Radio frequency resource adjustment method, model training method, equipment, device and medium

Abstract

The embodiment of the invention provides a radio frequency resource adjustment method, a model training method, equipment, a device and a medium, and relates to the field of wireless communication, and the method comprises the steps: determining a resource adjustment action to be executed by an AP in a wireless network according to a first index through employing a resource adjustment model; simulating a simulation network state of the whole wireless networking after the AP is adjusted according to the resource adjustment action, and predicting a second index reflecting the simulation network state; obtaining a network quality test result by using the evaluation model according to the second index; if the network quality test result does not reach the preset standard, taking the second index as a new first index, and returning to the step of determining the resource adjustment action to be executed by the access point AP in the wireless network according to the first index; and if the network quality test result reaches the preset standard, controlling the AP to adjust, so that the wireless networking is updated to a simulation network state reaching the preset standard. By applying the scheme, the radio frequency resources can be quickly adjusted.

Description

Technical Field

[0001] The present application relates to the field of wireless communication technology, and in particular to a radio frequency resource adjustment method, model training method, equipment, device and medium. Background Art

[0002] Radio Resource Management (RRM) is a scalable radio frequency management solution that rationally allocates or adjusts radio frequency resources, enabling wireless networks to rapidly adapt to changes in the wireless environment and maintain optimal radio resource status, thereby improving the transmission capacity of the entire network. Power, channel, and bandwidth are the three most important physical parameters in radio frequency resource management. Therefore, properly adjusting the power, channel, and bandwidth of each access point (AP) in a wireless network is crucial to improving the transmission capacity of the entire network.

[0003] In related technologies, RRM solutions use thresholds to trigger adjustments to the radio resources of each AP in a wireless network. Specifically, during each monitoring cycle, adjustments are made to APs whose monitoring indicators exceed the thresholds. During these adjustments, the optimal power, channel, or bandwidth is selected for these APs based on the current wireless network conditions.

[0004] However, when the above solution adjusts radio frequency resources, only the local network of the wireless network is adjusted within a monitoring cycle. This requires multiple rounds of adjustments to stabilize the entire network, and cannot achieve the effect of quickly improving the performance of the entire network. Summary of the Invention

[0005] The purpose of the embodiments of the present application is to provide a radio frequency resource adjustment method, a model training method, a device, an apparatus and a medium, so as to achieve a stable network environment after one round of adjustment when adjusting radio frequency resources.

[0006] The specific technical solutions are as follows:

[0007] In a first aspect, an embodiment of the present application provides a method for adjusting radio frequency resources, the method comprising:

[0008] Determining, using a pre-trained resource adjustment model, a resource adjustment action to be performed by an access point (AP) in a wireless network based on a first indicator, wherein the first indicator reflects the entire network status of the wireless network, and the resource adjustment action of the AP is: adjusting at least one of power, channel, and bandwidth of the AP;

[0009] Simulating a simulated network state of the entire wireless network after the AP is adjusted according to the resource adjustment action, and predicting a second indicator reflecting the simulated network state;

[0010] Obtaining a network quality test result according to the second indicator using a pre-trained evaluation model, wherein the network quality test result represents the network quality of the wireless network in the simulated network state;

[0011] If the network quality test result does not meet the preset standard, the second indicator is used as a new first indicator, and the process returns to the step of determining the resource adjustment action to be performed by the access point AP in the wireless network based on the first indicator;

[0012] If the network quality test result reaches the preset standard, the AP is controlled to adjust so that the wireless networking is updated to a simulated network state that reaches the preset standard.

[0013] In one embodiment of the present application, the first indicator includes: a first quality indicator reflecting the overall network quality of the wireless network and / or a first interference indicator reflecting the interference status between APs in the wireless network;

[0014] and / or

[0015] The second indicator includes: a second quality indicator reflecting the overall network quality of the wireless network after the AP is adjusted according to the resource adjustment action and / or a second interference indicator reflecting the interference status between the APs in the wireless network after the AP is adjusted according to the resource adjustment action.

[0016] In a second aspect, an embodiment of the present application provides a model training method, the method comprising:

[0017] Inputting the third indicator into a pre-trained basic resource adjustment model to obtain a first output result, wherein the third indicator reflects the entire network state of the first wireless networking environment, and the first output result represents the probability of an AP in the wireless network performing different resource adjustment actions, wherein the basic resource adjustment model is trained based on a virtual second wireless networking environment, and the first wireless networking environment has the same network structure as the wireless network;

[0018] determining a simulated resource adjustment action to be performed by the AP based on the first output result;

[0019] simulating a network state of the first wireless networking environment after adjusting the AP according to the simulated resource adjustment action, and predicting a fourth indicator reflecting the network state;

[0020] Inputting the fourth indicator into a pre-trained basic evaluation model to obtain a second output result, where the second output result reflects the network quality of the first wireless networking environment after the AP performs the simulated resource adjustment action, and the basic evaluation model is trained based on the second wireless networking environment;

[0021] If the second output result does not meet the preset standard, the parameters of the basic resource adjustment model and the basic evaluation model are adjusted, the fourth indicator is used as the new third indicator, and the process returns to the step of inputting the third indicator into the basic resource adjustment model to obtain the first output result, until the second output result meets the preset standard, thereby obtaining the trained resource adjustment model and evaluation model;

[0022] In one embodiment of the present application, the basic resource adjustment model and the basic evaluation model are trained based on the following method:

[0023] Build a virtual second wireless networking environment;

[0024] Obtaining a fifth indicator reflecting the entire network status of the second wireless networking environment;

[0025] Inputting the fifth indicator into the initial resource adjustment model to obtain a third output result, wherein the third output result represents the probability of the AP in the wireless network performing different resource adjustment actions;

[0026] determining a simulated resource adjustment action to be performed by the AP based on the first output result;

[0027] simulating a network state of the second wireless networking environment after adjusting the AP according to the simulated resource adjustment action, and predicting a sixth indicator reflecting the network state;

[0028] Inputting the sixth indicator into the initial evaluation model to obtain a fourth output result, wherein the fourth output result reflects the network quality of the second wireless networking environment after the AP performs the simulated resource adjustment action;

[0029] If the second output result does not meet the preset standard, the parameters of the initial resource adjustment model and the initial evaluation model are adjusted, the sixth indicator is used as the new fifth indicator, and the step of inputting the sixth indicator into the basic resource adjustment model to obtain the third output result is returned to execute, until the fourth output result meets the preset standard, and the trained basic resource adjustment model and basic evaluation model are obtained.

[0030] In a third aspect, an embodiment of the present application further provides an electronic device, the device comprising:

[0031] processor;

[0032] transceiver;

[0033] A machine-readable storage medium storing machine-executable instructions capable of being executed by the processor; the machine-executable instructions prompting the processor to perform the following steps:

[0034] Determining, using a pre-trained resource adjustment model, a resource adjustment action to be performed by an access point (AP) in the wireless network based on a first indicator, wherein the first indicator reflects the entire network status of the wireless network, and the resource adjustment action of the AP is: adjusting at least one of power, channel, and bandwidth of the AP;

[0035] Simulating a simulated network state of the entire wireless network after the AP is adjusted according to the resource adjustment action, and predicting a second indicator reflecting the simulated network state;

[0036] Obtaining a network quality test result according to the second indicator using a pre-trained evaluation model, wherein the network quality test result represents the network quality of the wireless network in the simulated network state;

[0037] If the network quality test result does not meet the preset standard, the second indicator is used as a new first indicator, and the process returns to the step of determining the resource adjustment action to be performed by the access point AP in the wireless network based on the first indicator and starts execution;

[0038] If the network quality test result reaches the preset standard, the AP is controlled to adjust so that the wireless networking is updated to a simulated network state that reaches the preset standard.

[0039] In one embodiment of the present application, the first indicator includes: a first quality indicator reflecting the overall network quality of the wireless network and / or a first interference indicator reflecting the interference status between APs in the wireless network;

[0040] and / or

[0041] The second indicator includes: a second quality indicator reflecting the overall network quality of the wireless network after the AP is adjusted according to the resource adjustment action and / or a second interference indicator reflecting the interference status between the APs in the wireless network after the AP is adjusted according to the resource adjustment action.

[0042] In a fourth aspect, an embodiment of the present application provides an electronic device, comprising:

[0043] processor;

[0044] transceiver;

[0045] A machine-readable storage medium storing machine-executable instructions capable of being executed by the processor; the machine-executable instructions prompting the processor to perform the following steps:

[0046] Inputting the third indicator into a pre-trained basic resource adjustment model to obtain a first output result, wherein the third indicator reflects the entire network state of the first wireless networking environment, and the first output result represents the probability of an AP in the wireless network performing different resource adjustment actions, wherein the basic resource adjustment model is trained based on a virtual second wireless networking environment, and the first wireless networking environment has the same network structure as the wireless network;

[0047] determining a simulated resource adjustment action to be performed by the AP based on the first output result;

[0048] simulating a network state of the first wireless networking environment after adjusting the AP according to the simulated resource adjustment action, and predicting a fourth indicator reflecting the network state;

[0049] Inputting the fourth indicator into a pre-trained basic evaluation model to obtain a second output result, where the second output result reflects the network quality of the first wireless networking environment after the AP performs the simulated resource adjustment action, and the basic evaluation model is trained based on the second wireless networking environment;

[0050] If the second output result does not meet the preset standard, the parameters of the basic resource adjustment model and the basic evaluation model are adjusted, the fourth indicator is used as the new third indicator, and the process returns to the step of inputting the third indicator into the basic resource adjustment model to obtain the first output result, until the second output result meets the preset standard, and the trained resource adjustment model and evaluation model are obtained.

[0051] In one embodiment of the present application, the machine executable instructions further cause the processor to train the basic resource adjustment model and the basic evaluation model based on the following method:

[0052] Inputting the fifth indicator into the initial resource adjustment model to obtain a third output result, wherein the fifth indicator reflects the entire network state of the second wireless networking environment, and the third output result represents the probability of the AP in the wireless network performing different resource adjustment actions;

[0053] determining a simulated resource adjustment action to be performed by the AP based on the first output result;

[0054] simulating a network state of the second wireless networking environment after adjusting the AP according to the simulated resource adjustment action, and predicting a sixth indicator reflecting the network state;

[0055] Inputting the sixth indicator into the initial evaluation model to obtain a fourth output result, wherein the fourth output result reflects the network quality of the second wireless networking environment after the AP performs the simulated resource adjustment action;

[0056] If the second output result does not meet the preset standard, the parameters of the initial resource adjustment model and the initial evaluation model are adjusted, the sixth indicator is used as the new fifth indicator, and the step of inputting the sixth indicator into the basic resource adjustment model to obtain the third output result is returned to execute, until the fourth output result meets the preset standard, and the trained basic resource adjustment model and basic evaluation model are obtained.

[0057] In a fifth aspect, an embodiment of the present application further provides a radio frequency resource adjustment device, the device comprising:

[0058] a resource adjustment action determination module, configured to determine, using a pre-trained resource adjustment model and based on a first indicator, a resource adjustment action to be performed by an access point (AP) in a wireless network, wherein the first indicator reflects the entire network status of the wireless network, and the resource adjustment action of the AP is an action of adjusting at least one of the power, channel, and bandwidth of the AP;

[0059] A second indicator prediction module is used to simulate the simulated network state of the entire wireless network after the AP is adjusted according to the resource adjustment action, and predict a second indicator reflecting the simulated network state;

[0060] a network quality test result acquisition module, configured to obtain a network quality test result based on the second indicator using a pre-trained evaluation model, wherein the network quality test result represents the network quality of the wireless network in the simulated network state;

[0061] a first indicator updating module configured to, if the network quality test result does not meet a preset standard, use the second indicator as a new first indicator, and prompt the resource adjustment action determining module to start executing the resource adjustment action to be performed by the access point AP in the wireless network from the time when the resource adjustment action is determined based on the first indicator;

[0062] The adjustment module is used to control the AP adjustment if the network quality test result reaches the preset standard, so that the wireless networking is updated to a simulated network state that reaches the preset standard.

[0063] In a sixth aspect, an embodiment of the present application provides a model training device, the device comprising:

[0064] a first output result obtaining module, configured to input a third indicator into a pre-trained basic resource adjustment model to obtain a first output result, wherein the third indicator reflects the entire network state of the first wireless networking environment, and the first output result represents a probability of an AP in the wireless network performing different resource adjustment actions, wherein the basic resource adjustment model is trained based on a virtual second wireless networking environment, and the first wireless networking environment has the same network structure as the wireless network;

[0065] a first simulated resource adjustment action determining module, configured to determine a simulated resource adjustment action to be performed by the AP based on the first output result;

[0066] a fourth indicator prediction module, configured to simulate a network state of the first wireless networking environment after the AP is adjusted according to the simulated resource adjustment action, and predict a fourth indicator reflecting the network state;

[0067] a second output result obtaining module, configured to input the fourth indicator into a pre-trained basic evaluation model to obtain a second output result, wherein the second output result reflects the network quality of the first wireless networking environment after the AP performs the simulated resource adjustment action, and the basic evaluation model is trained based on the second wireless networking environment;

[0068] The model determination module is used to adjust the parameters of the basic resource adjustment model and the basic evaluation model if the second output result does not meet the preset standard, use the fourth indicator as the new third indicator, and prompt the first output result acquisition module to return to execute the step of inputting the third indicator into the basic resource adjustment model to obtain the first output result, until the second output result meets the preset standard, and the trained resource adjustment model and evaluation model are obtained.

[0069] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements any of the above-mentioned radio frequency resource adjustment methods or model training methods.

[0070] An embodiment of the present application also provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute any of the above-mentioned radio frequency resource adjustment methods or model training methods.

[0071] Beneficial effects of the embodiments of the present application:

[0072] In the technical solution provided by the embodiment of the present application, based on a first indicator reflecting the network state of the entire network, after determining the resource adjustment action to be performed by the AP in the wireless network, a simulated network state of the entire wireless network after adjusting the AP according to the resource adjustment action is simulated, and it is determined whether the network quality of the wireless network in the simulated network state meets the preset standard. If the preset standard is not met, the first indicator is updated and the above steps are repeated. If the preset standard is met, the AP is controlled to make adjustments so that the network quality of the wireless network after radio frequency resource optimization can meet the preset standard. That is, in the technical solution provided by the embodiment of the present application, based on at least one round of radio frequency resource optimization simulation, a final radio frequency resource optimization plan is determined, and radio frequency optimization is performed on the actual wireless network based on the final radio frequency resource optimization plan. In this way, radio frequency resource optimization of the entire wireless network can be achieved based on only one round of actual adjustment, and the network quality of the wireless network after radio frequency resource optimization can be guaranteed to meet the preset standard, thereby achieving the effect of rapidly improving the performance of the entire network. Furthermore, when adjusting AP radio resources in a wireless network, at least one of the following parameters—channel, bandwidth, and power—is optimized for the entire AP network based on the overall network status. This optimization ensures that the results match the overall wireless network status, rather than the local network status, achieving consistent and optimal communication across the entire network. Furthermore, each time a resource adjustment action is determined, the AP's channel, bandwidth, and power may be adjusted simultaneously, preventing the mutual impact of independent adjustments. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other embodiments can also be obtained based on these drawings.

[0074] Figure 1 A schematic diagram of a wireless network topology provided in an embodiment of the present application;

[0075] Figure 2 A flowchart of a radio frequency resource adjustment method provided in an embodiment of the present application;

[0076] Figure 3 A flowchart of a resource adjustment model and evaluation model training method provided in an embodiment of the present application;

[0077] Figure 4 A flowchart of a basic resource adjustment model and a basic evaluation model training method provided in an embodiment of the present application;

[0078] Figure 5A schematic diagram of radio frequency resource adjustment based on a reinforcement learning network provided in an embodiment of the present application;

[0079] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application;

[0080] Figure 7 A schematic structural diagram of another electronic device provided in an embodiment of the present application;

[0081] Figure 8 A schematic diagram of the structure of a radio frequency resource adjustment device provided in an embodiment of the present application;

[0082] Figure 9 A schematic diagram of the structure of a model training device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0083] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field based on this application are within the scope of protection of this application.

[0084] RRM is a scalable radio frequency management solution that rationally allocates or adjusts radio frequency resources, enabling wireless networks to rapidly adapt to changes in the wireless environment and maintain optimal radio resource status, thereby improving the transmission capacity of the entire network. Power, channel, and bandwidth are the three most important physical parameters for radio frequency resource management. Therefore, properly adjusting the power, channel, and bandwidth of each AP in a wireless network is crucial to improving the transmission capacity of the entire network.

[0085] In related technologies, RRM solutions use thresholds to trigger adjustments to the radio resources of each AP in a wireless network. Specifically, during each monitoring cycle, adjustments are made to APs whose monitoring indicators exceed the thresholds. During these adjustments, the optimal power, channel, or bandwidth is selected for these APs based on the current wireless network conditions.

[0086] However, the above solution has the following problems when adjusting radio resources:

[0087] (1) During a monitoring cycle, only local adjustments can be made to the wireless network, or even only one AP in the wireless network can be adjusted individually. This requires multiple rounds of adjustments to keep the entire wireless network stable. Even after multiple rounds of adjustments, the stability of the entire wireless network cannot be guaranteed, and the effect of quickly improving the performance of the entire network cannot be achieved.

[0088] (2) Each adjustment is based on the network indicators of a single AP and individually adjusts the channel, bandwidth, or frequency of the AP in the local network. Multiple adjustments may affect each other and cannot guarantee that the adjustments will optimize the final indicators of the entire network.

[0089] (3) Based on the monitoring cycle and threshold adjustment, it is unable to cope with the rapid changes of network topology in wireless networking.

[0090] In order to solve at least one of the above technical problems, embodiments of the present application provide a radio frequency resource adjustment method, electronic device, apparatus, and medium.

[0091] First, the network topology of the application scenario of this application is described.

[0092] See also Figure 1 , is a network topology diagram of a wireless network provided in an embodiment of the present application.

[0093] Figure 1 The system includes: AP, switch, and access controller (AC). It should be noted that the number of APs shown in the figure is 3, which is not limiting. The circles in the figure represent the signal coverage of the AP.

[0094] In one embodiment of the present application, the AP can collect network status data and neighbor relationships between surrounding APs in real time, and send it to the AC based on the switching device. The AC analyzes the data collected by the AP and, based on the analysis results, comprehensively allocates the channels, transmission power and bandwidth used by each AP. That is, the AC can execute the radio frequency resource adjustment method provided in the embodiment of the present application.

[0095] In another embodiment of the present application, the network topology may further include a computing device, such as an artificial intelligence (AI) server (not shown). The AP can collect network status data in real time and send it to the AC based on the switching device. The AC then forwards this data to the AI ​​server via the switching device. The AI ​​server analyzes the data collected by the AP and, based on the analysis results, coordinates the allocation of channels, transmit power, and bandwidth used by each AP. In other words, the radio frequency resource adjustment method provided in the embodiment of the present application can be executed by the computing device.

[0096] The radio frequency resource adjustment method provided in the embodiment of the present application is explained below with reference to specific embodiments.

[0097] See also Figure 2, which is a flowchart of a radio frequency resource adjustment method provided in an embodiment of the present application. This method can be applied to ACs in wireless networks and computing devices such as AI servers, without limitation. For ease of description, the following description uses the AC as the execution entity, without limitation. The method includes steps S201 through S205.

[0098] S201: Determine a resource adjustment action to be performed by an AP in a wireless network according to a first indicator by using a pre-trained resource adjustment model.

[0099] Among them, the above-mentioned first indicator reflects the entire network status of the wireless network.

[0100] Specifically, the first indicator may be a first quality indicator reflecting the overall network quality of the wireless network and / or a first interference indicator reflecting the interference status between APs in the wireless network.

[0101] The first quality indicator may include: the maximum carryable traffic of the entire network within a preset duration, the maximum bandwidth of the entire network, the minimum latency of the entire network, etc. Specifically, an AP in the wireless network can report its maximum carryable traffic within the preset duration to the AC, which can then sum the maximum carryable traffic of each AP to obtain the maximum carryable traffic of the entire network. An AP in the wireless network can also report its minimum latency within the preset duration to the AC, which can then calculate the average or median of the minimum latency of each AP to obtain the minimum latency of the entire network.

[0102] The first interference indicator may include: channel occupancy, channel packet loss rate, and neighbor relationships between APs within a preset duration. Specifically, APs in a wireless network may send their channel occupancy within a preset duration to the AC, which then analyzes this data to determine the occupancy of each channel. APs in a wireless network may report their packet loss rate on each channel within a preset duration to the AC, which then analyzes this data to determine the packet loss rate of each channel. APs in a wireless network establish neighbor relationships between APs through neighbor detection and report these neighbor relationships to the AC. Neighbor relationships between APs may include, for example, the transmit power of neighboring APs.

[0103] In an embodiment of the present application, the AC can obtain a first indicator reflecting the network status of the entire wireless network, and based on the first indicator, determine the resource adjustment action to be performed by the AP in the wireless network, that is, taking into account the network status of the entire network, and providing a basis for the subsequent optimization of the radio frequency resources of the entire network, which can be completed with one adjustment.

[0104] The resource adjustment action of the AP may be an action of adjusting at least one of the power, channel, and bandwidth of the AP.

[0105] Setting the AP's transmit power to its maximum value can extend its signal coverage, but excessive power can cause unnecessary interference to other APs. Therefore, it's important to select an appropriate power level for the AP that balances coverage and usage requirements.

[0106] For wireless LANs, channels are a very scarce resource. Each AP can only work on a very limited number of channels, so allocating the best channel to the AP is very critical.

[0107] Each channel has a limited bandwidth (e.g., 20 MHz). Channel bonding can be used to expand the AP's bandwidth by combining adjacent channels, allowing one AP to occupy multiple channels. This technology increases the AP's data transmission rate, but also increases the signal's susceptibility to interference, reducing data transmission stability. Therefore, it's necessary to dynamically adjust the AP's bandwidth based on the actual wireless network environment.

[0108] In an embodiment of the present application, after obtaining the first indicator, the AC can determine the resource adjustment action to be performed by the AP in the wireless network. Here, the AC can determine the resource adjustment action to be performed by all APs in the wireless network, or can determine the resource adjustment action to be performed by some APs in the wireless network, depending on actual conditions. The resource adjustment action to be performed by the AP determined by the AC can be: adjusting one or more of the power, channel, and bandwidth of the AP.

[0109] In one embodiment of the present application, the AC can input the first indicator into a pre-trained resource adjustment model and, based on the output of the resource adjustment model, determine the resource adjustment action to be performed by the AP in the wireless network. This approach, due to the generalizability of the resource adjustment model, can adapt to changes in the network topology in the wireless network.

[0110] In another embodiment of the present application, the AC can determine the resource adjustment actions to be performed by the APs in the wireless network based on the first indicator and in combination with a dynamic search algorithm. For example, based on the Dynamic Channel Assignment (DCA) algorithm, the APs in the wireless network are divided into multiple groups, and optimal channels are allocated to the APs through exhaustive or iterative methods. Based on the Transmit Power Control (TPC) algorithm, an appropriate transmit power is selected for the AP to meet the AP's coverage requirements while not causing significant interference to neighboring APs.

[0111] The embodiment of the present application does not specifically limit the implementation method of determining the resource adjustment action to be performed by the AP in the wireless network based on the first indicator.

[0112] S202: Simulate a simulated network state of the entire wireless network after the AP is adjusted according to the resource adjustment action, and predict a second indicator reflecting the simulated network state.

[0113] The above-mentioned second indicator has a similar meaning to the first indicator. The only difference is that the first indicator reflects the entire network status of the wireless network, while the second indicator reflects the simulated network status of the entire wireless network. Please refer to the relevant description of the first indicator in the previous article.

[0114] In an embodiment of the present application, after the AC determines the resource adjustment action to be performed by the AP, it does not directly issue an adjustment instruction to the AP that is to perform the resource adjustment action, allowing the AP to directly perform the resource adjustment action. Instead, it simulates the simulated network state of the entire wireless network after the AP is adjusted according to the resource adjustment action, and predicts a second indicator reflecting the simulated network state.

[0115] Specifically, the AC can issue instructions to a computing device equipped with a simulation program, allowing it to build a simulated wireless networking environment with the same network structure as the current wireless networking on the simulation program, and let the AP in the simulated wireless networking environment perform resource adjustment actions, thereby simulating the network status of the entire wireless network after the AP is adjusted according to the resource adjustment action.

[0116] S203: Obtain a network quality test result according to the second indicator using the pre-trained evaluation model.

[0117] The above network quality test result indicates the network quality of the wireless network in the simulated network state.

[0118] In one embodiment of the present application, the AC may input the second indicator into a pre-trained evaluation model to obtain a network quality test result output by the evaluation model.

[0119] In another embodiment of the present application, the AC may normalize the second indicator to obtain a normalized value, and determine which evaluation interval the normalized value falls into. Different evaluation intervals correspond to different levels of network quality test results, thereby determining the network quality test results.

[0120] S204: If the network quality test result does not meet the preset standard, the second indicator is used as a new first indicator.

[0121] Return to step S201.

[0122] The above-mentioned preset standards can be set based on actual needs. The preset standards can indicate that the network quality of the wireless network is at a good level. For example, the preset standards can include the network's maximum bearer traffic being greater than a first threshold, the network's minimum latency being less than a second threshold, and the channel's packet loss rate being less than a third threshold. The specific values ​​of the first, second, and third thresholds can be set based on actual needs.

[0123] In this embodiment of the present application, when the AC determines that the network quality test result does not meet the preset standard, it indicates that after the simulated AP performs the resource adjustment action, the network quality of the wireless network in the simulated network state does not meet the requirements. If the AP in the real wireless network environment is adjusted according to the determined resource adjustment action, the network quality of the adjusted wireless network will also not meet the requirements. Therefore, the AC will use the second indicator as the new first indicator and return to the above step S202 to start.

[0124] S205: If the network quality test result reaches the preset standard, the AP is controlled to adjust so that the wireless network is updated to a simulated network state that reaches the preset standard.

[0125] In an embodiment of the present application, when the AC determines that the network quality test result meets the preset standard, it means that after the simulated AP performs the resource adjustment action, the network quality of the wireless network in the simulated network state can meet the requirements. Based on this, the AC can issue a global adjustment instruction to the AP that needs to perform the resource adjustment action in the real wireless networking environment to control the AP to perform the resource adjustment action, so that the network state of the real wireless network is updated to the simulated network state that meets the preset standard, that is, the network quality of the updated wireless network can meet the preset standard.

[0126] In the embodiment of the present application, the AC may determine a final global adjustment instruction based on each resource adjustment action to be performed by the AP, and issue the final global adjustment instruction to the AP.

[0127] For example, the AC first determines that the resource adjustment action to be performed by the AP is to adjust AP1 to use channel 1. After simulation, the network quality of the simulated wireless network state does not meet the preset standard, so the resource adjustment action to be performed by the AP needs to be further determined.

[0128] The AC determines the second resource adjustment action for the APs: adjust AP1 to channel 11 and reduce AP2's transmit power. After the simulation, the wireless network quality of the simulated network still does not meet the preset standard, so the AC needs to further determine the resource adjustment action for the APs.

[0129] The AC determines the third resource adjustment action to be performed by the APs: adjust AP1 to use channel 1 and increase AP3's bandwidth. After the simulation, the network quality of the simulated wireless network meets the preset standard.

[0130] At this time, the AC determines the final global adjustment instructions based on the resource adjustment actions to be performed by the three APs: switch AP1 to channel 1; reduce AP2's transmit power; and increase AP3's bandwidth.

[0131] As can be seen, in the embodiments of the present application, a final radio frequency resource optimization scheme is determined based on at least one round of radio frequency resource optimization simulation, and radio frequency optimization is performed on the actual wireless network based on the final radio frequency resource optimization scheme. In this way, radio frequency resource optimization of the wireless network can be achieved based on only one round of actual adjustment, and the network quality of the wireless network after radio frequency resource optimization can be guaranteed to meet preset standards. In addition, when adjusting the radio frequency resources of the AP in the wireless network, at least one of the channel, bandwidth, and power of the AP in the entire network is optimized based on the network status of the entire network, so that the optimization result matches the network status of the entire wireless network rather than the local network status, thereby achieving a unified and better communication effect for the entire network. Moreover, each time a resource adjustment action is determined, the channel, bandwidth, and power of the AP may all be adjusted, which can avoid the mutual influence of the three independent adjustments.

[0132] For ease of understanding, the following Figure 1 The network topology diagram shown in FIG. 1 illustrates the radio frequency resource adjustment solution provided in the embodiment of the present application. Assume that at the current moment, AP1 and AP2 both use channel 6, and AP3 uses channel 1.

[0133] The AC receives network status data reported by AP1-AP3 and neighbor relationships between APs based on the switching device, and determines a first indicator based on the data.

[0134] Based on the first indicator, the AC determines that the resource adjustment action to be performed by AP1 is: adjusting AP1 to use channel 1.

[0135] The AC simulates and adjusts AP1 to use channel 1, and predicts the second indicator based on the simulated network status of the entire wireless network after the adjustment.

[0136] The AC obtains a network quality test result of a simulated network status of the entire wireless networking based on the second indicator.

[0137] The AC determines that the network quality test result does not meet the preset standard, updates the second indicator to the first indicator, and based on the updated first indicator, determines that the resource adjustment action to be performed by AP1 is: adjusting AP1 to use channel 11; and determines that the resource adjustment action to be performed by AP2 is: reducing the transmission power of AP2.

[0138] The AC simulates adjusting AP1 to use channel 11, reducing the transmit power of AP2, and predicts the second indicator based on the simulated network status of the entire wireless network after the adjustment.

[0139] The AC obtains a network quality test result of a simulated network status of the entire wireless networking based on the second indicator.

[0140] The AC determines that the network quality test result does not meet the preset standard, updates the second indicator to the first indicator, and based on the updated first indicator, determines that the resource adjustment action to be performed by AP1 is: adjusting AP1 to use channel 1; and determines that the resource adjustment action to be performed by AP3 is: increasing AP3's bandwidth.

[0141] The AC simulates adjusting AP1 to use channel 1, increases the bandwidth of AP3, and predicts the second indicator based on the simulated network status of the entire wireless network after the adjustment.

[0142] The AC obtains a network quality test result of a simulated network status of the entire wireless networking based on the second indicator.

[0143] The AC determines that the network quality test results meet the preset standards and controls the APs in the wireless network to perform resource adjustment actions, such as switching AP1 to channel 1; reducing the transmit power of AP2; and increasing the bandwidth of AP3.

[0144] In the technical solution provided by the embodiment of the present application, based on a first indicator reflecting the network state of the entire network, after determining the resource adjustment action to be performed by the AP in the wireless network, a simulated network state of the entire wireless network after adjusting the AP according to the resource adjustment action is simulated, and it is determined whether the network quality of the wireless network in the simulated network state meets the preset standard. If the preset standard is not met, the first indicator is updated and the above steps are repeated. If the preset standard is met, the AP is controlled to make adjustments so that the network quality of the wireless network after radio frequency resource optimization can meet the preset standard. That is, in the technical solution provided by the embodiment of the present application, based on at least one round of radio frequency resource optimization simulation, a final radio frequency resource optimization plan is determined, and radio frequency optimization is performed on the actual wireless network based on the final radio frequency resource optimization plan. In this way, radio frequency resource optimization of the wireless network can be achieved based on only one round of actual adjustment, and the network quality of the wireless network after radio frequency resource optimization can be guaranteed to meet the preset standard. Furthermore, when adjusting AP radio resources in a wireless network, at least one of the following parameters—channel, bandwidth, and power—is optimized for the entire AP network based on the overall network status. This optimization ensures that the results match the overall wireless network status, rather than the local network status, achieving consistent and optimal communication across the entire network. Furthermore, each time a resource adjustment action is determined, the AP's channel, bandwidth, and power may be adjusted simultaneously, preventing the mutual impact of independent adjustments.

[0145] As mentioned above, the AC can input the first indicator into a pre-trained resource adjustment model and, based on the output of the resource adjustment model, determine the resource adjustment action to be performed by the AP in the wireless network. Furthermore, the AC can input the second indicator into a pre-trained evaluation model to obtain the network quality test results output by the evaluation model. For ease of understanding, the following describes the training methods for the resource adjustment model and evaluation model provided in the embodiments of this application, with reference to specific examples.

[0146] See also Figure 3 , which is a flow chart of a resource adjustment model and evaluation model training method provided in an embodiment of the present application. The method can be applied to electronic devices and also to AC. The following description uses AC as the execution subject and does not serve as a limitation. The method includes steps S301 to S306.

[0147] S301: Input the third indicator into a pre-trained basic resource adjustment model to obtain a first output result.

[0148] The third indicator reflects the entire network status of the first wireless networking environment. The third indicator has a similar meaning to the first indicator. The process of obtaining the third indicator can refer to the process of obtaining the first indicator above.

[0149] The first wireless networking environment can be understood as a virtual wireless networking environment having the same network structure as the current real wireless network. The same network structure can be understood as the same number of APs in the network and the same neighbor relationships between APs.

[0150] In an embodiment of the present application, the AC can first obtain the network structure of the wireless networking, that is, the AC has learned the neighbor relationship of the AP in the wireless networking. When constructing the first wireless networking environment, it first randomly determines the geographical location of any first AP, and based on the neighbor relationship between the first APs, determines the geographical location of the second AP adjacent to the first AP, and repeats the step of determining the geographical location of the AP based on the neighbor relationship until a first wireless networking environment with the same network structure as the wireless networking is obtained.

[0151] In the embodiment of the present application, the reason for constructing a first wireless networking environment with the same network structure as the wireless networking is that the network topology of the wireless networking will change dynamically. When a new AP comes online or an AP goes offline, the network topology of the wireless networking will change. If the network topology of the wireless networking changes, by constructing a first wireless networking environment with the same network structure as the wireless networking, and training the basic resource adjustment model and the basic evaluation model in the first wireless networking environment to obtain a trained resource adjustment model and evaluation model, in this way, by training in the first wireless networking environment, the change information of the network topology is transmitted to the resource adjustment model and the evaluation model, so that when the resource adjustment model and the evaluation model are used to perform radio frequency resource optimization on the AP in the wireless networking, a better optimization effect can be obtained.

[0152] The first output result indicates the probability of the AP in the wireless network performing different resource adjustment actions.

[0153] The above-mentioned basic resource adjustment model is trained based on a virtual second wireless networking environment. Specifically, a relatively general basic resource adjustment model is trained in the second wireless networking environment. This basic resource adjustment model is then further trained in the first wireless networking environment to obtain a resource adjustment model suitable for the current wireless networking scenario. The training process for the basic resource adjustment model is described below.

[0154] In an embodiment of the present application, the AC may input the third indicator into a pre-trained basic resource adjustment model to obtain a first output result.

[0155] S302: Determine a simulated resource adjustment action to be performed by the AP based on the first output result.

[0156] In one embodiment of the present application, the AC may determine the resource adjustment action to be performed by the AP with the highest probability in the first output result as the simulated resource adjustment action.

[0157] S303: Simulate the network state of the first wireless networking environment after the AP is adjusted according to the simulated resource adjustment action, and predict a fourth indicator reflecting the network state.

[0158] For this step, please refer to the above description of step S202.

[0159] S304: Input the fourth indicator into the pre-trained basic evaluation model to obtain a second output result.

[0160] The second output reflects the network quality of the first wireless networking environment after the AP performs the simulated resource adjustment action. The basic evaluation model is trained based on the second wireless networking environment. Specifically, a relatively general basic evaluation model is trained in the second wireless networking environment, and then further trained in the first wireless networking environment to obtain an evaluation model suitable for the current wireless networking scenario. The training process of the basic evaluation model is described below.

[0161] In an embodiment of the present application, the AC may input the fourth indicator into a pre-trained basic evaluation model to obtain a second output result.

[0162] S305: If the second output result does not meet the preset standard, the parameters of the basic resource adjustment model and the basic evaluation model are adjusted, and the fourth indicator is used as a new third indicator.

[0163] Return to step S301 and execute again.

[0164] S306: If the second output result meets the preset standard, the trained resource adjustment model and evaluation model are obtained.

[0165] In an embodiment of the present application, the AC determines whether the second output result meets the preset standard. If not, the parameters of the basic resource adjustment model and the basic evaluation model are adjusted, and the fourth indicator is used as the new third indicator. The process returns to step S303 until the second output result meets the preset standard, thereby obtaining the trained resource adjustment model and evaluation model.

[0166] As can be seen from the above embodiments, in the technical solution provided by the embodiments of the present application, by constructing a first wireless networking environment with the same network structure as the above-mentioned wireless networking, the relatively general basic resource adjustment model and basic evaluation model obtained in advance are deployed in the first wireless networking environment, and are retrained using the third indicator and the fourth indicator determined based on the first wireless networking environment to obtain the final resource adjustment model and evaluation model. In this way, in the face of the changing network topology of the wireless network, by constructing a first wireless networking environment with the same network structure as the wireless network, the basic resource adjustment model and the basic evaluation model are retrained in the first wireless networking environment, and the change information of the network topology is transmitted to the resource adjustment model and the evaluation model to adapt to the change of the network topology, so that when the trained resource adjustment model and the evaluation model are used to perform radio frequency resource optimization on the AP in the wireless network, better optimization effect can be obtained.

[0167] The following combination Figure 4 The training process of the basic resource adjustment model and the basic evaluation model provided in the embodiments of the present application is described in detail.

[0168] See also Figure 4 , which is a flow chart of a basic resource adjustment model and basic evaluation model training method provided in an embodiment of the present application. The method can be applied to electronic devices and also to AC. The following description uses AC as the execution subject and does not serve as a limitation. The method includes steps S401 to S406.

[0169] S401: Input the fifth indicator into the initial resource adjustment model to obtain a third output result.

[0170] Among them, the fifth indicator reflects the entire network status of the second wireless networking environment.

[0171] Since the training of the basic resource adjustment model and the basic evaluation model takes a long time and will affect the actual use of the wireless network, the embodiment of the present application adopts a centralized training method to obtain a better basic resource adjustment model and basic evaluation model by training in multiple virtual second wireless networking environments.

[0172] In a real-world wireless network deployment, all radios (an AP can have multiple radios, but if only one radio is enabled, the AP is considered a single radio) can be considered a node. If radios can sense each other, an edge exists between the two nodes. Based on this, the network topology of the wireless network can be represented as a directed graph.

[0173] In an embodiment of the present application, a virtual second wireless networking environment is constructed, that is, a process of generating a directed graph of the network topology corresponding to the second wireless networking environment. The directed graph can be generated based on a random algorithm. Specifically, the transmission power and attenuation of each radio in the second wireless networking environment are obtained by a normal distribution, the interference relationship between the two radios is obtained by the geographical location and power, attenuation, according to the logarithmic distance loss model, and the load of the radio is obtained by adding a certain random number to a plurality of fixed patterns. After obtaining the interference relationship between each radio in the second wireless networking environment, the geographical location of any first AP is determined by evenly distributing points in a two-dimensional plane, and the geographical location of the second AP adjacent to the first AP is determined based on the neighbor relationship between the first APs, and the step of determining the geographical location of the AP based on the neighbor relationship is repeated until a virtual second wireless networking environment is constructed. In the embodiment of the present application, there is no specific limitation on the manner of constructing the second wireless networking environment.

[0174] The third output result indicates the probability of the AP in the wireless network performing different resource adjustment actions.

[0175] S402: Determine a simulated resource adjustment action to be performed by the AP based on the third output result.

[0176] S403: Simulate the network state of the second wireless networking environment after the AP is adjusted according to the simulated resource adjustment action, and predict a sixth indicator reflecting the network state.

[0177] S404: Input the sixth indicator into the initial evaluation model to obtain a fourth output result.

[0178] The fourth output result reflects the network quality of the second wireless networking environment after the AP performs the simulated resource adjustment action.

[0179] S405: If the fourth output result does not meet the preset standard, the parameters of the initial resource adjustment model and the initial evaluation model are adjusted, and the sixth indicator is used as the new fifth indicator.

[0180] Return to step S401.

[0181] S406: If the fourth output result meets the preset standard, the trained basic resource adjustment model and basic evaluation model are obtained.

[0182] Steps S401 to S406 are basically similar to the aforementioned steps S301 to S306, and reference may be made to the relevant description above.

[0183] In one embodiment of the present application, the initial resource adjustment model and the initial evaluation model select reinforcement learning models, and the training of the initial resource adjustment model and the initial evaluation model is realized based on the idea of ​​the soft action-evaluation (Soft Actor-Critic) algorithm. The initial resource adjustment model can be an actor (action) network, which is used to determine the adjustment action according to the network status of the wireless network, and the initial evaluation model can be a critic (evaluation) network, which is used to give an evaluation of the current network quality according to the network status of the wireless network after the adjustment action. The same model is used for all APs in the entire wireless network to decide on resource adjustment actions. When the number of APs is large, the time complexity of the radio frequency resource adjustment scheme based on reinforcement learning proposed in the embodiment of the present application increases linearly with the number of APs. Compared with the radio frequency resource adjustment scheme based on the dynamic search algorithm, the time complexity increases exponentially with the number of APs, reduces the correlation with factors such as the number of APs, and can be deployed in wireless networks of different scales.

[0184] As can be seen from the above embodiments, in the technical solution provided by the embodiments of the present application, by constructing a virtual second wireless networking environment, the initial resource adjustment model and the initial evaluation model are deployed in the second wireless networking environment, and the fifth indicator and the fifth indicator determined based on the second wireless networking environment are used to perform secondary training on them to obtain the basic resource adjustment model and the basic evaluation model. Since the basic resource adjustment model and the basic evaluation model are trained in a virtual networking environment, they will not affect the use of the existing network. Moreover, when using the basic resource adjustment model and the basic evaluation model to adjust radio frequency resources, the time complexity increases linearly with the number of APs, and can be deployed in wireless networks of different sizes.

[0185] The following combination Figure 5 The radio frequency resource adjustment method provided in the embodiment of the present application is further explained.

[0186] See also Figure 5 , is a schematic diagram of radio frequency resource adjustment based on a reinforcement learning network provided in an embodiment of the present application.

[0187] Figure 5 The left side of the figure represents a real wireless network environment, where RF resources need to be adjusted to improve network quality. This wireless network includes four APs. Arrows between APs indicate interference between them. These four APs can operate on channels 1, 6, and 11.

[0188] Figure 5The middle part is a first wireless networking environment that is constructed with the same network structure as the current real wireless networking. The AC can obtain a third indicator reflecting the entire network status of the first wireless networking environment. Figure 5 The occupancy rates of each channel are shown in (channel 1 occupancy rate is 90%, i.e. 30% + 30% + 20% + 10%, channel 6 occupancy rate is 90%, i.e. 30% + 20% + 20% + 20%, channel 11 occupancy rate is 0%), and based on the third indicator, the pre-trained basic resource adjustment model and the basic evaluation model are retrained.

[0189] Figure 5 The middle right half shows the process of training the resource adjustment model and the evaluation model based on the reinforcement learning algorithm. After the third indicator is input into the resource adjustment model, the simulated resource adjustment action to be performed by the AP is obtained ( Figure 5 The figure shows that after the top AP is switched from channel 1 to channel 11 (i.e., channel 1> channel 11), the network state of the first wireless networking environment after the AP is adjusted according to the simulated resource adjustment is simulated, and the network quality of the wireless network in the simulated network state is judged based on the basic evaluation model to determine whether it meets the preset standard. If it does not meet the preset standard, it returns to Figure 5 Repeat the training in the middle part of the process. When the preset standards are met, the trained resource adjustment model and evaluation model are obtained. The trained resource adjustment model and evaluation model can then be used to determine the RF tuning solution for the current real wireless network.

[0190] In the technical solution provided by the embodiment of the present application, based on a first indicator reflecting the network state of the entire network, after determining the resource adjustment action to be performed by the AP in the wireless network, a simulated network state of the entire wireless network after adjusting the AP according to the resource adjustment action is simulated, and it is determined whether the network quality of the wireless network in the simulated network state meets the preset standard. If the preset standard is not met, the first indicator is updated and the above steps are repeated. If the preset standard is met, the AP is controlled to make adjustments so that the network quality of the wireless network after radio frequency resource optimization can meet the preset standard. That is, in the technical solution provided by the embodiment of the present application, based on at least one round of radio frequency resource optimization simulation, a final radio frequency resource optimization plan is determined, and radio frequency optimization is performed on the actual wireless network based on the final radio frequency resource optimization plan. In this way, radio frequency resource optimization of the wireless network can be achieved based on only one round of actual adjustment, and the network quality of the wireless network after radio frequency resource optimization can be guaranteed to meet the preset standard. Furthermore, when adjusting AP radio resources in a wireless network, at least one of the following parameters—channel, bandwidth, and power—is optimized for the entire AP network based on the overall network status. This optimization ensures that the results match the overall wireless network status, rather than the local network status, achieving consistent and optimal communication across the entire network. Furthermore, each time a resource adjustment action is determined, the AP's channel, bandwidth, and power may be adjusted simultaneously, preventing the mutual impact of independent adjustments.

[0191] Based on the same inventive concept, an embodiment of the present application also provides an electronic device.

[0192] See also Figure 6 , is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application, the device comprising:

[0193] Processor 601;

[0194] transceiver 604;

[0195] A machine-readable storage medium 602 stores machine-executable instructions that can be executed by the processor; the machine-executable instructions prompt the processor 601 to execute any of the above-mentioned radio frequency resource adjustment methods.

[0196] like Figure 6As shown, the electronic device may further include a communication bus 603. The processor 601, the machine-readable storage medium 602, and the transceiver 604 communicate with each other via the communication bus 603. The communication bus 603 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The communication bus 603 may be divided into an address bus, a data bus, a control bus, and the like.

[0197] The transceiver 604 may be a wireless communication module. Under the control of the processor 601 , the transceiver 604 exchanges data with other devices.

[0198] The machine-readable storage medium 602 may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Alternatively, the machine-readable storage medium 602 may be at least one storage device located remote from the processor.

[0199] The processor 601 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components.

[0200] In the technical solution provided by the embodiment of the present application, based on a first indicator reflecting the network state of the entire network, after determining the resource adjustment action to be performed by the AP in the wireless network, a simulated network state of the entire wireless network after adjusting the AP according to the resource adjustment action is simulated, and it is determined whether the network quality of the wireless network in the simulated network state meets the preset standard. If the preset standard is not met, the first indicator is updated and the above steps are repeated. If the preset standard is met, the AP is controlled to make adjustments so that the network quality of the wireless network after radio frequency resource optimization can meet the preset standard. That is, in the technical solution provided by the embodiment of the present application, based on at least one round of radio frequency resource optimization simulation, a final radio frequency resource optimization plan is determined, and radio frequency optimization is performed on the actual wireless network based on the final radio frequency resource optimization plan. In this way, radio frequency resource optimization of the wireless network can be achieved based on only one round of actual adjustment, and the network quality of the wireless network after radio frequency resource optimization can be guaranteed to meet the preset standard. Furthermore, when adjusting AP radio resources in a wireless network, at least one of the following parameters—channel, bandwidth, and power—is optimized for the entire AP network based on the overall network status. This optimization ensures that the results match the overall wireless network status, rather than the local network status, achieving consistent and optimal communication across the entire network. Furthermore, each time a resource adjustment action is determined, the AP's channel, bandwidth, and power may be adjusted simultaneously, preventing the mutual impact of independent adjustments.

[0201] Based on the same inventive concept, an embodiment of the present application further provides an electronic device,

[0202] See also Figure 7 , is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application, the device comprising:

[0203] Processor 701;

[0204] transceiver 704;

[0205] A machine-readable storage medium 702 stores machine-executable instructions that can be executed by the processor; the machine-executable instructions prompt the processor 701 to execute any of the above-mentioned model training methods.

[0206] like Figure 7As shown, the electronic device may further include a communication bus 703. The processor 701, the machine-readable storage medium 702, and the transceiver 704 communicate with each other via the communication bus 703. The communication bus 703 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The communication bus 703 may be divided into an address bus, a data bus, a control bus, and the like.

[0207] The transceiver 704 may be a wireless communication module. Under the control of the processor 701 , the transceiver 704 exchanges data with other devices.

[0208] The machine-readable storage medium 702 may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Alternatively, the machine-readable storage medium 702 may be at least one storage device located remote from the processor.

[0209] The processor 701 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0210] As can be seen from the above embodiments, in the technical solution provided by the embodiments of the present application, by constructing a first wireless networking environment with the same network structure as the above-mentioned wireless networking, the relatively general basic resource adjustment model and basic evaluation model obtained in advance are deployed in the first wireless networking environment, and are retrained using the third indicator and the fourth indicator determined based on the first wireless networking environment to obtain the final resource adjustment model and evaluation model. In this way, in the face of the changing network topology of the wireless network, by constructing a first wireless networking environment with the same network structure as the wireless network, the basic resource adjustment model and the basic evaluation model are retrained in the first wireless networking environment, and the change information of the network topology is transmitted to the resource adjustment model and the evaluation model to adapt to the change of the network topology, so that when the trained resource adjustment model and the evaluation model are used to perform radio frequency resource optimization on the AP in the wireless network, better optimization effect can be obtained.

[0211] Based on the same inventive concept, an embodiment of the present application further provides a radio frequency resource adjustment device.

[0212] See also Figure 8 , is a structural diagram of a radio frequency resource adjustment device provided in an embodiment of the present application, the device comprising:

[0213] A resource adjustment action determination module 801 is configured to determine, using a pre-trained resource adjustment model and based on a first indicator, a resource adjustment action to be performed by an access point (AP) in a wireless network, wherein the first indicator reflects the entire network status of the wireless network, and the resource adjustment action of the AP is to adjust at least one of the power, channel, and bandwidth of the AP;

[0214] The second indicator prediction module 802 is configured to simulate a simulated network state of the entire wireless network after the AP is adjusted according to the resource adjustment action, and predict a second indicator reflecting the simulated network state;

[0215] A network quality test result acquisition module 803 is configured to acquire a network quality test result based on the second indicator using a pre-trained evaluation model, wherein the network quality test result represents the network quality of the wireless network in the simulated network state;

[0216] a first indicator updating module 804 configured to, if the network quality test result does not meet a preset standard, use the second indicator as a new first indicator, prompting the resource adjustment action determining module to return to the step of determining the resource adjustment action to be performed by the access point AP in the wireless network based on the first indicator;

[0217] The adjustment module 805 is configured to control the AP adjustment if the network quality test result reaches a preset standard, so that the wireless networking is updated to a simulated network state that reaches the preset standard.

[0218] In one embodiment of the present application, the first indicator includes: a first quality indicator reflecting the overall network quality of the wireless network and / or a first interference indicator reflecting the interference status between APs in the wireless network;

[0219] and / or

[0220] The second indicator includes: a second quality indicator reflecting the overall network quality of the wireless network after the AP is adjusted according to the resource adjustment action and / or a second interference indicator reflecting the interference status between the APs in the wireless network after the AP is adjusted according to the resource adjustment action.

[0221] In the technical solution provided by the embodiment of the present application, based on a first indicator reflecting the network state of the entire network, after determining the resource adjustment action to be performed by the AP in the wireless network, a simulated network state of the entire wireless network after adjusting the AP according to the resource adjustment action is simulated, and it is determined whether the network quality of the wireless network in the simulated network state meets the preset standard. If the preset standard is not met, the first indicator is updated and the above steps are repeated. If the preset standard is met, the AP is controlled to make adjustments so that the network quality of the wireless network after radio frequency resource optimization can meet the preset standard. That is, in the technical solution provided by the embodiment of the present application, based on at least one round of radio frequency resource optimization simulation, a final radio frequency resource optimization plan is determined, and radio frequency optimization is performed on the actual wireless network based on the final radio frequency resource optimization plan. In this way, radio frequency resource optimization of the wireless network can be achieved based on one round of adjustment, and the network quality of the wireless network after radio frequency resource optimization can be guaranteed to meet the preset standard.

[0222] Based on the same inventive concept, an embodiment of the present application also provides a model training device.

[0223] See also Figure 9 , is a schematic diagram of the structure of a model training device provided in an embodiment of the present application, the device comprising:

[0224] A first output result obtaining module 901 is configured to input a third indicator into a pre-trained basic resource adjustment model to obtain a first output result, wherein the third indicator reflects the entire network state of the first wireless networking environment, and the first output result represents the probability of an AP in the wireless network performing different resource adjustment actions. The basic resource adjustment model is trained based on a virtual second wireless networking environment, and the first wireless networking environment has the same network structure as the wireless network.

[0225] a first simulated resource adjustment action determining module 902, configured to determine a simulated resource adjustment action to be performed by the AP based on the first output result;

[0226] a fourth indicator prediction module 903, configured to simulate a network state of the first wireless networking environment after the AP is adjusted according to the simulated resource adjustment action, and predict a fourth indicator reflecting the network state;

[0227] A second output result obtaining module 904 is configured to input the fourth indicator into a pre-trained basic evaluation model to obtain a second output result, where the second output result reflects the network quality of the first wireless networking environment after the AP performs the simulated resource adjustment action, and the basic evaluation model is trained based on the second wireless networking environment;

[0228] The model determination module 905 is used to adjust the parameters of the basic resource adjustment model and the basic evaluation model if the second output result does not meet the preset standard, and use the fourth indicator as the new third indicator to prompt the first output result obtaining module to return to execute the step of inputting the third indicator into the basic resource adjustment model to obtain the first output result, until the second output result meets the preset standard, and the trained resource adjustment model and evaluation model are obtained.

[0229] In one embodiment of the present application, the basic resource adjustment model and the basic evaluation model are trained based on the following modules:

[0230] a third output result obtaining module, configured to input the fifth indicator into the initial resource adjustment model to obtain a third output result, wherein the third output result represents a probability of the AP in the wireless network executing different resource adjustment actions;

[0231] a second simulated resource adjustment action determining module, configured to determine a simulated resource adjustment action to be performed by the AP based on the first output result;

[0232] a sixth indicator prediction module, configured to simulate a network state of the second wireless networking environment after adjusting the AP according to the simulated resource adjustment action, and predict a sixth indicator reflecting the network state;

[0233] a fourth output result obtaining module, configured to input the sixth indicator into an initial evaluation model to obtain a fourth output result, wherein the fourth output result reflects the network quality of the second wireless networking environment after the AP performs the simulated resource adjustment action;

[0234] The basic model determination module is used to adjust the parameters of the initial resource adjustment model and the initial evaluation model if the second output result does not meet the preset standard, use the sixth indicator as the new fifth indicator, return to execute the step of inputting the sixth indicator into the basic resource adjustment model to obtain the third output result, until the fourth output result meets the preset standard, and obtain the trained basic resource adjustment model and basic evaluation model.

[0235] In another embodiment provided in the present application, a computer-readable storage medium is further provided, wherein a computer program is stored in the computer-readable storage medium. When the computer program is executed by a processor, the steps of any of the above-mentioned radio frequency resource adjustment methods are implemented.

[0236] In another embodiment provided by the present application, a computer program product including instructions is further provided, which, when executed on a computer, enables the computer to execute any of the radio frequency resource adjustment methods in the above embodiments.

[0237] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When software is used for implementation, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website, computer, server or data center to another website, computer, server or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrations. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).

[0238] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0239] Each embodiment in this specification is described in a related manner. Similar portions between the embodiments can be referenced to each other. Each embodiment focuses on the differences from other embodiments. In particular, the electronic device and apparatus embodiments are generally similar to the method embodiments, so their descriptions are relatively simple. For related portions, reference can be made to the descriptions of the method embodiments.

[0240] The above description is only a preferred embodiment of the present application and is not intended to limit the scope of protection of the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application are included in the scope of protection of the present application.

Claims

1. A radio frequency resource adjustment method, characterized in that: The method comprises: Determining, using a pre-trained resource adjustment model, a resource adjustment action to be performed by an access point (AP) in a wireless network based on a first indicator, wherein the first indicator reflects the entire network status of the wireless network, and the resource adjustment action of the AP is: adjusting at least one of power, channel, and bandwidth of the AP; Simulating a simulated network state of the entire wireless network after the AP is adjusted according to the resource adjustment action, and predicting a second indicator reflecting the simulated network state; Obtaining a network quality test result according to the second indicator using a pre-trained evaluation model, wherein the network quality test result represents the network quality of the wireless network in the simulated network state; If the network quality test result does not meet the preset standard, the second indicator is used as the new first indicator, and the process returns to the step of determining the resource adjustment action to be performed by the access point AP in the wireless network based on the first indicator; If the network quality test result reaches the preset standard, the AP is controlled to adjust so that the wireless networking is updated to a simulated network state that reaches the preset standard.

2. The method according to claim 1, characterized in that The first indicator includes: a first quality indicator reflecting the overall network quality of the wireless network and / or a first interference indicator reflecting the interference status between the APs in the wireless network; and / or The second indicator includes: a second quality indicator reflecting the overall network quality of the wireless network after the AP is adjusted according to the resource adjustment action and / or a second interference indicator reflecting the interference status between the APs in the wireless network after the AP is adjusted according to the resource adjustment action.

3. A model training method, characterized in that: The method comprises: Inputting the third indicator into a pre-trained basic resource adjustment model to obtain a first output result, wherein the third indicator reflects the entire network state of the first wireless networking environment, and the first output result represents the probability of an AP in the wireless network performing different resource adjustment actions, wherein the basic resource adjustment model is trained based on a virtual second wireless networking environment, and the first wireless networking environment has the same network structure as the wireless network; determining a simulated resource adjustment action to be performed by the AP based on the first output result; simulating a network state of the first wireless networking environment after adjusting the AP according to the simulated resource adjustment action, and predicting a fourth indicator reflecting the network state; Inputting the fourth indicator into a pre-trained basic evaluation model to obtain a second output result, where the second output result reflects the network quality of the first wireless networking environment after the AP performs the simulated resource adjustment action, and the basic evaluation model is trained based on the second wireless networking environment; If the second output result does not meet the preset standard, the parameters of the basic resource adjustment model and the basic evaluation model are adjusted, the fourth indicator is used as the new third indicator, and the process returns to the step of inputting the third indicator into the basic resource adjustment model to obtain the first output result, until the second output result meets the preset standard, and the trained resource adjustment model and evaluation model are obtained.

4. The method according to claim 3, characterized in that The basic resource adjustment model and the basic evaluation model are trained based on the following methods: Inputting the fifth indicator into the initial resource adjustment model to obtain a third output result, wherein the fifth indicator reflects the entire network state of the second wireless networking environment, and the third output result represents the probability of the AP in the wireless network performing different resource adjustment actions; determining a simulated resource adjustment action to be performed by the AP based on the third output result; simulating a network state of the second wireless networking environment after adjusting the AP according to the simulated resource adjustment action, and predicting a sixth indicator reflecting the network state; Inputting the sixth indicator into the initial evaluation model to obtain a fourth output result, wherein the fourth output result reflects the network quality of the second wireless networking environment after the AP performs the simulated resource adjustment action; If the fourth output result does not meet the preset standard, the parameters of the initial resource adjustment model and the initial evaluation model are adjusted, the sixth indicator is used as the new fifth indicator, and the process returns to the step of inputting the sixth indicator into the basic resource adjustment model to obtain the third output result, until the fourth output result meets the preset standard, and the trained basic resource adjustment model and basic evaluation model are obtained.

5. An electronic device, characterized in that: The device comprises: processor; transceiver; A machine-readable storage medium storing machine-executable instructions capable of being executed by the processor; the machine-executable instructions prompting the processor to perform the following steps: Determining, using a pre-trained resource adjustment model, a resource adjustment action to be performed by an access point (AP) in a wireless network based on a first indicator, wherein the first indicator reflects the entire network status of the wireless network, and the resource adjustment action of the AP is: adjusting at least one of power, channel, and bandwidth of the AP; Simulating a simulated network state of the entire wireless network after the AP is adjusted according to the resource adjustment action, and predicting a second indicator reflecting the simulated network state; Obtaining a network quality test result according to the second indicator using a pre-trained evaluation model, wherein the network quality test result represents the network quality of the wireless network in the simulated network state; If the network quality test result does not meet the preset standard, the second indicator is used as a new first indicator, and the process returns to the step of determining the resource adjustment action to be performed by the access point AP in the wireless network based on the first indicator; If the network quality test result reaches the preset standard, the AP is controlled to adjust so that the wireless networking is updated to a simulated network state that reaches the preset standard.

6. The electronic device according to claim 5, characterized in that The first indicator includes: a first quality indicator reflecting the overall network quality of the wireless network and / or a first interference indicator reflecting the interference status between the APs in the wireless network; and / or The second indicator includes: a second quality indicator reflecting the overall network quality of the wireless network after the AP is adjusted according to the resource adjustment action and / or a second interference indicator reflecting the interference status between the APs in the wireless network after the AP is adjusted according to the resource adjustment action.

7. An electronic device, characterized in that: The device comprises: processor; transceiver; A machine-readable storage medium storing machine-executable instructions capable of being executed by the processor; the machine-executable instructions prompting the processor to perform the following steps: Inputting the third indicator into a pre-trained basic resource adjustment model to obtain a first output result, wherein the third indicator reflects the entire network state of the first wireless networking environment, and the first output result represents the probability of an AP in the wireless network performing different resource adjustment actions, wherein the basic resource adjustment model is trained based on a virtual second wireless networking environment, and the first wireless networking environment has the same network structure as the wireless network; determining a simulated resource adjustment action to be performed by the AP based on the first output result; simulating a network state of the first wireless networking environment after adjusting the AP according to the simulated resource adjustment action, and predicting a fourth indicator reflecting the network state; Inputting the fourth indicator into a pre-trained basic evaluation model to obtain a second output result, where the second output result reflects the network quality of the first wireless networking environment after the AP performs the simulated resource adjustment action, and the basic evaluation model is trained based on the second wireless networking environment; If the second output result does not meet the preset standard, the parameters of the basic resource adjustment model and the basic evaluation model are adjusted, the fourth indicator is used as the new third indicator, and the process returns to the step of inputting the third indicator into the basic resource adjustment model to obtain the first output result, until the second output result meets the preset standard, and the trained resource adjustment model and evaluation model are obtained.

8. A radio frequency resource adjustment device, characterized in that: The device comprises: a resource adjustment action determination module, configured to determine, using a pre-trained resource adjustment model and based on a first indicator, a resource adjustment action to be performed by an access point (AP) in a wireless network, wherein the first indicator reflects the entire network status of the wireless network, and the resource adjustment action of the AP is an action of adjusting at least one of the power, channel, and bandwidth of the AP; A second indicator prediction module is used to simulate the simulated network state of the entire wireless network after the AP is adjusted according to the resource adjustment action, and predict a second indicator reflecting the simulated network state; a network quality test result acquisition module, configured to obtain a network quality test result based on the second indicator using a pre-trained evaluation model, wherein the first indicator reflects the entire network status of the wireless network, and the network quality test result represents the network quality of the wireless network in the simulated network state; a first indicator updating module configured to, if the network quality test result does not meet a preset standard, use the second indicator as a new first indicator, prompting the resource adjustment action determining module to return to the step of determining the resource adjustment action to be performed by the access point AP in the wireless network based on the first indicator; The adjustment module is used to control the AP adjustment if the network quality test result reaches the preset standard, so that the wireless networking is updated to a simulated network state that reaches the preset standard.

9. A model training device, characterized in that: The device comprises: a first output result obtaining module, configured to input a third indicator into a pre-trained basic resource adjustment model to obtain a first output result, wherein the third indicator reflects the entire network state of the first wireless networking environment, and the first output result represents a probability of an AP in the wireless network performing different resource adjustment actions, wherein the basic resource adjustment model is trained based on a virtual second wireless networking environment, and the first wireless networking environment has the same network structure as the wireless network; a first simulated resource adjustment action determining module, configured to determine a simulated resource adjustment action to be performed by the AP based on the first output result; a fourth indicator prediction module, configured to simulate a network state of the first wireless networking environment after the AP is adjusted according to the simulated resource adjustment action, and predict a fourth indicator reflecting the network state; a second output result obtaining module, configured to input the fourth indicator into a pre-trained basic evaluation model to obtain a second output result, wherein the second output result reflects the network quality of the first wireless networking environment after the AP performs the simulated resource adjustment action, and the basic evaluation model is trained based on the second wireless networking environment; The model determination module is used to adjust the parameters of the basic resource adjustment model and the basic evaluation model if the second output result does not meet the preset standard, use the fourth indicator as the new third indicator, and prompt the first output result acquisition module to return to execute the step of inputting the third indicator into the basic resource adjustment model to obtain the first output result, until the second output result meets the preset standard, and the trained resource adjustment model and evaluation model are obtained.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method steps described in any one of claims 1-2 or 3-4 are implemented.

Images