Wireless network traffic prediction method and apparatus
By constructing a multi-input traffic prediction model that considers the interaction between cells, the problem of information loss caused by isolated cell prediction in existing technologies is solved, and more accurate network traffic prediction is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA MOBILE GRP HEILONGJIANG CO LTD
- Filing Date
- 2022-05-20
- Publication Date
- 2026-06-16
Smart Images
Figure CN117135649B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of wireless communication technology, and in particular to a method and apparatus for predicting wireless network traffic. Background Technology
[0002] Currently, network traffic and capacity prediction is a crucial research topic in wireless network optimization. Accurate network traffic prediction can more efficiently guide capacity expansion, carrier scheduling, wireless planning, and the assurance of critical scenarios. Existing wireless network traffic prediction technologies can be broadly categorized into two types: time-series data prediction based on machine learning and deep neural network technologies, and time-series processing and prediction based on traditional statistical methods such as moving averages and autoregressive models.
[0003] However, existing prediction techniques are based on the processing of traffic data from a single cell, which means that the cell is isolated from the communication network for traffic prediction, resulting in the loss of a lot of useful information. Summary of the Invention
[0004] This application provides a wireless network traffic prediction method and apparatus to solve the technical problem in the prior art that prediction is based on the processing of traffic data of a single cell, that is, the cell is isolated from the communication network for traffic prediction, which results in the loss of a lot of useful information.
[0005] In a first aspect, embodiments of this application provide a wireless network traffic prediction method, including:
[0006] Obtain handover information between the cell and all its neighboring cells within a target time period, obtain the influence weight of each neighboring cell on the cell, and stratify all neighboring cells of the cell according to the influence weight to obtain stratified neighboring cells;
[0007] Calculate the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time;
[0008] Input the first traffic transfer data and the second traffic transfer data into the traffic prediction model to obtain the traffic prediction result output by the traffic prediction model; wherein, the traffic prediction model includes a first network structure and a second network structure, the first network structure is used to obtain the traffic prediction result of the cell based on the first traffic transfer data, and the second network structure is used to obtain the traffic prediction result of the hierarchical neighboring cells based on the second traffic transfer data.
[0009] In one embodiment, obtaining handover information between the cell and all its neighboring cells within a target time period, and obtaining the influence weight of each neighboring cell on the cell, includes:
[0010] Obtain the first handover information from the cell to the neighboring cell within the target time period;
[0011] Obtain the second handover information from the neighboring cell to the cell within the target time period;
[0012] The influence weight of the neighboring cell on the cell is obtained based on the first handover information, the second handover information, and the distance between the cell and the neighboring cell.
[0013] In one embodiment, the step of stratifying all neighboring cells of the cell according to the influence weight to obtain stratified neighboring cells includes:
[0014] Based on the aforementioned influence weights, all neighboring cells of the cell are sorted.
[0015] Based on the target ratio, all sorted neighboring regions are divided into layers to obtain layered neighboring regions.
[0016] In one embodiment, calculating the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time includes:
[0017] Based on the cell's traffic at the target time, the cell's total number of users, and the number of successful handovers at the target time, the first traffic transfer data of the cell at the target time is obtained.
[0018] Based on the traffic of the hierarchical neighboring cells at the target time and the handover ratio of the hierarchical neighboring cells to the cell, the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time is obtained.
[0019] In one embodiment, the first network structure includes several cascaded RNN networks, and the second network structure is determined based on the influence weights of the hierarchical neighboring regions.
[0020] In one embodiment, the traffic prediction model further includes network aggregation and a fully connected network. The network aggregation is used to aggregate the outputs of the first network structure and the second network structure, and the fully connected network is used to filter the output of the network aggregation to obtain the traffic prediction result.
[0021] In one embodiment, the traffic prediction result includes traffic prediction data at several time points.
[0022] Secondly, embodiments of this application provide a wireless network traffic prediction device, comprising:
[0023] The hierarchical module is used to obtain handover information between the cell and all its neighboring cells within a target time period, obtain the influence weight of each neighboring cell on the cell, and hierarchically divide all the neighboring cells of the cell according to the influence weight to obtain hierarchical neighboring cells;
[0024] The traffic calculation module is used to calculate the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time;
[0025] The prediction module is used to input the first traffic transfer data and the second traffic transfer data into the traffic prediction model to obtain the traffic prediction result output by the traffic prediction model; wherein, the traffic prediction model includes a first network structure and a second network structure, the first network structure is used to obtain the traffic prediction result of the cell based on the first traffic transfer data, and the second network structure is used to obtain the traffic prediction result of the hierarchical neighboring cells based on the second traffic transfer data.
[0026] Thirdly, embodiments of this application provide an electronic device, including a processor and a memory storing a computer program, wherein the processor executes the program to implement the steps of the wireless network traffic prediction method described in the first aspect.
[0027] Fourthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the steps of the wireless network traffic prediction method described in the first or second aspect.
[0028] The wireless network traffic prediction method and apparatus provided in this application classify the neighboring cells of a cell by influence weights and construct a multi-input traffic prediction model based on the cell and the layered neighboring cells. It does not process the traffic data of a single cell, but takes into account the interaction and influence between cells, so as not to lose useful information. Attached Figure Description
[0029] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0030] Figure 1 This is a flowchart illustrating the wireless network traffic prediction method provided in an embodiment of this application;
[0031] Figure 2 This is provided by the embodiments of this application. Figure 1 A flowchart illustrating step S1.
[0032] Figure 3 This is provided by the embodiments of this application. Figure 1 A flowchart illustrating step S2;
[0033] Figure 4 This is a schematic diagram of the traffic prediction model provided in the embodiments of this application;
[0034] Figure 5 This is a schematic diagram of the process of traffic prediction based on a traffic prediction model provided in an embodiment of this application;
[0035] Figure 6 This is a training and testing loss curve of the traffic prediction model provided in the embodiments of this application;
[0036] Figure 7 It is a curve of the data normalization prediction results;
[0037] Figure 8 This is a schematic diagram of the wireless network traffic prediction device provided in the embodiments of this application;
[0038] Figure 9 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation
[0039] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0040] Figure 1 This is one of the flowcharts illustrating the wireless network traffic prediction method provided in an embodiment of this application. (Refer to...) Figure 1 This application provides a wireless network traffic prediction method, which may include:
[0041] S1, obtain the handover information between the cell and all its neighboring cells within the target time period, obtain the influence weight of each neighboring cell on the cell, and divide all the neighboring cells of the cell into layers according to the influence weight to obtain layered neighboring cells.
[0042] S2, calculate the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time.
[0043] S3, input the first traffic transfer data and the second traffic transfer data into the traffic prediction model to obtain the traffic prediction result output by the traffic prediction model; wherein, the traffic prediction model includes a first network structure and a second network structure, the first network structure is used to obtain the traffic prediction result of the cell based on the first traffic transfer data, and the second network structure is used to obtain the traffic prediction result of the hierarchical neighboring cell based on the second traffic transfer data.
[0044] In step S1, the target time period is generally at the hourly level, and this application does not specify a particular time period. The traffic changes in a cell within a certain time period are mainly affected by two factors: 1) the traffic generated by users flowing into the cell from neighboring cells; and 2) the traffic generated by existing users in the cell. The strength of the influence between neighboring cells can be measured using handover information.
[0045] Based on the magnitude of the influence weight, all neighboring areas can be hierarchically divided, resulting in at least two types of hierarchical neighboring areas, such as first-level neighboring areas and second-level neighboring areas, with each type of hierarchical neighboring area including several neighboring areas.
[0046] In step S3, the number of second network structures corresponds to the number of types of hierarchical neighbor cells; that is, each first-level neighbor cell corresponds to one type of second network structure, and each second-level neighbor cell corresponds to one type of second network structure. The second traffic transfer data from the first-level and second-level neighbor cells to the cell at the target time is input into their respective corresponding second network structures.
[0047] It is understood that the embodiments of this application hierarchically classify the neighboring cells of a cell by influence weights and construct a multi-input traffic prediction model based on the cell and hierarchical neighboring cells. This is not a single cell traffic data processing, but takes into account the interaction and influence between cells, so as not to lose useful information and improve the prediction accuracy.
[0048] Based on the above embodiments, as a preferred embodiment, such as Figure 2 As shown, the step of obtaining handover information between the cell and all its neighboring cells within the target time period, and obtaining the influence weight of each neighboring cell on the cell, includes:
[0049] S210, Obtain the first handover information from the cell to the neighboring cell within the target time period. The first handover information includes the number of handovers h from cell i to its neighboring cell j. ij Switching success rate p ij .
[0050] S220, Obtain the second handover information from the neighboring cell to the cell within the target time period. The second handover information includes the number of handovers h from neighboring cell j to cell i. ji Switching success rate p ji .
[0051] S230, based on the first handover information, the second handover information, and the distance between the cell and the neighboring cell, the influence weight of the neighboring cell on the cell is obtained. The formula for calculating the influence weight is as follows:
[0052]
[0053] Where, d ji v is the distance between cells i and j. ji With v ij These represent the correlation strength between cell i and cell j, respectively, with v ji For example, the higher the handover success rate from cell j to cell i, and the higher the proportion of handovers from cell j to cell i to the total number of handovers from cell j, the greater the probability of users moving from cell j to cell i, and the greater the impact of cell j on traffic changes in cell i. Simultaneously considering the user outflow issue from cell i, a v... ij The meaning of v ji Similarly, the calculation formulas for both are as follows:
[0054]
[0055] It is understood that the embodiments of this application simultaneously consider both cell-to-neighbor cell handover and neighbor cell-to-cell handover, and combine the distance between the cell and the neighbor cell to obtain the influence weight, thereby improving the accuracy of subsequent neighbor cell stratification.
[0056] Based on the above embodiments, as a preferred embodiment, the step of stratifying all neighboring cells of the cell according to the influence weight to obtain stratified neighboring cells includes:
[0057] Based on the influence weight, all neighboring cells of the cell are sorted, and can be sorted from largest to smallest according to the influence weight.
[0058] Based on the target ratio, all sorted neighboring cells are stratified into tiered neighboring cells. This stratification can be done at a 1:1 ratio or by setting an influence weight threshold. This application uses a proportional division for specific details: the first-level neighboring cells are the top 50%, and the second-level neighboring cells are the bottom 50%. The first-level neighboring cell table L1: [cell1, cell2, ..., cell...] x+1 ], Second-level neighbor table L2: [cell x+2 cell x+3 , ..., cell n ].
[0059] It is understood that the embodiments of this application fully consider the special characteristics of communication networks, namely the interaction and influence between cells on traffic changes, which can improve the accuracy of prediction results.
[0060] Based on the above embodiments, as a preferred embodiment, such as Figure 3 As shown, the calculation of the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time includes:
[0061] S310, based on the cell's traffic at the target time, the total number of users in the cell, and the number of successful handovers in the cell at the target time, the first traffic transfer data of the cell at the target time is obtained.
[0062] Assume the traffic flow in cell i at time t is The total number of users in community i is The number of successful switches at time t is Then, at time t, the transfer of subsequent traffic within cell i itself.
[0063]
[0064] The range of values is Average number of handovers per user in cell i It is inversely proportional, meaning that the more times each user switches out on average, the smaller their contribution to the subsequent increase in traffic.
[0065] S320, based on the traffic of the hierarchical neighboring cell at the target time and the handover ratio of the hierarchical neighboring cell to the cell, the second traffic transfer data of the hierarchical neighboring cell to the cell at the target time is obtained.
[0066] For cell j, a neighboring cell of cell i, assume that the current traffic of cell j at time t is... The number of handover requests from j to cell k is The traffic transfer from neighboring cell j to cell i at time t can be expressed as follows: This serves as the final input for the neighboring cell flow in the subsequent prediction model:
[0067]
[0068] Understandably, calculating traffic transfer data for cells and hierarchical neighbor cells for subsequent traffic prediction can improve the accuracy of traffic prediction.
[0069] Optionally, the traffic prediction results include traffic prediction data at several time points.
[0070] In terms of traffic impact and switching metrics, data blocks with different lookback and delay lengths are constructed for switching data. Here, lookback represents the length of the input historical data; lookback=7 means the input data is from the past 7 time points. Delay represents the time length of the predicted traffic; delay=1 means predicting the traffic at the next time point, and delay=5 means predicting the traffic 5 time points later. Here, lookback=7 can be fixed, and delay can take values of 1, 2, 3, 4, and 5 sequentially. For each cell, the following data format is obtained:
[0071] Neighboring cell j:
[0072] (7,1)->Input: Output:
[0073] (7,2)->Input: Output:
[0074] ...
[0075] (7,5)->Enter: Output:
[0076] Similarly, the input and output data for cell i are obtained:
[0077] (7,1)->Input: Output:
[0078] (7,2)->Input: Output:
[0079] ...
[0080] (7,5)->Enter: Output:
[0081] It is understood that the embodiments of this application can achieve multiple inputs and multiple outputs, thereby improving prediction capabilities.
[0082] Based on the above embodiments, as a preferred embodiment, the first network structure includes several cascaded RNN networks, and the second network structure is determined based on the influence weights of the hierarchical neighboring regions.
[0083] The g values of the local cell network, the first-level neighbor cell network, and the second-level neighbor cell network are used as inputs to construct a traffic prediction model. The model structure diagram is shown in Figure 4.
[0084] The model consists of three interconnected networks, which can be abstractly expressed as follows:
[0085] (1) The network corresponding to cell i is composed of two-level RNN networks connected in series, used to characterize the temporal changes of the cell's own traffic. i :
[0086] out i =RNN(RNN([g ii ]))
[0087] (2) Level 1 neighbor network L1, composed of RNN+DNN networks connected in series, is used to characterize the time-series traffic impact of Level 1 neighbor networks on cell i:
[0088] out L1 =RNN(RNN([g L1 ]))
[0089] (3) The L2 neighbor network corresponding to the L2 neighbor cell consists of a single-layer RNN network, used to characterize the impact of the L2 neighbor cell on the time-series traffic of cell i. Since the impact of the L2 neighbor cell on cell i is relatively weak, it can be characterized by a single RNN:
[0090] out L2 =RNN(RNN([g L2 ]))
[0091] Optionally, the traffic prediction model further includes network aggregation and a fully connected network. The network aggregation is used to aggregate the outputs of the first network structure and the second network structure, and the fully connected network is used to filter the output of the network aggregation to obtain the traffic prediction result.
[0092] The overall structure of the traffic prediction model is as follows:
[0093] out = RNN(concate(out) i out L1 out L2 ))
[0094] Among them, RNN, DNN, and concatenate correspond to recurrent neural networks, fully connected networks, and network convergence, respectively.
[0095] The model is explained in detail below:
[0096] (1) Model input: the time series lengths corresponding to 7 time periods from t=1 to 7.
[0097] (2) Model output: The input of the model is a multi-output DNN structure, which can complete the prediction of 5 time periods from t=8 to 12 at one time. If you want to predict a longer time period, you can adjust the time length of the input, but the accuracy will be gradually lost.
[0098] (3) For cell i, generally speaking, the state of the cell itself has the greatest impact on the subsequent changes in its own traffic. Therefore, this cell adopts a two-level RNN in series, and then merges it with the output of the neighboring cells of levels 1 and 2, and then performs DNN filtering, in order to mine the temporal information of the cell itself as much as possible through the deep structure.
[0099] (4) For Level 1 neighbor cell L1, its impact on cell i is higher than that of Level 2 neighbor cell but lower than that of the original cell. Therefore, an RNN plus DNN cascaded structure is adopted.
[0100] (5) For Level 2 neighbor cells L2, the impact on the traffic changes of the original cell is minimal, so only Level 1 RNN structure is used to extract its time series features.
[0101] Here, RNN nodes can use LSTM or GRU networks, and DNN networks can use a structure with one hidden layer. The number of nodes or states is generally set to 64 or 96.
[0102] like Figure 5 As shown, when performing traffic prediction based on the traffic prediction model, it is necessary to collect network traffic data in real time. Considering the processing capacity of the network management system and the network optimization requirements, it is generally sufficient to collect data at the hourly level.
[0103] To determine whether to update the hierarchical neighboring cells, machine learning training and prediction methods can be used. The hierarchical division of neighboring cells only needs to be updated once a month; otherwise, time-series data transformation can be performed directly.
[0104] The time-series data transformation is the process in step S2, which involves transforming the data of a given cell and its neighboring cells and calculating the g value.
[0105] If a new neighboring segmentation is implemented, the model needs to be retrained; otherwise, traffic prediction is performed directly based on the latest available data.
[0106] A test was conducted on a specific residential community. Hourly traffic and handover data for the community and its neighboring communities were collected over four months, totaling 190,000 data points. Data preprocessing was performed according to the described steps, and the training and test sets were divided in a 4:1 ratio. The mean absolute error (MAE) was used to calculate the loss. After 200 iterations, the model's prediction results stabilized, with the normalized MAE around 0.15. The training and testing loss curves are shown below. Figure 6 As shown.
[0107] 5-hour data normalized prediction results are as follows Figure 7 As shown, the traffic trend can be matched well, demonstrating that the embodiments of this application have high prediction accuracy.
[0108] The wireless network traffic prediction device provided in the embodiments of this application is described below. The wireless network traffic prediction device described below can be referred to in correspondence with the wireless network traffic prediction method described above.
[0109] like Figure 8 As shown, this application provides a wireless network traffic prediction device, including:
[0110] The layering module 810 is used to obtain handover information between the cell and all its neighboring cells within a target time period, obtain the influence weight of each neighboring cell on the cell, and layer all the neighboring cells of the cell according to the influence weight to obtain layered neighboring cells.
[0111] The traffic calculation module 820 is used to calculate the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time.
[0112] The prediction module 830 is used to input the first traffic transfer data and the second traffic transfer data into the traffic prediction model to obtain the traffic prediction result output by the traffic prediction model; wherein, the traffic prediction model includes a first network structure and a second network structure, the first network structure is used to obtain the traffic prediction result of the cell based on the first traffic transfer data, and the second network structure is used to obtain the traffic prediction result of the hierarchical neighboring cells based on the second traffic transfer data.
[0113] In one embodiment, the layered module 810 is used for:
[0114] Obtain the first handover information from the cell to the neighboring cell within the target time period.
[0115] Obtain the second handover information from the neighboring cell to the cell within the target time period.
[0116] The influence weight of the neighboring cell on the cell is obtained based on the first handover information, the second handover information, and the distance between the cell and the neighboring cell.
[0117] In one embodiment, the layering module 810 is further configured to:
[0118] Based on the influence weights, all neighboring cells of the cell are sorted.
[0119] Based on the target ratio, all sorted neighboring regions are divided into layers to obtain layered neighboring regions.
[0120] In one embodiment, the traffic calculation module 820 is used for:
[0121] The first traffic transfer data of the cell at the target time is obtained based on the cell's traffic at the target time, the total number of users in the cell, and the number of successful handovers in the cell at the target time.
[0122] Based on the traffic of the hierarchical neighboring cells at the target time and the handover ratio of the hierarchical neighboring cells to the cell, the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time is obtained.
[0123] In one embodiment, the first network structure includes several cascaded RNN networks, and the second network structure is determined based on the influence weights of the hierarchical neighboring regions.
[0124] In one embodiment, the traffic prediction model further includes network aggregation and a fully connected network. The network aggregation is used to aggregate the outputs of the first network structure and the second network structure, and the fully connected network is used to filter the output of the network aggregation to obtain the traffic prediction result.
[0125] In one embodiment, the traffic prediction result includes traffic prediction data at several time points.
[0126] Figure 9 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 9 As shown, the electronic device may include a processor 910, a communication interface 920, a memory 930, and a communication bus 940, wherein the processor 910, the communication interface 920, and the memory 930 communicate with each other via the communication bus 940. The processor 910 can call a computer program stored in the memory 930 to execute steps of a wireless network traffic prediction method, such as including:
[0127] The handover information between the cell and all its neighboring cells within the target time period is obtained, and the influence weight of each neighboring cell on the cell is obtained. Based on the influence weight, all neighboring cells of the cell are divided into layers to obtain layered neighboring cells.
[0128] Calculate the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time.
[0129] Input the first traffic transfer data and the second traffic transfer data into the traffic prediction model to obtain the traffic prediction result output by the traffic prediction model; wherein, the traffic prediction model includes a first network structure and a second network structure, the first network structure is used to obtain the traffic prediction result of the cell based on the first traffic transfer data, and the second network structure is used to obtain the traffic prediction result of the hierarchical neighboring cells based on the second traffic transfer data.
[0130] Furthermore, the logical instructions in the aforementioned memory 930 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0131] On the other hand, embodiments of this application also provide a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can perform the steps of the wireless network traffic prediction method provided in the above embodiments, such as including:
[0132] The handover information between the cell and all its neighboring cells within the target time period is obtained, and the influence weight of each neighboring cell on the cell is obtained. Based on the influence weight, all neighboring cells of the cell are divided into layers to obtain layered neighboring cells.
[0133] Calculate the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time.
[0134] Input the first traffic transfer data and the second traffic transfer data into the traffic prediction model to obtain the traffic prediction result output by the traffic prediction model; wherein, the traffic prediction model includes a first network structure and a second network structure, the first network structure is used to obtain the traffic prediction result of the cell based on the first traffic transfer data, and the second network structure is used to obtain the traffic prediction result of the hierarchical neighboring cells based on the second traffic transfer data.
[0135] On the other hand, embodiments of this application also provide a processor-readable storage medium storing a computer program for causing a processor to perform the steps of the methods provided in the above embodiments, such as including:
[0136] The handover information between the cell and all its neighboring cells within the target time period is obtained, and the influence weight of each neighboring cell on the cell is obtained. Based on the influence weight, all neighboring cells of the cell are divided into layers to obtain layered neighboring cells.
[0137] Calculate the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time.
[0138] Input the first traffic transfer data and the second traffic transfer data into the traffic prediction model to obtain the traffic prediction result output by the traffic prediction model; wherein, the traffic prediction model includes a first network structure and a second network structure, the first network structure is used to obtain the traffic prediction result of the cell based on the first traffic transfer data, and the second network structure is used to obtain the traffic prediction result of the hierarchical neighboring cells based on the second traffic transfer data.
[0139] The processor-readable storage medium can be any available medium or data storage device that the processor can access, including but not limited to magnetic memory (e.g., floppy disk, hard disk, magnetic tape, magneto-optical disk (MO)), optical memory (e.g., CD, DVD, BD, HVD), and semiconductor memory (e.g., ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid-state drive (SSD)).
[0140] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0141] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, 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 can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0142] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.
Claims
1. A method for predicting wireless network traffic, characterized in that, include: Obtain handover information between the cell and all its neighboring cells within a target time period, obtain the influence weight of each neighboring cell on the cell, and stratify all neighboring cells of the cell according to the influence weight to obtain stratified neighboring cells; Calculate the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time; Input the first traffic transfer data and the second traffic transfer data into the traffic prediction model to obtain the traffic prediction result output by the traffic prediction model; wherein, the traffic prediction model includes a first network structure and a second network structure, the first network structure is used to obtain the traffic prediction result of the cell based on the first traffic transfer data, and the second network structure is used to obtain the traffic prediction result of the hierarchical neighboring cells based on the second traffic transfer data.
2. The wireless network traffic prediction method according to claim 1, characterized in that, The step of obtaining handover information between the cell and all its neighboring cells within a target time period, and obtaining the influence weight of each neighboring cell on the cell, includes: Obtain the first handover information from the cell to the neighboring cell within the target time period; Obtain the second handover information from the neighboring cell to the cell within the target time period; The influence weight of the neighboring cell on the cell is obtained based on the first handover information, the second handover information, and the distance between the cell and the neighboring cell.
3. The wireless network traffic prediction method according to claim 1, characterized in that, The step of stratifying all neighboring cells of the cell according to the influence weight to obtain stratified neighboring cells includes: Based on the aforementioned influence weights, all neighboring cells of the cell are sorted. Based on the target ratio, all sorted neighboring regions are divided into layers to obtain layered neighboring regions.
4. The wireless network traffic prediction method according to claim 1, characterized in that, The calculation of the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time includes: Based on the cell's traffic at the target time, the cell's total number of users, and the number of successful handovers at the target time, the first traffic transfer data of the cell at the target time is obtained. Based on the traffic of the hierarchical neighboring cells at the target time and the handover ratio of the hierarchical neighboring cells to the cell, the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time is obtained.
5. The wireless network traffic prediction method according to claim 1, characterized in that, The first network structure includes several cascaded RNN networks, and the second network structure is determined based on the influence weights of the hierarchical neighboring regions.
6. The wireless network traffic prediction method according to claim 1, characterized in that, The traffic prediction model also includes network aggregation and a fully connected network. The network aggregation is used to aggregate the outputs of the first network structure and the second network structure, and the fully connected network is used to filter the output of the network aggregation to obtain the traffic prediction result.
7. The wireless network traffic prediction method according to claim 1, characterized in that, The traffic prediction results include traffic prediction data at several points in time.
8. A wireless network traffic prediction device, characterized in that, include: The hierarchical module is used to obtain handover information between the cell and all its neighboring cells within a target time period, obtain the influence weight of each neighboring cell on the cell, and hierarchically divide all the neighboring cells of the cell according to the influence weight to obtain hierarchical neighboring cells; The traffic calculation module is used to calculate the first traffic transfer data of the cell at the target time and the second traffic transfer data of the hierarchical neighboring cells to the cell at the target time; The prediction module is used to input the first traffic transfer data and the second traffic transfer data into the traffic prediction model to obtain the traffic prediction result output by the traffic prediction model; wherein, the traffic prediction model includes a first network structure and a second network structure, the first network structure is used to obtain the traffic prediction result of the cell based on the first traffic transfer data, and the second network structure is used to obtain the traffic prediction result of the hierarchical neighboring cells based on the second traffic transfer data.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the wireless network traffic prediction method as described in any one of claims 1 to 7.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the wireless network traffic prediction method as described in any one of claims 1 to 7.