An antenna parameter optimization method, device and related equipment

Abstract

The application provides an antenna parameter optimization method and device and related equipment, and relates to the technical field of data processing. The method comprises the following steps: acquiring network data of a plurality of grids, wherein the plurality of grids are grids included in a first region, and the first region is a region in which at least one antenna provides network services; respectively calculating a first score of each grid based on the network data of each grid; constructing a graph structure based on the network data of the plurality of grids, and calculating a second score of each grid based on the graph structure, wherein the second score is used to represent a network resource bearing condition of the corresponding grid; respectively calculating an imbalance index of each grid based on the first score and the second score of each grid; determining an associated cell of a target grid, wherein the target grid is a grid with an imbalance index greater than a set index threshold; and generating an adjustment strategy of the associated cell based on the network data of the target grid, wherein the adjustment strategy is used to perform parameter adjustment on an antenna of the associated cell. The application can improve the optimization effect of the antenna parameters.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to an antenna parameter optimization method, apparatus, and related equipment. Background Technology

[0002] Wireless networks are a crucial technology in communication networks. Signal quality is determined by network performance indicators, user distribution, and base station operating parameters to ascertain whether antenna adjustments are necessary. Antenna adjustments in related technologies include power adjustment, azimuth adjustment, and parameter adjustment. However, current adjustment methods address antennas only when a single parameter degradation occurs. They cannot accurately identify and optimize network parameter degradation that may occur in different areas, requiring adjustments only when significant anomalies are observed, resulting in poor antenna parameter optimization.

[0003] It is evident that the optimization of antenna parameters in the relevant technologies is ineffective. Summary of the Invention

[0004] This invention provides an antenna parameter optimization method, apparatus, and related equipment to address the problem of poor antenna parameter optimization performance in related technologies.

[0005] To solve the above problems, the present invention is implemented as follows: In a first aspect, embodiments of the present invention provide an antenna parameter optimization method, comprising: Acquire network data from multiple grids, wherein the multiple grids are grids included in a first region, and the first region is an area where at least one antenna provides network services; A first score is calculated for each grid based on the network data of each grid, and the first score is used to characterize the degree of network resource demand of the corresponding grid. A graph structure is constructed based on the network data of the multiple grids, and a second score is calculated for each grid based on the graph structure. The second score is used to characterize the network resource carrying capacity of the corresponding grid, and the graph structure is used to characterize the correlation between the features of the network data of each grid. Based on the first score and the second score of each grid, the imbalance index of each grid is calculated respectively. The imbalance index is used to characterize the balance between the network resource carrying capacity and demand of the corresponding grid. Identify the associated cells of the target grid, wherein the target grid is the grid whose imbalance index is greater than a set index threshold; An adjustment strategy for the associated cell is generated based on the network data of the target grid. The adjustment strategy is used to adjust the antenna parameters of the associated cell.

[0006] Secondly, embodiments of the present invention also provide an antenna parameter optimization device, comprising: The acquisition module is used to acquire network data from multiple grids, wherein the multiple grids are grids included in a first region, and the first region is an area where at least one antenna provides network services; The first calculation module is used to calculate a first score for each grid based on the network data of each grid, and the first score is used to characterize the degree of network resource demand of the corresponding grid. The second calculation module is used to construct a graph structure based on the network data of the multiple grids, and to calculate a second score for each grid based on the graph structure. The second score is used to characterize the network resource carrying capacity of the corresponding grid, and the graph structure is used to characterize the correlation between the features of the network data of each grid. The third calculation module is used to calculate the imbalance index of each grid based on the first score and the second score of each grid. The imbalance index is used to characterize the balance between the network resource carrying capacity and demand of the corresponding grid. The determination module is used to determine the associated cells of the target grid, wherein the target grid is a grid whose imbalance index is greater than a set index threshold; A generation module is used to generate an adjustment strategy for the associated cell based on the network data of the target grid. The adjustment strategy is used to adjust the antenna parameters of the associated cell.

[0007] Thirdly, embodiments of the present invention also provide an electronic device, including a transceiver and a processor. The transceiver is used to acquire network data from multiple grids, the multiple grids being grids included in a first region, the first region being an area where at least one antenna provides network services; The processor is configured to calculate a first score for each grid based on the network data of each grid, wherein the first score is used to characterize the degree of network resource demand of the corresponding grid. The processor is further configured to construct a graph structure based on the network data of the plurality of grids, and calculate a second score for each grid based on the graph structure. The second score is used to characterize the network resource carrying capacity of the corresponding grid, and the graph structure is used to characterize the correlation between the features of the network data of each grid. The processor is further configured to calculate an imbalance index for each grid based on the first score and the second score of each grid, wherein the imbalance index is used to characterize the balance between the network resource carrying capacity and demand of the corresponding grid. The processor is further configured to determine the associated cells of the target grid, wherein the target grid is a grid whose imbalance index is greater than a set index threshold; The processor is further configured to generate an adjustment strategy for the associated cell based on the network data of the target grid, the adjustment strategy being used to adjust the antenna parameters of the associated cell.

[0008] Fourthly, embodiments of the present invention provide an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, it implements the steps of the antenna parameter optimization method described in the first aspect.

[0009] Fifthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the antenna parameter optimization method described in the first aspect.

[0010] In a sixth aspect, the present invention also provides a computer program product, including computer instructions, which, when executed by a processor, implement the steps of the antenna parameter optimization method described in the first aspect.

[0011] In this embodiment of the invention, network data from multiple grids are acquired, wherein the multiple grids are grids included in a first region, and the first region is an area where at least one antenna provides network services; a first score is calculated for each grid based on the network data of each grid, the first score being used to characterize the degree of network resource demand of the corresponding grid; a graph structure is constructed based on the network data of the multiple grids, and a second score is calculated for each grid based on the graph structure, the second score being used to characterize the network resource carrying capacity of the corresponding grid, and the graph structure being used to characterize the correlation between the features of the network data of each grid; an imbalance index is calculated for each grid based on the first score and the second score of each grid, the imbalance index being used to characterize the balance between the carrying capacity and demand of network resources of the corresponding grid; associated cells of a target grid are determined, the target grid being a grid whose imbalance index is greater than a set index threshold; an adjustment strategy for the associated cells is generated based on the network data of the target grid, the adjustment strategy being used to adjust the antenna parameters of the associated cells. In this way, by calculating the first and second scores of each grid through network data, and calculating the imbalance index based on the first and second scores, network demand and network capacity can be combined to identify whether antenna adjustments are needed, thus improving the identification accuracy. Furthermore, target grids with imbalance indices greater than a set index threshold are identified, and associated cells of the target grids are identified so that the antenna parameters of the associated cells can be adjusted, thereby effectively improving the optimization effect of antenna parameters. Attached Figure Description

[0012] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a flowchart of an antenna parameter optimization method provided in an embodiment of the present invention; Figure 2 This is an overall schematic diagram of an antenna parameter optimization method provided in an embodiment of the present invention; Figure 3 This is an example heat map provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the GNN model calculating the second score provided in an embodiment of the present invention; Figure 5 This is a flowchart of the output adjustment strategy provided in an embodiment of the present invention; Figure 6 This is a structural diagram of an antenna parameter optimization device provided in an embodiment of the present invention; Figure 7 This is a structural diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Please see Figure 1 and Figure 2 , Figure 1 This is a flowchart of an antenna parameter optimization method provided in an embodiment of the present invention. Figure 2 This is an overall schematic diagram of an antenna parameter optimization method provided by an embodiment of the present invention. The antenna parameter optimization method includes the following steps: Step 101: Obtain network data from multiple grids, wherein the multiple grids are grids included in a first region, and the first region is an area where at least one antenna provides network services.

[0016] The aforementioned multiple grids constitute the grids included in the first region. The first region is an area where at least one antenna provides network service. By acquiring the grid data of the multiple grids included in the first region, it is determined whether the antenna parameters providing network service to the first region need to be adjusted. In some embodiments, the first region may be a business district, transportation hub, campus, hospital, residential area, office area, event venue, or a key protection area pre-designated by the operator. Since a stable network service is required for the first region, it is necessary to monitor the multiple grids included in the first region to determine whether antenna parameters need to be adjusted.

[0017] The size and dimensions of the multiple grids can be the same or different, and the multiple grids can be obtained by dividing the first region according to a preset size rule.

[0018] The aforementioned network data is used to characterize the wireless network status of each grid. By using the network data, the network resource requirements and carrying capacity of each grid can be determined, thereby determining whether it is necessary to adjust the antennas providing network services to the grid.

[0019] In some implementations, network data may include parameters such as grid traffic intensity, number of active users, Reference Signal Received Power (RSRP), Reference Signal Received Quality (RSRQ), Signal to Interference plus Noise Ratio (SINR), Physical Resource Block (PRB) utilization, uplink and downlink rates, access success rate, call drop rate, handover success rate, number of complaints, Measurement Report (MR) data, base station operating parameters, and cell coverage relationships.

[0020] Among them, grid traffic intensity reflects the density of service usage within the target area, and the number of active users reflects the scale of users engaging in service activities within the corresponding area; RSRP is used to characterize coverage strength, while RSRQ and SINR are used to characterize signal quality and interference levels; PRB utilization reflects the occupancy of radio resources, and uplink and downlink rates reflect user service experience; access success rate, drop rate, and handover success rate are used to describe the stability during connection establishment, connection maintenance, and mobility processes; complaint volume serves as a supplementary observation for user-perceived issues; MR data provides terminal-side measurement results; base station engineering parameters include information such as base station location, antenna height, azimuth angle, downtilt angle, transmit power, frequency, bandwidth, cell identifier, and neighbor cell relationships; cell coverage relationships are used to describe the coverage or influence relationships between the grid and serving cells, neighboring cells, and boundary cells.

[0021] Step 102: Calculate the first score for each grid based on the network data of each grid. The first score is used to characterize the degree of network resource demand of the corresponding grid.

[0022] The first score mentioned above is used to characterize the degree of network resource demand of the corresponding grid. The first score is calculated for each grid so that the grid's demand for network resources can be evaluated through the first score.

[0023] In some implementations, the first score for each grid is calculated based on the network data of each grid. This can be done by comparing the network data of each grid with historical data, calculating the change value between the network data and the historical data, and then calculating the first score based on the change value. Alternatively, the change value can be normalized to obtain the first score.

[0024] Step 103: Construct a graph structure based on the network data of the multiple grids, and calculate a second score for each grid based on the graph structure. The second score is used to characterize the network resource carrying capacity of the corresponding grid, and the graph structure is used to characterize the correlation between the features of the network data of each grid.

[0025] The graph structure described above is used to characterize the relationships between the features of network data in each grid. By constructing the graph structure, the relationships between network data in different grids can be determined. Furthermore, a second score can be calculated using the graph structure, which in turn allows the network resource carrying capacity of the corresponding grid to be determined.

[0026] In some implementations, a scoring model can be trained. After the graph structure is constructed, the graph structure is input into the scoring model, and the scoring model outputs a second score for each grid.

[0027] Step 104: Calculate the imbalance index of each grid based on the first score and the second score of each grid. The imbalance index is used to characterize the balance between the network resource carrying capacity and demand of the corresponding grid.

[0028] It should be understood that the first score is used to characterize the degree of demand for network resources of the corresponding grid, i.e. carrying capacity, and the second score is used to characterize the network resource carrying situation of the corresponding grid, i.e. actual carrying situation. The imbalance index of each grid can be calculated through the first and second scores. The imbalance index can determine the balance between the carrying capacity and demand of the network resources of the corresponding grid, i.e. whether the network resources carried by the grid match the carrying capacity of the grid.

[0029] In some implementations, the difference between the first score and the second score can be used as an imbalance index.

[0030] In other implementations, the imbalance index can be calculated using the following formula: B i (t) = D i (t) × [1-C i (t)]; In the formula, B i (t) represents the imbalance index of the i-th grid, D i (t) represents the first score of the i-th grid, C i (t) represents the second score of the i-th grid. Thus, when the first score increases and the second score decreases, the imbalance index increases accordingly, indicating that the network resources carried by the grid are more mismatched with the grid's carrying capacity, and the antenna parameters need to be adjusted more. Conversely, when the first score decreases or the second score increases, the imbalance index decreases accordingly, indicating that the network resources carried by the grid are more matched with the grid's carrying capacity, and the antenna parameters do not need to be adjusted more.

[0031] Step 105: Determine the associated cells of the target grid, wherein the target grid is the grid whose imbalance index is greater than a set index threshold.

[0032] It should be understood that after calculating the imbalance index, key protection grids (i.e., target grids) can be identified based on the imbalance index. In some implementations, if a grid shows a high-heat area in the flow heatmap (i.e., the first score is higher than the preset level) and the second score is lower than the preset level, the grid can be identified as a high-priority key protection area. In this case, the parameters of the antennas providing network services to the grid need to be adjusted.

[0033] In some implementations, if a grid has high traffic demand (i.e., a high first score) and strong carrying capacity (i.e., a high second score), it can be treated as a key monitoring area. Specifically, the detection frequency of network data for that grid can be increased. For example, the detection frequency of an ordinary grid is to calculate the imbalance index once every 10 minutes, while the detection frequency of a grid in a key monitoring area can be adjusted to calculate the imbalance index once every 5 minutes. This allows for timely adjustments when the grid's imbalance index exceeds a set threshold.

[0034] Furthermore, if a grid has low traffic demand but a low capacity score, its inclusion in a low-priority problem area can be determined by combining the number of complaints and the guarantee level. For example, if the number of complaints exceeds a set number, or the guarantee level is higher than a set level, it will not be included in a low-priority problem area; if the number of complaints is less than a set number, and / or the guarantee level is lower than a set level, it will be included in a low-priority problem area. For grids included in a low-priority problem area, the detection frequency can be reduced.

[0035] Furthermore, problem types can be categorized based on specific indicator status. If the first score is high and the RSRP is low, the grid can be marked as a high-demand, weak-coverage area. If the first score is high and the SINR or RSRQ is low, it can be marked as a high-demand, high-interference area. If the first score is high, PRB utilization is high, and uplink / downlink rates are low, it can be marked as a high-load, low-rate area. If the access success rate, drop rate, or handover success rate is abnormal, it can be marked as an area with connection stability or mobility experience issues. If complaints are concentrated and the capacity score is low, it can be treated as a key area for user perception issues.

[0036] For example, a certain business district grid consistently shows a high activity level (highest first score) in the evening peak traffic heatmap, with a significant increase in active users, while the second score is low. Further observation of the grid's metrics reveals high PRB utilization in the primary serving cell, low downlink speed, and strong overlap in neighboring cells leading to a decrease in SINR. Based on this, the grid can be identified as a high-priority grid requiring priority protection, and the problem type can be classified as a high-load, superimposed interference issue.

[0037] Based on this, this embodiment sets a set index threshold to intuitively determine whether the antenna parameters of the associated cells of the grid need to be adjusted. The set index threshold is a pre-configured threshold. If the imbalance index is greater than the set index threshold, it is considered that the network capacity of the target grid does not match the network demand, and the network stability is poor. Therefore, it is necessary to identify the associated cells of the target grid and adjust their antenna parameters. Conversely, if the imbalance index is less than or equal to the set index threshold, it is considered that the network capacity of the grid matches the network demand, and the network stability is good. Therefore, no adjustment is needed.

[0038] The aforementioned associated cells are cells that provide network services to the target grid or that may interfere with the network of the target grid. By adjusting the antenna parameters of the associated cells, the network requirements and / or network carrying capacity of the target grid can be adjusted, thereby achieving network optimization of the target grid.

[0039] Step 106: Generate an adjustment strategy for the associated cell based on the network data of the target grid. The adjustment strategy is used to adjust the antenna parameters of the associated cell.

[0040] The above adjustment strategy is used to adjust the parameters of the associated antennas, thereby optimizing the network data of the target grid so that the network requirements of the target grid match the network carrying capacity.

[0041] In some implementations, generating adjustment strategies for the associated cells based on the network data of the target grid can involve first classifying the target grid based on the network data to obtain its category, and then determining the adjustment strategy according to the specific category. For example, the classification based on network data may include categories such as key protection areas, key monitoring areas, low-priority problem areas, high-demand weak coverage areas, high-demand high-interference areas, and high-load low-rate areas. Adjustment strategies are then pre-configured for different categories, allowing the target grid's category to be determined based on its network data, thereby obtaining the adjustment strategy.

[0042] In this embodiment of the invention, network data from multiple grids are acquired, wherein the multiple grids are grids included in a first region, and the first region is an area where at least one antenna provides network services; a first score is calculated for each grid based on the network data of each grid, the first score being used to characterize the degree of network resource demand of the corresponding grid; a graph structure is constructed based on the network data of the multiple grids, and a second score is calculated for each grid based on the graph structure, the second score being used to characterize the network resource carrying capacity of the corresponding grid, and the graph structure being used to characterize the correlation between the features of the network data of each grid; an imbalance index is calculated for each grid based on the first score and the second score of each grid, the imbalance index being used to characterize the balance between the carrying capacity and demand of network resources of the corresponding grid; associated cells of a target grid are determined, the target grid being a grid whose imbalance index is greater than a set index threshold; an adjustment strategy for the associated cells is generated based on the network data of the target grid, the adjustment strategy being used to adjust the antenna parameters of the associated cells. In this way, by calculating the first and second scores of each grid through network data, and calculating the imbalance index based on the first and second scores, network demand and network capacity can be combined to identify whether antenna adjustments are needed, thus improving the identification accuracy. Furthermore, target grids with imbalance indices greater than a set index threshold are identified, and associated cells of the target grids are identified so that the antenna parameters of the associated cells can be adjusted, thereby effectively improving the optimization effect of antenna parameters.

[0043] In one embodiment, the network data includes first data and / or second data, wherein the first data is data including latitude and longitude information, and the second data is data collected through multiple cells, wherein the multiple cells are cells that provide network data to the first area; The acquisition of network data from multiple grids includes: Acquire first initial data and / or second initial data for the first region, wherein the first initial data includes data with latitude and longitude information, and the second initial data is data collected through the plurality of cells; The first region is divided into the multiple grids to obtain the position information of each grid. Based on the latitude and longitude information of the first initial data and the location information of the multiple grids, the first initial data is divided into the multiple grids to obtain the first data for each grid; and / or, Based on the location information of each grid, the weight coefficient of each grid relative to each of the plurality of cells is calculated, and the second data of each grid is calculated based on the weight coefficient of each cell and the second initial data.

[0044] It should be noted that network data is inconsistent in terms of source, temporal granularity, and spatial granularity. Therefore, the target area needs to be divided into several spatial grids, and data from different sources need to be mapped to a unified grid space. Thus, in this embodiment, the first initial data and / or the second initial data of the first cell are first obtained, and then the first initial data and / or the second initial data are mapped to each grid to obtain the first data and the second data of each grid.

[0045] For data containing latitude and longitude coordinates or location information (i.e., the first initial data), it can be directly matched to the corresponding grid according to spatial coordinates. For data with cells as the statistical unit (i.e., the second initial data), weighting coefficients can be determined based on parameters such as cell coverage relationship, MR sampling point distribution, serving cell ratio, user dwell time relationship, neighbor cell relationship, or signal strength contribution. Then, the data is weighted and mapped to the relevant grid according to the weighting coefficients. After this processing, base station, cell, terminal measurement, complaint events, and service statistics are organized into the same spatial scale, enabling subsequent models to analyze network requirements and network carrying capacity within a unified spatial framework.

[0046] Specifically, for the first initial data, based on the latitude and longitude information and the location information of multiple grids, the first initial data is divided into multiple grids to obtain the first data for each grid; while for the second initial data, based on the location information of each grid, the weight coefficient of each grid relative to each cell in the multiple cells is calculated, and the second data of each grid is calculated based on the weight coefficient of each cell and the second initial data.

[0047] The aforementioned location information includes the location of multiple regions and grids, such as cell coverage probability, MR sampling point affiliation ratio, user dwell ratio, RSRP strength, SINR contribution, neighbor cell relationship, and / or handover relationship. This location information is used to determine the weight coefficient of each grid relative to each cell, facilitating the mapping of the second initial data to different grids to obtain the second data for each grid. For example, when a grid is mainly covered by a single cell, and that cell has a high MR sampling point ratio within that grid, that cell receives a higher weight. Similarly, when a neighboring cell generates a strong signal within a grid and may cause interference, that neighboring cell also receives a corresponding weight. Through weighted mapping, the second initial data, such as cell-level indicators like PRB utilization, access success rate, drop rate, handover success rate, and uplink / downlink speeds, can be transformed into grid-level indicators, forming a unified grid-level data representation with grid traffic intensity, active user count, complaint volume, and MR sampling statistics.

[0048] In some implementations, after obtaining the weighting coefficients, the second data for each grid is calculated based on the weighting coefficients of each cell and the second initial data, which can be expressed by the following formula: ; In the formula, X i (t) represents the second data of the i-th grid at time t, X k (t) represents the second initial data of the k-th cell at time t, S(i) represents the set of cells that have coverage, service, or interference relationships with the i-th grid, w ik (t) represents the weight coefficient of the i-th grid relative to the k-th cell.

[0049] To ensure the comparability of contributions from different cells to the same grid, the system can normalize the influence weights of cells within the same grid to satisfy the following: .

[0050] In one embodiment, the plurality of grids includes a first grid, the network data includes at least one of traffic intensity and number of active users, the network data further includes traffic data, and the calculation of a first score for each grid based on the network data of each grid includes: Obtain historical traffic data and guarantee parameters for the first grid, wherein the guarantee parameters are used to characterize whether the current time is within a preset time period of the first grid; Based on the traffic data and historical traffic data of the first grid, a trend parameter of the first grid is calculated, and the trend parameter is used to characterize the traffic change of the first grid. The first score of the first grid is obtained by weighting at least one of the traffic intensity and the number of active users of the first grid, as well as the guarantee parameter and the trend parameter.

[0051] The first grid mentioned above is one of multiple grids. In this embodiment, the calculation of the first score of the first grid is used as an example. The calculation method of the first score of other grids in the multiple grids is the same as that of the first grid, and will not be repeated here.

[0052] The historical traffic data for the first grid refers to the traffic data acquired historically for that grid, representing the network status of the first grid over a historical period. In this embodiment, the change in traffic data can be calculated using the historical traffic data and the current traffic data for the first grid. This change can then be used to determine a trend parameter, allowing the trend parameter of the first grid to characterize its traffic changes.

[0053] The guarantee parameter of the first grid is used to characterize whether it is within the preset time period of the first grid. The preset time period of the first grid is a critical time period that needs to ensure the stability of network data, such as the weekday peak time period of transportation hubs, the school class time period in school areas, etc. The same or different preset time periods can be configured for each grid in multiple grids. When calculating the first score, the guarantee parameter of the grid is used to determine whether it is currently within the preset time period, so as to improve the accuracy of the calculated first score.

[0054] In some implementations, heatmaps can be generated based on grid traffic intensity, number of active users, historical traffic data, and trend parameters, such as... Figure 3 As shown, this heatmap is used to describe the spatial distribution of regional business needs and can be used for front-end display and other purposes. For example, heatmaps with different time granularities can be generated according to hours, morning and evening peak hours, weekdays, weekends, holidays, or key guarantee periods, and heatmaps with different levels can also be generated according to spatial granularities such as 50 meters, 100 meters, 250 meters, or 500 meters.

[0055] In some implementations, the first score can be calculated using the following formula: ; In the formula, D i (t) represents the first score of the i-th grid at time t; F i (t) represents the normalized grid flow intensity; U i (t) represents the normalized number of active users; G i (t) represents the trend parameter relative to the same historical period; E i (t) represents the guarantee parameter; These are the weighting coefficients.

[0056] If the first score is high, it indicates that the grid has a strong network demand during the corresponding time period. However, this does not directly mean that the grid has network problems. It is also necessary to combine the second score of network capacity to determine whether the area has sufficient coverage, capacity and experience support.

[0057] In one embodiment, constructing a graph structure based on the network data of the plurality of grids includes: Obtain the grid relationship for each grid, which is used to characterize the positional relationship between each grid and other grids in the plurality of grids; The graph structure is constructed based on the network data and grid relationships of the multiple grids. The graph structure includes multiple nodes and multiple edges. The multiple nodes correspond one-to-one with the multiple grids, and the multiple edges correspond one-to-one with the grid relationships of the multiple grids. The features of each node are used to characterize the network data of the grid corresponding to the node. The calculation of the second score for each grid based on the graph structure includes: The graph structure is input into a graph neural network (GNN) model, and a second score for each grid is output based on the GNN model. The GNN model is trained to output the second score for different nodes.

[0058] It should be noted that the optimization targets in a wireless network are not independent point units. Changes in the antenna downtilt angle, transmit power, azimuth angle, or load state of a cell may simultaneously affect the coverage quality and interference level of its serving grid, neighboring grids, and neighboring cells. Ordinary tabular models or single-point thresholds are insufficient to express this mutual influence relationship. Therefore, this embodiment employs a graph neural network to characterize the topology between this type of cell-grid-neighboring cell.

[0059] Specifically, the grid relationships of each grid are obtained, and a graph structure is constructed based on the network data and grid relationships of multiple grids. The graph structure includes multiple nodes and multiple edges. Multiple nodes correspond one-to-one with multiple grids, and multiple edges correspond one-to-one with the grid relationships of multiple grids, so that nodes can represent the network data of the corresponding grid, and edges can represent the corresponding grid relationships.

[0060] For example, the first region is represented as a graph structure G=(V,E). Here, V represents the set of nodes, which includes multiple grid nodes; E represents the set of edges, which represent the grid relationships between nodes, including coverage relationships, neighbor relationships, spatial adjacency relationships, switching relationships, overlapping coverage relationships, or interference influence relationships.

[0061] Among them, nodes correspond to grids, and the characteristics of nodes may include grid traffic intensity, number of active users, traffic heat level, RSRP, RSRQ, SINR, uplink and downlink rates, number of complaints and MR sampling statistics, as well as PRB utilization, number of active users, cell uplink and downlink traffic, access success rate, drop rate, handover success rate, current antenna downtilt angle, transmit power, azimuth angle, frequency point, bandwidth and number of neighboring cells.

[0062] The edges in the graph structure are used to describe the actual wireless influence relationships between nodes. For example, edges between grids represent coverage and service relationships; or neighboring cells, handover, or interference relationships; or spatial adjacency relationships or local continuity of service popularity. Edge weights can be determined based on cell coverage probability, MR sampling point affiliation ratio, neighboring cell signal strength, handover frequency, overlapping coverage degree, or interference intensity.

[0063] After constructing the graph structure, according to... Figure 4 The GNN model shown outputs a second score. The GNN model comprises four parts: graph construction, node feature encoding, graph information aggregation, and score output. The graph construction part generates a grid-based graph; the node feature encoding part maps the original indices of different types of nodes into vector representations of a unified dimension; the graph information aggregation part enables the target node to absorb information from its neighboring nodes; and the score output part provides a second score based on the final representation of the target node.

[0064] In one implementation, the graph information aggregation process of the l-th layer in the GNN model can be represented as: ; In the formula, This represents the feature representation of node i at layer l. N(i) represents the feature representation of node i at layer l+1; N(i) represents the set of neighboring nodes of node i; a ij W represents the weight of the influence of node j on node i. o and W l σ represents the learnable parameters of the model; σ() represents the nonlinear activation function.

[0065] After several layers of graph information aggregation, the scoring output layer outputs a second score based on the final representation of the target node: ; In the formula, C i(t) represents the second score of the i-th grid at time t; hᵢᴸ represents the node representation after aggregation of L layers of graph information; W and b are learnable parameters of the model. The second score can be normalized to between 0 and 1. A higher score indicates better coverage, capacity, and user experience in the area; a lower score indicates potential problems such as weak coverage, high interference, insufficient capacity, reduced speed, abnormal access, disconnection, or abnormal handover.

[0066] In this way, the carrying capacity evaluation of a grid no longer relies solely on its own RSRP, SINR, or rate metrics through the GNN model. The model simultaneously considers the load status of its primary serving cell, the interference impact of neighboring cells, the coverage relationship of boundary cells, and the service activity of surrounding grids. This modeling approach more closely approximates the actual causal relationships in wireless network optimization, effectively improving the accuracy of the second score.

[0067] In one embodiment, generating the adjustment strategy for the associated cells based on the network data of the target grid includes: The network data of the target grid, the first score, and the second score are input into a preset reinforcement learning model, and the adjustment strategy of the associated cell is output based on the preset reinforcement learning model. The preset reinforcement learning model is trained in the following manner: Obtain sample network data of the sample grid and the sample score corresponding to the sample network data. The sample score is used to characterize the network resource carrying capacity of the sample grid. Multiple intermediate learning models are generated based on an initial reinforcement learning model, and the parameters of the multiple intermediate learning models are different. The initial reinforcement learning model is constructed as an output adjustment strategy. The sample network data is processed based on each intermediate learning model to obtain the intermediate strategy of each intermediate learning model; The intermediate strategies of each intermediate learning model are adjusted to obtain the adjusted network data corresponding to each intermediate learning model. Calculate the intermediate score for each intermediate learning model based on the adjusted network data corresponding to each intermediate learning model. Based on the adjusted network data and intermediate scores corresponding to each intermediate learning model, as well as the sample network data and sample scores, the reward value of each intermediate learning model is calculated respectively. The model with the largest reward value among the multiple intermediate learning models is set as the preset reinforcement learning model.

[0068] It should be understood that when adjustments to antennas in associated cells are necessary, specific adjustment strategies need to be generated based on the problem type of the key protected grid, the degree of supply-demand imbalance, the relationship between associated cells, and the current network parameter status. Among these, optimizable antenna parameters include antenna downtilt angle, transmit power, azimuth angle, handover parameters, and load balancing parameters.

[0069] For example, for high-demand, weak-coverage areas, it is recommended to reduce the downtilt angle of the target cell, adjust the azimuth angle, or appropriately increase the transmit power to enhance coverage of key grids; for high-demand, high-interference areas, it is recommended to adjust the downtilt angle, azimuth angle, or transmit power of the target cell or interfering cell to reduce coverage spillover and neighboring cell interference; for high-load, low-rate areas, it is recommended to adjust load balancing parameters or handover parameters to divert some users to neighboring cells with carrying capacity; for areas with abnormal handover, it is recommended to optimize neighboring cell relationships, handover thresholds, or handover hysteresis parameters to improve connection stability during mobility.

[0070] In this embodiment of the invention, a preset reinforcement learning model generates adjustment strategies to improve the accuracy of the generated adjustment strategies. The inputs to the preset reinforcement learning model are network data, a first score, and a second score. The model outputs adjustment strategies including the first score, the second score, and parameters such as the imbalance index, current antenna downtilt angle, transmit power, azimuth angle, PRB utilization, RSRP, SINR, uplink and downlink rates, the status of associated cells, and the status of boundary cells. These output adjustment strategies may include downtilt angle adjustment, transmit power adjustment, azimuth angle adjustment, handover parameter adjustment, load balancing parameter adjustment, or remaining unchanged.

[0071] Furthermore, the preset reinforcement learning model of the present invention is obtained by training an initial reinforcement learning model, such as... Figure 5 As shown, sample network data and the corresponding sample scores are input into the initial reinforcement learning model. An adjustment strategy is then output and adjusted according to the strategy to obtain intermediate scores. The reward value is then calculated through key scores to verify the adjustment strategy, thereby optimizing the initial reinforcement learning model.

[0072] In some implementations, the reward value can be calculated using the following formula: ; In the formula, This indicates the increase in median ratings relative to the sample ratings; This indicates the increase in uplink / downlink speed or throughput. Indicates SINR or coverage quality improvement; Indicates the degree of degradation of indicators in surrounding grids or boundary cells; P indicates the parameter adjustment range or penalty for frequent adjustments; to The weighting coefficients are used to design the reward value so that the GNN model prioritizes adjustment strategies that improve key protected grids without significantly harming surrounding areas, effectively improving the accuracy of the output adjustment strategies.

[0073] In one embodiment, the associated cells include a primary serving cell, interfering cells, and boundary cells. The primary serving cell is used to provide services to the target grid beyond a set coverage area, or the primary serving cell is used to provide network data to the target grid beyond a set network traffic limit. The interfering cells are cells whose antennas generate neighboring cell signals, overlapping coverage, or interference effects in the target grid. The boundary cells are cells that experience coverage, interference, or load changes after the antennas of the associated cells are adjusted. The process of generating the adjustment strategy for the associated cells based on the network data of the target grid includes: The initial strategy for the primary serving cell and / or the interfering cell is generated based on the network data of the target grid; Predict the network change parameters of the boundary cell based on the initial strategy; If the network change parameter is less than the set change threshold, the initial strategy is set as the adjustment strategy.

[0074] It should be noted that the target grid only points to the specific location where the problem occurs, while antenna parameter optimization needs to be applied to specific cells or sectors. After determining the target grid, it is necessary to combine base station operating parameters, cell coverage relationships, MR data, user dwell relationships, and neighbor cell relationships to map the grid to the associated cells.

[0075] Among them, associated cells include primary serving cells, interfering cells, and boundary cells. A primary serving cell is a cell that provides primary coverage or carries primary service traffic to the target grid, such as providing service to the target grid beyond a set coverage area, or the primary serving cell is used to provide network data to the target grid beyond a set network traffic. An interfering cell is a cell that generates neighboring cell signals, overlapping coverage, or interference within the grid. A boundary cell is a surrounding cell that is affected by coverage, interference, or load after parameter adjustments.

[0076] For example, if a target grid is mainly served by cell A, and cell A has a high proportion of MR sampling points, cell A can be identified as the primary serving cell. If cells B and C generate strong neighboring cell signals within the grid, and the grid's SINR is low, cells B and C can be identified as interfering cells. If adjusting the downtilt angle or transmit power of cell A might change the coverage boundary of the surrounding grid, then the relevant neighboring cells are included in the boundary cell set. Through this mapping, the grid-level problem is transformed into an executable cell-level optimization object.

[0077] Furthermore, adjusting the antenna parameters may change the coverage boundaries and interference status of surrounding cells. Therefore, in this embodiment, the adjustment strategy is verified before outputting the adjustment strategy to improve the accuracy of the output adjustment strategy.

[0078] In some implementations, the verification includes whether the parameter adjustment range exceeds the allowable range of the project, whether the weak coverage of surrounding cells may increase, whether the SINR or PRB utilization of boundary cells may deteriorate, whether the candidate strategy violates key protection requirements, and whether it conflicts with existing network optimization strategies or manually locked parameters.

[0079] In this embodiment of the invention, network change parameters and a set change threshold are introduced. Whether the adjustment strategy can be output is determined by whether the network change parameters are less than the set change threshold. The network change parameters are predicted based on the initial strategy, such as changes in network data or the number of complaints. If the network change parameters are less than the set change threshold, it indicates that the impact on the boundary cells is relatively small, and the adjustment strategy can be output.

[0080] In some implementations, verified adjustment strategies can be output to optimization personnel for confirmation, or they can be implemented in a phased rollout in scenarios with automation capabilities. After the strategy is implemented, changes in network metrics, user experience, and complaints after the adjustment are collected, and the first score, second score, and imbalance index are recalculated. If the target grid metrics improve and there is no significant degradation in the surrounding areas, the adjustment strategy is retained; if the effect after implementation does not meet expectations, a rollback or regeneration of the adjustment is triggered.

[0081] Please see Figure 6 , Figure 6 This is a structural diagram of an antenna parameter optimization device provided in an embodiment of the present invention, as shown below. Figure 6 As shown, the antenna parameter optimization device 600 includes: The acquisition module 601 is used to acquire network data of multiple grids, wherein the multiple grids are grids included in a first region, and the first region is an area where at least one antenna provides network services; The first calculation module 602 is used to calculate a first score for each grid based on the network data of each grid, and the first score is used to characterize the degree of network resource demand of the corresponding grid. The second calculation module 603 is used to construct a graph structure based on the network data of the multiple grids, and to calculate a second score for each grid based on the graph structure. The second score is used to characterize the network resource carrying capacity of the corresponding grid, and the graph structure is used to characterize the correlation between the features of the network data of each grid. The third calculation module 604 is used to calculate the imbalance index of each grid based on the first score and the second score of each grid, and the imbalance index is used to characterize the balance between the network resource carrying capacity and demand of the corresponding grid. The determination module 605 is used to determine the associated cells of the target grid, wherein the target grid is a grid whose imbalance index is greater than a set index threshold; The generation module 606 is used to generate an adjustment strategy for the associated cell based on the network data of the target grid. The adjustment strategy is used to adjust the parameters of the antennas of the associated cell.

[0082] In one embodiment, the network data includes first data and / or second data, wherein the first data is data including latitude and longitude information, and the second data is data collected through multiple cells, wherein the multiple cells are cells that provide network data to the first area; The acquisition module 601 includes: The first acquisition unit is configured to acquire first initial data and / or second initial data of the first area, wherein the first initial data includes data with latitude and longitude information, and the second initial data is data collected through the plurality of cells; A partitioning unit is used to divide the first region into the plurality of grids to obtain the position information of each grid. A first processing unit is configured to divide the first initial data into the multiple grids based on the latitude and longitude information of the first initial data and the location information of the multiple grids, thereby obtaining first data for each grid; and / or, The second processing unit is used to calculate the weight coefficient of each grid relative to each cell in the plurality of cells based on the location information of each grid, and to calculate the second data of each grid based on the weight coefficient of each cell and the second initial data.

[0083] In one embodiment, the plurality of grids includes a first grid, the network data includes at least one of traffic intensity and number of active users, the network data further includes traffic data, and the first computing module 602 includes: The second acquisition unit is used to acquire historical traffic data and guarantee parameters of the first grid, wherein the guarantee parameters are used to characterize whether the current time is within a preset time period of the first grid; The first calculation unit is used to calculate the trend parameters of the first grid based on the traffic data and the historical traffic data of the first grid, and the trend parameters are used to characterize the traffic changes of the first grid. The second calculation unit is used to weight at least one of the traffic intensity and the number of active users of the first grid, as well as the guarantee parameter and the trend parameter, to obtain the first score of the first grid.

[0084] In one embodiment, the second computing module 603 includes: The third acquisition unit is used to acquire the grid relationship of each grid, wherein the grid relationship is used to characterize the positional relationship between each grid and other grids in the plurality of grids; A construction unit is used to construct the graph structure based on the network data and the grid relationships of the multiple grids. The graph structure includes multiple nodes and multiple edges. The multiple nodes correspond one-to-one with the multiple grids, and the multiple edges correspond one-to-one with the grid relationships of the multiple grids. The features of each node in the multiple nodes are used to characterize the network data of the grid corresponding to the corresponding node. The output unit is used to input the graph structure into the graph neural network (GNN) model and output a second score for each grid based on the GNN model, wherein the GNN model is trained to output the second score for different nodes.

[0085] In one embodiment, the generation module 606 includes: The output unit is used to input the network data of the target grid, the first score and the second score into a preset reinforcement learning model, and output the adjustment strategy of the associated cell based on the preset reinforcement learning model; The preset reinforcement learning model is trained in the following manner: Obtain sample network data of the sample grid and the sample score corresponding to the sample network data. The sample score is used to characterize the network resource carrying capacity of the sample grid. Multiple intermediate learning models are generated based on an initial reinforcement learning model, and the parameters of the multiple intermediate learning models are different. The initial reinforcement learning model is constructed as an output adjustment strategy. The sample network data is processed based on each intermediate learning model to obtain the intermediate strategy of each intermediate learning model; The intermediate strategies of each intermediate learning model are adjusted to obtain the adjusted network data corresponding to each intermediate learning model. Calculate the intermediate score for each intermediate learning model based on the adjusted network data corresponding to each intermediate learning model. Based on the adjusted network data and intermediate scores corresponding to each intermediate learning model, as well as the sample network data and sample scores, the reward value of each intermediate learning model is calculated respectively. The model with the largest reward value among the multiple intermediate learning models is set as the preset reinforcement learning model.

[0086] In one embodiment, the associated cells include a primary serving cell, interfering cells, and boundary cells. The primary serving cell is used to provide services to the target grid beyond a set coverage area, or the primary serving cell is used to provide network data to the target grid beyond a set network traffic limit. The interfering cells are cells whose antennas generate neighboring cell signals, overlapping coverage, or interference effects in the target grid. The boundary cells are cells that experience coverage, interference, or load changes after the antennas of the associated cells are adjusted. The generation module 606 includes: A generation unit is used to generate an initial strategy for the primary serving cell and / or the interfering cell based on the network data of the target grid; The prediction unit is used to predict the network change parameters of the boundary cell based on the initial strategy. The setting unit is used to set the initial strategy as the adjustment strategy when the network change parameter is less than a set change threshold.

[0087] The antenna parameter optimization device provided in this embodiment of the invention can realize each process of each embodiment of the above-described antenna parameter optimization method. The technical features are one-to-one and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0088] It should be noted that the antenna parameter optimization device in the embodiments of the present invention can be a device, or it can be a component, integrated circuit, or chip in an electronic device.

[0089] This invention also provides an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the above-described functionality. Figure 1 The various processes of the antenna parameter optimization method embodiment shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0090] For details, see Figure 7 As shown, this embodiment of the invention also provides an electronic device, including a bus 701, a transceiver 702, an antenna 703, a bus interface 704, a processor 705, and a memory 706.

[0091] The transceiver 702 is used to acquire network data from multiple grids, wherein the multiple grids are grids included in a first region, and the first region is an area where at least one antenna provides network services; The processor 705 is used to calculate a first score for each grid based on the network data of each grid, and the first score is used to characterize the degree of network resource demand of the corresponding grid. The processor 705 is further configured to construct a graph structure based on the network data of the plurality of grids, and calculate a second score for each grid based on the graph structure. The second score is used to characterize the network resource carrying capacity of the corresponding grid, and the graph structure is used to characterize the correlation between the features of the network data of each grid. The processor 705 is further configured to calculate an imbalance index for each grid based on the first score and the second score of each grid, wherein the imbalance index is used to characterize the balance between the network resource carrying capacity and demand of the corresponding grid. The processor 705 is further configured to determine the associated cells of the target grid, wherein the target grid is a grid whose imbalance index is greater than a set index threshold; The processor 705 is further configured to generate an adjustment strategy for the associated cell based on the network data of the target grid, the adjustment strategy being used to adjust the antenna parameters of the associated cell.

[0092] In one embodiment, the network data includes first data and / or second data, wherein the first data is data including latitude and longitude information, and the second data is data collected through multiple cells, wherein the multiple cells are cells that provide network data to the first area; The acquisition of network data from multiple grids includes: Acquire first initial data and / or second initial data for the first region, wherein the first initial data includes data with latitude and longitude information, and the second initial data is data collected through the plurality of cells; The first region is divided into the multiple grids to obtain the position information of each grid. Based on the latitude and longitude information of the first initial data and the location information of the multiple grids, the first initial data is divided into the multiple grids to obtain the first data for each grid; and / or, Based on the location information of each grid, the weight coefficient of each grid relative to each of the plurality of cells is calculated, and the second data of each grid is calculated based on the weight coefficient of each cell and the second initial data.

[0093] In one embodiment, the plurality of grids includes a first grid, the network data includes at least one of traffic intensity and number of active users, the network data further includes traffic data, and the calculation of a first score for each grid based on the network data of each grid includes: Obtain historical traffic data and guarantee parameters for the first grid, wherein the guarantee parameters are used to characterize whether the current time is within a preset time period of the first grid; Based on the traffic data and historical traffic data of the first grid, a trend parameter of the first grid is calculated, and the trend parameter is used to characterize the traffic change of the first grid. The first score of the first grid is obtained by weighting at least one of the traffic intensity and the number of active users of the first grid, as well as the guarantee parameter and the trend parameter.

[0094] In one embodiment, constructing a graph structure based on the network data of the plurality of grids includes: Obtain the grid relationship for each grid, which is used to characterize the positional relationship between each grid and other grids in the plurality of grids; The graph structure is constructed based on the network data and grid relationships of the multiple grids. The graph structure includes multiple nodes and multiple edges. The multiple nodes correspond one-to-one with the multiple grids, and the multiple edges correspond one-to-one with the grid relationships of the multiple grids. The features of each node are used to characterize the network data of the grid corresponding to the node. The calculation of the second score for each grid based on the graph structure includes: The graph structure is input into a graph neural network (GNN) model, and a second score for each grid is output based on the GNN model. The GNN model is trained to output the second score for different nodes.

[0095] In one embodiment, generating the adjustment strategy for the associated cells based on the network data of the target grid includes: The network data of the target grid, the first score, and the second score are input into a preset reinforcement learning model, and the adjustment strategy of the associated cell is output based on the preset reinforcement learning model. The preset reinforcement learning model is trained in the following manner: Obtain sample network data of the sample grid and the sample score corresponding to the sample network data. The sample score is used to characterize the network resource carrying capacity of the sample grid. Multiple intermediate learning models are generated based on an initial reinforcement learning model, and the parameters of the multiple intermediate learning models are different. The initial reinforcement learning model is constructed as an output adjustment strategy. The sample network data is processed based on each intermediate learning model to obtain the intermediate strategy of each intermediate learning model; The intermediate strategies of each intermediate learning model are adjusted to obtain the adjusted network data corresponding to each intermediate learning model. Calculate the intermediate score for each intermediate learning model based on the adjusted network data corresponding to each intermediate learning model. Based on the adjusted network data and intermediate scores corresponding to each intermediate learning model, as well as the sample network data and sample scores, the reward value of each intermediate learning model is calculated respectively. The model with the largest reward value among the multiple intermediate learning models is set as the preset reinforcement learning model.

[0096] In one embodiment, the associated cells include a primary serving cell, interfering cells, and boundary cells. The primary serving cell is used to provide services to the target grid beyond a set coverage area, or the primary serving cell is used to provide network data to the target grid beyond a set network traffic limit. The interfering cells are cells whose antennas generate neighboring cell signals, overlapping coverage, or interference effects in the target grid. The boundary cells are cells that experience coverage, interference, or load changes after the antennas of the associated cells are adjusted. The process of generating the adjustment strategy for the associated cells based on the network data of the target grid includes: The initial strategy for the primary serving cell and / or the interfering cell is generated based on the network data of the target grid; Predict the network change parameters of the boundary cell based on the initial strategy; If the network change parameter is less than the set change threshold, the initial strategy is set as the adjustment strategy.

[0097] exist Figure 7 In this document, a bus architecture (represented by bus 701) is used. Bus 701 can include any number of interconnected buses and bridges, linking various circuits including one or more processors represented by processor 705 and memory represented by memory 706. Bus 701 can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. Bus interface 704 provides an interface between bus 701 and transceiver 702. Transceiver 702 can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 705 is transmitted over a wireless medium via antenna 703, which further receives data and transmits data to processor 705.

[0098] Processor 705 manages bus 701 and general processing, and also provides various functions, including timing, peripheral interface, voltage regulation, power management, and other control functions. Memory 706 can be used to store data used by processor 705 during operation.

[0099] Optionally, the processor 705 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a graphics processing unit (GPU).

[0100] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the above-described functions. Figure 1 The various processes of the corresponding antenna parameter optimization method embodiments, which can achieve the same technical effect, will not be described again here to avoid repetition. The computer-readable storage medium mentioned includes, for example, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0101] The present invention also provides a computer program product, including computer instructions that, when executed by a processor, implement the above-described... Figure 1 The various processes of the corresponding antenna parameter optimization method embodiments can achieve the same technical effect, and will not be described again here to avoid repetition.

[0102] In the embodiments of this invention, the terms "first," "second," etc., are used to distinguish similar object parameters and are not necessarily used to describe a specific order or sequence. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices. Additionally, the use of "and / or" in this application indicates at least one of the connected object parameters, such as A and / or B and / or C, representing seven possibilities: A alone, B alone, C alone, both A and B present, both B and C present, both A and C present, and A, B, and C present.

[0103] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0104] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or second terminal device, etc.) to execute the methods of the various embodiments of this application.

[0105] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for optimizing antenna parameters, characterized in that, include: Acquire network data from multiple grids, wherein the multiple grids are grids included in a first region, and the first region is an area where at least one antenna provides network services; A first score is calculated for each grid based on the network data of each grid, and the first score is used to characterize the degree of network resource demand of the corresponding grid. A graph structure is constructed based on the network data of the multiple grids, and a second score is calculated for each grid based on the graph structure. The second score is used to characterize the network resource carrying capacity of the corresponding grid, and the graph structure is used to characterize the correlation between the features of the network data of each grid. Based on the first score and the second score of each grid, the imbalance index of each grid is calculated respectively. The imbalance index is used to characterize the balance between the network resource carrying capacity and demand of the corresponding grid. Identify the associated cells of the target grid, wherein the target grid is the grid whose imbalance index is greater than a set index threshold; An adjustment strategy for the associated cell is generated based on the network data of the target grid. The adjustment strategy is used to adjust the antenna parameters of the associated cell.

2. The method as described in claim 1, characterized in that, The network data includes first data and / or second data, wherein the first data is data including latitude and longitude information, and the second data is data collected through multiple cells, wherein the multiple cells are cells that provide network data to the first area; The acquisition of network data from multiple grids includes: Acquire first initial data and / or second initial data for the first region, wherein the first initial data includes data with latitude and longitude information, and the second initial data is data collected through the plurality of cells; The first region is divided into the multiple grids to obtain the position information of each grid. Based on the latitude and longitude information of the first initial data and the location information of the multiple grids, the first initial data is divided into the multiple grids to obtain the first data for each grid; and / or, Based on the location information of each grid, the weight coefficient of each grid relative to each of the plurality of cells is calculated, and the second data of each grid is calculated based on the weight coefficient of each cell and the second initial data.

3. The method as described in claim 1 or 2, characterized in that, The plurality of grids includes a first grid, the network data includes at least one of traffic intensity and number of active users, the network data also includes traffic data, and the calculation of a first score for each grid based on the network data of each grid includes: Obtain historical traffic data and guarantee parameters for the first grid, wherein the guarantee parameters are used to characterize whether the current time is within a preset time period of the first grid; Based on the traffic data and historical traffic data of the first grid, a trend parameter of the first grid is calculated, and the trend parameter is used to characterize the traffic change of the first grid. The first score of the first grid is obtained by weighting at least one of the traffic intensity and the number of active users of the first grid, as well as the guarantee parameter and the trend parameter.

4. The method as described in claim 1 or 2, characterized in that, The construction of the graph structure based on the network data of the multiple grids includes: Obtain the grid relationship for each grid, which is used to characterize the positional relationship between each grid and other grids in the plurality of grids; The graph structure is constructed based on the network data and grid relationships of the multiple grids. The graph structure includes multiple nodes and multiple edges. The multiple nodes correspond one-to-one with the multiple grids, and the multiple edges correspond one-to-one with the grid relationships of the multiple grids. The features of each node are used to characterize the network data of the grid corresponding to the node. The calculation of the second score for each grid based on the graph structure includes: The graph structure is input into a graph neural network (GNN) model, and a second score for each grid is output based on the GNN model. The GNN model is trained to output the second score for different nodes.

5. The method as described in claim 1 or 2, characterized in that, The process of generating the adjustment strategy for the associated cells based on the network data of the target grid includes: The network data of the target grid, the first score, and the second score are input into a preset reinforcement learning model, and the adjustment strategy of the associated cell is output based on the preset reinforcement learning model. The preset reinforcement learning model is trained in the following manner: Obtain sample network data of the sample grid and the sample score corresponding to the sample network data. The sample score is used to characterize the network resource carrying capacity of the sample grid. Multiple intermediate learning models are generated based on an initial reinforcement learning model, and the parameters of the multiple intermediate learning models are different. The initial reinforcement learning model is constructed as an output adjustment strategy. The sample network data is processed based on each intermediate learning model to obtain the intermediate strategy of each intermediate learning model; The intermediate strategies of each intermediate learning model are adjusted to obtain the adjusted network data corresponding to each intermediate learning model. Calculate the intermediate score for each intermediate learning model based on the adjusted network data corresponding to each intermediate learning model. Based on the adjusted network data and intermediate scores corresponding to each intermediate learning model, as well as the sample network data and sample scores, the reward value of each intermediate learning model is calculated respectively. The model with the largest reward value among the multiple intermediate learning models is set as the preset reinforcement learning model.

6. The method as described in claim 5, characterized in that, The associated cells include primary serving cells, interfering cells, and boundary cells. The primary serving cell is used to provide services to the target grid beyond a set coverage area, or the primary serving cell is used to provide network data to the target grid beyond a set network traffic. The interfering cell is a cell whose antenna generates neighbor cell signals, overlapping coverage, or interference in the target grid. The boundary cell is a cell that experiences coverage, interference, or load changes after the antenna of the associated cell is adjusted. The process of generating the adjustment strategy for the associated cells based on the network data of the target grid includes: The initial strategy for the primary serving cell and / or the interfering cell is generated based on the network data of the target grid; Predict the network change parameters of the boundary cell based on the initial strategy; If the network change parameter is less than the set change threshold, the initial strategy is set as the adjustment strategy.

7. An antenna parameter optimization device, characterized in that, include: The acquisition module is used to acquire network data from multiple grids, wherein the multiple grids are grids included in a first region, and the first region is an area where at least one antenna provides network services; The first calculation module is used to calculate a first score for each grid based on the network data of each grid, and the first score is used to characterize the degree of network resource demand of the corresponding grid. The second calculation module is used to construct a graph structure based on the network data of the multiple grids, and to calculate a second score for each grid based on the graph structure. The second score is used to characterize the network resource carrying capacity of the corresponding grid, and the graph structure is used to characterize the correlation between the features of the network data of each grid. The third calculation module is used to calculate the imbalance index of each grid based on the first score and the second score of each grid. The imbalance index is used to characterize the balance between the network resource carrying capacity and demand of the corresponding grid. The determination module is used to determine the associated cells of the target grid, wherein the target grid is a grid whose imbalance index is greater than a set index threshold; A generation module is used to generate an adjustment strategy for the associated cell based on the network data of the target grid. The adjustment strategy is used to adjust the antenna parameters of the associated cell.

8. An electronic device, characterized in that, include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the antenna parameter optimization method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the antenna parameter optimization method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, It includes computer instructions that, when executed by a processor, implement the steps of the antenna parameter optimization method as described in any one of claims 1 to 6.

Images