based on the regional data, it may specifically include: Based on the regional network data and DEM data corresponding to the first network unit, a three-dimensional network coverage cell corresponding to the first network unit is constructed by using three-dimensional Thiessen polygons; wherein, the first network unit is one of multiple network units, and the boundary of the three-dimensional Thiessen polygon corresponding to the first network unit corresponds to the boundary of the three-dimensional network coverage cell corresponding to the first network unit.

[0117] In some embodiments, when the feature determination unit 702 performs boundary revisions on each three-dimensional network coverage cell and determines the actual network area covered by each network cell, it may specifically include: Based on the ray tracing model, multiple sampling points are set for the 3D network coverage cell corresponding to the first network unit; wherein, the first network unit is one of the multiple network units; Based on the network setting status of the first network unit, line of sight and beam determination are performed on each sampling point to obtain the determination result corresponding to each sampling point. Based on the judgment results, the superimposed path loss is calculated, and the actual network coverage status corresponding to each sampling point is determined according to the corresponding loss calculation results. Based on the actual network coverage status corresponding to each sampling point, the serviceable area corresponding to the first network unit is determined as the actual network area covered by the first network unit.

[0118] In some embodiments, the feature determination unit 702 described above may also be used for: Based on the three-dimensional network coverage area model and regional data, energy-saving characteristic service data corresponding to each network unit are obtained for the actual network area covered by each network unit. The energy-saving feature business data is periodically synchronized to obtain the spatiotemporal joint feature vector corresponding to each network unit, which serves as the regional network feature corresponding to each network unit.

[0119] As an optional implementation, the network area cluster energy-saving strategy adjustment device may further include a preprocessing unit (not shown), wherein, after the data acquisition unit 701 acquires the area data collected for multiple network units respectively, the preprocessing unit may be used for: Preprocessing is performed on regional data of different dimensions to obtain preprocessed regional data for each network unit.

[0120] Based on this, the aforementioned feature determination unit 702 can be specifically used for: A three-dimensional network coverage area model is constructed based on the preprocessed regional data; Based on the three-dimensional network coverage area model and the preprocessed regional data corresponding to each network unit, the regional network characteristics corresponding to each network unit are determined.

[0121] In some embodiments, when the preprocessing unit is used to preprocess regional data of different dimensions to obtain preprocessed regional data corresponding to each network unit, it may specifically include: For regional data of different dimensions, normalization is performed based on the corresponding data fields of each dimension. Missing value detection and outlier detection are performed on the normalized regional data to obtain the preprocessed regional data corresponding to each network unit.

[0122] In some embodiments, the region aggregation unit 703 described above can be specifically used for: Based on the regional network characteristics corresponding to each network unit, the corresponding time-series trend characteristics are determined respectively; Based on temporal trend characteristics, each network unit is clustered to obtain at least one class of network units with relevant regional network characteristics, and the actual network regions covered by different classes of network units are aggregated into corresponding target network region clusters.

[0123] In some embodiments, the aforementioned time-series trend features may include at least periodic features and / or burst features. Therefore, when the regional aggregation unit 703 determines the corresponding time-series trend features based on the regional network features corresponding to each network unit, it may specifically include: Based on the timestamps corresponding to the regional network features, time-series trend feature analysis is performed on the regional network features corresponding to each network unit, and the corresponding periodic scores and burst scores are calculated. If the periodic score meets the first judgment condition, the corresponding regional network feature is determined to correspond to the periodic feature; and, If the burst score meets the second judgment condition, the corresponding regional network feature is determined to correspond to the burst feature.

[0124] In some embodiments, when the region aggregation unit 703 is used to cluster various network units based on temporal trend characteristics to obtain at least one class of network units with relevant regional network characteristics, it may specifically include: Based on the regional network characteristics and corresponding temporal trend characteristics of each network unit, the density-based clustering algorithm DBSCAN is used to cluster each network unit according to the feature patterns, so as to obtain at least one class of network units with relevant regional network characteristics.

[0125] In some embodiments, the strategy adjustment unit 704 described above can be specifically used for: Based on the regional network characteristics corresponding to the network units contained in the target network region cluster, target energy-saving strategies applicable to the characteristic patterns corresponding to the regional network characteristics are determined from a pre-built energy-saving strategy library.

[0126] Based on this, the network area cluster energy-saving strategy adjustment device may further include a prediction unit (not shown), wherein the prediction unit may be used for: Load prediction is performed on the target network area clusters, and the target energy-saving strategy is dynamically adjusted based on the prediction results.

[0127] In some embodiments, the prediction unit described above can be specifically used for: Real-time network performance monitoring is performed on the target network area cluster, and load prediction is performed on the target network area cluster based on the real-time monitoring results to obtain short-term load prediction results. Based on short-term load forecasts, the target energy-saving strategy is dynamically adjusted, including changing the target energy-saving strategy or suspending all energy-saving strategies.

[0128] As can be seen, the network area cluster energy-saving strategy adjustment device described in the above embodiments can also perform three-dimensional network coverage cell modeling of network units using three-dimensional Thiessen polygons, and further revise the boundaries using a ray tracing model, thereby minimizing the error between the constructed model and the actual network area covered by the network unit. This helps improve the modeling accuracy for the network coverage space and further enhances the control precision of the energy-saving strategy. Furthermore, by monitoring the network performance of the network area cluster in real time, short-term prediction of the network load within the area can be achieved, which helps to adapt energy-saving strategies in advance, improve the response efficiency of network units, reduce the execution latency of energy-saving strategies, and further improve overall energy-saving efficiency.

[0129] Please see Figure 8 , Figure 8 This is a modular schematic diagram of an electronic device disclosed in an embodiment of this application. For example... Figure 8 As shown, the electronic device may include: Memory 801 storing executable program code; Processor 802 coupled to memory 801; The processor 802 can call the executable program code stored in the memory 801 to execute all or part of the steps in any of the network area cluster energy-saving strategy adjustment methods described in the above embodiments.

[0130] Furthermore, embodiments of this application disclose a computer-readable storage medium storing a computer program for electronic data interchange, wherein the computer program enables a computer to execute all or part of the steps in any of the network area cluster energy-saving strategy adjustment methods described in the above embodiments.

[0131] Furthermore, this application further discloses a computer program product that, when run on a computer, enables the computer to execute all or part of the steps in any of the network area cluster energy-saving strategy adjustment methods described in the above embodiments.

[0132] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compactdisc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0133] The foregoing has provided a detailed description of a network area cluster energy-saving strategy adjustment method, apparatus, and electronic device disclosed in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for adjusting energy-saving strategies for network area clusters, characterized in that, include: Acquire regional data collected separately for multiple network units; A three-dimensional network coverage area model is constructed based on the regional data, and the regional network characteristics corresponding to each network unit are determined based on the three-dimensional network coverage area model and the regional data. Based on the regional network characteristics corresponding to each network unit, the actual network regions covered by the network units with relevant regional network characteristics are aggregated into corresponding target network region clusters; For the network units contained in the target network region cluster, a corresponding target energy-saving strategy is determined.

2. The method according to claim 1, characterized in that, The step of constructing a three-dimensional network coverage area model based on the regional data includes: Based on the regional data, construct three-dimensional network coverage cells corresponding to each network unit; Each of the three-dimensional network coverage cells is revising its boundaries to determine the actual network area covered by each network unit, thus obtaining a three-dimensional network coverage area model corresponding to each network unit.

3. The method according to claim 2, characterized in that, The regional data includes at least regional network data and digital elevation model (DEM) data. The step of constructing three-dimensional network coverage cells corresponding to each network unit based on the regional data includes: Based on the regional network data and DEM data corresponding to the first network unit, a three-dimensional network coverage cell corresponding to the first network unit is constructed by using three-dimensional Thiessen polygons; wherein, the first network unit is one of a plurality of network units, and the boundary of the three-dimensional Thiessen polygon corresponding to the first network unit corresponds to the boundary of the three-dimensional network coverage cell corresponding to the first network unit.

4. The method according to claim 2, characterized in that, The step of revising the boundaries of each of the three-dimensional network coverage cells to determine the actual network area covered by each network element includes: Based on the ray tracing model, multiple sampling points are set for the three-dimensional network coverage cell corresponding to the first network unit; wherein, the first network unit is one of the multiple network units; Based on the network setting status of the first network unit, line of sight and beam determination are performed on each of the sampling points to obtain the determination result corresponding to each of the sampling points; Based on the determination result, the superimposed path loss is calculated, and the actual network coverage status corresponding to each sampling point is determined according to the corresponding loss calculation result. Based on the actual network coverage status corresponding to each sampling point, the serviceable area corresponding to the first network unit is determined as the actual network area covered by the first network unit.

5. The method according to claim 1, characterized in that, The step of determining the regional network characteristics corresponding to each network unit based on the three-dimensional network coverage area model and the regional data includes: Based on the three-dimensional network coverage area model and the area data, energy-saving characteristic service data corresponding to each network unit is obtained for the actual network area covered by each network unit. The energy-saving feature service data is periodically synchronized to obtain the spatiotemporal joint feature vector corresponding to each network unit, which serves as the regional network feature corresponding to each network unit.

6. The method according to any one of claims 1 to 5, characterized in that, After acquiring the regional data collected from multiple network units respectively, the method further includes: Preprocessing is performed on the regional data of different dimensions to obtain the preprocessed regional data corresponding to each network unit; The step of constructing a three-dimensional network coverage area model based on the regional data, and determining the regional network characteristics corresponding to each network unit based on the three-dimensional network coverage area model and the regional data, includes: A three-dimensional network coverage area model is constructed based on the preprocessed regional data; Based on the three-dimensional network coverage area model and the preprocessed regional data corresponding to each network unit, the regional network features corresponding to each network unit are determined respectively.

7. The method according to claim 6, characterized in that, The preprocessing of the region data for different dimensions to obtain preprocessed region data corresponding to each network unit includes: For the regional data of different dimensions, normalization processing is performed based on the corresponding data fields of each dimension; The normalized regional data is processed for missing value judgment and outlier detection to obtain the preprocessed regional data corresponding to each network unit.

8. The method according to any one of claims 1 to 5, characterized in that, The step of aggregating the actual network regions covered by network units with relevant regional network characteristics into corresponding target network region clusters based on the regional network characteristics corresponding to each network unit includes: Based on the regional network characteristics corresponding to each network unit, the corresponding time-series trend characteristics are determined respectively; Based on the time-series trend characteristics, each network unit is clustered to obtain at least one type of network unit with relevant regional network characteristics, and the actual network regions covered by different types of network units are respectively aggregated into corresponding target network region clusters.

9. The method according to claim 8, characterized in that, The time-series trend features include at least periodic features and / or burst features. The step of determining the corresponding time-series trend features based on the regional network features corresponding to each network unit includes: Based on the timestamps corresponding to the regional network features, time-series trend feature analysis is performed on the regional network features corresponding to each network unit, and the corresponding periodic scores and burst scores are calculated. If the periodic score meets the first judgment condition, the corresponding regional network feature is determined to correspond to the periodic feature; and, If the suddenness score meets the second judgment condition, the corresponding regional network feature is determined to correspond to the suddenness feature.

10. The method according to claim 8, characterized in that, The process of clustering the network units based on the temporal trend characteristics to obtain at least one class of network units with relevant regional network characteristics includes: Based on the regional network characteristics and corresponding temporal trend characteristics of each network unit, the density-based clustering algorithm DBSCAN is used to cluster each network unit according to the feature patterns, thereby obtaining at least one class of network units with relevant regional network characteristics.

11. The method according to any one of claims 1 to 5, characterized in that, The step of determining the corresponding target energy-saving strategy for the network elements contained in the target network region cluster includes: Based on the regional network characteristics corresponding to the network units contained in the target network region cluster, a target energy-saving strategy suitable for the characteristic rules corresponding to the regional network characteristics is determined from a pre-built energy-saving strategy library. The method further includes: Load prediction is performed on the target network region cluster, and the target energy-saving strategy is dynamically adjusted based on the prediction results.

12. The method according to claim 11, characterized in that, The step of performing load prediction on the target network region cluster and dynamically adjusting the target energy-saving strategy based on the prediction results includes: The network performance of the target network area cluster is monitored in real time, and the load of the target network area cluster is predicted based on the real-time monitoring results to obtain short-term load prediction results. Based on the short-term load forecast results, the target energy-saving strategy is dynamically adjusted, including changing the target energy-saving strategy or suspending all energy-saving strategies.

13. A network area cluster energy-saving strategy adjustment device, characterized in that, The energy-saving strategy adjustment device includes: The data acquisition unit is used to acquire regional data collected from multiple network units respectively. The feature determination unit is used to construct a three-dimensional network coverage area model based on the regional data, and to determine the regional network features corresponding to each network unit based on the three-dimensional network coverage area model and the regional data. The region aggregation unit is used to aggregate the actual network regions covered by the network units with relevant regional network characteristics into corresponding target network region clusters based on the regional network characteristics corresponding to each network unit. The strategy adjustment unit is used to determine the corresponding target energy-saving strategy for the network units contained in the target network region cluster.

14. An electronic device, characterized in that, The method includes a memory and a processor, wherein the memory stores a computer program that, when executed by the processor, causes the processor to perform the method as described in any one of claims 1 to 12.

15. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 12.

Images