Relay network resource scheduling method and apparatus, device, medium, and program product
By constructing a relay scheduling prediction model and utilizing graph convolutional networks and neural network technology, the model identifies communication devices with degraded performance and generates optimized relay scheduling schemes. This solves the problem of low relay scheduling accuracy in existing technologies and achieves more efficient network resource allocation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA MOBILE GROUP ZHEJIANG
- Filing Date
- 2024-07-30
- Publication Date
- 2026-06-16
AI Technical Summary
In existing technologies, relay scheduling is based on the operator's subjective judgment, which has low accuracy and leads to insufficient communication capacity in hot areas and during hot periods.
By acquiring basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area, a relay scheduling prediction model is constructed. Using graph convolutional network layers, convolutional neural network layers, Transformer layers, and stage prediction layers, a relay scheduling scheme is generated, and communication devices with degraded performance are identified and scheduling is optimized.
It improves the accuracy of relay scheduling, avoids insufficient communication capacity in hot areas and during hot periods, and optimizes network operating efficiency.
Smart Images
Figure CN119052859B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of network resource scheduling technology, and in particular to a relay network resource scheduling method, apparatus, equipment, medium, and program product. Background Technology
[0002] With the rapid development of wireless network technology, the application of 5G wireless technology is becoming increasingly widespread. 5G signal relay is an important component of 5G wireless technology, helping to expand the coverage of 5G signals and enabling 5G users to obtain better network services over a wider area. However, current technologies rely on operator judgment for relay scheduling to address regional communication surges, resulting in low accuracy and potential communication capacity shortages in hotspot areas and during peak times. Summary of the Invention
[0003] This invention provides a relay network resource scheduling method, apparatus, equipment, medium, and program product to solve the defects of the prior art in which relay scheduling is based on the subjective judgment of the operator, resulting in low accuracy and easy occurrence of insufficient communication capacity in hot areas and hot periods.
[0004] In a first aspect, the present invention provides a relay network resource scheduling method, comprising:
[0005] During the data collection period, basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area are acquired to determine the future impact parameters of the target area during the prediction period.
[0006] A relay scheduling prediction model is constructed by inputting the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the relay scheduling prediction model to obtain the relay scheduling scheme within the prediction time period output by the relay scheduling prediction model.
[0007] The relay scheduling prediction model is trained based on the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters of the sample area during the sample time period, as well as the relay scheduling scheme label corresponding to the sample area.
[0008] In some embodiments, the relay scheduling prediction model includes a graph convolutional network layer, a convolutional neural network layer, a Transformer layer, and a stage prediction layer;
[0009] Correspondingly, the construction of the relay scheduling prediction model involves inputting the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the relay scheduling prediction model to obtain the relay scheduling scheme output by the relay scheduling prediction model, including:
[0010] Construct a network connection topology diagram of the target area;
[0011] The network connection topology diagram of the target area is input into the graph convolutional network layer to obtain the structural features of the communication devices in the target area output by the graph convolutional network layer.
[0012] The basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters are input into the convolutional neural network layer to obtain the time-domain and frequency-domain features corresponding to each parameter output by the convolutional neural network layer.
[0013] The time-domain features, frequency-domain features, and structural features are input into the Transformer layer to obtain the fused features output by the Transformer layer;
[0014] The fused features are dynamically clustered using a clustering algorithm to identify target communication devices with degraded performance in the target region, thereby obtaining the identification result.
[0015] Based on the identification results, the time-domain features, frequency-domain features, and structural features are filtered to obtain the target time-domain features, target frequency-domain features, and target structural features corresponding to the target communication device;
[0016] The target time-domain features, target frequency-domain features, and target structural features are input into the stage prediction layer to obtain the degradation stage of the target communication device output by the stage prediction layer.
[0017] The relay scheduling scheme is generated based on the degradation stage of the target communication device.
[0018] In some embodiments, constructing the network connection topology diagram of the target area includes:
[0019] Obtain basic information about communication devices in the target area, including device type, function, location, and connection links between communication devices;
[0020] Based on the basic information of the communication devices, a network connection topology diagram of the target area is constructed, with the communication devices as nodes and the connection links between the communication devices as edges, and the weights of the nodes and edges and the directions of the edges are determined.
[0021] In some embodiments, the step of inputting the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the convolutional neural network layer to obtain the time-domain and frequency-domain features corresponding to each parameter output by the convolutional neural network layer includes:
[0022] The time-series data of the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters are input into the convolutional neural network layer. The time-series data are then convolved using a one-dimensional convolutional layer to extract the temporal features of the time-series data.
[0023] The nonlinearity of the convolutional neural network layer is increased by using a nonlinear activation function, and the dimensionality of the temporal features is reduced by using a pooling layer.
[0024] Perform a Fast Fourier Transform on the time-series data to convert the time-series data into frequency domain data;
[0025] A one-dimensional convolutional layer is used to perform convolution operations on the frequency domain data to extract the frequency domain features of the frequency domain data, and a pooling layer is used to reduce the dimensionality of the frequency domain features.
[0026] In some embodiments, the basic communication parameters include communication demand distribution parameters, communication traffic parameters, communication traffic peak parameters, and communication traffic valley parameters; the base station power parameters include base station basic parameters, base station coverage parameters, base station uplink frequency parameters, base station downlink frequency parameters, and base station bandwidth overlap range parameters; the future influencing parameters include event factor parameters, weather factor parameters, activity factor parameters, and holiday factor parameters.
[0027] In some embodiments, acquiring the basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area within the acquisition time period includes:
[0028] The target area is divided into grids to obtain a grid map of the target area;
[0029] During the collection period, collect communication traffic parameters of the target area to determine the peak and valley communication traffic parameters.
[0030] Based on the communication traffic parameters and the target area grid map, a communication traffic grid map of the target area is obtained;
[0031] Based on the communication traffic grid, the weights of the communication traffic peak parameters and the communication traffic valley parameters are determined.
[0032] The communication fluctuation parameter is obtained based on the communication traffic peak parameter and the communication traffic valley parameter, as well as the weights of the communication traffic peak parameter and the communication traffic valley parameter.
[0033] In some embodiments, the process of determining the relay scheduling prediction model includes:
[0034] Obtain the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters of the sample area within the sample time period;
[0035] Determine the relay scheduling scheme label corresponding to the sample area;
[0036] Using the sample communication basic parameters, sample communication fluctuation parameters, sample base station power parameters, and sample future impact parameters as training samples, and using the relay scheduling scheme label corresponding to the sample area as sample label, an initial relay scheduling prediction model is trained.
[0037] The parameters of the initial relay scheduling prediction model are iteratively optimized to obtain the relay scheduling prediction model.
[0038] The parameters of the relay scheduling prediction model include the weights of basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters.
[0039] Secondly, the present invention also provides a relay network resource scheduling device, comprising:
[0040] The acquisition unit is used to acquire basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area during the acquisition time period, and to determine the future impact parameters of the target area during the prediction time period.
[0041] The prediction unit is used to construct a relay scheduling prediction model. The basic communication parameters, communication fluctuation parameters, base station power parameters and future impact parameters are input into the relay scheduling prediction model to obtain the relay scheduling scheme within the prediction time period output by the relay scheduling prediction model.
[0042] The relay scheduling prediction model is trained based on the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters of the sample area during the sample time period, as well as the relay scheduling scheme label corresponding to the sample area.
[0043] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the relay network resource scheduling method described above.
[0044] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the relay network resource scheduling method as described above.
[0045] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the relay network resource scheduling method described above.
[0046] The relay network resource scheduling method, apparatus, equipment, medium, and program products provided by this invention obtain basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area within a collection time period, determine the future impact parameters of the target area within a prediction time period, construct a relay scheduling prediction model, input each parameter into the relay scheduling prediction model, and obtain a relay scheduling scheme within the prediction time period. This improves the accuracy of relay scheduling and can avoid insufficient communication capacity in hot areas and hot periods. Attached Figure Description
[0047] To more clearly illustrate the technical solutions in this invention 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 invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0048] Figure 1 This is a flowchart illustrating the relay network resource scheduling method provided in an embodiment of the present invention;
[0049] Figure 2 This is a flowchart illustrating the determination process of the relay scheduling prediction model provided in this embodiment of the invention.
[0050] Figure 3 This is a schematic diagram of the structure of the relay network resource scheduling device provided in an embodiment of the present invention;
[0051] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation
[0052] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this 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 this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0053] Figure 1 This is a flowchart illustrating the relay network resource scheduling method provided in an embodiment of the present invention. Figure 1As shown, a relay network resource scheduling method is provided, including the following steps: step 110 and step 120. This method's steps are merely one possible implementation of the present invention.
[0054] Step 110: Acquire basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area within the collection period, and determine the future impact parameters of the target area within the prediction period.
[0055] Among them, the base station power parameter reflects the network coverage and signal strength, while the communication fluctuation parameter reflects the changes in network load.
[0056] In some embodiments, the basic communication parameters include communication demand distribution parameters, communication traffic parameters, communication traffic peak parameters, and communication traffic valley parameters; the base station power parameters include base station basic parameters, base station coverage parameters, base station uplink frequency parameters, base station downlink frequency parameters, and base station bandwidth overlap range parameters; and the future influencing parameters include event factor parameters, weather factor parameters, activity factor parameters, and holiday factor parameters.
[0057] Understandably, by acquiring basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area within the collection period, the future impact parameters of the target area within the prediction period can be determined. This provides comprehensive data support for the prediction of relay scheduling schemes, which helps the relay scheduling prediction model to more accurately predict the communication network demand and resource allocation in the future period. This can optimize the relay scheduling scheme and improve network operation efficiency.
[0058] In some embodiments, acquiring basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area during the acquisition time period includes:
[0059] The target area is divided into grids to obtain a grid map of the target area;
[0060] Collect communication traffic parameters of the target area during the collection period, and determine the peak and valley communication traffic parameters.
[0061] Based on communication traffic parameters and the target area grid map, a communication traffic grid map of the target area is obtained;
[0062] Based on the communication traffic grid diagram, determine the weights of the peak communication traffic parameter and the valley communication traffic parameter.
[0063] Based on the peak and valley parameters of communication traffic, as well as the weights of the peak and valley parameters, the communication fluctuation parameters are obtained.
[0064] Optionally, based on the communication traffic parameters, the target area grid map is marked with light and dark colors to obtain the communication traffic grid map.
[0065] Optionally, the data collection period is 30 to 45 days, and a daily communication traffic grid map of the target area is obtained based on the daily communication traffic parameters and the target area grid map.
[0066] Optionally, event factor parameters can be obtained based on special events within the prediction period; weather factor parameters can be obtained based on daily weather conditions within the prediction period; activity factor parameters can be obtained based on the expected number of participants within the prediction period; and holiday factor parameters can be obtained based on holiday communication fluctuation parameters within the prediction period.
[0067] Optionally, the event factor parameters, weather factor parameters, activity factor parameters, and holiday factor parameters can be integrated to construct a numerical matrix and obtain future impact parameters.
[0068] Step 120: Construct a relay scheduling prediction model. Input basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the relay scheduling prediction model to obtain the relay scheduling scheme within the prediction time period output by the relay scheduling prediction model.
[0069] The relay scheduling prediction model is trained based on the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters of the sample area during the sample time period, as well as the relay scheduling scheme label corresponding to the sample area.
[0070] It is understandable that by constructing a relay scheduling prediction model, and then using the relay scheduling prediction model to make predictions based on basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters, a relay scheduling scheme can be obtained, thereby improving the accuracy and efficiency of the scheduling scheme.
[0071] In some embodiments, the relay scheduling prediction model includes a graph convolutional network layer, a convolutional neural network layer, a Transformer layer, and a stage prediction layer;
[0072] Correspondingly, step 120 constructs a relay scheduling prediction model by inputting basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the relay scheduling prediction model to obtain the relay scheduling scheme output by the relay scheduling prediction model, including:
[0073] Step 121: Construct a network connection topology diagram for the target area;
[0074] Step 122: Input the network connection topology diagram of the target area into the graph convolutional network layer to obtain the structural features of the communication devices in the target area output by the graph convolutional network layer;
[0075] Step 123: Input the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the convolutional neural network layer to obtain the time-domain and frequency-domain features corresponding to each parameter output by the convolutional neural network layer;
[0076] Step 124: Input the time domain features, frequency domain features, and structural features into the Transformer layer to obtain the fused features output by the Transformer layer;
[0077] Step 125: Dynamically cluster the fused features using a clustering algorithm to identify target communication devices with degraded performance in the target area and obtain the identification results;
[0078] Step 126: Based on the recognition results, the time domain features, frequency domain features, and structural features are filtered to obtain the target time domain features, target frequency domain features, and target structural features corresponding to the target communication device;
[0079] Step 127: Input the target time domain features, target frequency domain features and target structural features into the stage prediction layer to obtain the degradation stage of the target communication device output by the stage prediction layer;
[0080] Step 128: Generate a relay scheduling scheme based on the degradation stage of the target communication device.
[0081] The Transformer layer consists of an encoder layer, a decoder layer, and a multi-head self-attention layer. The Transformer layer learns the dependencies between feature vectors through a self-attention mechanism.
[0082] Among them, the degradation stage of the target communication device indicates the severity of the performance degradation of the target communication device, which can be determined based on a preset threshold.
[0083] Optionally, a multi-dimensional feature parameter set {Par} is constructed based on the target's time-domain features, frequency-domain features, and structural features:
[0084] {Par}={{T a (t)},{F b (t)},{S c (t)}};
[0085] {T a (t)}={T1(t),T2(t),…,T A (t)};
[0086] {F b F(t)}={F1(t),F2(t),…,F B (t)};
[0087] {S c F(t)}={F1(t),F2(t),…,F B (t)};
[0088] Among them, {T a (t)}、{F b (t)} and {S c (t)} represents the time-domain subset, frequency-domain subset, and topological subset of the multidimensional feature parameter set {Par} at time t, respectively; a is the time-domain parameter index, A is the time-domain parameter dimension, b is the frequency-domain parameter index, B is the frequency-domain parameter dimension, c is the topological parameter index, and C is the topological parameter dimension.
[0089] Optionally, the target time-domain feature vector is constructed based on the time-domain subset, the frequency-domain subset, and the topological structure subset, respectively. Target frequency domain feature vector and target structure feature vector as follows:
[0090]
[0091] in, This represents the time-domain eigenvector at time t. This represents the frequency domain eigenvector at time t. This represents the structural eigenvector at time t.
[0092] Optionally, the target time-domain feature vector Target frequency domain feature vector and target structure feature vector The input is fed into the stage prediction layer to obtain the degradation stage of the target communication device output by the stage prediction layer.
[0093] Optionally, based on the degradation stage of the target communication equipment, the performance of each communication equipment in the target area is classified into different levels. For example, according to the performance from high to low, they can be divided into level 1, level 2, and level 3. Based on the performance level of the communication equipment, a relay scheduling scheme is generated, and communication equipment with high performance level is scheduled first.
[0094] Understandably, by extracting features from basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters, time-domain and frequency-domain features are obtained. By extracting features from the network connection topology diagram of the target area, the structural features of the communication equipment in the target area are obtained. The time-domain, frequency-domain, and structural features are fused to obtain fused features. Based on the fused features, target communication equipment with degraded performance is initially identified. Based on the identification results, the time-domain, frequency-domain, and structural features are filtered to obtain the target time-domain, target frequency-domain, and target structural features corresponding to the target communication equipment. Furthermore, based on the target time-domain, target frequency-domain, and target structural features, the degree of performance degradation of the target equipment can be predicted. Thus, based on the prediction results, an optimized relay scheduling scheme can be generated to reasonably schedule the relay communication equipment.
[0095] In this embodiment of the invention, by acquiring the basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area within the collection time period, the future impact parameters of the target area within the prediction time period are determined, a relay scheduling prediction model is constructed, and each parameter is input into the relay scheduling prediction model to obtain the relay scheduling scheme within the prediction time period. This improves the accuracy of relay scheduling and can avoid insufficient communication capacity in hotspot areas and hotspot periods.
[0096] In some embodiments, basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters are input into a convolutional neural network layer to obtain the time-domain and frequency-domain features corresponding to each parameter output by the convolutional neural network layer, including:
[0097] The time-series data of basic communication parameters, communication fluctuation parameters, base station power parameters and future impact parameters are input into the convolutional neural network layer. The one-dimensional convolutional layer is used to perform convolution operations on the time-series data to extract the temporal features of the time-series data.
[0098] Nonlinear activation functions are used to increase the nonlinearity of convolutional neural network layers, and pooling layers are used to reduce the dimensionality of temporal features;
[0099] Perform a Fast Fourier Transform on the time-series data to convert it into frequency domain data;
[0100] One-dimensional convolutional layers are used to perform convolution operations on frequency domain data to extract frequency domain features, and pooling layers are used to reduce the dimensionality of frequency domain features.
[0101] The convolutional neural network layer includes a one-dimensional convolutional layer and a pooling layer.
[0102] It is understandable that using one-dimensional convolutional layers to perform convolution operations on time-series data can effectively capture local patterns and trends in the data. Using non-linear activation functions to increase the non-linearity of convolutional neural network layers enables them to learn more complex time-domain features. By using pooling layers to reduce the dimensionality of time-domain and frequency-domain features, the most significant feature information can be retained, the computational complexity can be reduced, and overfitting can be prevented.
[0103] Figure 2 This is a flowchart illustrating the process of determining the relay scheduling prediction model provided in an embodiment of the present invention. In some embodiments, the process of determining the relay scheduling prediction model includes:
[0104] Step 210: Obtain the basic parameters of sample communication, the fluctuation parameters of sample communication, the power parameters of sample base stations, and the parameters of sample future impact for the sample area during the sample time period;
[0105] Step 220: Determine the relay scheduling scheme label corresponding to the sample area;
[0106] Step 230: Using the basic parameters of sample communication, the fluctuation parameters of sample communication, the power parameters of sample base stations, and the future impact parameters of samples as training samples, and using the relay scheduling scheme label corresponding to the sample area as the sample label, train the initial relay scheduling prediction model.
[0107] Step 240: Iteratively optimize the parameters of the initial relay scheduling prediction model to obtain the relay scheduling prediction model;
[0108] The parameters of the relay scheduling prediction model include the weights of basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters.
[0109] Optionally, the basic parameters of sample communication, the fluctuation parameters of sample communication, the power parameters of sample base stations, and the future impact parameters of samples are input into the initial relay scheduling prediction model to obtain the relay scheduling prediction scheme output by the initial relay scheduling prediction model. The loss function value is calculated based on the relay scheduling prediction scheme and the relay scheduling scheme label. Based on the loss function value, the parameters of the initial relay scheduling prediction model are iteratively optimized.
[0110] Optionally, the expression for the relay scheduling prediction model is as follows:
[0111]
[0112] Among them, C max C represents the optimal relay scheduling prediction model. 0i Indicates basic communication parameters, a i F represents the weight of the basic communication parameters. 0nB represents the communication fluctuation parameter, d represents the weight of the communication fluctuation parameter, and B represents the weight of the communication fluctuation parameter. si Indicates base station power parameters, b i E represents the weight of the base station power parameter. 0i c represents the parameter that will influence future events. i This represents the weights that will influence the parameters in the future, where i represents the parameter dimension and n is a positive integer greater than or equal to 1.
[0113] The relay network resource scheduling device provided in the embodiments of the present invention is described below. The relay network resource scheduling device described below can be referred to in correspondence with the relay network resource scheduling method described above.
[0114] Figure 3 This is a schematic diagram of the structure of the relay network resource scheduling device provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the relay network resource scheduling device 300 includes:
[0115] The acquisition unit 310 is used to acquire basic communication parameters, communication fluctuation parameters and base station power parameters of the target area during the acquisition time period, and determine the future impact parameters of the target area during the prediction time period.
[0116] The prediction unit 320 is used to construct a relay scheduling prediction model. It inputs basic communication parameters, communication fluctuation parameters, base station power parameters and future impact parameters into the relay scheduling prediction model to obtain the relay scheduling scheme within the prediction time period output by the relay scheduling prediction model.
[0117] The relay scheduling prediction model is trained based on the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters of the sample area during the sample time period, as well as the relay scheduling scheme label corresponding to the sample area.
[0118] Optionally, the relay scheduling prediction model includes graph convolutional network layers, convolutional neural network layers, Transformer layers, and stage prediction layers;
[0119] Correspondingly, a relay scheduling prediction model is constructed. Basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters are input into the relay scheduling prediction model to obtain the relay scheduling scheme output by the model, including:
[0120] Construct a network connection topology diagram for the target area;
[0121] The network connection topology diagram of the target region is input into the graph convolutional network layer to obtain the structural features of the communication devices in the target region output by the graph convolutional network layer.
[0122] The basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters are input into the convolutional neural network layer to obtain the time-domain and frequency-domain features corresponding to each parameter output by the convolutional neural network layer.
[0123] Temporal features, frequency features, and structural features are input into the Transformer layer to obtain the fused features output by the Transformer layer;
[0124] By dynamically clustering the fused features using a clustering algorithm, the target communication devices with degraded performance in the target area are identified, and the identification results are obtained.
[0125] Based on the identification results, time-domain features, frequency-domain features, and structural features are filtered to obtain the target time-domain features, target frequency-domain features, and target structural features corresponding to the target communication device;
[0126] The target time-domain features, target frequency-domain features, and target structural features are input into the stage prediction layer to obtain the degradation stage of the target communication device output by the stage prediction layer.
[0127] A relay scheduling scheme is generated based on the degradation stage of the target communication device.
[0128] Optionally, a network connection topology diagram of the target area is constructed, including:
[0129] Obtain basic information about communication devices in the target area. This information includes device type, function, location, and connection links between communication devices.
[0130] Based on the basic information of the communication equipment, a network connection topology diagram of the target area is constructed, with the communication equipment as nodes and the connection links between the communication equipment as edges, and the weights of the nodes and edges and the direction of the edges are determined.
[0131] Optionally, basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters are input into a convolutional neural network layer to obtain the time-domain and frequency-domain features corresponding to each parameter output by the convolutional neural network layer, including:
[0132] The time-series data of basic communication parameters, communication fluctuation parameters, base station power parameters and future impact parameters are input into the convolutional neural network layer. The one-dimensional convolutional layer is used to perform convolution operations on the time-series data to extract the temporal features of the time-series data.
[0133] Nonlinear activation functions are used to increase the nonlinearity of convolutional neural network layers, and pooling layers are used to reduce the dimensionality of temporal features;
[0134] Perform a Fast Fourier Transform on the time-series data to convert it into frequency domain data;
[0135] One-dimensional convolutional layers are used to perform convolution operations on frequency domain data to extract frequency domain features, and pooling layers are used to reduce the dimensionality of frequency domain features.
[0136] Optionally, the basic communication parameters include communication demand distribution parameters, communication traffic parameters, communication traffic peak parameters, and communication traffic valley parameters; the base station power parameters include base station basic parameters, base station coverage parameters, base station uplink frequency parameters, base station downlink frequency parameters, and base station bandwidth overlap range parameters; and the future influencing parameters include event factor parameters, weather factor parameters, activity factor parameters, and holiday factor parameters.
[0137] Optionally, during the data collection period, basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area are acquired, including:
[0138] The target area is divided into grids to obtain a grid map of the target area;
[0139] Collect communication traffic parameters of the target area during the collection period, and determine the peak and valley communication traffic parameters.
[0140] Based on communication traffic parameters and the target area grid map, a communication traffic grid map of the target area is obtained;
[0141] Based on the communication traffic grid diagram, determine the weights of the peak communication traffic parameter and the valley communication traffic parameter.
[0142] Based on the peak and valley parameters of communication traffic, as well as the weights of the peak and valley parameters, the communication fluctuation parameters are obtained.
[0143] Optionally, the process of determining the relay scheduling prediction model includes:
[0144] Obtain the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters of the sample area within the sample time period;
[0145] Determine the relay scheduling scheme label corresponding to the sample area;
[0146] The initial relay scheduling prediction model is trained using the basic parameters of sample communication, the fluctuation parameters of sample communication, the power parameters of sample base stations, and the future impact parameters of samples, and the relay scheduling scheme label corresponding to the sample area as the sample label.
[0147] The parameters of the initial relay scheduling prediction model are iteratively optimized to obtain the relay scheduling prediction model.
[0148] The parameters of the relay scheduling prediction model include the weights of basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters.
[0149] It should be noted that the relay network resource scheduling device provided in this embodiment of the invention can implement all the method steps implemented in the above-described relay network resource scheduling method embodiment, and can achieve the same technical effect. Here, the parts that are the same as those in the method embodiment and the beneficial effects will not be described in detail.
[0150] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, such as... Figure 4 As shown, the electronic device may include a processor 410, a communications interface 420, a memory 430, and a communication bus 440. The processor 410, communications interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute a relay network resource scheduling method. This method includes: acquiring basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area within a collection time period; determining the future impact parameters of the target area within a prediction time period; constructing a relay scheduling prediction model; inputting the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the relay scheduling prediction model; and obtaining the relay scheduling scheme for the prediction time period output by the relay scheduling prediction model. The relay scheduling prediction model is trained based on sample basic communication parameters, sample communication fluctuation parameters, sample base station power parameters, and sample future impact parameters of the sample area within a sample time period, as well as the relay scheduling scheme label corresponding to the sample area.
[0151] Furthermore, the logical instructions in the aforementioned memory 430 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 the present invention, essentially, 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 the present invention. 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.
[0152] On the other hand, the present invention also provides 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 execute the relay network resource scheduling method provided by the above methods. The method includes: acquiring basic communication parameters, communication fluctuation parameters, and base station power parameters of a target area during a collection period, and determining the future impact parameters of the target area during a prediction period; constructing a relay scheduling prediction model, inputting the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the relay scheduling prediction model, and obtaining the relay scheduling scheme for the prediction period output by the relay scheduling prediction model; wherein, the relay scheduling prediction model is trained based on sample basic communication parameters, sample communication fluctuation parameters, sample base station power parameters, and sample future impact parameters of a sample area during a sample period, as well as the relay scheduling scheme label corresponding to the sample area.
[0153] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program is implemented to perform the relay network resource scheduling method provided by the above methods. The method includes: acquiring basic communication parameters, communication fluctuation parameters, and base station power parameters of a target area during a collection period, and determining the future impact parameters of the target area during a prediction period; constructing a relay scheduling prediction model, inputting the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the relay scheduling prediction model, and obtaining the relay scheduling scheme for the prediction period output by the relay scheduling prediction model; wherein the relay scheduling prediction model is trained based on sample basic communication parameters, sample communication fluctuation parameters, sample base station power parameters, and sample future impact parameters of a sample area during a sample period, as well as the relay scheduling scheme label corresponding to the sample area.
[0154] 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.
[0155] 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.
[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention 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; and these 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 the present invention.
Claims
1. A relay network resource scheduling method, characterized in that, include: During the data collection period, basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area are acquired to determine the future impact parameters of the target area during the prediction period. A relay scheduling prediction model is constructed by inputting the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the relay scheduling prediction model to obtain the relay scheduling scheme within the prediction time period output by the relay scheduling prediction model. The relay scheduling prediction model is trained based on the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters of the sample area during the sample time period, as well as the relay scheduling scheme label corresponding to the sample area. The relay scheduling prediction model includes a graph convolutional network layer, a convolutional neural network layer, a Transformer layer, and a stage prediction layer; Correspondingly, the construction of the relay scheduling prediction model involves inputting the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the relay scheduling prediction model to obtain the relay scheduling scheme output by the relay scheduling prediction model, including: Construct a network connection topology diagram of the target area; The network connection topology diagram of the target area is input into the graph convolutional network layer to obtain the structural features of the communication devices in the target area output by the graph convolutional network layer. The basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters are input into the convolutional neural network layer to obtain the time-domain and frequency-domain features corresponding to each parameter output by the convolutional neural network layer. The time-domain features, frequency-domain features, and structural features are input into the Transformer layer to obtain the fused features output by the Transformer layer; The fused features are dynamically clustered using a clustering algorithm to identify target communication devices with degraded performance in the target region, thereby obtaining the identification result. Based on the identification results, the time-domain features, frequency-domain features, and structural features are filtered to obtain the target time-domain features, target frequency-domain features, and target structural features corresponding to the target communication device; The target time-domain features, target frequency-domain features, and target structural features are input into the stage prediction layer to obtain the degradation stage of the target communication device output by the stage prediction layer. The relay scheduling scheme is generated based on the degradation stage of the target communication device.
2. The relay network resource scheduling method according to claim 1, characterized in that, The construction of the network connection topology diagram of the target area includes: Obtain basic information about communication devices in the target area, including device type, function, location, and connection links between communication devices; Based on the basic information of the communication devices, a network connection topology diagram of the target area is constructed, with the communication devices as nodes and the connection links between the communication devices as edges, and the weights of the nodes and edges and the directions of the edges are determined.
3. The relay network resource scheduling method according to claim 1, characterized in that, The step of inputting the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the convolutional neural network layer to obtain the time-domain and frequency-domain features corresponding to each parameter output by the convolutional neural network layer includes: The time-series data of the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters are input into the convolutional neural network layer. The time-series data are then convolved using a one-dimensional convolutional layer to extract the temporal features of the time-series data. The nonlinearity of the convolutional neural network layer is increased by using a nonlinear activation function, and the dimensionality of the temporal features is reduced by using a pooling layer. Perform a Fast Fourier Transform on the time-series data to convert the time-series data into frequency domain data; A one-dimensional convolutional layer is used to perform convolution operations on the frequency domain data to extract the frequency domain features of the frequency domain data, and a pooling layer is used to reduce the dimensionality of the frequency domain features.
4. The relay network resource scheduling method according to claim 1, characterized in that, The basic communication parameters include communication demand distribution parameters, communication traffic parameters, communication traffic peak parameters, and communication traffic valley parameters; the base station power parameters include base station basic parameters, base station coverage parameters, base station uplink frequency parameters, base station downlink frequency parameters, and base station bandwidth overlap range parameters; the future influencing parameters include event factor parameters, weather factor parameters, activity factor parameters, and holiday factor parameters.
5. The relay network resource scheduling method according to claim 4, characterized in that, The acquisition of basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area during the data collection period includes: The target area is divided into grids to obtain a grid map of the target area; During the collection period, collect communication traffic parameters of the target area to determine the peak and valley communication traffic parameters. Based on the communication traffic parameters and the target area grid map, a communication traffic grid map of the target area is obtained; Based on the communication traffic grid, the weights of the communication traffic peak parameters and the communication traffic valley parameters are determined. The communication fluctuation parameter is obtained based on the communication traffic peak parameter and the communication traffic valley parameter, as well as the weights of the communication traffic peak parameter and the communication traffic valley parameter.
6. The relay network resource scheduling method according to claim 1, characterized in that, The process of determining the relay scheduling prediction model includes: Obtain the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters of the sample area within the sample time period; Determine the relay scheduling scheme label corresponding to the sample area; Using the sample communication basic parameters, sample communication fluctuation parameters, sample base station power parameters, and sample future impact parameters as training samples, and using the relay scheduling scheme label corresponding to the sample area as sample label, an initial relay scheduling prediction model is trained. The parameters of the initial relay scheduling prediction model are iteratively optimized to obtain the relay scheduling prediction model. The parameters of the relay scheduling prediction model include the weights of basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters.
7. A relay network resource scheduling device, characterized in that, include: The acquisition unit is used to acquire basic communication parameters, communication fluctuation parameters, and base station power parameters of the target area during the acquisition time period, and to determine the future impact parameters of the target area during the prediction time period. The prediction unit is used to construct a relay scheduling prediction model. The basic communication parameters, communication fluctuation parameters, base station power parameters and future impact parameters are input into the relay scheduling prediction model to obtain the relay scheduling scheme within the prediction time period output by the relay scheduling prediction model. The relay scheduling prediction model is trained based on the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters of the sample area during the sample time period, as well as the relay scheduling scheme label corresponding to the sample area. The relay scheduling prediction model includes a graph convolutional network layer, a convolutional neural network layer, a Transformer layer, and a stage prediction layer; Correspondingly, the construction of the relay scheduling prediction model involves inputting the basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters into the relay scheduling prediction model to obtain the relay scheduling scheme output by the relay scheduling prediction model, including: Construct a network connection topology diagram of the target area; The network connection topology diagram of the target area is input into the graph convolutional network layer to obtain the structural features of the communication devices in the target area output by the graph convolutional network layer. The basic communication parameters, communication fluctuation parameters, base station power parameters, and future impact parameters are input into the convolutional neural network layer to obtain the time-domain and frequency-domain features corresponding to each parameter output by the convolutional neural network layer. The time-domain features, frequency-domain features, and structural features are input into the Transformer layer to obtain the fused features output by the Transformer layer; The fused features are dynamically clustered using a clustering algorithm to identify target communication devices with degraded performance in the target region, thereby obtaining the identification result. Based on the identification results, the time-domain features, frequency-domain features, and structural features are filtered to obtain the target time-domain features, target frequency-domain features, and target structural features corresponding to the target communication device; The target time-domain features, target frequency-domain features, and target structural features are input into the stage prediction layer to obtain the degradation stage of the target communication device output by the stage prediction layer. The relay scheduling scheme is generated based on the degradation stage of the target communication device.
8. 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 relay network resource scheduling method as described in any one of claims 1 to 6.
9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the relay network resource scheduling method as described in any one of claims 1 to 6.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the relay network resource scheduling method as described in any one of claims 1 to 6.