Method of implementing a self-organizing network of multiple access network devices and electronic device performing the method

Abstract

A method of adjusting a state of an access network comprising at least one cell, the method comprising: obtaining state history information of the access network at a first point in time; determining, based on the obtained state history information, a first time period required for adjusting the state of the access network; determining, based on the first point in time and the first time period, a second point in time, the second point in time being a reference point in time at which the state of the access network is to be adjusted; estimating, based on the obtained state history information, a state of the access network at the second point in time; determining, based on the estimated state of the access network at the second point in time, a value of a state control parameter for adjusting the state of the access network, and sending the determined value of the state control parameter to the access network.

Description

Technical Field

[0001] This disclosure relates to a method for implementing a self-organizing network of access network devices and an electronic device for performing the method. Background Technology

[0002] With the development of communication technology, each user can perform various types of operations by communicating with the base station using user equipment (UE). For example, users can use their UE to send and receive voice or text messages, reproduce sound sources or videos, or use the Internet.

[0003] A base station can establish an access network with multiple UEs to provide radio communication services to them. As an access network device, the base station can include multiple sectors, and each sector can include multiple cells based on frequency bands. Multiple UEs can connect to each cell, and radio communication services can be provided to each of the multiple UEs.

[0004] For example, a self-organizing network (SON) for access network devices can refer to a network that can self-configure considering its surrounding environment and can be used to economically control network coverage and increase transmission capacity, as well as improve the economic efficiency of network operation and management. To implement a self-organizing network, a central server capable of collecting and analyzing data from multiple access network devices and determining the optimal configuration may be required.

[0005] When implementing a self-organizing network with multiple access network devices using a central server, significant time may be required to collect and analyze the necessary data from these devices. In this scenario, a substantial time lag may exist between the point in time when data is collected from the access network devices and the point in time when the central server determines the optimal configuration. Therefore, the state of the access network at the point in time when the central server determines the optimal configuration may differ significantly from the state at the point in time of data collection, due to various changes occurring within each state. The optimal configuration determined by the central server may be inappropriate at the point in time when it is applied, thus potentially hindering the effective implementation of a self-organizing network. Summary of the Invention

[0006] Technical solution

[0007] Embodiments of this disclosure provide a method for effectively implementing a self-organizing network and an electronic device for performing the method, thereby solving the problem caused by the time difference between the time point of collecting data from the access network device and the time point of determining the optimal configuration.

[0008] Other aspects will be partly described in the following description, and in part can be seen from the description. Attached Figure Description

[0009] The above and other aspects, features, and advantages of certain embodiments of this disclosure will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, wherein:

[0010] Figure 1 This is a diagram illustrating an example communication system according to various embodiments, including multiple access network devices and a server for adjusting the status of the multiple access networks;

[0011] Figure 2 This is a block diagram illustrating an example electronic device for adjusting the state of multiple access networks according to various embodiments;

[0012] Figure 3 This is a block diagram illustrating example access network devices according to various embodiments;

[0013] Figure 4 This is a flowchart illustrating example methods for adjusting the state of an access network according to various embodiments;

[0014] Figure 5 This is a signal flow diagram illustrating example operations of an electronic device for adjusting the state of an access network according to various embodiments;

[0015] Figure 6 This is a diagram illustrating example methods for calculating the difference between the estimated state and the actual state of an access network according to various embodiments;

[0016] Figure 7a This is a diagram illustrating examples of methods for adjusting the state of an access network according to various embodiments;

[0017] Figure 7b This is a diagram illustrating examples of methods for adjusting the state of an access network according to various embodiments;

[0018] Figure 8 These are diagrams illustrating example methods for training a first artificial intelligence learning model according to various embodiments; and

[0019] Figure 9 This is a diagram illustrating example methods for training a first artificial intelligence learning model and a second artificial intelligence learning model according to various embodiments.

[0020] Regarding the description of the accompanying drawings, similar reference numerals may be used for similar elements. Detailed Implementation

[0021] According to an example embodiment of this disclosure, a method performed by an electronic device to adjust the state of an access network comprising at least one cell may include: acquiring state history information of the access network at a first time point; determining a first time period required to adjust the state of the access network based on the acquired state history information; determining a second time point based on the first time point and the first time period, wherein the second time point is a reference time point at which the state of the access network is to be adjusted; estimating the state of the access network at the second time point based on the acquired state history information; determining a value of a state control parameter for adjusting the state of the access network based on the estimated state of the access network at the second time point; and sending the determined value of the state control parameter to the access network, wherein the value of the state control parameter sent to the access network is applied to the state control parameter of the access network at the second time point to adjust the state of the access network.

[0022] According to an example embodiment of this disclosure, an electronic device for adjusting the state of an access network including at least one cell may include: at least one processor and a memory connected to the at least one processor, wherein the at least one processor may be configured to control the electronic device to: acquire state history information of the access network at a first time point; determine a first time period required to adjust the state of the access network based on the acquired state history information; determine a second time point based on the first time point and the first time period, the second time point being a reference time point for adjusting the state of the access network; estimate the state of the access network at a second time point based on the acquired state history information; determine a value of a state control parameter for adjusting the state of the access network based on the estimated state of the access network at the second time point; and send the determined value of the state control parameter to the access network, wherein the value of the state control parameter sent to the access network is applied to the state control parameter of the access network at the second time point to adjust the state of the access network.

[0023] Figure 1 The diagram illustrates an example communication system 100 according to various embodiments, which includes a plurality of access network devices 110 and a server 101 for adjusting the status of the plurality of access network devices 110.

[0024] Reference Figure 1 The communication system 100 may include a server 101 and multiple access network devices 110, and the multiple access network devices 110 may include a first access network device 111, a second access network device 112, and a third access network device 113. According to various embodiments of this disclosure, the number of access network devices connected to the server 101 is not limited to... Figure 1 The quantities shown.

[0025] According to embodiments of this disclosure, server 101 can send data to and receive data from each of a plurality of access network devices 110. For example, server 101 can receive information about the status of the plurality of access networks from each of the plurality of access network devices 110 at a preset time point. For example, server 101 can receive information about the status of the plurality of access networks from each of the plurality of access network devices 110 at a time point from t0 to t1. As another example, server 101 can send information about parameter values ​​for adjusting the status of each of the plurality of access networks to each of the plurality of access network devices 110. For example, server 101 can send information about parameter values ​​for adjusting the status of each of the plurality of access networks at a time point t2 to each of the plurality of access network devices 110. In embodiments of this disclosure, time point t2 may, for example, be after time point t1, and time point t1 may, for example, be after time point t0.

[0026] According to various embodiments of this disclosure, server 101 can determine parameter values ​​suitable for each of the plurality of access network devices 110 to perform adjustments to the state of each of the plurality of access networks. For example, first access network device 111 can adjust the state of its corresponding access network so that load balancing can be performed on a plurality of cells included in first access network device 111, and for this purpose, server 101 can determine parameter values ​​suitable for performing load balancing on a plurality of cells based on information about the state of the access network of first access network device 111. As another example, second access network device 112 can adjust the state of its corresponding access network to reduce power consumption, and for this purpose, server 101 can determine parameter values ​​suitable for reducing power consumption based on information about the state of the access network of second access network device 112.

[0027] According to various embodiments of this disclosure, each of the plurality of access network devices 110 may correspond to, for example, a base station, or may include, but is not limited to, a central unit (CU), a distributed unit (DU), a digital unit (DU), or a radio unit (RU), which include elements of a base station. According to embodiments of this disclosure, each of the plurality of access network devices 110 may include a plurality of cells. For example, a sector, which is a regional unit of a base station, may include a plurality of cells according to a frequency band, and each of the plurality of UEs may connect to any one of the cells selected from the plurality of cells.

[0028] According to various embodiments of this disclosure, multiple access network devices 110 can adjust the states of multiple access networks to balance the load of multiple cells, or adjust the states to reduce the power consumption of multiple cells. In various embodiments of this disclosure, the multiple access network devices 110 can change parameter values ​​related to adjusting the states of the multiple access networks. For example, the multiple access network devices 110 can receive parameter values ​​determined by server 101 from server 101, and can adjust the state of each of the multiple access networks by changing the current value of the parameter to the received value.

[0029] According to embodiments of this disclosure, each of the plurality of access network devices 110 may send information about the status of the access network to the server 101 at a preset time point or within a preset time period. For example, each of the plurality of access network devices 110 may send information about the status of the access network to the server 101 at preset time intervals. As another example, each of the plurality of access network devices 110 may send information about the status of the access network to the server 101 in response to a request message received from the server 101.

[0030] Figure 2 This is a block diagram illustrating an example electronic device for adjusting the state of multiple access networks 110 according to various embodiments.

[0031] Reference Figure 2 The electronic device 200 may include a processor (e.g., including processing circuitry) 210, an input unit (e.g., including input circuitry) 220, an output unit (e.g., including output circuitry) 230, a memory 240, and a module unit (e.g., including various processing circuitry and / or executable program elements) 250 comprising multiple operable modules. According to various embodiments of this disclosure, the configuration of the electronic device 200 is not limited to... Figure 2 The configuration shown, and may also include Figure 2 Configurations not shown in the diagram may be omitted. Figure 2 Part of the configuration shown. For example, although Figure 2 All of the multiple operable modules shown are included in module unit 250, but at least some of the modules can also be implemented as software modules stored in memory 240. For example, such as Figure 2 As shown, the status information acquisition module 251 may not be implemented as a separate hardware module included in the module unit 250, but can be stored as a software module in the memory 240 and executed by the processor 210. According to embodiments of this disclosure, the electronic device 200 can be understood to have... Figure 1The configuration of server 101 shown is the same as or similar to that shown.

[0032] Processor 210 may include various processing circuits and be electrically connected to components included in electronic device 200 to perform computational or data processing for controlling and / or communicating with components included in electronic device 200. According to embodiments of this disclosure, processor 210 may load commands or data received from at least one of the other components into memory 240, process the commands or data, and store the resulting data in memory 240.

[0033] Input unit 220 and output unit 230 may include various circuits and are electrically connected to processor 210 to be configured for outputting to another electronic device outside electronic device 200 (e.g., Figure 1 Each access network device 110 shown in the diagram has an interface for sending or receiving data from it. According to embodiments of this disclosure, input unit 220 may include various input circuits and receive information about the status of the access network, such as information received from the access network device via data communication module 255, and output unit 230 may include various output circuits and output values, such as status control parameters to be sent to the access network device via data communication module 255.

[0034] The memory 240 may be electrically connected to the processor 210 and may store commands or data related to the operation of components included in the electronic device 200. According to various embodiments of this disclosure, the memory 240 may store information about the status of network access acquired using the status information acquisition module 251, values ​​of status control parameters acquired using the parameter value determination module 254, or instructions for the aforementioned operations. According to embodiments of this disclosure, where at least some modules included in the module unit 250 are implemented as software to be executed by the processor 210, the memory 240 may store instructions for executing at least some modules.

[0035] Module unit 250 may include multiple modules having various processing circuits and / or executable program elements for implementing multiple operations performed by electronic device 200. According to various embodiments of this disclosure, module unit 250 can be understood as being used to implement... Figure 2 The hardware configuration of at least some of the multiple modules shown, or conceptually understood as including a configuration of multiple hardware units for implementing each module. According to various embodiments of this disclosure, the configuration of module unit 250 is not limited to... Figure 2 The configuration shown in the document, and included Figure 2 At least some of the multiple modules in the module unit 250 shown can be implemented as software to be stored in the memory 240.

[0036] The status information acquisition module 251 may include various processing circuits and / or executable program elements, and is executed by the processor 210 to acquire information about the status of the access network. According to embodiments of this disclosure, the access network device may include at least one cell, and each cell may allocate radio resources to multiple user equipments (UEs) connected to that cell. In various embodiments of this disclosure, radio resources can be understood as frequency resources used for radio communication, which are limited resources that can only be shared by a limited number of users within a preset time period. In embodiments of this disclosure, radio resources may refer to resource blocks (RBs) in orthogonal frequency division multiplexing (OFDM) radio communication.

[0037] According to embodiments of this disclosure, the state of the access network may relate to the allocation of radio resources to multiple UEs by each cell included in the access network device. For example, for each cell included in at least one cell in the access network device, the state of the access network may relate to at least one of the following: the number of UEs currently connected to that cell, physical resource block (PRB) utilization or usage, Internet Protocol (IP) throughput, or downlink (DL) data volume.

[0038] In embodiments of this disclosure, the access network device may include four cells, and the state of the access network may represent the number of UEs currently connected to the first cell, the number of UEs currently connected to the second cell, the number of UEs currently connected to the third cell, and the number of UEs currently connected to the fourth cell. In embodiments of this disclosure, the access network device may include four cells, and the state of the access network may represent the ratio of the total available radio resources allocated to the first cell to the radio resources currently allocated to the multiple UEs connected to the first cell, the ratio of the total available radio resources allocated to the second cell to the radio resources currently allocated to the multiple UEs connected to the second cell, the ratio of the total available radio resources allocated to the third cell to the radio resources currently allocated to the multiple UEs connected to the third cell, and the ratio of the total available radio resources allocated to the fourth cell to the radio resources currently allocated to the multiple UEs connected to the fourth cell. In embodiments of this disclosure, similar to the above examples, the access network device may include four cells, and the state of the access network may represent the IP throughput of the multiple UEs connected to each cell or the downlink data volume of each cell.

[0039] According to embodiments of this disclosure, electronic device 200 can be connected to multiple access network devices, and status information acquisition module 251 can acquire information about the status of the corresponding access network for each of the multiple access network devices.

[0040] According to embodiments of this disclosure, the status information acquisition module 251 can acquire information about the status of the access network at a given time point. For example, the status information acquisition module 251 can acquire information about the status of the access network at a first time point. According to embodiments of this disclosure, the status information acquisition module 251 can acquire information about the status of the access network at a given time point at preset time intervals. For example, the status information acquisition module 251 can acquire information about the status of the access network at a first time point, and can acquire information about the status of the access network at a second time point after a preset time interval following the first time point.

[0041] According to embodiments of this disclosure, the state information acquisition module 251 can acquire the state history information of the access network at a first time point, which is the time point at which the acquisition is performed. For example, the state history information may include the state information of the access network at each of one or more time points within a preset time period earlier than the first time point (i.e., the time point at which the acquisition is performed). For example, the state history information may only include the state of the access network at the first time point, or it may include the state of the access network at the first time point and the state information of the access network at time points at a preset time period earlier than the first time point. In various embodiments of this disclosure, the preset time period may be set to a preset value, which may be set by a user or an artificial intelligence learning model, or it may be changed from the current value.

[0042] The estimated time point determination module 252 may include various processing circuitry and / or executable program elements, and is executed by the processor 210 to determine a first time period required to adjust the state of the access network. According to various embodiments of this disclosure, the first time period may be the sum of at least one of the following: a time period required to obtain information about the state of the access network from the access network device; a time period required to collect the obtained information based on a reference time point; a time period required to estimate the state of the access network at a preset time; a time period required to determine parameter values ​​for optimizing the state of the access network; or a time period required to send the determined parameter values ​​to the access network device.

[0043] In embodiments of this disclosure, the estimation time point determination module 252 can determine a first time period based on the state history information acquired by the state information acquisition module 251 at a first time point. In embodiments of this disclosure, the first time period can be increased based on the amount of state history information. In various embodiments of this disclosure, the amount of state history information can be increased based on the number of access networks, based on the number of cells included in each access network, or based on one or more time points included in the state history information. As the amount of acquired state history information increases, the time period required to analyze and process the state history information increases, and the time period required to determine parameter values ​​for adjusting the state of the access network based on the state history information also increases.

[0044] According to embodiments of this disclosure, the estimated time point determination module 252 can compare the acquired amount of state history information with a preset size to determine a first time period required to adjust the state of the access network. For example, information about the amount of regularly acquired state history information and the corresponding required time period can be input by a user. The estimated time point determination module 252 can compare the acquired amount of state history information with the user-inputted information and then determine the first time period by calculating their ratio. As another example, a table of the amount of state history information and the corresponding estimated first time periods can be pre-stored in the memory 240, and the estimated time point determination module 252 can determine the first time period by selecting the amount closest to the acquired amount of state history information from the table, and then retrieving the first time period corresponding to the selected amount. As another example, the estimated time point determination module 252 can use an artificial intelligence learning model to determine the first time period. The artificial intelligence learning model can be trained using previously acquired state history information and the actual time period spent adjusting the state of the access network, and can be used to determine the first time period.

[0045] According to embodiments of this disclosure, the estimated time point determination module 252 may consider a first time period previously determined under the same or similar conditions to determine the first time period. For example, the estimated time point determination module 252 may consider conditions such as the following same or similar conditions to determine the first time period as one of the previously determined first time periods: the same or similar time zone, the same or similar day of the week, the same or similar season, the same or similar weather, the same or similar events, etc. In various embodiments of this disclosure, the estimated time point determination module 252 may use past statistical data under the same or similar conditions to reduce the time period required to determine the first time period, thereby reducing the load applied to the electronic device 200.

[0046] In embodiments of this disclosure, the estimated time point determination module 252 can determine the first time period as a first time period determined at the same time period of the previous day. For example, the amount of state history information acquired at 1 p.m. today can be equal to or similar to the amount of state history information acquired at 1 p.m. yesterday. Therefore, the estimated time point determination module 252 can reduce the time period or workload required to determine the first time period by determining it as a first time period determined at 1 p.m. yesterday.

[0047] In embodiments of this disclosure, the estimated time point determination module 252 can determine the first time period as the same time period determined on the same day of the previous week. For example, the amount of state history information acquired at 1 p.m. on Saturday can be the same as or similar to the amount of state history information acquired at 1 p.m. on the previous Saturday. Therefore, the estimated time point determination module 252 can reduce the time period or workload required to determine the first time period by determining it as the first time period determined at 1 p.m. on the previous Saturday.

[0048] In embodiments of this disclosure, the estimated time point determination module 252 can determine the first time period as a time period determined at the time when the same or similar events occur. For example, the amount of state history information acquired at the time of a concert, commemorative event, or gathering may be the same or similar to the amount of state history information previously acquired at the time of a concert, commemorative event, or gathering of the same or similar scale. Therefore, the estimated time point determination module 252 can reduce the time period or workload required to determine the first time period by determining it as a first time period previously determined at the time of a concert, commemorative event, or gathering of the same or similar scale.

[0049] According to embodiments of this disclosure, the estimated time point determination module 252 can determine a second time point based on a first time point for acquiring state history information and a determined first time period. This second time point is a reference time point for adjusting the state of the access network. In embodiments of this disclosure, the second time point can be a time point later than the first time point by the first time period. For example, the second time point can be a time point after the first time point, after the first time period and the time period used to determine the first time period. According to various embodiments of this disclosure, because the first time period determined by the estimated time point determination module 252 may differ from the actual time period spent adjusting the state of the access network, the estimated time point determination module 252 can determine the second time point as a time point after the first time point, after the first time period and a sufficiently long time period.

[0050] According to embodiments of this disclosure, the estimated time point determination module 252 can first determine a second time point, which is a reference time point for adjusting the state of the access network, and then determine a first time point for acquiring state history information based on the second time point. In embodiments of this disclosure, the estimated time point determination module 252 can determine the second time point based on a predefined plan. For example, the estimated time point determination module 252 can determine the second time point as a preset time point so that the state of the access network can be adjusted at a preset time point on a certain day of the week or at a preset time point every day. As another example, the estimated time point determination module 252 can determine the second time point at preset time intervals so that the state of the access network can be adjusted according to the preset time intervals. As another example, the estimated time point determination module 252 can determine the second time point as the time point when an event (e.g., a meeting or gathering) is to be held so that the state of the access network can be adjusted at the time of that specific event.

[0051] According to embodiments of this disclosure, when the second time point is determined before the first time point for which state history information is to be acquired, the estimated time point determination module 252 can determine the first time point based on the determined second time point. For example, the estimated time point determination module 252 can determine the first time period by considering a first time period previously determined under the same or similar conditions, and can determine the first time point based on the second time point and the first time period. In embodiments of this disclosure, the first time point can be a time point that is a first time period earlier than the second time point. For example, the first time point can be a time point that is a first time period earlier than the second time point and the sum of the time period used to determine the first time period. According to various embodiments of this disclosure, because the first time period determined by the estimated time point determination module 252 may be different from the actual time period used to adjust the state of the access network, the estimated time point determination module 252 can determine the first time point as a time period that is a sufficiently long time earlier than a time point (which is a first time period earlier than the second time point).

[0052] The state estimation module 253 may include various processing circuits and / or executable program elements, and may be executed by the processor 210 to estimate the state of the access network at a second time point determined by the estimation time point determination module 252. According to embodiments of this disclosure, the state estimation module 253 may estimate the state of the access network at the second time point based on the state information acquisition module 251's acquired state history information at a first time point, using a first artificial intelligence learning model. According to various embodiments of this disclosure, the state estimation module 253 may estimate the state of the access network at multiple time points within a preset time period. For example, the state estimation module 253 may estimate the state of the access network at multiple time points within a preset time period, including the second time point. In embodiments of this disclosure, the multiple time points may include time points earlier than the second time point by a preset time period or time points later than the second time point by a preset time period. The estimated state of the access network at multiple time points may be compared with the actual state of the access network at multiple time points for training the first artificial intelligence learning model. For example, the estimated state at multiple time points may be compared with the actual state, and multiple feedbacks from the first artificial intelligence learning model may be obtained.

[0053] In embodiments of this disclosure, the state estimation module 253 can use historical state information and the time difference between a first time point and a second time point as input to the first artificial intelligence learning model. The state estimation module 253 can use the first artificial intelligence learning model to estimate possible state changes in the access network between the first and second time points, and can obtain the estimated state of the access network at the second time point as the output of the first artificial intelligence learning model. In embodiments of this disclosure, the first and second time points, rather than the time difference between the first and second time points, can be used as input to the first artificial intelligence learning model.

[0054] According to various embodiments of this disclosure, the state of the access network can change over time. For example, the number of UEs connected to each cell in the access network device can increase or decrease over time. Using a first artificial intelligence learning model, the state estimation module 253 can estimate the number of UEs connected to each cell at a second time point based on the historical number of UEs connected to each cell obtained from a first time point and the time interval between the first time point and a second time point.

[0055] According to embodiments of this disclosure, the state estimation module 253 can further utilize the state of the access network at at least one third time point between the first and second time points when estimating the state of the access network at a second time point. For example, after acquiring the state history information at the first time point, before the state estimation module 253 estimates the state of the access network at the second time point, the state information acquisition module 251 can further acquire the state of the access network at at least one third time point. The state estimation module 253 can use the acquired state history information at the first time point and the acquired state of the access network at the third time point to estimate the state of the access network at the second time point. Therefore, the state estimation module 253 can further improve the accuracy of estimating the state of the access network at the second time point.

[0056] According to embodiments of this disclosure, when estimating the state of the access network at a second time point, the state estimation module 253 may further use the acquired state of the access network at at least one third time point between the first and second time points and the value of the state control parameter determined by the parameter value determination module 254 at the third time point. For example, after acquiring the state history information at the first time point, before the state estimation module 253 estimates the state of the access network at the second time point, the state information acquisition module 251 may further acquire the state of the access network at at least one third time point. The parameter value determination module 254 may determine the value of the state control parameter used to adjust the state of the access network based on the acquired state of the access network at the third time point. In embodiments of this disclosure, the state estimation module 253 may use the acquired state of the access network at the third time point and the determined value of the state control parameter at the third time point to estimate the adjusted state of the access network at the third time point. According to embodiments of this disclosure, the state estimation module 253 may use the acquired state history information at the first time point and the adjusted state of the access network at the third time point to estimate the state of the access network at the second time point. Therefore, the state estimation module 253 can further improve the accuracy of estimating the state of the access network at the second time point.

[0057] According to embodiments of this disclosure, when estimating the state of the access network at a second time point, the state estimation module 253 may further use the estimated state of the access network at at least one third time point between the first and second time points. For example, unlike the state estimation of the access network at the second time point, the state estimation module 253 may estimate the state of the access network at the third time point at a time point earlier than the first time point. In embodiments of this disclosure, the state estimation module 253 may use the acquired state history information from the first time point and the estimated state of the access network at the third time point to estimate the state of the access network at the second time point. Therefore, the state estimation module 253 can save the time required to obtain the state of the access network at the third time point from the access network device and can further improve the accuracy of the state estimation of the access network at the second time point.

[0058] According to embodiments of this disclosure, when estimating the state of the access network at a second time point, the state estimation module 253 may further use the estimated state of the access network at at least one third time point between the first and second time points and the value of the state control parameter determined by the parameter value determination module 254 at the third time point. For example, unlike the state estimation of the access network at the second time point, the state estimation module 253 may estimate the state of the access network at the third time point at a time point earlier than the first time point. The parameter value determination module 254 may determine the value of the state control parameter used to adjust the state of the access network based on the estimated state of the access network at the third time point. In embodiments of this disclosure, the state estimation module 253 may use the acquired state of the access network at the third time point and the determined value of the state control parameter at the third time point to estimate the adjusted state of the access network at the third time point. According to embodiments of this disclosure, the state estimation module 253 may use the acquired state history information at the first time point and the adjusted state of the access network at the third time point to estimate the state of the access network at the second time point. Therefore, the state estimation module 253 can save the time required to obtain the state of the access network at the third time point from the access network device, and can further improve the estimation accuracy of the state of the access network at the second time point.

[0059] According to various embodiments of this disclosure, the first artificial intelligence learning model used by the state estimation module 253 can be trained using at least one of, but not limited to, reinforcement learning, supervised learning, machine learning, continuous learning, joint learning, and deep learning. In embodiments of this disclosure, the first artificial intelligence learning model can be trained using, for example, but not limited to, supervised learning, and the state estimation module 253 can use a supervised learning scheme to train the first artificial intelligence learning model. For example, the state estimation module 253 can compare the estimated state of the access network at a second time point with the actual state of the access network at the second time point obtained by the state information acquisition module 251, and can train the first artificial intelligence learning model by providing the difference between the estimated state and the actual state as feedback. The state estimation module 253 can use the first artificial intelligence learning model to estimate the state of the access network at different time points, and can repeatedly compare the estimated state with the actual state obtained by the state information acquisition module 251. The state estimation module 253 can improve the accuracy of the first artificial intelligence learning model by repeating this process.

[0060] According to various embodiments of this disclosure, the state estimation module 253 can train a first artificial intelligence learning model using information about historical states, based on specific conditions of the access network. For example, the state estimation module 253 can use the historical states of the access network on a certain day of the week or at a certain point in time to train the first artificial intelligence learning model. As another example, the state estimation module 253 can use the historical states of the access network to train the first artificial intelligence learning model based on an event (e.g., but not limited to, a meeting, a party, a public holiday, etc.).

[0061] According to various embodiments of this disclosure, the first artificial intelligence learning model may be implemented as a separate module included in module unit 250, stored in memory 240, or included in another electronic device external to electronic device 200, but is not limited thereto. For example, the first artificial intelligence learning model may be included in another electronic device external to electronic device 200 that communicates with data communication module 255 via wired or wireless means, or may be stored in a storage device external to electronic device 200, which is connected to electronic device 200 via input unit 220 or output unit 230 using various interfaces (e.g., USB).

[0062] The parameter value determination module 254 may include various processing circuitry and / or executable program elements, and may be executed by the processor 210, to determine values ​​of state control parameters for adjusting the state of the access network based on the state of the access network estimated by the state estimation module 253 at a second time point. State control parameters can be understood to include parameters that affect the state of the access network. For example, state control parameters can adjust the state of the access network by influencing the adjustment of load applied to a cell or the opening or closing of a cell for multiple cells in the access network equipment. According to embodiments of this disclosure, state control parameters may include, but are not limited to, at least one of, a Reference Signal Received Power (RSRP) threshold, a Reference Signal Received Quality (RSRQ) threshold, a Physical Resource Block (PRB) ratio threshold, and a cell reselection priority.

[0063] According to embodiments of this disclosure, the RSRP threshold or RSRQ threshold can affect the adjustment of cell coverage. For example, regarding a first cell among a plurality of cells, as the RSRP threshold or RSRQ threshold increases, the cell coverage area of ​​the first cell can decrease, while as the RSRP threshold or RSRQ threshold decreases, the cell coverage area of ​​the first cell can increase. As the cell coverage area of ​​the first cell increases, the number of UEs connected to the first cell also increases, and as the cell coverage area of ​​the first cell decreases, the number of UEs connected to the first cell also decreases.

[0064] According to embodiments of this disclosure, the PRB ratio threshold can affect the determination of whether a cell is enabled or disabled, or whether the power amplifier of the access network device 110 is enabled or disabled. In embodiments of this disclosure, if the proportion of radio resources allocated by the cell to a UE connected to the cell is less than the PRB ratio threshold, the power amplifier of the cell or the access network device 110 can be set to disabled. In embodiments of this disclosure, if the proportion of radio resources allocated by a first cell among a plurality of cells is a specific value and the PRB ratio threshold of the first cell is set to be greater than that specific value, the first cell can be disabled.

[0065] According to embodiments of this disclosure, cell reselection priority can affect the number of UEs connected to a cell. For example, as the cell reselection priority of a first cell among a plurality of cells increases, the number of UEs connected to the first cell also increases. As another example, as the cell reselection priority of the first cell decreases, the number of UEs connected to the first cell also decreases.

[0066] According to embodiments of this disclosure, the value of the state control parameter can be determined between a first time point and a second time point. The state information acquisition module 251 acquires the historical state information of the access network at the first time point, and the second time point is a reference time point determined by the estimated time point determination module 252 for which the state of the access network needs to be adjusted.

[0067] According to various embodiments of this disclosure, the values ​​of the state control parameters can be determined based on the purpose of adjusting the state of the access network. For example, the values ​​of the state control parameters can be determined for the purpose of load balancing for each cell of at least one cell included in the access network device. In embodiments of this disclosure, for at least one cell included in the access network device, the values ​​of the state control parameters can be determined such that the number of UEs currently connected to these cells or the proportion of radio resources allocated to these cells is consistent. In embodiments of this disclosure, for at least one cell included in the access network device, the values ​​of the state control parameters can be determined such that the IP throughput or downlink data volume is consistent by adjusting the number of UEs connected to the cell. In various embodiments of this disclosure, it is understood that load balancing for each cell in at least one cell will be used to adjust the state of the access network to reduce the standard deviation of the cell state.

[0068] As another example, the value of the state control parameter can be determined to adjust the power consumption of at least one unit included in the access network device. In embodiments of this disclosure, the value of the state control parameter can adjust the state of the access network to reduce the power consumption of all units in at least one unit. For example, when the number of UEs currently connected to at least one cell is less than a preset level and the total power consumption can be reduced, the value of the state control parameter can be determined such that: UEs currently connected to at least one cell are connected to another cell, and the power amplifier of at least one cell or access network device 110 is turned off. For example, when the proportion of radio resources allocated by at least one cell is less than a preset level and the total power consumption can be reduced, the value of the state control parameter can be determined such that: the radio resource allocation for UEs connected to at least one cell is performed by another cell, and the power amplifier of at least one cell or access network device 110 is turned off.

[0069] According to embodiments of this disclosure, the parameter value determination module 254 can determine the parameter value application time period and the value of the state control parameter. The parameter value application time period can be understood as, for example, the shortest time period required to apply the determined state control parameter value to the access network, or the shortest time period required to apply the determined state control parameter value to the access network without any further updates. In embodiments of this disclosure, the parameter value determination module 254 can determine the parameter value application time period based on the state of the access network estimated by the state estimation module 253 at a second time point. In embodiments of this disclosure, the parameter value determination module 254 can determine the parameter value application time period based on the state history information obtained by the state information acquisition module 251.

[0070] In embodiments of this disclosure, the determined values ​​of the state control parameters and the determined application time period of the parameter values ​​can be sent to the access network device 110. In embodiments of this disclosure, the parameter value determination module 254 may not determine the new value of the state control parameters until the determined application time period has elapsed after the second time point. In embodiments of this disclosure, when the electronic device 200 receives a message from the access network device 110 requesting a new value for the state control parameters, the parameter value determination module 254 can determine the new value of the state control parameters even if the determined application time period has elapsed after the second time point. In embodiments of this disclosure, when the state estimation module 253 has estimated the state of the access network at a new time point according to a preset plan or a new event, the parameter value determination module 254 can determine the new value of the state control parameters even if the determined application time period has elapsed after the second time point.

[0071] According to embodiments of this disclosure, the parameter value determination module 254 can use a second artificial intelligence learning model to determine the values ​​of state control parameters for adjusting the state of the access network, based on the state of the access network estimated by the state estimation module 253 at a second time point. According to embodiments of this disclosure, the parameter value determination module 254 can use the estimated state of the access network at the second time point as input to the second artificial intelligence learning model. Considering purposes such as load balancing and power consumption adjustment, the parameter value determination module 254 can use the second artificial intelligence learning model to estimate and determine a combination of values ​​for the state control parameters used to achieve the purpose of adjusting the state of the access network.

[0072] According to various embodiments of this disclosure, the second artificial intelligence learning model used by the parameter value determination module 254 can be trained using at least one of, for example, but not limited to, reinforcement learning, supervised learning, machine learning, continuous learning, joint learning, and deep learning. In embodiments of this disclosure, the second artificial intelligence learning model can be trained using reinforcement learning, and the parameter value determination module 254 can use a reinforcement learning scheme to train the second artificial intelligence learning model. For example, the parameter value determination module 254 can use the state of the access network obtained by the state information acquisition module 251 at a first time point as the state variable of the reinforcement learning model, use the value of the state control parameter output by the second artificial intelligence learning model as the action variable of the reinforcement learning model, and then use the state information acquisition module 251 to obtain the state of the access network that has applied the determined state control parameter value to calculate the value of the reward variable. The parameter value determination module 254 can train the second artificial intelligence learning model by repeating the above operations at different time points.

[0073] According to various embodiments of this disclosure, the second artificial intelligence learning model can be implemented as a separate module included in module unit 250, stored in memory 240, or included in another electronic device external to electronic device 200. For example, the second artificial intelligence learning model can be included in another electronic device external to electronic device 200 that communicates with data communication module 255 via wired or wireless means, or it can be stored in a storage device external to electronic device 200, which is connected to electronic device 200 via input unit 220 or output unit 230 using various interfaces (e.g., USB).

[0074] The data communication module 255 may include various processing circuitry and / or executable program elements, and may be executed by the processor 210 to send data to or receive data from other electronic devices external to the electronic device 200. For example, the data communication module 255 may receive information about the status of multiple access networks from multiple access network devices. As another example, the data communication module 255 may send information about the values ​​of status control parameters for adjusting the status of multiple access networks to multiple access network devices.

[0075] Figure 3 This is a block diagram illustrating example access network devices according to various embodiments.

[0076] Reference Figure 3The access network device 300 may include a processor (e.g., including processing circuitry) 310, an input unit (e.g., including input circuitry) 320, an output unit (e.g., including output circuitry) 330, a memory 340, and a module unit (e.g., including processing circuitry and / or executable program elements) 350 comprising multiple operable modules. According to various embodiments of this disclosure, the configuration of the access network device 300 is not limited to... Figure 3 The configuration shown, and may also include Figure 3 Configurations not shown in the diagram, or which may be omitted. Figure 3 Part of the configuration shown. For example, although Figure 3 Multiple operable modules are shown to be included in module unit 350, but at least some of these modules can also be implemented as software modules stored in memory 340. For example, parameter control module 351 may be less... Figure 3 Instead of being implemented as a separate hardware module included in module unit 350 as shown, it can be stored as a software module in memory 340 and executed by processor 310 to perform its functions.

[0077] According to embodiments of this disclosure, the access network device 300 can be understood to have the same characteristics as... Figure 1 The first, second, and third access network devices 111, 112, and 113 shown have the same or similar configurations and can be understood as being the same as... Figure 2 The electronic device 200 shown is electrically connected. According to embodiments of this disclosure, the access network device 300 may include at least one cell, and each cell may be connected to one or more UEs to respond to requests from the one or more UEs and to allocate radio resources to the one or more UEs. Figure 2 The descriptions of processor 210, input unit 220, output unit 230 and module unit 250 may be applied in the same or similar manner to processor 310, input unit 320, output unit 330 and module unit 350.

[0078] The parameter control module 351 may include various processing circuits and / or executable program elements, and may be executed by the processor 310 to change the values ​​of state control parameters, thereby adjusting the state of network access. For example, the parameter control module 351 may control the data exchanged with a server (e.g., via the data communication module 355). Figure 2 The value received by the electronic device 200 is applied to the state control parameter to change the value of the state control parameter.

[0079] According to embodiments of this disclosure, the parameter control module 351 can change the value of the state control parameter based on specific conditions. For example, the parameter control module 351 can change the value of the state control parameter based on the result of the state comparison module 353 comparing the estimated state with the actual state of the access network. In embodiments of this disclosure, if the state comparison module 353 has determined that the difference between the estimated state and the actual state is less than a preset level, the parameter control module 351 can change the value of the state control parameter to the value received from the server. In embodiments of this disclosure, if the state comparison module 353 has determined that the difference between the estimated state and the actual state is greater than or equal to a preset level, the parameter control module 351 can maintain the value of the state control parameter without changing the value of the state control parameter to the value received from the server.

[0080] The status monitoring module 352 may include various processing circuits and / or executable program elements, and may be executed by the processor 310 to monitor the status of the access network and obtain the status of the access network at a preset time point. For example, for each cell included in at least one cell in the access network device 300, the status of the access network may be related to at least one of the following: the number of UEs currently connected to the cell, physical resource block (PRB) utilization or usage, Internet Protocol (IP) throughput, downlink (DL) data volume, etc.

[0081] According to embodiments of this disclosure, the status monitoring module 352 can acquire information about the status of the access network at preset time intervals. For example, the status monitoring module 352 can acquire information about the status of the access network at preset time intervals within a preset time period, and can use the acquired information to store historical status information within the preset time period in the memory 340. According to embodiments of this disclosure, the status monitoring module 352 can acquire information about the status of the access network based on a request received from a server. According to embodiments of this disclosure, the status monitoring module 352 can acquire information about the status of the access network based on a request from another module included in the access network device 300. For example, the status monitoring module 352 can acquire information about the status of the access network based on a request from the parameter control module 351 or the status comparison module 353.

[0082] The state comparison module 353 may include various processing circuits and / or executable program elements, and may be executed by the processor 310 to compare the estimated state of the access network at a preset time point with the actual state of the access network at the preset time point. According to embodiments of this disclosure, the estimated state may be based on information obtained from a server. According to embodiments of this disclosure, the actual state may be obtained by the state monitoring module 352.

[0083] According to embodiments of this disclosure, the state comparison module 353 may use separate parameters to represent the difference between the estimated state and the actual state. For example, the state comparison module 353 may generate a vector representing the difference between the estimated state and the actual state. In embodiments of this disclosure, where the access network device 300 includes four cells and the number of UEs currently connected to each cell is used to represent the state of each of the four cells, the vector may include four elements, each representing the difference between the estimated number of UEs currently connected to each cell and the actual number of UEs currently connected to each cell. In embodiments of this disclosure, where the access network device 300 includes four cells and the number of UEs currently connected to each cell and the IP throughput of each cell are used to represent the state of each of the four cells, the vector may include eight elements, representing the difference between the estimated number of UEs currently connected to the cell and the actual number of UEs currently connected to the cell, and the difference between the estimated IP throughput of the cell and the actual IP throughput of the cell. In various embodiments of this disclosure, the number of elements in the vector may depend on the number of cells included in the access network device 300 and the number of variables representing the state of the access network. According to embodiments of this disclosure, the state comparison module 353 can convert the elements of a vector into magnitudes and compare these magnitudes with a threshold. For example, the state comparison module 353 can calculate the mean square or root mean square of the elements of the vector and compare the calculated value with a threshold to determine whether the difference between the estimated state and the actual state is less than a preset level.

[0084] According to embodiments of this disclosure, the state comparison module 353 can send a comparison result between the estimated state and the actual state to the parameter control module 351. For example, the state comparison module 353 can send a result comparing the difference between the estimated state and the actual state with a threshold to the parameter control module 351.

[0085] The threshold setting module 354 may include various processing circuitry and / or executable program elements, and may be executed by the processor 310 to set the threshold used by the state comparison module 353. For example, the threshold may be set to a user-predefined value. As another example, the threshold may be set based on the stability or load level of the access network. For instance, with a relatively high threshold, the frequency of changes in the state control parameter values ​​may be relatively high, which could affect the stability or load level of the access network. On the other hand, with a relatively low threshold, the frequency of changes in the state control parameter values ​​may be relatively low, which could lead to a lack of proper adjustment. As another example, the threshold may be set using a separate artificial intelligence learning model.

[0086] Data communication module 355 may include various processing circuits and / or executable program elements, and may be executed by processor 310 to communicate with electronic devices external to network device 300 (e.g., Figure 2 The electronic device 200 sends data or receives data from it. For example, the data communication module 355 can send information about the status of the access network to the server. As another example, the data communication module 355 can receive information from the server about the values ​​of status control parameters used to adjust the status of the access network.

[0087] Figure 4 This is a flowchart illustrating example methods for adjusting the state of an access network according to various embodiments.

[0088] Reference Figure 4 According to embodiments of this disclosure, electronic devices (e.g., Figure 2 The method for adjusting the state of an access network, performed by an electronic device 200, may include operations 401 to 411. In embodiments of this disclosure, operations 401 to 411 may be performed by a processor of the electronic device (e.g., Figure 2 The processor 210) executes the method. According to various embodiments of this disclosure, the method for adjusting the state of the access network executed by the electronic device is not limited to... Figure 4 The method shown herein may also include... Figure 4 Operations not shown or omitted Figure 4 At least one operation shown in the diagram.

[0089] In Operation 401, the electronic device can access the network from the network device (e.g., at the first point in time). Figure 3 The access network device 300 acquires status history information. This status history information may include, for example, information about the status of the access network at one or more points in time within a preset time period earlier than the first time point. For example, for at least one cell included in the access network device, the information about the status of the access network may relate to at least one of the following: the number of UEs currently connected to the cell, physical resource block (PRB) utilization or usage, Internet Protocol (IP) throughput, downlink (DL) data volume, etc.

[0090] In operation 403, the electronic device may determine a first time period required to adjust the state of the access network. According to various embodiments of this disclosure, the first time period may be the sum of at least one of the following: a time period required to obtain information about the state of the access network from the access network device; a time period required to collect the obtained information based on a reference time point; a time period required to estimate the state of the access network at a preset time point; a time period required to determine parameter values ​​for optimizing the state of the access network; or a time period required to send the determined parameter values ​​to the access network device.

[0091] In operation 405, the electronic device can determine a second time point based on a first time point and a first time period. The second time point can be understood as a reference time point for adjusting the state of network access. In embodiments of this disclosure, the second time point can be later than a time point after the first time point but after the first time period has elapsed. For example, the second time point can be later than a time point after the first time point but after the first time period and the time period used to determine the first time period.

[0092] In operation 407, the electronic device can estimate the state of the access network at a second time point. According to embodiments of this disclosure, the electronic device can estimate the state of the access network at the second time point based on the state history information acquired in operation 401 and a first artificial intelligence learning model. For example, the first artificial intelligence learning model may include an artificial intelligence model for estimating the state of the access network at a time point after a preset time period, and the state history information of the access network may be input into the first artificial intelligence learning model. In embodiments of this disclosure, information about the first time point and information about the second time point may be input into the first artificial intelligence learning model to input a preset time period. In embodiments of this disclosure, information about the time difference between the first and second time points may be input into the first artificial intelligence learning model to input a preset time period.

[0093] In operation 409, the electronic device can determine the value of a state control parameter for adjusting the state of the access network based on the state of the access network estimated in operation 407 at a second time point. According to various embodiments of this disclosure, the electronic device can determine the value of the state control parameter for the purpose of load balancing multiple cells or adjusting the power consumption of multiple cells, based on the estimated state of the access network at a second time point. For example, but not limited to, the electronic device can determine an RSRP threshold, an RSRQ threshold, or a cell reselection priority for a specific cell to distribute the load applied to multiple cells. As another example, the electronic device can determine a PRB ratio threshold for a specific cell to adjust the power consumption of multiple cells.

[0094] According to embodiments of this disclosure, the value of the state control parameter can be determined based on an estimated state of the access network at a second time point and a second artificial intelligence learning model. For example, the second artificial intelligence learning model may include an artificial intelligence model for determining the value of the state control parameter, which is adapted for load balancing or power adjustment for a specific state of the access network. According to embodiments of this disclosure, the second artificial intelligence learning model may receive an estimated state of the access network at a second time point as input to determine the value of the state control parameter for maximizing rewards (e.g., maximizing the amount of load or power reduction allocated through load balancing) in the resulting state of the already adjusted access network.

[0095] In operation 411, the electronic device can send the values ​​of the status control parameters determined in operation 409 to the access network device. The values ​​of the status control parameters can be sent to the access network device to be applied to multiple cells included in the access network device, thereby adjusting the status of the access network.

[0096] Figure 5 This is a signal flow diagram illustrating example operations of an electronic device for adjusting the state of an access network according to various embodiments.

[0097] Reference Figure 5 Electronic devices according to embodiments of the present disclosure (e.g., Figure 2 Electronic devices 200 and Figure 3 The method for adjusting the state of the access network (access network device 300) may include operations 501 to 512. According to embodiments of this disclosure, Figure 5 The SON (Self-Organizing Network) server 200 shown in the figure can be Figure 2 The electronic device 200 is shown in the figure. According to embodiments of this disclosure, operations 501 to 512 can be performed by the processor of the electronic device (e.g., Figure 2 Processor 210 or Figure 3 The processor 310) executes the process. According to various embodiments of this disclosure, the method by which an electronic device adjusts the state of its access network is not limited to... Figure 5 The method shown herein may also include... Figure 5 Operations not shown or omitted Figure 5 At least one operation shown in the diagram.

[0098] In operation 501, the SON server 200 may request the access network device 300 to provide status history information at a first point in time. In embodiments of this disclosure, a request message may be sent from the SON server 200 to the access network device 300, and the access network device 300 may respond to the request message by sending the status history information at the first point in time to the SON server 200. According to embodiments of this disclosure, the access network device 300 may store the status of the access network in a memory at preset time intervals, and may generate status history information based on the status of the access network at a preset number of time points. According to various embodiments of this disclosure, the access network device 300 may periodically generate or update status history information according to preset time intervals, or may generate status history information in response to request messages from the SON server 200. According to embodiments of this disclosure, operation 501 may be omitted. For example, the SON server 200 may not request the access network device 300 to provide status history information, and the access network device 300 may periodically send status history information to the SON server 200 regardless of the request message.

[0099] In operation 503, SON server 200 can determine a first time period required to adjust the state of the access network. In embodiments of this disclosure, the first time period can be determined based on state history information requested by SON server 200 and sent to SON server 200 in operation 501. Figure 4 The description of operation 403 can be applied exactly or similarly to operation 503.

[0100] In Operation 505, SON server 200 can determine a second time point based on a first time point and a first time period. The second time point can be understood as a reference time point for adjusting the status of the access network. Figure 4 The description of Operation 405 can be applied exactly or similarly to Operation 505.

[0101] In operation 507, the SON server 200 may request the access network device 300 to provide the status of the access network at a third time point. According to embodiments of this disclosure, the third time point may be a time point between a first time point and a second time point, and the third time point may be at least one time point. In embodiments of this disclosure, the third time point may be a time point earlier than the time point at which operation 509 is performed. According to embodiments of this disclosure, a request message may be sent from the SON server 200 to the access network device 300, and the access network device 300 may respond to the request message by sending information about the status of the access network at the third time point to the SON server 200. In embodiments of this disclosure, if multiple third time points exist, the access network device 300 may send information about multiple statuses of the access network to the SON server 200. According to embodiments of this disclosure, operation 507 may be omitted. For example, the access network device 300 may periodically send the status of the access network to the SON server 200 at preset time intervals, independent of the request message. In this case, the SON server 200 may not request the access network device 300 to provide the status of the access network at the third time point.

[0102] In operation 509, the SON server 200 can estimate the state of the access network at a second time point. According to embodiments of this disclosure, the SON server 200 can estimate the state of the access network at the second time point based on the historical state information acquired in operation 501 and a first artificial intelligence learning model. Figure 4 The description of operation 407 can be applied in the same or similar way to operation 509.

[0103] According to embodiments of this disclosure, SON server 200 can estimate the state of the access network at a second time point based on the state history information obtained in operation 501, the state of the access network at a third time point obtained in operation 507, and a first artificial intelligence learning model. Because the state of the access network at the third time point is information obtained at a later time point than the first time point when the state history information was obtained, SON server 200 can more accurately estimate the state of the access network at the second time point using the state of the access network at the third time point.

[0104] According to embodiments of this disclosure, information regarding the state of the access network estimated in operation 509 at a second point in time can be sent to the access network device 300.

[0105] In operation 510, the access network device 300 can obtain (e.g., receive) the estimated state of the access network at a second time point from the SON server 200, and can compare the estimated state of the access network at the second time point with the actual state of the access network at the second time point. According to embodiments of this disclosure, the access network device 300 can use separate parameters to represent the difference between the estimated state and the actual state. For example, the access network device 300 can generate a vector representing the difference between the estimated state and the actual state. In embodiments of this disclosure, the elements of the vector can be converted into amplitudes (magnitudes), and these amplitudes can be compared with a preset threshold. In embodiments of this disclosure, if the amplitude is less than the preset threshold, the estimated state can be determined to be reliable, while if the amplitude is greater than or equal to the preset threshold, the estimated state can be determined to be unreliable.

[0106] In operation 511, the SON server 200 can determine the values ​​of state control parameters for adjusting the state of the access network based on the state of the access network estimated in operation 509 at a second time point. According to embodiments of this disclosure, the values ​​of the state control parameters can be determined based on the estimated state of the access network at the second time point and a second artificial intelligence learning model. Figure 4 The description of operation 409 can be applied in the same or similar way to operation 511.

[0107] According to an embodiment of this disclosure, the value of the state control parameter determined in operation 511 can be sent to the access network device 300.

[0108] In operation 512, access network device 300 may apply the value of the state control parameter based on a comparison between the state of the access network estimated in operation 510 at a second time point and the actual state of the access network at the second time point. In embodiments of this disclosure, if the estimated state is determined to be reliable based on the comparison result in operation 510, access network device 300 may apply the value of the state control parameter obtained from SON server 200 to multiple cells. In embodiments of this disclosure, if the estimated state is determined to be unreliable based on the comparison result in operation 510, access network device 300 may maintain the existing value of the state control parameter and not apply the value of the state control parameter obtained from SON server 200.

[0109] According to various embodiments of this disclosure, the order of operations 501 to 512 is not limited to... Figure 5The order shown is as follows. For example, although the information about the estimated state at the second time point as a result of operation 509 is illustrated as being sent from the SON server 200 to the access network device 300 before operations 510 and 511, the information about the estimated state at the second time point can be sent to the access network device 300 after operation 511 along with the determined values ​​of the state control parameters. In this case, operation 510 can be performed after operation 511.

[0110] According to various embodiments of this disclosure, although in Figure 5 Although not shown in the diagram, the SON server 200 can perform the operations of training the first artificial intelligence learning model and / or the second artificial intelligence learning model independently of operations 501 to 512. For example, the SON server 200 can train the first artificial intelligence learning model and / or the second artificial intelligence learning model between operations 501 to 512, before operation 501, or after operation 512. The following will refer to... Figure 8 and Figure 9 The methods for training the first and second artificial intelligence learning models are described in more detail.

[0111] Figure 6 This is a diagram illustrating example methods for calculating the difference between the estimated state and the actual state of an access network according to various embodiments.

[0112] Reference Figure 6 SON server 200 (for example, Figure 2 The electronic device 200 can estimate the state of the access network at time point t+Δ, and the access network device 300 can acquire the actual (e.g., real) state of the access network at time point t+Δ. In embodiments of this disclosure, t can be understood as a first time point described throughout the disclosure, Δ can be understood as a first time period described throughout the disclosure, and t+Δ can be understood as a second time point described throughout the disclosure.

[0113] According to embodiments of this disclosure, the estimated state of the access network at a second time point can be determined by vector 610. This indicates that the actual state of the access network at the second time point can be represented by vector 620. Representation. For example, vector 610 The number of elements and the vector 620 The number of elements can be the same, and the vector is 610. Sum of vectors 620 The number of elements n for each element in the vector can be based on the number of cells included in the access network device 300 and the number of elements related to the access network status. For example, if the access network device 300 includes four cells and the access network status information includes two elements for each cell (these two elements are the number of UEs currently connected to each cell and the proportion of radio resources allocated to each cell), then the vector... 610 and vector The number of elements n for each of 620 can be 8. As another example, in the case where the access network device 300 includes six cells and the information about the access network's state includes only one element for each cell (this element is the number of UEs currently connected to each cell), the vector... 610 and vector The number of elements n for each of the 620 can be 6.

[0114] The estimated state and actual state of the access network at a second time point can be compared to determine the reliability of the first artificial intelligence learning model that estimates the state of the access network at that second time point. For example, this can be achieved by calculating vectors. 610 and vectors The vector is calculated by the difference between each pair of elements with the same index in 620. 630. In embodiments of this disclosure, vector The elements of 630 can be converted into amplitudes, and these amplitudes can be compared to a threshold. For example, a vector The amplitude of 630 can be determined by vector. The elements of 630 are represented by their mean square or root mean square, and can be compared with a threshold.

[0115] Figure 7a This is a diagram illustrating examples of methods for adjusting the state of an access network according to various embodiments. Figure 7b This is a diagram illustrating examples of methods for adjusting the state of an access network according to various embodiments.

[0116] Reference Figure 7a and Figure 7b The SON server 200 and the access network device 300 may have information about the state of the access network at each point in time. According to embodiments of this disclosure, information about the state of the access network at each point in time can be sent from the SON server 200 and the access network device 300, and can be used to adjust the state of the access network.

[0117] According to various embodiments of this disclosure Figure 7a or Figure 7b The adjustment of the access network state shown is just an example, and the methods for adjusting the access network state are not limited to this. Figure 7a or Figure 7b The methods shown in the example. For example, although Figure 7a and Figure 7b The example shows that the number of multiple units is 4, but the number of multiple units is not limited to this and can be any positive integer. As another example, although... Figure 7a and Figure 7b The number of UEs currently connected to each cell is shown as being used to represent the state of the access network. However, the state of the access network can be represented using different types of variables, and therefore the number of variables can be determined accordingly. For example, for each cell, the state of the access network may include at least one of the following: the number of UEs currently connected to that cell, the utilization or usage of physical resource blocks (PRBs), the throughput of Internet Protocol (IP), or the amount of downlink (DL) data.

[0118] According to embodiments of this disclosure, the access network device 300 can obtain the status 701 of the access network at a first time point, and the status 701 of the access network at the first time point can be sent to the SON server 200. According to embodiments of this disclosure, at the first time point, the number of UEs connected to the first cell can be 80, the number of UEs connected to the second cell can be 50, the number of UEs connected to the third cell can be 50, and the number of UEs connected to the fourth cell can be 20.

[0119] According to embodiments of this disclosure, SON server 200 can estimate the state 710 of the access network at a second time point based on the state 701 of the access network at a first time point. In embodiments of this disclosure, SON server 200 can estimate the number of UEs connected to the first cell at the second time point as 45, the number of UEs connected to the second cell at the second time point as 50, the number of UEs connected to the third cell at the second time point as 50, and the number of UEs connected to the fourth cell at the second time point as 55.

[0120] According to embodiments of this disclosure, the SON server 200 can determine the values ​​of state control parameters, thereby enabling load balancing of multiple cells based on the estimated state 710 of the access network at a second time point. In embodiments of this disclosure, the SON server 200 can determine a cell state adjustment value 711 based on the estimated state 710 of the access network at a second time point. The values ​​of the state control parameters can be determined based on the state adjustment value 711 determined by the SON server 200. According to embodiments of this disclosure, the SON server 200 can determine the values ​​of the state control parameters to increase the number of UEs connected to the first cell by 5 and decrease the number of UEs connected to the fourth cell by 5.

[0121] According to embodiments of this disclosure, SON server 200 can send the values ​​of state control parameters determined based on the estimated state 710 and state adjustment value 711 of the access network at a second point in time to access network device 300.

[0122] In embodiments of this disclosure, reference is made to Figure 7a The access network device 300 can obtain the actual state 720a of the access network at a second time point. According to embodiments of this disclosure, the access network device 300 can compare the estimated state 710 of the access network at the second time point with the actual state 720a of the access network at the second time point, and can calculate a vector representing the difference between the two. 730a. According to embodiments of this disclosure, a vector representing the difference between estimated state 710 and actual state 720a. The elements in 730a can be converted into amplitudes, and these amplitudes can be compared to thresholds. For example, vectors The amplitude of the 730a can be a vector. The root mean square of the elements in 730a, and the threshold can be set to, for example, 5. In this case, the vector The root mean square of the elements in 730a can be And it can be determined to be below a threshold of 5. In embodiments of this disclosure, in vector If the amplitude of 730a is less than the threshold, the access network device 300 can determine that the information obtained from the SON server 200 is reliable. The access network device 300 can apply the values ​​of the state control parameters obtained from the SON server 200 to adjust the number of UEs connected to each cell. In embodiments of this disclosure, the final state 740a of the access network can be changed by adjusting the state 720a of the access network at a second time point. For example, the number of UEs connected to the first cell can be adjusted from 40 to 45, and the number of UEs connected to the fourth cell can be adjusted from 60 to 55.

[0123] In embodiments of this disclosure, reference is made to Figure 7b The access network device 300 can obtain the actual state 720b of the access network at a second time point. According to embodiments of this disclosure, the access network device 300 can compare the estimated state 710 of the access network at the second time point with the actual state 720b of the access network at the second time point, and can calculate a vector representing the difference between the two. 730b. According to embodiments of this disclosure, a vector representing the difference between estimated state 710 and actual state 720b. The elements of 730b can be converted into amplitudes, and these amplitudes can be compared to a threshold. For example, a vector The amplitude of the 730b can be a vector. The root mean square of the elements of 730b, and the threshold can be set to, for example, 5. In this case, the vector The root mean square of the elements in 730b can be And it can be determined to be greater than a threshold of 5. In embodiments of this disclosure, in vector If the amplitude of 730b exceeds the threshold, the access network device 300 can determine that the information obtained from the SON server 200 is unreliable. Therefore, the access network device 300 may not apply the values ​​of the state control parameters obtained from the SON server 200, and can maintain the number of UEs connected to each cell in the final state 740b of the access network. In this case, the access network device 300 can send a message requesting the SON server 200 to provide new values ​​for the state control parameters.

[0124] Figure 8 This is a diagram illustrating example methods for training a first artificial intelligence learning model according to various embodiments.

[0125] According to embodiments of this disclosure, the first artificial intelligence learning model 801 can use at least a portion of the state history information 810 as input. According to various embodiments of this disclosure, the state history information 810 input to the first artificial intelligence learning model 801 may include state history information containing the state of the access network acquired within a preset time period (e.g., within a week) and / or information about the state of the access network under predetermined conditions acquired within the preset time period, such as state history information containing the state of the access network on a certain day of the week or at a preset time point, acquired last month. According to embodiments of this disclosure, the state history information input to the first artificial intelligence learning model 801 may include state history information containing the state of the access network according to an event (e.g., a meeting, party, or holiday).

[0126] According to embodiments of this disclosure, a first artificial intelligence learning model 801 can be repeatedly trained and tested by receiving various state history information 810 as input. For example, the first artificial intelligence learning model 801 can use information 811 extracted from the state history information 810 regarding the state of the access network during a certain time period to estimate the state of the access network at a preset time point after that time period. The first artificial intelligence learning model 801 can obtain the actual state 812 of the access network at the preset time point from the state history information 810. The first artificial intelligence learning model 801 can be tested by comparing the estimated state 820 with the obtained actual state 812. The comparison result (e.g., DIFF) 830 can be used as feedback, and accordingly, the first artificial intelligence learning model 801 can be trained. The first artificial intelligence learning model 801 can be trained by repeatedly using various time periods and the corresponding states of these time periods in the state history information 810 to estimate and test the state of the access network at various time points.

[0127] Figure 9 This is a diagram illustrating example methods for training a first artificial intelligence learning model and a second artificial intelligence learning model according to various embodiments.

[0128] According to embodiments of this disclosure, as referred to above... Figure 8 As described, a first artificial intelligence learning model 801 can be trained by acquiring state history information from the access network device 300 and estimating the state of the access network. For example, the first artificial intelligence learning model 801 can be trained by using partial state history information to estimate the state of the access network at a certain point in time and comparing the estimated state with the actual state of the access network at that point in time in the state history information.

[0129] According to embodiments of this disclosure, the state of the access network estimated by the first artificial intelligence learning model 801 can be input as a state variable into the second artificial intelligence learning model 901, thereby enabling training of the second artificial intelligence learning model 901. According to embodiments of this disclosure, the second artificial intelligence learning model 901 can determine the value of a state control parameter used to adjust the estimated state as an action variable based on the state of the access network estimated by the first artificial intelligence learning model 801 at a certain point in time. The second artificial intelligence learning model 901 can calculate a reward variable by obtaining the state of the access network from the access network device 300 at a time later than the specific time point. The reward variable may include, for example, the amount of load allocated or the amount of power reduced after adjusting the state of the access network. In embodiments of this disclosure, the second artificial intelligence learning model 901 can determine the value of the state control parameter to increase the reward variable, preferably to maximize the reward variable, and can be trained by repeating the above process.

[0130] According to an example embodiment of this disclosure, a method performed by an electronic device to adjust the state of an access network comprising at least one cell may include: acquiring state history information of the access network at a first time point; determining a first time period required to adjust the state of the access network based on the acquired state history information; determining a second time point based on the first time point and the first time period, the second time point being a reference time point at which the state of the access network is to be adjusted; estimating the state of the access network at the second time point based on the acquired state history information; determining a value of a state control parameter for adjusting the state of the access network based on the estimated state of the access network at the second time point, and sending the determined value of the state control parameter to the access network, wherein the value of the state control parameter sent to the access network is applied to the state control parameter of the access network at the second time point to adjust the state of the access network.

[0131] According to an example embodiment of this disclosure, the method may further include obtaining the state of the access network at a third time point between the first time point and the second time point, and the estimation of the state of the access network at the second time point may be performed by estimating the state of the access network at the second time point based on the obtained state history information and the obtained state of the access network at the third time point.

[0132] According to an example embodiment of this disclosure, the state history information may correspond to the first state history information. The method may further include: acquiring the second state history information of the access network at a fourth time point earlier than the first time point, and estimating the state of the access network at at least one third time point between the first time point and the second time point based on the second state history information. The estimation of the state of the access network at the second time point may be performed by estimating the state of the access network at the second time point based on the acquired first state history information and the estimated state of the access network at the third time point.

[0133] According to an example embodiment of this disclosure, the method may further include: determining the value of a state control parameter for adjusting the state of the access network at a third time point based on the estimated state of the access network at a third time point; and estimating the adjusted state of the access network at a third time point based on the estimated state of the access network at the third time point and the determined value of the state control parameter for adjusting the state of the access network at the third time point, wherein the estimation of the state of the access network at a second time point can be performed by estimating the state of the access network at the second time point based on the acquired first state history information and the estimated adjusted state of the access network at the third time point.

[0134] According to an example embodiment of this disclosure, the value of the state control parameter can be determined by the following operation: determining the value of the state control parameter during a time period between a first time point and a second time point.

[0135] According to an example embodiment of this disclosure, the method may further include: sending information about the estimated state of the access network at a second time point to the access network; the access network comparing the estimated state of the access network at the second time point with the actual state of the access network at the second time point; and based on the comparison result, applying the value of the state control parameter to the state control parameter of the access network at the second time point.

[0136] According to an example embodiment of this disclosure, the method may further include: calculating an error value, which is the difference between the estimated state of the access network at a second time point and the actual state of the access network at the second time point, wherein applying the value of the state control parameter can be performed by: applying the value of the state control parameter to the state control parameter of the access network at the second time point based on the calculated error value being less than a preset threshold.

[0137] According to an example embodiment of this disclosure, for at least one cell, the state of the access network may include at least one of the following: the number of UEs currently connected to the cell, physical resource block (PRB) utilization or usage, Internet Protocol (IP) throughput, or downlink (DL) data volume.

[0138] According to an example embodiment of this disclosure, the estimation of the state of the access network at a second time point can be performed using a first artificial intelligence learning model, the determination of the value of the state control parameter can be performed using a second artificial intelligence learning model, and each of the first and second artificial intelligence learning models can be trained by at least one of reinforcement learning, supervised learning, machine learning, continuous learning, joint learning or deep learning.

[0139] According to an example embodiment of this disclosure, the value of the state control parameter can be determined by: determining the value of the state control parameter used for load balancing of each cell in at least one cell in the access network.

[0140] According to an example embodiment of this disclosure, the value of the state control parameter can be determined by: determining the value of the state control parameter used to adjust the power consumption of each cell in at least one cell in the access network.

[0141] According to an example embodiment of this disclosure, the method may further include: acquiring the state of the access network at a second time point, wherein the first artificial intelligence learning model may be trained based on historical state information, the estimated state of the access network at the second time point, and the acquired state of the access network at the second time point.

[0142] According to an example embodiment of this disclosure, the method may further include: acquiring the state of the access network at a fourth time point later than the second time point, wherein the second artificial intelligence learning model may be trained based on the estimated state of the access network at the second time point, the determined values ​​of the state control parameters, and the acquired state of the access network at the fourth time point.

[0143] According to example embodiments of this disclosure, state control parameters may include at least one of a Reference Signal Received Power (RSRP) threshold, a Reference Signal Received Quality (RSRQ) threshold, a Physical Radio Block (PRB) ratio threshold, or a cell reselection priority.

[0144] According to an example embodiment of this disclosure, an electronic device configured to adjust the state of an access network including at least one cell may include: at least one processor and a memory connected to the at least one processor, wherein the at least one processor may be configured to control the electronic device to: acquire state history information of the access network at a first time point; determine a first time period required to adjust the state of the access network based on the acquired state history information; determine a second time point based on the first time point and the first time period, wherein the second time point is a reference time point for adjusting the state of the access network; estimate the state of the access network at a second time point based on the acquired state history information; determine a value of a state control parameter for adjusting the state of the access network based on the estimated state of the access network at the second time point; and send the determined value of the state control parameter to the access network, wherein the value of the state control parameter sent to the access network is applied to the state control parameter of the access network at the second time point to adjust the state of the access network.

[0145] According to an example embodiment of this disclosure, at least one processor may also be configured to control electronic devices to: acquire the state of the access network at a third time point between a first time point and a second time point, and estimate the state of the access network at a second time point based on the acquired state history information and the acquired state of the access network at the third time point.

[0146] According to an example embodiment of this disclosure, the state history information may correspond to the first state history information, and at least one processor may also be configured to control electronic devices to: acquire second state history information of the access network at a fourth time point earlier than the first time point, estimate the state of the access network at at least one third time point between the first time point and the second time point based on the second state history information, and estimate the state of the access network at a second time point based on the acquired first state history information and the estimated state of the access network at the third time point.

[0147] According to an example embodiment of this disclosure, at least one processor may also be configured to: determine the value of a state control parameter for adjusting the state of the access network at the third time point based on the estimated state of the access network at the third time point; estimate the adjusted state of the access network at the third time point based on the estimated state of the access network at the third time point and the determined value of the state control parameter for adjusting the state of the access network at the third time point; and estimate the state of the access network at a second time point based on the acquired first state history information and the adjusted estimated state of the access network at the third time point.

[0148] According to an example embodiment of this disclosure, the at least one processor may also be configured to determine the value of the state control parameter during a time period between a first time point and a second time point.

[0149] According to an example embodiment of this disclosure, at least one processor may also be configured to control electronic devices to: send information to the access network about an estimated state of the access network at a second time point; the access network may compare the estimated state of the access network at the second time point with the actual state of the access network at the second time point; and based on the comparison result, apply the value of a state control parameter to the state control parameter of the access network at the second time point.

[0150] According to an example embodiment of this disclosure, an error value can be calculated, which is the difference between the estimated state of the access network at a second time point and the actual state of the access network at the second time point. Based on the fact that the calculated error value is less than a preset threshold, the value of the state control parameter can be applied to the state control parameter of the access network at the second time point.

[0151] According to an example embodiment of this disclosure, for at least one cell, the state of the access network may include at least one of the following: the number of UEs currently connected to the cell, physical resource block (PRB) utilization or usage, Internet Protocol (IP) throughput, or downlink (DL) data volume.

[0152] According to an example embodiment of this disclosure, a first artificial intelligence learning model can be used to estimate the state of the access network at a second time point, a second artificial intelligence learning model can be used to determine the value of the state control parameter, and each of the first and second artificial intelligence learning models can be trained by at least one of reinforcement learning, supervised learning, machine learning, continuous learning, joint learning or deep learning.

[0153] According to an example embodiment of this disclosure, at least one processor may also be configured to control electronic devices to: acquire the state of the access network at a second time point, wherein a first artificial intelligence learning model may be trained based on state history information, the estimated state of the access network at the second time point, and the acquired state of the access network at the second time point.

[0154] According to an example embodiment of this disclosure, at least one processor may also be configured to control electronic devices to: acquire the state of the access network at a fourth time point later than the second time point, and a second artificial intelligence learning model may be trained based on the estimated state of the access network at the second time point, the determined values ​​of the state control parameters, and the acquired state of the access network at the fourth time point.

[0155] According to example embodiments of this disclosure, state control parameters may include at least one of a Reference Signal Received Power (RSRP) threshold, a Reference Signal Received Quality (RSRQ) threshold, a Physical Radio Block (PRB) ratio threshold, or a cell reselection priority.

[0156] The various embodiments of this disclosure and the terminology used therein are not intended to limit the technology disclosed herein to the specific embodiments, and they should be construed as including all changes, equivalents, or substitutions to the embodiments of this disclosure. Regarding the description of the drawings, similar reference numerals may be used for similar elements. Singular expressions also include plural meanings, unless inconsistent with the context. In this disclosure, expressions such as “A or B,” “at least one of A and / or B,” “A, B, or C,” and “at least one of A, B, and C” include all possible combinations of the listed items. Expressions such as “first” and “second” as used in this disclosure may denote corresponding elements regardless of order and / or importance, and may be used to distinguish one element from another, but do not limit these elements. When an element (e.g., a first element) is described as “operably or communicatively” “coupled,” “coupled to another element,” “connected to another element,” or “connected to another element,” it should be understood that the element can be directly connected to the other element or connected to the other element through another element (e.g., a third element).

[0157] As used in this disclosure, the term "module" can include a unit embodied as hardware, software, or firmware, or any combination thereof, and can be used in conjunction with terms such as logic, logic block, component, circuit, or similar terms. A "module" can be a component of a whole configuration or a minimum unit or part thereof that performs one or more functions. For example, a module can be embodied as an application-specific integrated circuit (ASIC).

[0158] Various embodiments of this disclosure can be embodied as software including instructions stored on a machine-readable storage medium (e.g., internal or external memory) that are readable by a machine (e.g., a computer). The machine can be a device that invokes instructions stored in the storage medium and can operate according to the invoked instructions, including electronic devices according to the disclosed embodiments. When the instructions are executed by a processor, the processor can directly perform the function corresponding to the instructions, or perform the function corresponding to the instructions using other components under the control of the processor. The instructions may include code generated by a compiler or code executed by an interpreter. The machine-readable storage medium may be provided in the form of a non-transitory storage medium. A non-transitory storage medium does not refer to a transient electrical signal, but rather to a tangible medium, and does not distinguish whether data is stored semi-permanently or temporarily on the storage medium.

[0159] According to embodiments of this disclosure, methods according to various embodiments of this disclosure can be provided in a computer program product. The computer program product can be traded as a commodity between a seller and a buyer. The computer program product can be distributed in the form of a machine-readable storage medium (e.g., an optical disc read-only memory (CD-ROM)) or through an app store (e.g., Google Play). TM Online distribution. In the case of online distribution, at least a portion of the computer program product may be temporarily stored or temporarily created on storage media such as the manufacturer's server, the application store's server, or the storage of a relay server.

[0160] Each element (e.g., a module or program) according to various embodiments of this disclosure may include a single entity or multiple entities. Some sub-elements of the aforementioned sub-elements may be omitted, or other elements may be further included in the various embodiments. Alternatively, or additionally, some elements (e.g., modules or programs) may be integrated into one entity to perform the same or similar functions performed by each element prior to integration. According to various embodiments of this disclosure, operations performed by modules, programs, or other elements may be performed sequentially, in parallel, repeatedly, or heuristically, or at least some operations may be performed in a different order, or omitted, or another operation may be added.

[0161] The artificial intelligence-related functions disclosed herein can be executed by a processor and memory. The processor may include one or more processors. In this case, one or more processors may include, for example, a general-purpose processor (such as a central processing unit (CPU), application processor (AP), or digital signal processor (DSP)), a dedicated graphics processor (such as a graphics processing unit (GPU), a vision processing unit (VPU)), or a dedicated artificial intelligence processor (such as a neural processing unit (NPU)). One or more processors can control input data to process the input data according to predefined operating rules or an artificial intelligence model stored in memory. When one or more processors are dedicated artificial intelligence processors, the dedicated artificial intelligence processors can be designed with hardware architectures specifically designed to process a particular artificial intelligence model.

[0162] Predefined operating rules or artificial intelligence models can be generated through a training process. This can refer, for example, to predefined operating rules or artificial intelligence models configured to operate according to desired features (or purposes), generated by training a basic artificial intelligence model using a learning algorithm that utilizes a large amount of training data. The training process can be performed by a device performing artificial intelligence or a separate server and / or system. Examples of learning algorithms can include, but are not limited to, supervised learning, unsupervised learning, semi-supervised learning, reinforcement learning, etc.

[0163] Artificial intelligence models can include multiple neural network layers. Each neural network layer can include multiple weight values, and neural network arithmetic operations can be performed by arithmetic operations between the results of arithmetic operations on the previous layer and the multiple weight values. The multiple weight values ​​in each neural network layer can be optimized based on the results of training the artificial intelligence model. For example, the multiple weight values ​​can be refined to reduce or minimize the loss or cost of the artificial intelligence model during training. Artificial neural networks can include, for example, deep neural networks (DNNs), and can include, for example, convolutional neural networks (CNNs), deep neural networks (DNNs), recurrent neural networks (RNNs), restricted Boltzmann machines (RBMs), deep belief networks (DBNs), bidirectional recurrent deep neural networks (BRDNNs), deep Q-networks (DQNs), etc., but are not limited to these.

[0164] In a method for adjusting the state of an access network comprising at least one cell according to the electronic device of this disclosure, an artificial intelligence model can be used to optimize the state or state control parameter values ​​of the access network at a second time point using historical state information of the access network, inferring or estimating the state or state control parameter of the access network. The processor can preprocess the data to transform it into a form suitable as input to the artificial intelligence model. The artificial intelligence model can be generated through a training process. This can refer, for example, to predefined operating rules configured to operate according to desired features (or purposes), or the artificial intelligence model can be generated by training a basic artificial intelligence model using a learning algorithm utilizing a large amount of training data. The artificial intelligence model can include multiple neural network layers. Each neural network layer can include multiple weight values, and neural network arithmetic operations can be performed through arithmetic operations between the arithmetic results of the previous layer and the multiple weight values.

[0165] Reasoning / prediction can refer to information judgment techniques used for logical reasoning and prediction, and includes knowledge / probability-based reasoning, optimization prediction, preference-based planning, recommendation, etc.

[0166] According to embodiments of this disclosure, the impact of changes in the state of the access network can be reduced, which may be caused by the time difference between the point in time when data is collected from the access network device and the point in time when the optimal configuration is determined. Therefore, the reliability of the determined configuration can be improved, and the load balancing effect or power consumption effect of the access network device can be improved.

[0167] While this disclosure has been described and illustrated with reference to various exemplary embodiments, it is to be understood that these exemplary embodiments are illustrative and not restrictive. Those skilled in the art will further understand that various changes in form and detail may be made without departing from the true spirit and full scope of this disclosure (including the appended claims and their equivalents).

Claims

1. A method performed by an electronic device to adjust the state of an access network comprising at least one cell, the method comprising: Obtain the historical status information of the access network at the first point in time; Based on the acquired historical state information, a first time period required to adjust the state of the access network is determined; A second time point is determined based on the first time point and the first time period. The second time point is a reference time point for adjusting the state of the access network. Based on the acquired historical state information, estimate the state of the access network at the second point in time; Based on the estimated state of the access network at the second time point, the values ​​of the state control parameters used to adjust the state of the access network are determined; as well as The determined value of the state control parameter is sent to the access network. The value of the state control parameter sent to the access network is applied to the state control parameter of the access network at the second time point to adjust the state of the access network.

2. The method according to claim 1, further comprising: Obtain the status of the access network at a third time point between the first time point and the second time point, and The estimation of the state of the access network at the second time point is performed by estimating the state of the access network at the second time point based on the acquired state history information and the acquired state of the access network at the third time point.

3. The method according to claim 1, wherein, The state history information corresponds to the first state history information, and The method further includes: Obtain the second state history information of the access network at a fourth time point earlier than the first time point; and Based on the second state historical information, the state of the access network at at least one third time point between the first time point and the second time point is estimated, and The estimation of the state of the access network at the second time point is performed by the following operation: based on the acquired first state history information and the estimated state of the access network at the third time point, the state of the access network at the second time point is estimated.

4. The method according to claim 3, further comprising: Based on the estimated state of the access network at the third time point, the value of the state control parameter used to adjust the state of the access network at the third time point is determined; as well as Based on the estimated state of the access network at the third time point and the determined values ​​of the state control parameters used to adjust the state of the access network at the third time point, the adjusted state of the access network at the third time point is estimated. The estimation of the state of the access network at the second time point is performed by the following operation: based on the acquired first state history information and the estimated adjusted state of the access network at the third time point, the state of the access network at the second time point is estimated.

5. The method according to claim 1, wherein, The value of the state control parameter is determined by performing the following operation: determining the value of the state control parameter within the time period between the first time point and the second time point.

6. The method according to claim 5, further comprising: Send information about the estimated state of the access network at the second point in time to the access network; The estimated state of the access network at the second time point is compared with the actual state of the access network at the second time point. as well as Based on the comparison results, the value of the state control parameter is applied to the state control parameter of the access network at the second time point.

7. The method according to claim 6, further comprising: Calculate the error value, which is the difference between the estimated state of the access network at the second time point and the actual state of the access network at the second time point. The application of the state control parameter value includes: applying the state control parameter value to the state control parameter of the access network at the second time point based on the calculated error value being less than a preset threshold.

8. The method according to claim 1, wherein, For each of the at least one cells, the state of the access network includes at least one of the following: the number of user equipment (UE) currently connected to the cell, physical resource block (PRB) utilization or usage, Internet Protocol (IP) throughput, or downlink DL data volume.

9. The method according to claim 1, wherein, The state of the access network at the second time point was estimated using a first artificial intelligence learning model. The values ​​of the state control parameters are determined using a second artificial intelligence learning model, and Both the first artificial intelligence learning model and the second artificial intelligence learning model have been trained using at least one of reinforcement learning, supervised learning, machine learning, continuous learning, joint learning, or deep learning.

10. The method according to claim 1, wherein, The value of the state control parameter is determined by the following operation: determining the value of the state control parameter for load balancing each of the at least one cell in the access network.

11. The method according to claim 1, wherein, The value of the state control parameter is determined by the following operation: determining the value of the state control parameter for adjusting the power consumption of each of the at least one cell in the access network.

12. The method according to claim 9, further comprising: Obtain the status of the access network at the second time point. The first artificial intelligence learning model is trained based on the state history information, the estimated state of the access network at the second time point, and the acquired state of the access network at the second time point.

13. The method according to claim 9, further comprising: Obtain the status of the access network at a fourth time point that is later than the second time point. The second artificial intelligence learning model is trained based on the estimated state of the access network at the second time point, the determined values ​​of the state control parameters, and the acquired state of the access network at the fourth time point.

14. The method according to claim 1, wherein, The state control parameters include at least one of the following: Reference Signal Received Power (RSRP) threshold, Reference Signal Received Quality (RSRQ) threshold, Physical Resource Block (PRB) ratio threshold, or cell reselection priority.

15. An electronic device configured to adjust the state of an access network comprising at least one cell, the electronic device comprising: At least one processor; as well as Memory connected to the at least one processor, The at least one processor is configured to control the electronic device to: Obtain the historical status information of the access network at the first point in time; Based on the acquired historical state information, a first time period required to adjust the state of the access network is determined; A second time point is determined based on the first time point and the first time period. The second time point is a reference time point for adjusting the state of the access network. Based on the acquired historical state information, estimate the state of the access network at the second point in time; Based on the estimated state of the access network at the second time point, determine the values ​​of the state control parameters used to adjust the state of the access network; and The determined value of the state control parameter is sent to the access network. The value of the state control parameter sent to the access network is applied to the state control parameter of the access network at the second time point to adjust the state of the access network.

16. The electronic device according to claim 15, wherein, The at least one processor is configured to control the electronic device to: Obtain the status of the access network at a third time point between the first time point and the second time point, and The estimation of the state of the access network at the second time point is performed by estimating the state of the access network at the second time point based on the acquired state history information and the acquired state of the access network at the third time point.

17. The electronic device according to claim 15, in, The state history information corresponds to the first state history information, and The at least one processor is configured to control the electronic device to: Obtain the second state history information of the access network at a fourth time point earlier than the first time point; and Based on the second state historical information, the state of the access network at at least one third time point between the first time point and the second time point is estimated, and The estimation of the state of the access network at the second time point is performed by the following operation: based on the acquired first state history information and the estimated state of the access network at the third time point, the state of the access network at the second time point is estimated.

18. The electronic device according to claim 17, in, The at least one processor is configured to control the electronic device to: Based on the estimated state of the access network at the third time point, the value of the state control parameter used to adjust the state of the access network at the third time point is determined; as well as Based on the estimated state of the access network at the third time point and the determined values ​​of the state control parameters used to adjust the state of the access network at the third time point, the adjusted state of the access network at the third time point is estimated. The estimation of the state of the access network at the second time point is performed by the following operation: based on the acquired first state history information and the estimated adjusted state of the access network at the third time point, the state of the access network at the second time point is estimated.

19. The electronic device according to claim 15, wherein, The value of the state control parameter is determined by performing the following operation: determining the value of the state control parameter within the time period between the first time point and the second time point.

20. The electronic device according to claim 19, in, The at least one processor is configured to control the electronic device to: Send information about the estimated state of the access network at the second point in time to the access network; as well as Based on the comparison between the estimated state of the access network at the second time point and the actual state of the access network at the second time point, the value of the state control parameter is applied to the state control parameter of the access network at the second time point.

21. The electronic device according to claim 20, wherein, The at least one processor is configured to control the electronic device to: Calculate the error value, which is the difference between the estimated state of the access network at the second time point and the actual state of the access network at the second time point. The application of the state control parameter value includes: applying the state control parameter value to the state control parameter of the access network at the second time point based on the calculated error value being less than a preset threshold.

22. The electronic device according to claim 15, wherein, For each of the at least one cells, the state of the access network includes at least one of the following: the number of user equipment (UE) currently connected to the cell, physical resource block (PRB) utilization or usage, Internet Protocol (IP) throughput, or downlink DL data volume.

23. The electronic device according to claim 15, wherein, The state of the access network at the second time point was estimated using a first artificial intelligence learning model. The values ​​of the state control parameters are determined using a second artificial intelligence learning model, and Both the first artificial intelligence learning model and the second artificial intelligence learning model have been trained using at least one of reinforcement learning, supervised learning, machine learning, continuous learning, joint learning, or deep learning.

24. The electronic device according to claim 15, wherein, The value of the state control parameter is determined by the following operation: determining the value of the state control parameter for load balancing each of the at least one cell in the access network.

25. The electronic device according to claim 15, wherein, The value of the state control parameter is determined by the following operation: determining the value of the state control parameter for adjusting the power consumption of each of the at least one cell in the access network.

26. The electronic device according to claim 23, wherein, The at least one processor is configured to control the electronic device to: Obtain the status of the access network at the second time point. The first artificial intelligence learning model is trained based on the state history information, the estimated state of the access network at the second time point, and the acquired state of the access network at the second time point.

27. The electronic device according to claim 23, wherein, The at least one processor is configured to control the electronic device to: Obtain the status of the access network at a fourth time point that is later than the second time point. The second artificial intelligence learning model is trained based on the estimated state of the access network at the second time point, the determined values ​​of the state control parameters, and the acquired state of the access network at the fourth time point.

28. The electronic device according to claim 15, wherein, The state control parameters include at least one of the following: Reference Signal Received Power (RSRP) threshold, Reference Signal Received Quality (RSRQ) threshold, Physical Resource Block (PRB) ratio threshold, or cell reselection priority.

Images