Network scheduling method, network equipment and computer readable storage medium

Abstract

The invention provides a network scheduling method, network equipment and a computer readable storage medium. The network scheduling method comprises the following steps: acquiring current terminal feature information of user equipment (UE) in a target network, and obtaining a predicted user service demand based on the current terminal feature information and a preset user portrait; determining a base station energy consumption prediction set according to a preset base station energy consumption portrait and predicted user service requirements; and performing energy-saving scheduling processing on the base stations in the target network according to the base station energy consumption prediction set. According to the scheme of the embodiment of the invention, the user perception and the energy consumption of the base station can be coordinated, the energy consumption of the base station is reduced under the condition of ensuring the user perception, and the purposes of energy conservation and emission reduction are achieved.

Description

Technical Field

[0001] The embodiments of the present application relate to but are not limited to the field of communication technology, and in particular, to a network scheduling method, a network device, and a computer-readable storage medium. Background Art

[0002] With the continuous development of wireless network technology, the transmission rate of wireless networks is constantly improving, the transmission delay is constantly decreasing, the transmission bandwidth is also constantly increasing, and the types of services that can be carried are also constantly enriched. As a result, users' requirements for service perception are increasing. In the existing technology, in order to allow users to have a better user experience, only attention is paid to whether the user's perception requirements are met, and the energy consumption problem of the network is often ignored, resulting in high energy consumption of the existing network, which cannot achieve the purpose of energy saving and emission reduction. Summary of the invention

[0003] The following is a summary of the subject matter described in detail herein. This summary is not intended to limit the scope of the claims.

[0004] The embodiments of the present application provide a network scheduling method, a network device, and a computer-readable storage medium, which can coordinate user perception with base station energy consumption, reduce base station energy consumption while ensuring user perception, and achieve the purpose of energy conservation and emission reduction.

[0005] In a first aspect, an embodiment of the present application provides a network scheduling method, the method comprising:

[0006] Acquire current terminal feature information of a user terminal UE in the target network, and obtain predicted user service needs based on the current terminal feature information and a preset user portrait;

[0007] Determine a base station energy consumption prediction set according to a preset base station energy consumption profile and the predicted user service demand;

[0008] Energy-saving scheduling is performed on base stations in the target network according to the base station energy consumption prediction set.

[0009] In a second aspect, an embodiment of the present application further provides a network device, including:

[0010] at least one processor;

[0011] at least one memory for storing at least one program;

[0012] When at least one of the programs is executed by at least one of the processors, the network scheduling method as described above is implemented.

[0013] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are used to execute the network scheduling method as described above.

[0014] The embodiment of the present application includes: obtaining the current terminal feature information of the user terminal in the target network, and determining the predicted user service demand based on the current terminal feature information and the preset user portrait; then determining the base station energy consumption prediction set based on the preset base station energy consumption portrait and the predicted user service demand; finally, performing energy-saving scheduling processing on the base station in the target network based on the base station energy consumption prediction set. According to the technical solution provided in the embodiment of the present application, user perception and base station energy consumption are coordinated to reduce the base station energy consumption while ensuring user perception, thereby achieving the purpose of energy saving and emission reduction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings are used to provide further understanding of the technical solution of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the technical solution of the present application and do not constitute a limitation on the technical solution of the present application.

[0016] Figure 1 is a schematic diagram of a system architecture for executing a network method provided by an embodiment of the present application;

[0017] Figure 2 is a flow chart of a network scheduling method provided by an embodiment of the present application;

[0018] Figure 3 is a flowchart of determining a base station energy consumption prediction set provided by an embodiment of the present application;

[0019] Figure 4 This is a flowchart of building a user portrait provided by an embodiment of the present application;

[0020] Figure 5 It is a parameter diagram of a user portrait provided by an embodiment of the present application;

[0021] Figure 6 This is a flowchart of constructing a network-level logical grid image provided by an embodiment of the present application;

[0022] Figure 7 It is a parameter diagram of a network-level logical grid image provided by an embodiment of the present application;

[0023] Figure 8 is a flowchart of determining a cell-level logical grid provided by an embodiment of the present application;

[0024] Fig. 9 is a flowchart of constructing an initial logic grid provided by an embodiment of the present application;

[0025] Fig.10 is a schematic diagram of the construction of an initial logic grid provided by an embodiment of the present application;

[0026] Fig.11 is a schematic diagram of the construction of an initial logic grid provided by another embodiment of the present application;

[0027] Fig.12 is a specific flow chart for determining a cell-level logical grid provided by an embodiment of the present application;

[0028] Fig.13 This is a flowchart of constructing a cell-level grid image provided by an embodiment of the present application;

[0029] Fig.14 This is a schematic diagram of a specific process of constructing a cell-level grid provided by an embodiment of the present application;

[0030] Fig.15 This is a flow chart of constructing a base station energy consumption profile provided by an embodiment of the present application;

[0031] Fig.16 It is a parameter diagram of a base station energy consumption profile provided by an embodiment of the present application;

[0032] Fig.17 is a flowchart of determining a base station energy consumption prediction set provided by an embodiment of the present application;

[0033] Fig.18 It is a flow chart of constructing a segmented energy consumption model of a base station provided by an embodiment of the present application;

[0034] Fig.19 This is a schematic diagram of base station energy consumption reasoning provided by an embodiment of the present application;

[0035] Fig. 20 This is a schematic diagram of building a base station energy consumption model provided by an embodiment of the present application;

[0036] Fig.21 This is a flowchart of determining a set of cells in which a user can reside, provided by an embodiment of the present application;

[0037] Fig. 22 This is a flowchart of performing segmented prediction on a set of cells where a user can reside, provided by an embodiment of the present application;

[0038] Fig.23 is a flowchart of energy-saving scheduling processing for base stations in a target network provided by an embodiment of the present application;

[0039] Fig.24 This is a specific scheduling diagram of a network scheduling method provided by an embodiment of the present application;

[0040] Fig.25 It is a schematic diagram of the structure of a network device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0042] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0043] The present application provides a network scheduling method, a network device, and a computer-readable storage medium, which obtains the current terminal feature information of the user terminal in the target network, and determines the predicted user service demand based on the current terminal feature information and the preset user portrait; then determines the base station energy consumption prediction set based on the preset base station energy consumption portrait and the predicted user service demand; finally, performs energy-saving scheduling processing on the base station in the target network based on the base station energy consumption prediction set. According to the technical solution provided in the embodiment of the present application, user perception and base station energy consumption are coordinated, and the base station energy consumption is reduced while ensuring user perception, thereby achieving the purpose of energy saving and emission reduction.

[0044] The embodiments of the present application are further described below in conjunction with the accompanying drawings.

[0045] like Figure 1 As shown, Figure 1 Schematic diagram of a system architecture for executing a network scheduling method provided by an embodiment of the present application. Figure 1 In the example, the device includes at least one base station (Base Stat ion, BS) and at least one user terminal (User Equ ipment, UE), and the two can be connected to data through a wireless connection; one base station can carry multiple user terminals, and multiple user terminals can communicate and obtain related services through the same base station.

[0046] In some embodiments of the present application, the terminal may be a mobile phone, a tablet computer, an Internet of Things device, etc., and the above devices may be connected to the base station via a wireless connection.

[0047] The system architecture and application scenarios described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of the system architecture and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are equally applicable to similar technical problems.

[0048] It can be understood by those skilled in the art that Figure 1 The system architecture shown in the figure does not constitute a limitation on the embodiments of the present application, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0049] Based on the structure of the above system architecture, various embodiments of the scheduling method of the present application are proposed.

[0050] like Figure 2 As shown, Figure 2 1 is a flow chart of a network scheduling method provided by an embodiment of the present application. The method includes but is not limited to step S100, step S200 and step S300.

[0051] Step S100, obtaining current terminal feature information of a user terminal UE in the target network, and obtaining predicted user service requirements based on the current terminal feature information and a preset user portrait;

[0052] Step S200, determining a base station energy consumption prediction set according to a preset base station energy consumption profile and predicted user service requirements;

[0053] Step S300: performing energy-saving scheduling processing on base stations in the target network according to the base station energy consumption prediction set.

[0054] In an embodiment of the present application, the current terminal feature information of the user terminal in the target network is obtained, and the predicted user service demand is determined based on the current terminal feature information and the preset user portrait; then the base station energy consumption prediction set is determined based on the preset base station energy consumption portrait and the predicted user service demand; finally, the base station in the target network is energy-efficiently scheduled based on the base station energy consumption prediction set. According to the technical solution provided in the embodiment of the present application, user perception is coordinated with base station energy consumption, and base station energy consumption is reduced while ensuring user perception, thereby achieving the purpose of energy saving and emission reduction.

[0055] It is worth noting that the target network may be a network that needs to be scheduled, and the target network may be composed of multiple base stations, each base station may include multiple cells, and each cell may include multiple user terminals; the current terminal feature information of the user terminals in the target network is collected and processed, wherein the current terminal feature information may be understood as the real-time terminal status information of the user terminals, and the terminal status information may include real-time user attribute information, real-time user service attribute information, and real-time user behavior attribute information.

[0056] It can be understood that the user portrait can be a database model obtained based on deep learning of historical user terminal feature information. The model has the function of inference and prediction. Therefore, the current terminal feature information obtained can be matched and inferred with the user portrait, and the prediction result of the user's business needs can be obtained, which is the predicted user business needs in the embodiment of the present application; exemplarily, the current terminal feature information obtained includes moving speed information, and then the user's moving speed information and user portrait can be combined to predict the cell location information where the user terminal is located at the next moment; or, within a fixed time period every day, a user terminal will play a video. Therefore, it can be inferred based on the combination of the user portrait and the collected current terminal feature information to obtain the user's need to play the video and maintain a large bandwidth business requirement.

[0057] It is understandable that the base station energy consumption profile can be a database model obtained based on deep learning of historical base station energy consumption data. The model also has the function of inference and prediction. According to the base station energy consumption profile and predicted user business needs, the base station energy consumption prediction set can be determined. Finally, according to the base station energy consumption prediction set, energy-saving scheduling can be implemented for the base stations in the target network. Among them, in the process of implementing energy-saving scheduling for the base stations in the target network, it is ensured that all user terminals in the target network meet their respective business needs, and the energy consumption of each base station in the target network can be minimized, thereby achieving the purpose of energy saving and emission reduction.

[0058] like Figure 3 As shown, the above step S200 may include but is not limited to step S210 and step S220.

[0059] Step S210, obtaining a set of cells where the user can reside based on the predicted user service demand and a preset network-level logical grid profile;

[0060] Step S220, inferring the user's possible camping cell set based on the base station energy consumption profile to obtain a base station energy consumption prediction set.

[0061] In an embodiment of the present application, in the process of determining a base station energy consumption prediction set, first, a set of user-available cell sets is obtained based on predicted user service needs and a preset network-level logical grid portrait; then, the energy consumption of the user-available cell set is inferred by combining the service volume prediction of the user-available cell set and the base station energy consumption portrait to obtain a base station energy consumption prediction set of the base stations where the user-available cell sets are located.

[0062] It can be understood that the network-level logical grid portrait can be a database model obtained based on deep learning of historical network-level logical grid data, and the model has the function of inference and prediction; the user-residence cell set can be a set of cells where the user terminal may reside when the business needs are met; the base station energy consumption prediction set can be obtained by inferring the user-residence cell set according to the base station energy consumption portrait, wherein the base station energy consumption prediction set is to perform base station energy consumption prediction processing on each user-residence cell information in the user-residence cell set while meeting the business needs of the user terminal, so as to obtain multiple base station energy consumption prediction information, and the combination of multiple base station energy consumption prediction information is the base station energy consumption prediction set.

[0063] like Figure 4 As shown, the user portrait can be obtained through but not limited to step S110, step S120, step S130, step S140 and step S150.

[0064] Step S110, obtaining user attribute information, user service attribute information and user behavior attribute information of the UE;

[0065] Step S120, performing data cleaning and classification on the user attribute information, the user service attribute information and the user behavior attribute information to obtain a first training set and a first test set;

[0066] Step S130, adjusting the model parameters of the preset first portrait model based on the first training set;

[0067] Step S140, verifying the adjusted first portrait model based on the first test set;

[0068] Step S150: If the verification passes, the first portrait model is determined as the user portrait.

[0069] In an embodiment of the present application, in the process of constructing a user portrait, first, the user attribute information, user business attribute information and user behavior attribute information of the user terminal are obtained; then, the user attribute information, user business attribute information and user behavior attribute information are cleansed and classified to obtain a first training set and a first test set; then, based on the first training set, the model parameters of the preset first portrait model are adjusted, and then, based on the first test set, the adjusted first portrait model is verified; if the verification passes, the first portrait model can be determined as the user portrait.

[0070] It is worth noting that by mining and deep learning the user attribute information, user business attribute information and user behavior attribute information of the user terminal, it is possible to build a user portrait, manage the characteristics of the user terminal, and facilitate refined management of the user terminal.

[0071] For example, Figure 5 As shown, the user portrait may include but is not limited to contract information, service cell information, user service attributes, user location information and user mobility attributes; wherein the contract information may include user level, user number and contracted service, the service cell information may include wireless standard and wireless bandwidth, the user service attributes may include service type, service rate, service duration and service quality, the user location information may include geographic location, cell and grid, and the user mobility attributes may include moving direction and moving speed.

[0072] For example, in the process of building a user portrait, through the accumulation and learning of data on the user's own attributes, user business attributes, and user behavior attributes; the whole process can be as follows: first, data collection is carried out, mainly the collection and preliminary preprocessing of definition data; then data sorting is carried out, mainly including the elimination of abnormal data and data with extremely low probability. Clustering is then performed to calculate the converged center position data set The legal image dataset U in the grid satisfies the distance to the center location dataset is (μ U , σ U ), where refers to the “center of mass” of the data set, (μ U , σ U ) represent the mean and variance of the distance from each group of data to the centroid, and the distance here can be calculated using the second-order moment; then perform baseline data verification, The error analysis is performed with the subsequently collected image data U to satisfy P(U|μ U ,σ U )≥η U (η Uis a pre-defined probability threshold), the baseline profile data is constructed successfully; the user profile data also needs to be updated periodically to avoid changes in user profiles caused by changes in user behavior and network changes.

[0073] like Figure 6 As shown, the network-level logical grid portrait can be obtained through but not limited to steps S211, S212, S213 and S214.

[0074] Step S211, obtaining wireless signal strength information between the base station and the UE;

[0075] Step S212, determining a cell-level logical grid according to the wireless signal strength information;

[0076] Step S213, obtaining a cell-level grid portrait according to the cell information corresponding to the cell-level logical grid and the UE behavior data;

[0077] Step S214, constructing a network-level logical grid image based on the cell-level logical grid and the cell-level grid image.

[0078] In an embodiment of the present application, in the process of constructing a network-level logical grid portrait, the wireless signal strength information between the base station and the user terminal is first obtained; then the cell-level logical grid is determined based on the wireless signal strength information; then the cell-level grid portrait is determined based on the cell information corresponding to the cell-level logical grid and the user terminal behavior data; finally, the network-level logical grid portrait can be constructed based on the cell-level logical grid and the cell-level grid portrait.

[0079] For example, Figure 7 As shown, the network-level logical grid portrait may include but is not limited to cell information and portrait information, wherein the cell information may include service cell, same-frequency neighboring cell, different-frequency neighboring cell and service cell load information; the portrait information may include service portrait, channel portrait and mobility portrait; the service portrait includes service delay and service rate, the channel portrait includes spectrum efficiency and channel quality, and the mobility portrait includes the target cell switching success rate.

[0080] It is worth noting that the network-level logical grid portrait divides the target network into grids, and each grid is regarded as an independent wireless resource scheduling and user management unit to achieve refined control of network resources and users. Data mining, data maintenance, grid construction, and grid updates of network-level grid portraits can be deployed on open application models or on dedicated servers. Taking into account the hardware capabilities of the base station and the required computing resources, the maintenance and calculation of the logical grid and portrait data involved in the embodiments of the present application can be deployed on a base station, an open application model, or a separate server.

[0081] It should be noted that the wireless signal strength information between the base station and the user terminal may be wireless signal strength information reported by the user terminal.

[0082] A network-level grid portrait is constructed based on the cell-level logical grids, cell-level grid portraits and cell coverage relationships within the target network. The network-level logical grid and network-level logical grid portrait also need to be updated periodically to avoid deviations in the logical grid and grid portrait caused by network changes.

[0083] like Figure 8 As shown, the above step S212 may include but is not limited to step S2121 and step S2122.

[0084] Step S2121, constructing an initial logical grid according to the wireless signal strength information;

[0085] Step S2122: determining a cell-level logical grid according to the measurement signal in the initial logical grid.

[0086] In the embodiment of the present application, in the process of constructing the cell-level logical grid, the initial logical grid is first constructed according to the wireless signal strength information, and then the cell-level logical grid is determined according to the measurement signal in the initial logical grid.

[0087] It should be noted that the measurement signal can be obtained through, but is not limited to, the user terminal's measurement of the base station's downlink broadcast, or the base station's measurement of the user terminal's uplink signal.

[0088] like Fig. 9 As shown, the above step S2121 may include but is not limited to step S2123 and step S2124.

[0089] Step S2123: when the number of cells where the UE is located is not less than three, the wireless signal strength information is the first signal strength of the serving cell and the neighboring cell, and the first signal strength of the serving cell and the neighboring cell is divided and processed according to a preset strength interval range to obtain an initial logical grid;

[0090] Step S2124, when the number of cells where the UE is located is less than three, the wireless signal strength information is the second signal strength of the serving cell, and the second signal strength of the serving cell and the radial direction of the serving cell are divided according to a preset strength interval range to obtain an initial logical grid.

[0091] In an embodiment of the present application, when the number of cells where the user terminal is located is not less than three, the wireless signal strength information is the first signal strength of the service cell and the neighboring cell, and then the first signal strength of the service cell and the neighboring cell is divided and processed according to a preset strength interval range to obtain an initial logical grid; when the number of cells where the user terminal is located is less than three, the wireless signal strength information is the second signal strength of the service cell, and then the second signal strength of the service cell and the radial direction of the service cell are divided and processed according to a preset strength interval range to obtain an initial logical grid.

[0092] For example, Fig.10 As shown in the figure, the wireless signal coverage areas of service cell A, neighboring cell A1, and neighboring cell A2 are divided into continuous intervals with a step size of ΔdB according to the decreasing direction of signal strength; the signal intervals of service cell A, neighboring cell A1, and neighboring cell A2 with a step size of ΔdB can form polygonal grids one by one, and each grid is called an initial logical grid. Fig.11 As shown, the wireless signal coverage area of ​​service cell A is also divided into continuous intervals with a step size of ΔdB according to the direction of decreasing signal strength; at the same time, the coverage area of ​​service cell A is divided into multiple sector-shaped areas along the radial direction; the axial interval and radial interval of service cell A also constitute multiple initial logical grids. Fig.10 Assuming that the cell ID of the service cell A is 101, the cell ID of the neighboring cell A1 is 201, and the cell ID of the neighboring cell A2 is 301, after the initial logical grid is constructed, the cell IDs of the service cell and the neighboring cell corresponding to different initial logical grids can be shown in the following table.

[0093] Grid ID Serving cell ID Neighborhood 1 Neighborhood 2 a1 201 301 101 a2 101 201 301 a3 101 201 301 a4 301 101 201 a5 201 101 301

[0094] like Fig.12 As shown, the above step S2122 may include but is not limited to step S2125 and step S2126.

[0095] Step S2125, collecting measurement signals in the initial logic grid within a preset time interval;

[0096] Step S2126: When the number of times that the measured signal is not lower than the preset signal threshold reaches a set number, the corresponding initial logical grid is determined as a cell-level logical grid.

[0097] In an embodiment of the present application, after constructing the initial logical grid, in the subsequent process of determining to construct the cell-level logical grid, the measurement signal in the initial logical grid can be collected within a preset time interval. When the number of times the measurement signal is not lower than the preset signal threshold reaches a set number of times, the corresponding initial logical grid will be determined as a cell-level logical grid.

[0098] Exemplarily, the measurement signal may be a reference signal receiving power signal covered by the base station, and the measurement signal is continuously collected for N days (N≥14), and the number of valid measurement results collected (signal strength is not less than the signal threshold) is not less than the preset number of times, then the corresponding initial logical grid is determined as the cell-level logical grid. The determination of the signal threshold and the preset number of times can be obtained through experimental verification, or can be determined based on experience or theoretical analysis.

[0099] like Fig.13 As shown, the above step S213 may include but is not limited to step S2131, step S2132, step S2133, step S2134 and step S2135.

[0100] Step S2131, obtaining cell information and UE behavior data corresponding to the cell-level logical grid;

[0101] Step S2132, sorting the cell information and UE behavior data to obtain a second training set and a second test set;

[0102] Step S2133, adjusting the model parameters of the preset second portrait model based on the second training set;

[0103] Step S2134, verifying the adjusted second portrait model based on the second test set;

[0104] Step S2135: If the verification passes, the second portrait model is determined as a cell-level grid portrait.

[0105] In an embodiment of the present application, in the process of constructing a cell-level grid portrait, first, the cell information and user terminal behavior data corresponding to the cell-level logical grid are obtained; then, the cell information and user terminal behavior data are sorted to obtain a second training set and a second test set; then, based on the second training set, the model parameters of the preset second portrait model are adjusted; then, based on the second test set, the adjusted second portrait model is verified; if the verification passes, the second portrait model is determined as a cell-level grid portrait.

[0106] For example, Fig.14 As shown in the figure, in the process of constructing the cell-level grid image, the cell information and user terminal behavior data corresponding to each cell-level logical grid are first collected; then the data is filtered, cleaned, and abnormal data is removed, and then clustering is performed to calculate the converged center position data set. The legal image dataset D in the grid satisfies the distance to the center location dataset is (μ D , σ D), that is, the probability density function is During the baseline data verification process, The deviation analysis is performed with the subsequent collected image data D to satisfy P(D|μ D ,σ D )≥η D (η D is a predefined probability threshold), then the baseline profile data is constructed successfully, where refers to the “center of mass” of the data set, (μ D , σ D ) respectively represent the mean and variance of the distance from each group of data to the centroid; after the above steps, each cell has multiple logical grids, and each grid has also completed the construction of logical grid portrait data; upload the cell-level logical grid data and grid portrait data of each cell to the logical grid / portrait server. The cell-level logical grid and grid portrait data need to be updated periodically to avoid the deviation of logical grid and grid portrait caused by network changes; the logical grid and grid portrait are updated once in each update window.

[0107] like Fig.15 As shown, the base station energy consumption portrait can be obtained through but not limited to steps S221, S222, S223, S224 and S225.

[0108] Step S221, obtaining hardware feature data, configuration data and operation data of the base station;

[0109] Step S222, sorting the hardware feature data, configuration data and operation data to obtain a third training set and a third test set;

[0110] Step S223, adjusting the model parameters of the preset third portrait model based on the third training set;

[0111] Step S224, verifying the adjusted third portrait model based on the third test set;

[0112] Step S225: If the verification passes, the third portrait model is determined as the base station energy consumption portrait.

[0113] In an embodiment of the present application, in the process of constructing a base station energy consumption portrait, the hardware feature data, configuration data and operation data of the base station can be first obtained; then, the hardware feature data, configuration data and operation data can be sorted to obtain a third training set and a third test set; then, based on the third training set, the model parameters of the preset third portrait model are adjusted; then, based on the third test set, the adjusted third portrait model is verified; finally, if the verification passes, the third portrait model is determined as the base station energy consumption portrait.

[0114] For example, Fig.16 As shown, the base station energy consumption profile may include parameters of the baseband processing unit and parameters of the radio frequency unit; wherein the parameters of the baseband processing unit and the radio frequency unit may include hardware parameters, wireless configuration, energy-saving functions and energy consumption data; the parameters of the baseband processing unit may also include single-board load indicators and cell loads.

[0115] like Fig.17 As shown, the above step S220 may include but is not limited to step S226 and step S227.

[0116] Step S226, constructing a segmented energy consumption model of the base station based on the base station energy consumption portrait;

[0117] Step S227: performing segmented prediction processing on the user campable cell set based on the segmented energy consumption model of the base station to obtain a base station energy consumption prediction set.

[0118] In an embodiment of the present application, after obtaining the base station energy consumption portrait, it is also necessary to construct a base station energy consumption model based on the base station energy consumption portrait; then, based on the base station segmented energy consumption model, a segmented prediction processing is performed on the user's available cell set to obtain a base station energy consumption prediction set.

[0119] It is worth noting that the characteristic parameters, configuration parameters and runtime parameters of the base station (including the baseband processing unit and the radio frequency unit) are collected, and the collected data sets are analyzed and processed using data mining and deep learning technologies to construct the energy consumption profile data set of the base station. Compared with the construction of user portraits, the energy consumption profile of the base station is more complex, so the portrait data set alone cannot meet the prediction of the power consumption of the base station. This is mainly because: the power consumption of the radio frequency unit power amplifier will occupy 40%-65% of the power consumption of the radio frequency unit. Under different business load conditions, the working area of ​​the power amplifier is different, and the power amplifier efficiency is also different (power amplifier efficiency = power amplifier output power / power amplifier input power); in order to ensure the output power of the power amplifier, it is necessary to change the input power of the power amplifier. Different interference environments and different user distributions will affect the space division multiplexing capability and scheduling performance of the base station, which will affect the size of the air interface business load, and then affect the energy consumption of the radio frequency unit and the baseband processing unit. Different business loads will also affect the energy-saving functions adopted by the baseband processing unit and the radio frequency processing unit, which will also affect the overall energy consumption of the base station.

[0120] like Fig.18 As shown, the above-mentioned step S226 may include but is not limited to step S2261, step S2262, step S2263, step S2264, step S2265, step S2266 and step S2267.

[0121] Step S2261, obtaining hardware feature data, configuration data and operation data of the base station;

[0122] Step S2262, determining energy consumption prediction influencing parameters from hardware feature data, configuration data, and operation data;

[0123] Step S2263, determining a plurality of power consumption reasoning models according to the energy consumption prediction influencing parameters and the preset segment division range;

[0124] Step S2264, obtaining a fourth training set and a fourth test set based on the base station energy consumption profile;

[0125] Step S2265, adjusting model parameters of multiple power consumption inference models based on the fourth training set;

[0126] Step S2266, verifying the multiple power consumption reasoning models after adjustment based on the fourth test set;

[0127] Step S2267: If the verification is passed, multiple power consumption reasoning models are determined as base station segment energy consumption models.

[0128] In an embodiment of the present application, in the process of constructing a base station segmented energy consumption model based on the base station energy consumption portrait, the hardware feature data, configuration data and operation data of the base station are first obtained, and then the energy consumption prediction influencing parameters are determined from the hardware feature data, configuration data and operation data; then, multiple power consumption reasoning models are determined according to the energy consumption prediction influencing parameters and the preset segmentation range; then, a fourth training set and a fourth test set are obtained based on the base station energy consumption portrait; model parameters of multiple power consumption reasoning models are adjusted based on the fourth training set; then, the multiple power consumption reasoning models after adjustment are verified based on the fourth test set; if the verification passes, the multiple power consumption reasoning models are determined as the base station segmented energy consumption model.

[0129] For example, Fig.19 As shown, select the factors that have the greatest impact on the energy consumption of the base station (such as downlink PRB occupancy, number of downlink space division multiplexing layers, downlink modulation and coding method, etc.), and determine the impact of these factors or a combination of these factors on the energy consumption of the base station under different value ranges through laboratory / commercial field test analysis. Combined with the characteristics of the base station hardware, the segmentation range is reasonably defined and the judgment conditions for the segmentation are determined. In one embodiment provided in the present application, segmentation is performed according to load, and energy consumption reasoning is performed by dividing it into three sections: high load, medium load and low load. As shown in FIG. Fig. 20 As shown, in the process of building a segmented energy consumption model of a base station, the present application provides a specific example. Within different energy consumption reasoning segment ranges, a sufficient number of base station energy consumption portrait data are collected, and energy consumption models of different load segments are built based on machine learning; this process can be functionally divided into four sub-functions: "data maintenance, model training, model maintenance, and model distribution"; data maintenance module: mainly supports data collection, data maintenance, training set data / test set data provision, obtains and maintains base station data reporting and energy consumption prediction feedback, provides training set data to the model training module, and provides test set data to the model maintenance module; model training ... provides training set data to the model training module, and provides test set data to the model maintenance module; model training module: mainly supports data collection, data maintenance, training set data / test set data provision, provides training set data to the model training module, and provides test set data to the model maintenance module; model training module: mainly supports data collection, data maintenance, training set data / test set data provision, provides training set data to the model training module, and provides test set data to the model maintenance module; model training module: mainly supports data collection, data maintenance, training set data / test set data The main functions use the neural network algorithm, perform deep learning based on the training set data, send the generated energy consumption model to the model maintenance module, receive the error result feedback of the model maintenance module based on the test set data, and optimize the neural network algorithm based on the error result feedback of the model maintenance module; the model maintenance module: receives the model update of the model training module, uses the base station energy consumption model based on the test set data to infer the base station energy consumption, and feeds back the inference error to the model training module, and sends the determined energy consumption inference model to the model sending module; the model sending module: receives the energy consumption model sent by the model determination module, and sends the energy consumption model to the base station equipment. In the subsequent operation of the base station equipment, the energy consumption profile parameter set to be predicted is input into the energy consumption inference path corresponding to different loads, and the predicted base station energy consumption under the corresponding conditions is obtained through the corresponding energy consumption model inference.

[0130] like Fig.21 As shown, the above step S210 may include but is not limited to step S230 and step S240.

[0131] Step S230, determining information of a plurality of predicted resident cells that meet predicted user service requirements according to the network-level logical grid portrait;

[0132] Step S240: adjusting the information of the plurality of predicted camped cells according to the cell energy saving strategy in the target network to obtain a set of cells where the user can camp.

[0133] In an embodiment of the present application, in the process of determining a set of cells where users can reside, first, multiple predicted cell information that meets the predicted user service needs is determined based on a network-level logical grid portrait; then, the multiple predicted cell information is adjusted and processed according to the cell energy-saving strategy in the target network to obtain a set of cells where users can reside.

[0134] It is worth noting that some cells in the target network will also be configured with energy-saving strategies, and the cell may not be able to provide services at the next moment. Therefore, it is necessary to further screen and process the information of multiple predicted resident cells that meet the predicted user service needs to obtain a set of user-residing cells.

[0135] like Fig. 22 As shown, the above step S227 may include but is not limited to step S2271, step S2272 and step S2273.

[0136] Step S2271, selecting a corresponding power consumption inference model for each predicted camping cell information in the set of camping cells for the user;

[0137] Step S2272, predicting and processing the corresponding predicted resident cell information according to the power consumption inference model to obtain predicted power consumption information, wherein the predicted power consumption information corresponds to the predicted resident cell information one by one;

[0138] Step S2273: determine the multiple predicted power consumption information as a base station energy consumption prediction set.

[0139] In an embodiment of the present application, in the process of calculating the base station energy consumption prediction set, first, a corresponding power consumption reasoning model is selected for each predicted resident cell information in the user's possible resident cell set to prepare for the energy consumption segmentation prediction; then, the corresponding predicted resident cell information is predicted and processed according to the selected power consumption reasoning model to obtain the predicted power consumption information; finally, the obtained multiple predicted power consumption information is determined as the base station energy consumption prediction set.

[0140] It is worth noting that in the process of selecting the corresponding power consumption reasoning model for each predicted resident cell information in the user's resident cell set, each predicted resident cell information will correspond to a power consumption reasoning model. For example, if a certain predicted resident cell information corresponds to a high load situation, the high load power consumption reasoning model will be used for power consumption prediction processing later.

[0141] like Fig.23 As shown, the above step S300 may include but is not limited to step S310 and step S320.

[0142] Step S310, selecting predicted power consumption information with the lowest power consumption from the base station energy consumption prediction set;

[0143] Step S320: scheduling the base stations and UEs in the target network according to the predicted resident cell information corresponding to the lowest predicted power consumption information.

[0144] In an embodiment of the present application, during energy-saving scheduling of base stations in the target network, the predicted power consumption information with the lowest power consumption is first selected from the base station energy consumption prediction set, and then the base stations and user terminals in the target network can be scheduled based on the predicted resident cell information corresponding to the lowest predicted power consumption information.

[0145] It is worth noting that the process of energy-saving scheduling of base stations in the target network is essentially to achieve scheduling of user terminals and wireless resources of the target network, so that user terminals reside in suitable cells, meet the business needs of user terminals and keep the energy consumption of base stations in the target network at the lowest state.

[0146] In order to more clearly illustrate the scheduling process of the network scheduling method provided in the embodiment of the present application, a specific example is given below for illustration.

[0147] like Fig.24 As shown in the figure, base stations A1-A3 have enabled the F1 frequency band, and base stations B1-B2 have enabled the F2 frequency band. At time t1, terminal U1 resides in the F1 frequency band of base station A1, and terminal U2 resides in the F2 frequency band of base station B2.

[0148] Base station A1 predicts the user's location at time t2: At time t1, terminal U1 is in logical grid a1, and terminal U2 is in logical grid b1. The "user portrait" is queried to obtain the terminal's moving direction and speed, and combined with the terminal's historical portrait, the grid position of the terminal at time t2 is determined. The figure indicates that terminal U1 will be in grid a2 and terminal U2 will be in grid b2 at time t2.

[0149] Base station A1 determines the set of solutions where the user can reside: Based on the user portrait, the user's service attributes can be learned. Combined with the user's real-time service characteristics at time t1, the user's service needs at time t2 can be predicted; based on the user's service needs at time t2 and the user's grid position, the grid a2 and grid b2 positions are searched in the network-level grid portrait library, and the set of resident cells {Cell1, Cell2, ..., Cell n}, base station A1 sends prediction information that terminal U1 will reside in each cell in the set of cells that can reside at time t2. When each cell in the set of cells that can reside receives the prediction information that terminal U1 will reside in at time t2 sent by base station A1, each cell needs to make a judgment based on its own configured energy-saving strategy. k (1≤k≤n) determines that at time t2, the Cell will be unable to provide services due to energy saving. k (1≤k≤n) will record this information.

[0150] Each base station determines its own energy consumption at time t2: After predicting the set of cells where the terminal can reside at time t2, the energy consumption profile data of the base station where the cell resides at time t2 is predicted, and a segmented base station energy consumption reasoning method is used to predict the energy consumption set of each base station at time t2. In this example, it is {P BS_A1 ,P BS_A2 ,P BS_A3 ,P BS_B1 ,P BS_B2}.

[0151] Determine the residence plan with the lowest energy consumption: Each base station in the target network area (in this example, base stations A1-A3, B1-B2) will predict the energy consumption information at time t2 {P BS_A1 ,P BS_A2 ,P BS_A3 ,P BS_B1 ,P BS_B2}, and the base station power consumption P BS_j The cell information corresponding to the terminal resides in (j∈{A1,A2,A3,B1,B2}) is sent to the decision control server (DCS). The DCS calculates the power consumption set of the target network area at time t2 based on the power consumption information of each base station predicted at time t2 and the terminal residency information corresponding to the power consumption of each base station. (There are K base stations in the current network area and m optional resident solutions.) DCS selects the resident solution with the lowest energy consumption from the power consumption set; it should be noted that: Cell k(1≤k≤n) reports to DCS that it cannot meet the terminal's residence requirements. When DCS selects available residence solutions, it will remove the Cell k When selecting an available residence plan, DCS will combine the energy-saving strategies of each cell in the target network to avoid selecting a cell that is already in the energy-saving moment or is about to enter the energy-saving moment at time t2; in particular, energy-saving functions that will cause the cell to be decommissioned, such as carrier shutdown, deep sleep and other energy-saving functions.

[0152] DCS issues control commands: DCS issues corresponding control commands to each base station / cell in the area according to the selected residence plan at time t2, ensuring that each base station / cell in the area executes according to the plan selected by DCS, so that each terminal in the area resides in the designated cell at time t2. On the basis of satisfying user service perception, the energy consumption in the area is minimized.

[0153] In addition, if Fig.25 As shown, an embodiment of the present application further provides a network device 700, the network device 700 comprising:

[0154] The memory 720 , the processor 710 , and a computer program stored in the memory 720 and executable on the processor 710 .

[0155] The processor 710 and the memory 720 may be connected via a bus or in other ways.

[0156] It should be noted that the network device 700 in this embodiment and the network scheduling method in the above embodiments belong to the same inventive concept, so these embodiments have the same implementation principles and technical effects, which will not be described in detail here.

[0157] The non-transient software program and instructions required to implement the network scheduling method of the above embodiment are stored in the memory 720. When executed by the processor 710, the network scheduling method of the above embodiment is executed, for example, the above described Figure 2 The method steps S100 to S300, Figure 3 Steps S210 to S220 of the method, Figure 4 Steps S110 to S150 of the method, Figure 6 Steps S211 to S214 of the method, Figure 8 Steps S2121 to S2122 of the method, Fig. 9 Steps S2123 to S2124 of the method, Fig.12 Steps S2125 to S2126 of the method, Fig.13 Steps S2131 to S2135 of the method, Fig.15 Steps S221 to S225 of the method, Fig.17Steps S226 to S227 of the method, Fig.18 Steps S2261 to S2267 of the method, Fig.21 Steps S230 to S240 of the method, Fig. 22 Method steps S2271 to S2272 and Fig.23 Method steps S310 to S320 in .

[0158] In addition, an embodiment of the present application further provides a computer-readable storage medium, which stores computer-executable instructions. The computer-executable instructions are executed by a processor 710, for example, by a processor 710 in the above network device 700 embodiment, so that the processor 710 can execute the network scheduling method in the above embodiment, for example, execute the above described Figure 2 The method steps S100 to S300, Figure 3 Steps S210 to S220 of the method, Figure 4 Steps S110 to S150 of the method, Figure 6 Steps S211 to S214 of the method, Figure 8 Steps S2121 to S2122 of the method, Fig. 9 Steps S2123 to S2124 of the method, Fig.12 Steps S2125 to S2126 of the method, Fig.13 Steps S2131 to S2135 of the method, Fig.15 Steps S221 to S225 of the method, Fig.17 Steps S226 to S227 of the method, Fig.18 Steps S2261 to S2267 of the method, Fig.21 Steps S230 to S240 of the method, Fig. 22 Method steps S2271 to S2272 and Fig.23 Method steps S310 to S320 in .

[0159] It will be appreciated by those skilled in the art that all or some of the steps and systems in the disclosed method above may be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that may be used to store desired information and may be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0160] The above is a specific description of the preferred implementation of the present application, but the present application is not limited to the above-mentioned implementation mode. Technical personnel familiar with the field can also make various equivalent deformations or substitutions without violating the spirit of the present application. These equivalent deformations or substitutions are all included in the scope defined by the claims of the present application.

Claims

1. A network scheduling method, the method comprising: Acquire current terminal feature information of a user terminal UE in the target network, and obtain predicted user service needs based on the current terminal feature information and a preset user portrait; Determine a base station energy consumption prediction set according to a preset base station energy consumption profile and the predicted user service demand; Energy-saving scheduling is performed on base stations in the target network according to the base station energy consumption prediction set.

2. The network scheduling method according to claim 1, characterized in that: The determining of the base station energy consumption prediction set according to the preset base station energy consumption profile and the predicted user service demand includes: Obtaining a set of cells where the user can reside based on the predicted user service demand and a preset network-level logical grid portrait; The base station energy consumption prediction set is obtained by inferring the user-residence cell set according to the base station energy consumption profile.

3. The network scheduling method according to claim 1, characterized in that: The user portrait is obtained in the following way: Acquire user attribute information, user service attribute information, and user behavior attribute information of the UE; Performing data cleaning and classification on the user attribute information, the user service attribute information, and the user behavior attribute information to obtain a first training set and a first test set; Adjusting model parameters of a preset first portrait model based on the first training set; Performing verification processing on the adjusted first portrait model based on the first test set; If the verification passes, the first portrait model is determined as the user portrait.

4. The network scheduling method according to claim 2, characterized in that: The network-level logical grid image is obtained by: Acquire wireless signal strength information between the base station and the UE; Determine a cell-level logical grid according to the wireless signal strength information; Obtain a cell-level grid portrait according to the cell information corresponding to the cell-level logical grid and UE behavior data; The network-level logical grid portrait is constructed according to the cell-level logical grid and the cell-level grid portrait.

5. The network scheduling method according to claim 4, characterized in that: The determining of the cell-level logical grid according to the wireless signal strength information comprises: Constructing an initial logical grid according to the wireless signal strength information; The cell-level logical grid is determined according to the measurement signal in the initial logical grid.

6. The network scheduling method according to claim 5, characterized in that: The constructing an initial logical grid according to the wireless signal strength information comprises: When the number of cells where the UE is located is not less than three, the wireless signal strength information is the first signal strength of the serving cell and the neighboring cell, and the first signal strength of the serving cell and the neighboring cell is divided and processed according to a preset strength interval range to obtain the initial logical grid; When the number of cells where the UE is located is less than three, the wireless signal strength information is the second signal strength of the serving cell, and the second signal strength of the serving cell and the radial direction of the serving cell are divided according to a preset strength interval range to obtain the initial logical grid.

7. The network scheduling method according to claim 5, characterized in that: The determining of the cell-level logical grid according to the measurement signal in the initial logical grid comprises: Collecting measurement signals in the initial logic grid within a preset time interval; When the number of times that the measurement signal is not lower than the preset signal threshold reaches a set number, the corresponding initial logical grid is determined as the cell-level logical grid.

8. The network scheduling method according to claim 4, characterized in that: The obtaining of a cell-level grid portrait according to the cell information corresponding to the cell-level logical grid and the UE behavior data includes: Obtaining cell information and UE behavior data corresponding to the cell-level logical grid; Arranging the cell information and the UE behavior data to obtain a second training set and a second test set; Adjusting model parameters of a preset second portrait model based on the second training set; Performing verification processing on the adjusted second portrait model based on the second test set; If the verification passes, the second portrait model is determined as the cell-level grid portrait.

9. The network scheduling method according to claim 1, characterized in that: The base station energy consumption profile is obtained in the following manner: Acquiring hardware feature data, configuration data, and operation data of the base station; Arrange the hardware feature data, the configuration data, and the operation data to obtain a third training set and a third test set; Adjusting model parameters of a preset third portrait model based on the third training set; Performing verification processing on the adjusted third portrait model based on the third test set; If the verification passes, the third portrait model is determined as the base station energy consumption portrait.

10. The network scheduling method according to claim 2, characterized in that: The inference of the user-residenceable cell set according to the preset base station energy consumption profile to obtain a base station energy consumption prediction set includes: Constructing a segmented energy consumption model of a base station based on the energy consumption portrait of the base station; The base station energy consumption prediction set is obtained by performing segmented prediction processing on the user-residenceable cell set based on the segmented energy consumption model of the base station.

11. The network scheduling method according to claim 10, characterized in that: The constructing of a base station segmented energy consumption model based on the base station energy consumption portrait includes: Acquiring hardware feature data, configuration data, and operation data of the base station; Determining energy consumption prediction influencing parameters from the hardware feature data, the configuration data, and the operation data; Determining multiple power consumption reasoning models according to the energy consumption prediction influencing parameters and the preset segment division range; Obtaining a fourth training set and a fourth test set based on the base station energy consumption portrait; Adjusting model parameters of the plurality of power consumption reasoning models based on the fourth training set; Performing verification processing on the multiple power consumption reasoning models after adjustment based on the fourth test set; If the verification is passed, the multiple power consumption reasoning models are determined as the base station segment energy consumption model.

12. The network scheduling method according to claim 2, characterized in that: The obtaining of a set of cells where the user can reside according to the predicted user service demand and a preset network-level logical grid profile includes: Determine information of a plurality of predicted resident cells that meet the predicted user service demand according to the network-level logical grid portrait; The plurality of predicted resident cell information are adjusted and processed according to the cell energy saving strategy in the target network to obtain the user resident cell set.

13. The network scheduling method according to claim 10, characterized in that: The performing segmented prediction processing on the set of cells where the user can reside based on the segmented energy consumption model of the base station to obtain the base station energy consumption prediction set includes: Selecting the corresponding power consumption reasoning model for each predicted resident cell information in the set of resident cells where the user can reside; Performing prediction processing on the corresponding predicted resident cell information according to the power consumption inference model to obtain predicted power consumption information, wherein the predicted power consumption information corresponds to the predicted resident cell information one by one; The plurality of predicted power consumption information are determined as the base station energy consumption prediction set.

14. The network scheduling method according to claim 13, characterized in that: The performing energy-saving scheduling processing on the base stations in the target network according to the base station energy consumption prediction set includes: Selecting the predicted power consumption information with the lowest power consumption from the base station energy consumption prediction set; The base station and the UE in the target network are scheduled according to the predicted resident cell information corresponding to the lowest predicted power consumption information.

15. A network device, characterized in that: include: at least one processor; at least one memory for storing at least one program; When at least one of the programs is executed by at least one of the processors, the network scheduling method according to any one of claims 1 to 14 is implemented.

16. A computer-readable storage medium storing computer-executable instructions, characterized in that: The computer executable instructions are used to execute the network scheduling method described in any one of claims 1 to 14.

Images