Resource allocation method, device, server, and storage medium

Abstract

The present application relates to a technical field of communication, and to a resource allocation method, an apparatus, a server, and a storage medium. The resource allocation method includes the steps of: acquiring a first communication load sequence of a first past period of a specified DSS cell group, acquiring a first predicted communication load of the DSS cell group based on the first communication load sequence according to a preset prediction algorithm, predicting a communication load of the DSS cell group according to a preset communication load prediction model obtained by training based on a second communication load sequence of the DSS cell group in a second past period greater than the first past period, and acquiring a second predicted communication load of the DSS cell group, and determining a resource allocation policy of the DSS cell group based on a difference value between the first predicted communication load and the second predicted communication load, and performing resource allocation for the DSS cell group according to the resource allocation policy.

Description

[Technical Field]

[0001] (CROSS-REFERENCE TO RELATED APPLICATIONS) This application claims priority to a Chinese patent application filed on December 13, 2021, bearing application number 202111521921.9.

[0002] TECHNICAL FIELD The present disclosure relates to the technical field of communications, and more particularly to a resource allocation method, device, server, and storage medium. [Background technology]

[0003] Dynamic Spectrum Sharing (DSS) allows 4G Long Term Evolution (4G LTE) and 5G New Radio (5G NR) to share the same spectrum and dynamically allocate time and frequency resources to 4G and 5G users. Spectrum resources are dynamically and flexibly allocated to different technologies within the same frequency band. This is because the physical layer design of 5G NR is similar to that of 4G LTE, which is a prerequisite for realizing dynamic spectrum sharing between 4G and 5G. The principle behind this is that when scheduling NR users in LTE subframes under the same subcarrier spacing and similar time domain configuration, it is first necessary to ensure that the common channels in 4G and 5G networks are independent of each other and do not affect each other, for example, to ensure that 5G NR reference signals and LTE reference signals do not conflict in time and frequency resource allocation, and then insert 5G NR user data into LTE subframes. Currently, there are three resource allocation forms of spectrum resource sharing technologies: based on a multicast / broadcast single frequency network (abbreviated as MBSFN), based on 5G mini-slot, and based on rate matching.

[0004] However, when dynamically allocating spectrum resources as described above, the amount of resources to be allocated to 5G is typically predicted by learning past channel conditions and analyzing the activity patterns of 4G and 5G users to predict future spectrum resource usage. This prediction is smarter and more consistent with the definition of cognitive radio. However, due to the limited computing and storage resources of network elements, such predictions often adopt a real-time prediction method, using only data from a few periods prior to the current time for statistical evaluation to determine the next-stage resource allocation method. However, due to the long-term development trends of 4G and 5G users, the impact of unexpected incidents, and the cyclical changes in daily network traffic, this method inevitably leads to prediction failures, resulting in a mismatch between the predicted resources and actual service needs, which impacts user experience. Summary of the Invention [Problem to be solved by the invention]

[0005] The main objective of the embodiments of the present application is to propose a resource allocation method, device, server, and storage medium, which aims to improve the accuracy of the determined spectrum resource allocation scheme for the DSS cell group, thereby matching the predicted resource with the actual service needs. [Means for solving the problem]

[0006] To achieve the above object, an embodiment of the present application provides a resource allocation method, which includes the steps of: obtaining a first communication load array for a first past period of a specified dynamic spectrum sharing DSS cell group; obtaining a first predicted communication load for the DSS cell group based on the first communication load array in accordance with a predetermined prediction algorithm; predicting the communication load of the DSS cell group in accordance with a predetermined communication load prediction model obtained by training the communication load prediction model based on a second communication load array for the DSS cell group in a second past period that is greater than the first past period, and obtaining a second predicted communication load for the DSS cell group; and determining a resource allocation policy for the DSS cell group based on a difference value between the first predicted communication load and the second predicted communication load, and performing resource allocation for the DSS cell group in accordance with the resource allocation policy.

[0007] To achieve the above object, an embodiment of the present application further provides a resource allocation device, the resource allocation device including: an acquisition module used to acquire a first communication load array for a first past period of a specified dynamic spectrum sharing DSS cell group; a first prediction module used to acquire a first predicted communication load for the DSS cell group based on the first communication load array according to a predetermined prediction algorithm; a second prediction module used to predict the communication load of the DSS cell group according to a predetermined communication load prediction model obtained by training the communication load prediction model based on a second communication load array for the DSS cell group for a second past period greater than the first past period, and to acquire a second predicted communication load for the DSS cell group; and a resource allocation module used to determine a resource allocation policy for the DSS cell group based on a difference value between the first predicted communication load and the second predicted communication load, and to perform resource allocation for the DSS cell group in accordance with the resource allocation policy.

[0008] To achieve the above object, an embodiment of the present application further provides a server, the server including at least one processor and a memory communicatively connected to the at least one processor, the memory storing instructions executable by the at least one processor, the execution of the instructions by the at least one processor enabling the at least one processor to perform the resource allocation method.

[0009] To achieve the above object, according to an embodiment of the present application, there is further provided a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above resource allocation method.

[0010] According to the resource allocation method of the present application, in resource allocation for a DSS cell group, a first communication load array for a first past period of a specified dynamic spectrum sharing DSS cell group is obtained, a first predicted communication load for the DSS cell group is obtained based on the first communication load array in accordance with a predetermined prediction algorithm, the communication load for the DSS cell group is predicted in accordance with a predetermined communication load prediction model obtained by training based on a second communication load array for the DSS cell group in a second past period that is greater than the first past period, a second predicted communication load for the DSS cell group is obtained, a resource allocation policy for the DSS cell group is determined based on a difference value between the first predicted communication load and the second predicted communication load, and resource allocation is performed for the DSS cell group in accordance with the resource allocation policy. By determining a resource allocation method for a DSS cell group using a difference between a first predicted communication load obtained from short-term historical data and a second predicted communication load obtained from long-term historical data, the present application can improve the accuracy of the spectrum resource allocation method for the DSS cell group to ensure that the predicted resources match actual service needs. This solves the technical problem of the prior art, which relies only on multiple short-term historical data to evaluate a resource allocation method, resulting in low accuracy in resource allocation for the DSS cell group and a mismatch between the predicted resources and actual service needs, which affects user experience. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a flowchart of a resource allocation method provided by an embodiment of the present application; [Figure 2] 1 is a flowchart of step 104 of a resource allocation method provided by an embodiment of the present application; [Figure 3] 1 is a flowchart of a resource allocation method provided by an embodiment of the present application; [Figure 4] 1 is a flowchart of a resource allocation method provided by an embodiment of the present application; [Figure 5] FIG. 1 is a schematic diagram illustrating the configuration of a resource allocation device provided by an embodiment of the present application; [Figure 6] FIG. 2 is a schematic diagram illustrating the configuration of a server provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE INVENTION

[0012] In order to clarify the objectives, technical solutions, and advantages of the embodiments of the present application, the following detailed description of each embodiment will be provided in conjunction with the accompanying drawings. However, those skilled in the art will understand that many technical details are provided in each embodiment of the present application to help readers better understand the present application. However, the technical solutions claimed for protection of the present application can be realized without these technical details and various changes and modifications based on the following embodiments. The division of the following embodiments is made for the convenience of explanation and should not be construed as imposing any limitations on the specific embodiments of the present application. The embodiments may be combined or referenced with each other as long as they are not inconsistent.

[0013] Dynamic spectrum sharing technology is beneficial for the smooth evolution from 4G to 5G, reduces 5G investment costs, and solves the problem of having too many 4G users and too few 5G users in the early stages of 5G development. It also solves the problem of 4G radio waves over high 5G frequency bands and poor signal penetration. Traditional dynamic spectrum allocation methods typically perform multiple checks and evaluations based on the current number and load of 4G and 5G users, and then adjust the spectrum resources allocated to 4G and 5G users in the next stage for access by their terminals. This policy has two problems: 1. Switching delay: The next resource allocation requires multiple checks and evaluations of the current 4G and 5G users' telephone traffic and load over several periods, resulting in long switching delays. Currently, when facing long peak times of user telephone traffic and load, the long evaluation times result in significant losses in telephone traffic quality and a poor user experience. Furthermore, if the dynamic conversion time is too long, for example, reaching 100 ms, the DSS cell group scheduler will switch from 4G to 5G only after some 5G users' traffic has already switched from peak to valley, and the 5G users no longer need to be scheduled. 2. Accuracy Issues: This resource allocation method, which evaluates the next period in real time based on only a few previous periods, is relatively monotonous and has low accuracy, making it easy for the evaluated resources to not match the actual service needs, which will affect the user experience.

[0014] An embodiment of the present application relates to a resource allocation method applied to a base station, and as shown in FIG. 1, the resource allocation method includes the following steps:

[0015] In step 101, a first communication load array for a first past period of a designated dynamic spectrum sharing DSS cell group is obtained.

[0016] In one exemplary embodiment, the DSS cell group is designated by a computing management center of the base station. After the DSS cell group is designated, a first communication load array for a first past period of the designated DSS cell group may be obtained from a database storing past communication load arrays for the DSS cell group according to the cell group identifier of the designated DSS cell group. Here, the first past period may limit the number of past periods, for example, the first past period refers to the past N periods. The first past period may limit the length of the past period, for example, the first past period refers to all periods within the past three hours. This specification does not specifically limit the manner in which the first past period is limited. The first communication load array is actually composed of the communication loads for each past period within the first past period, and each communication load refers to a communication-related indicator of interest to the operator, such as telephone traffic volume, uplink and downlink traffic, or a weighted computing indicator for both.

[0017] In step 102, a first predicted communication load of a DSS cell group is obtained based on a first communication load arrangement according to a preset prediction algorithm.

[0018] In one exemplary embodiment, the prediction algorithm may calculate a communication load trend based on each communication load in the first communication load array and obtain a first predicted communication load for the DSS cell group based on the communication load trend. The prediction algorithm may also calculate an average value for each communication load in the first communication load array and use the calculated average value as the first predicted communication load for the DSS cell group. Here, the first predicted communication load for the DSS cell group refers to a communication load predicted value for the DSS cell group for the next period.

[0019] In step 103, the communication load of the DSS cell group is predicted according to a predetermined communication load prediction model obtained by training based on a second communication load array of the DSS cell group for a second past period that is greater than the first past period, and a second predicted communication load of the DSS cell group is obtained.

[0020] In one exemplary embodiment, the communication load prediction model is an intelligent model that can autonomously infer the communication load of the DSS cell group for a certain period or a certain time in the future. Therefore, by inferring the communication load of the DSS cell group using the communication load prediction model, a second predicted communication load of the DSS cell group can be obtained. Here, the second predicted communication load of the DSS cell group refers to a communication load predicted value of the DSS cell group for a next period. The first predicted communication load and the second predicted communication load are communication load predicted values ​​of the DSS cell group for the same period, and differ in that the first predicted communication load is inferred from data of a short past period (e.g., the past few periods) and may be referred to as a short-term inferred communication load, while the second predicted communication load is inferred from data of a long past period (e.g., the past few months or even years) and may be referred to as a long-term inferred communication load.

[0021] In one exemplary embodiment, the communication load prediction model is obtained by training based on a second communication load array of a second past period of the DSS cell group. Here, the second past period limits the length of the past period, for example, the second past period refers to all periods within the past three months or one year. When comparing the second past period with the first past period, the number of past periods included in the second past period is much greater than the number of past periods included in the first past period. The second communication load array is actually composed of the communication loads of each past period within the second past period.

[0022] In step 104, a resource allocation policy for the DSS cell group is determined based on the difference between the first predicted communication load and the second predicted communication load, and resource allocation is performed for the DSS cell group according to the resource allocation policy.

[0023] In one exemplary embodiment, determining a resource allocation policy for a DSS cell group based on a difference value between the first predicted communication load and the second predicted communication load is actually determining the accuracy of the first predicted communication load and the second predicted communication load based on the difference value. Specifically, the process of determining a resource allocation policy for a DSS cell group based on a difference value between the first predicted communication load and the second predicted communication load, as shown in Figure 2, includes the following steps:

[0024] In step 201, it is detected whether the differential value falls within a preset load differential value range.

[0025] In one exemplary embodiment, the load difference value range is actually an acceptable range of the difference between the first predicted communication load and the second predicted communication load, and if the difference value is within the load difference value range, step 202 is performed, and if the difference value is outside the load difference value range, step 205 is performed.

[0026] In step 202, one is added to the number of times the DSS cell group belongs to the DSS cell group.

[0027] In one exemplary embodiment, if the difference value between the first predicted communication load and the second predicted communication load falls within a load difference value range, it indicates that the values ​​predicted by the two prediction methods are relatively close. However, in this case, it is not possible to determine whether the first predicted communication load or the second predicted communication load should be used to perform resource allocation, and the value of the number of times of belonging to the DSS cell group needs to be incremented by 1. Here, the value of the number of times of belonging to the DSS cell group is always accumulated and is not cleared.

[0028] In step 203, it is detected whether the value of the number of memberships reaches a first threshold.

[0029] In one exemplary embodiment, a first threshold value is set for the number of times of attribution, and the first threshold value indicates the number of times an event occurs in which the difference value between the first predicted communication load and the second predicted communication load is within a load difference value range. If the value of the number of times of attribution reaches the first threshold value, execute step 204; if the value of the number of times of attribution does not reach the first threshold value, execute step 208.

[0030] In step 204, resource allocation is performed according to the second predicted communication load as a resource allocation policy.

[0031] In one exemplary embodiment, when the value of the number of associations reaches a first threshold, the corresponding resource allocation policy is to perform resource allocation for the DSS cell group according to a second predicted communication load.

[0032] In step 205, 1 is added to the value of the non-membership count of the DSS cell group.

[0033] In one exemplary embodiment, if the difference between the first predicted communication load and the second predicted communication load does not fall within the load difference range, it indicates that the values ​​predicted by the two prediction methods are far apart. However, in this case, it is not possible to determine whether to use the first predicted communication load or the second predicted communication load to perform resource allocation, and the value of the non-association count for the DSS cell group needs to be incremented by 1. Here, the value of the non-association count for the DSS cell group is always accumulated and is not cleared.

[0034] In step 206, it is detected whether the value of the non-attribute count reaches a second threshold.

[0035] In one exemplary embodiment, a second threshold value corresponding to the non-attribute count is set, and the second threshold value indicates the number of occurrences of an event in which the difference value between the first predicted communication load and the second predicted communication load is outside the load difference value range. If the value of the attribute count reaches the second threshold value, execute step 207; if the value of the attribute count does not reach the second threshold value, execute step 208.

[0036] In step 207, resource allocation is performed according to the first predicted communication load as a resource allocation policy.

[0037] In one exemplary embodiment, when the value of the number of associations reaches the second threshold, the corresponding resource allocation policy is to perform resource allocation for the DSS cell group according to the first predicted communication load.

[0038] In step 208, resource allocation is performed according to the original resource allocation policy of the DSS cell group.

[0039] In an exemplary embodiment, if the value of the belonging count does not reach the first threshold and the value of the non-belonging count reaches the second threshold, perform resource allocation according to the original resource allocation policy of the DSS cell group.

[0040] In one exemplary embodiment, after a resource allocation policy is determined, resource allocation is performed for spectrum resources of a DSS cell group according to the determined resource allocation policy. If the determined resource allocation policy is resource allocation according to a first predicted communication load, the value of the first resource allocation count for the DSS cell group needs to be incremented by 1. Furthermore, if the value of the first resource allocation count reaches a preset third threshold, it indicates that the communication load prediction model is outdated, and new training of the communication load prediction model needs to be initiated to update the communication load prediction model. Here, the third threshold refers to the total number of times the second predicted communication load and the first predicted communication load do not match overall within a certain time range. For example, if the number of mismatches within 100 times is 30, the communication load prediction model needs to be updated, and the value of the first resource allocation count is updated to 0 at the 101st prediction.

[0041] According to an embodiment of the present application, in resource allocation for a DSS cell group, a first communication load array for a specified dynamic spectrum sharing DSS cell group is obtained for a first past time period, a first predicted communication load for the DSS cell group is obtained based on the first communication load array in accordance with a preset prediction algorithm, a communication load for the DSS cell group is predicted based on a predetermined communication load prediction model obtained by training the second communication load array for the DSS cell group for a second past time period greater than the first past time period, a second predicted communication load for the DSS cell group is obtained, a resource allocation policy for the DSS cell group is determined based on a difference between the first predicted communication load and the second predicted communication load, and resource allocation for the DSS cell group is performed in accordance with the resource allocation policy. By determining a resource allocation method for the DSS cell group using a difference between the first predicted communication load obtained from short-term past data and the second predicted communication load obtained from long-term past data, the present application can improve the accuracy of the spectrum resource allocation method for the DSS cell group to better match the predicted resources with actual service needs. This solves the technical problem in the prior art, where resource allocation methods are evaluated based only on multiple short-term historical data, resulting in low accuracy in resource allocation for DSS cell groups, a mismatch between predicted resources and actual service needs, and an impact on user experience.

[0042] An embodiment of the present application relates to a resource allocation method applied to a base station, and as shown in FIG. 3, the resource allocation method includes the following steps:

[0043] In step 301, a first communication load array for a first past period of a specified dynamic spectrum sharing DSS cell group is obtained.

[0044] In one exemplary embodiment, this step is generally the same as step 101 in the present application example, and therefore will not be described here.

[0045] In step 302, a first predicted communication load of a DSS cell group is obtained based on a first communication load arrangement according to a preset prediction algorithm.

[0046] In one exemplary embodiment, this step is generally the same as step 102 in the present application example, and therefore will not be described again here.

[0047] In step 303, a cell group identifier of the DSS cell group is obtained, and a scenario identifier is obtained by identifying a scenario of the DSS cell group.

[0048] In one exemplary embodiment, the cell group identifier can uniquely identify the specified DSS cell group. After the DSS cell group is specified, scenario identification can be performed for the specified DSS cell group to obtain the scenario in which the current DSS cell group is located, and a scenario identifier can be generated. For example, the scenario identifier can be a hotspot venue.

[0049] In step 304, a communication load prediction model corresponding to the cell group identifier and the scenario identifier is obtained from a preset model database.

[0050] In an exemplary embodiment, the model database is located in the computing management center of the base station, and each communication load forecasting model is also generated through training by the computing management center of the base station. When generating each communication load forecasting model, first, a second communication load sequence of a second past period of the DSS cell group in each scenario is obtained, and then, for each scenario, the second communication load sequence corresponding to the scenario is input into the pre-trained model based on a preset loss function for training to generate a communication load forecasting model corresponding to the scenario. After generating the communication load forecasting model for each scenario, the communication load forecasting model corresponding to the scenario is stored in the model database according to the DSS cell group and scenario, where the loss function refers to the difference between the predicted communication load and the actual communication load, and the training process refers to iteratively processing the model parameters according to the value of the loss function.

[0051] In one exemplary embodiment, when obtaining a communication load forecasting model corresponding to a cell group identifier and a scenario identifier, the cell group identifier and the scenario identifier need to be sent to the computing management center of the base station, and the computing management center will obtain the corresponding model from the model database and send it back. On the other hand, if a communication load forecasting model corresponding to the cell group identifier and the scenario identifier is not obtained from the model database, it indicates that the communication load forecasting model corresponding to this cell group identifier and the scenario identifier is immature and cannot be used. In this case, a general-purpose communication load forecasting model can be obtained from the model database as the communication load forecasting model.

[0052] In step 305, the communication load of the DSS cell group is predicted according to a predetermined communication load prediction model obtained by training based on a second communication load array of the DSS cell group for a second past period that is greater than the first past period, and a second predicted communication load of the DSS cell group is obtained.

[0053] In one exemplary embodiment, this step is generally the same as step 103 in the present application example, and therefore will not be described here.

[0054] In step 306, a resource allocation policy for the DSS cell group is determined based on the difference between the first predicted communication load and the second predicted communication load, and resource allocation is performed for the DSS cell group according to the resource allocation policy.

[0055] In one exemplary embodiment, this step is generally the same as step 104 in the present application example, and therefore will not be described here.

[0056] In addition to other embodiments, the present embodiment identifies a scenario of a DSS cell group and selects a corresponding communication load prediction model according to the scenario of the DSS cell group, so that the second predicted communication load obtained can be more suitable to the actual application scenario and more accurate.

[0057] An embodiment of the present application relates to a resource allocation method applied to a base station, and as shown in FIG. 4, the resource allocation method includes the following steps:

[0058] In step 401, a first communication load array for a first past period of a specified dynamic spectrum sharing DSS cell group is obtained.

[0059] In one exemplary embodiment, this step is generally the same as step 101 in the present application example, and therefore will not be described here.

[0060] In step 402, a first predicted communication load of a DSS cell group is obtained based on a first communication load arrangement according to a preset prediction algorithm.

[0061] In one exemplary embodiment, this step is generally the same as step 102 in the present application example, and therefore will not be described again here.

[0062] In step 403, the communication load of the DSS cell group is predicted according to a predetermined communication load prediction model obtained by training based on a second communication load array of the DSS cell group for a second past period that is greater than the first past period, and a second predicted communication load of the DSS cell group is obtained.

[0063] In one exemplary embodiment, this step is generally the same as step 103 in the present application example, and therefore will not be described here.

[0064] In step 404, the current communication quality and / or current communication load of the DSS cell group before resource allocation is obtained.

[0065] In one exemplary embodiment, before performing resource allocation for a DSS cell group, the current communication quality and / or current communication load of the DSS cell group must be recorded.

[0066] In step 405, a resource allocation policy for the DSS cell group is determined based on the difference between the first predicted communication load and the second predicted communication load, and resource allocation is performed for the DSS cell group according to the resource allocation policy.

[0067] In one exemplary embodiment, this step is generally the same as step 104 in the present application example, and therefore will not be described here.

[0068] In step 406, a first communication quality and / or a first communication load after resource allocation of the DSS cell group is obtained.

[0069] In one exemplary embodiment, after resource allocation for the DSS cell group is completed, a first communication quality and / or a first communication load after resource allocation for the DSS cell group needs to be recorded.

[0070] In step 407, the difference between the current communication quality and the first communication quality is obtained, and / or the difference between the current communication load and the first communication load is obtained.

[0071] In one exemplary embodiment, a difference in quality of the DSS cell group before and after performing resource allocation is obtained based on the recorded current communication quality and the first communication quality, and / or a difference in load of the DSS cell group before and after performing resource allocation is obtained based on the recorded current communication load and the first communication load.

[0072] In step 408, if the quality difference and / or load difference meets the preset tolerance condition for difference, wait for the next resource allocation period of the DSS cell group; if not, send alarm information to the manager of the DSS cell group.

[0073] In one exemplary embodiment, after obtaining the quality difference and / or load difference, it is necessary to detect whether the quality difference and / or load difference meets a preset difference tolerance condition. If the quality difference and / or load difference meets the difference tolerance condition, it indicates that the change in communication quality and / or communication load before and after resource allocation of the DSS cell group is within a normal range, and the resource allocation method determination for the DSS cell group can continue and wait for the next resource allocation period for this DSS cell group. On the other hand, if the quality difference and / or load difference do not meet the difference tolerance condition, it indicates that the change in communication quality and / or communication load before and after resource allocation of the DSS cell group is abnormal. In this case, it is necessary to stop determining the resource allocation method for the DSS cell group and send warning information to the administrator of the DSS cell group so that the administrator can reset the resource allocation method acquisition method for the DSS cell group.

[0074] In addition to other examples, the embodiment of the present application can make the present application more intelligent by comparing the communication load and communication quality before and after resource allocation to determine whether the method for determining a resource allocation policy is reasonable.

[0075] The division of the steps in the above various methods is merely for the purpose of clarity, and when implemented, they may be combined into one step or some steps may be subdivided into multiple steps, and as long as they contain the same logical relationship, they all fall within the scope of protection of this application. Any insignificant modifications or insignificant design changes to the algorithms or processes without changing the core design of the algorithms and processes also fall within the scope of protection of this application.

[0076] Another embodiment of the present application relates to a resource allocation device, and the details of the resource allocation device of this embodiment will be described in detail below. However, the following content is provided to facilitate understanding of the implementation details provided and is not a necessary condition for implementing this embodiment. Figure 5 is a schematic diagram of the resource allocation device described in this embodiment, which includes an acquisition module 501, a first prediction module 502, a second prediction module 503, and a resource allocation module 504.

[0077] Here, the obtaining module 501 is used to obtain a first communication load array of a specified dynamic spectrum sharing DSS cell group for a first past period.

[0078] The first prediction module 502 is used to obtain a first predicted communication load of the DSS cell group based on the first communication load array according to a preset prediction algorithm.

[0079] The second prediction module 503 is used to predict the communication load of the DSS cell group according to a predetermined communication load prediction model obtained by training based on a second communication load array of the DSS cell group for a second past period greater than the first past period, and to obtain a second predicted communication load of the DSS cell group.

[0080] The resource allocation module 504 is used to determine a resource allocation policy for the DSS cell group based on a difference value between the first predicted communication load and the second predicted communication load, and to perform resource allocation for the DSS cell group according to the resource allocation policy.

[0081] This embodiment is a system embodiment corresponding to the above method embodiment, and it is easy to understand that this embodiment can be implemented in combination with the above method embodiment. The relevant technical details and technical effects mentioned in the above embodiment are also valid in this embodiment, so they will not be described here to reduce redundancy. Therefore, the relevant technical details described in this embodiment are also applicable to the above embodiment.

[0082] It should be noted that the present system embodiment mainly describes the resource allocation method provided by the method embodiment at the software implementation level, and its implementation must also rely on hardware support. For example, the functionality of the associated modules may be located on a processor so that the processor executes the corresponding functions. In particular, associated data generated by the execution may be stored in memory for subsequent inspection and use.

[0083] In addition, each module in this embodiment is a logical module, and in actual application, one logical unit may be one physical unit, may be part of one physical unit, or may be realized by a combination of multiple physical units. In addition, in order to highlight the creative aspects of this application, means that are not closely related to solving the technical problem raised in this application are not introduced in this embodiment, but this does not mean that other means are not present in this embodiment.

[0084] Another embodiment of the present invention relates to a server, as shown in Figure 6, which includes at least one processor 601 and a memory 602 communicatively connected to the at least one processor 601, wherein the memory 602 stores instructions executable by the at least one processor 601, and the instructions, when executed by the at least one processor 601, enable the at least one processor 601 to perform the resource allocation method of each of the above embodiments.

[0085] Here, the memory and the processor are connected via a bus system, which may include any number of interconnected buses and bridges, connecting various circuits of one or more processors and memories together. The bus may also connect various other circuits, such as peripherals, voltage regulators, and power management circuits, which are well known in the art and will not be further described herein. A bus interface provides an interface between the bus and a transceiver. The transceiver may be a single element or multiple elements, such as multiple receivers and transmitters, and provides a means for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over a wireless medium via an antenna, which also receives data and transmits it to the processor.

[0086] A processor may provide a variety of functions beyond managing the bus and general processing, including timing, peripheral interfacing, voltage regulation, power management, and other control functions. Memory, on the other hand, may be used to store data used when performing operations by a processor.

[0087] Another embodiment of the present application relates to a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method embodiments described above.

[0088] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by instructing related hardware through a program. The program is stored in a storage medium and includes several instructions for causing a device (which may be a one-chip computer, chip, etc.) or a processor to execute all or part of the steps of the methods described in each embodiment of the present application. Meanwhile, the storage medium includes various media capable of storing program code, such as a USB memory, a removable hard disk, a read-only memory (abbreviated as ROM), a random access memory (abbreviated as RAM), a magnetic disk, or an optical disk.

[0089] Those skilled in the art will understand that the above embodiments are specific examples for implementing the present application, and that in actual applications, various changes can be made in form and details without departing from the spirit and scope of the present application.

Claims

1. obtaining a first communication load array for a first past time period of a designated dynamic spectrum sharing DSS cell group; obtaining a first predicted communication load of the DSS cell group based on the first communication load array according to a preset prediction algorithm; predicting a communication load of the DSS cell group according to a predetermined communication load prediction model obtained by training the communication load prediction model based on a second communication load sequence of the DSS cell group for a second past period that is greater than the first past period, and obtaining a second predicted communication load of the DSS cell group; determining a resource allocation policy for the DSS cell group based on a difference value between the first predicted communication load and the second predicted communication load, and performing resource allocation for the DSS cell group in accordance with the resource allocation policy; A resource allocation method comprising:

2. the resource allocation policy includes resource allocation according to the first predicted communication load, resource allocation according to the second predicted communication load, and resource allocation according to an original resource allocation policy of the DSS cell group, and the method includes: adding 1 to a value of a first resource allocation count of the DSS cell group when the resource allocation policy of the DSS cell group is resource allocation according to the first predicted communication load; updating the communication load prediction model when the value of the number of times the first resource is allocated reaches a third threshold value set in advance; The resource allocation method of claim 1 , further comprising:

3. the step of determining a resource allocation policy for the DSS cell group based on a difference value between the first predicted communication load and the second predicted communication load, detecting whether the differential value falls within a preset load differential value range; If the difference value falls within the load difference value range, add 1 to the value of the number of associations of the DSS cell group, detect whether the value of the number of associations has reached a first threshold, and if the value of the number of associations has reached the first threshold, set the resource allocation policy to resource allocation according to the second predicted communication load, otherwise perform resource allocation according to the original resource allocation policy of the DSS cell group; or When the difference value does not fall within the load difference value range, adding 1 to the value of the non-association count of the DSS cell group, detecting whether the value of the non-association count has reached a second threshold, and when the value of the non-association count has reached the second threshold, setting the resource allocation policy to resource allocation according to the first predicted communication load, and otherwise performing resource allocation according to the original resource allocation policy; The resource allocation method of claim 1 , comprising:

4. Before the step of predicting the communication load of the DSS cell group according to a preset communication load prediction model, obtaining a cell group identifier of the DSS cell group and identifying a scenario of the DSS cell group to obtain a scenario identifier; Obtaining the communication load prediction model corresponding to the cell group identifier and the scenario identifier from a preset model database; The resource allocation method of claim 1 , comprising:

5. The method comprises: a step of acquiring a general-purpose communication load prediction model from the model database as the communication load prediction model when the communication load prediction model corresponding to the cell group identifier and the scenario identifier is not acquired from the model database. The resource allocation method of claim 4 further comprising:

6. The method comprises: obtaining a second communication load array for the DSS cell group for a second past period in each scenario; For each of the scenarios, based on a preset loss function, inputting the second communication load sequence corresponding to the scenario into a pre-trained model to perform training, and generating the communication load prediction model corresponding to the scenario; storing the communication load prediction model corresponding to the scenario in the model database according to the DSS cell group and the scenario; The resource allocation method of claim 4 further comprising:

7. before the step of performing resource allocation for the DSS cell group in accordance with the resource allocation policy, acquiring a current communication quality and / or a current communication load of the DSS cell group before resource allocation, After the step of performing resource allocation for the DSS cell group according to the resource allocation policy, acquiring a first communication quality and / or a first communication load after resource allocation of the DSS cell group; acquiring a difference in quality between the current communication quality and the first communication quality, and / or acquiring a difference in load between the current communication load and the first communication load; If the quality difference and / or the load difference meets a preset tolerance condition for difference, waiting for the next resource allocation period of the DSS cell group, and if not, sending warning information to an administrator of the DSS cell group; The resource allocation method of claim 1 , comprising:

8. an acquisition module configured to acquire a first communication load array for a first past time period of a designated dynamic spectrum sharing DSS cell group; a first prediction module configured to obtain a first predicted communication load of the DSS cell group based on the first communication load array according to a preset prediction algorithm; a second prediction module configured to predict a communication load of the DSS cell group according to a predetermined communication load prediction model obtained by training the communication load prediction model based on a second communication load sequence of the DSS cell group for a second past period greater than the first past period, and to obtain a second predicted communication load of the DSS cell group; a resource allocation module configured to determine a resource allocation policy for the DSS cell group based on a difference value between the first predicted communication load and the second predicted communication load, and to perform resource allocation for the DSS cell group according to the resource allocation policy; A resource allocation device including:

9. 1. A server including at least one processor and a memory communicatively coupled to the at least one processor, The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the resource allocation method of any one of claims 1 to 7. server.

10. A computer-readable storage medium storing a computer program, The computer program, when executed by a processor, implements the resource allocation method according to any one of claims 1 to 7. A computer-readable storage medium.

Images