Method and device for determining network energy consumption control strategy and electronic equipment

Abstract

The invention discloses a method and device for determining a network energy consumption control strategy and electronic equipment. The method comprises the following steps: acquiring feature data of a plurality of base stations in a communication network; analyzing the feature data by using a network energy consumption model to obtain an initial strategy; performing iterative optimization on the initial strategy under a multi-dimensional collaborative optimization framework to obtain a target strategy; and determining a network configuration parameter corresponding to the base station according to the target strategy, and configuring the communication network by using the network configuration parameter. According to the method and the device, the technical problem of relatively poor optimization effect caused by difficulty in realizing accurate control on the energy consumption of the network equipment when a network energy consumption optimization method adopted by the related technology faces a complex and changeable network environment is solved.

Description

Technical Field

[0001] The present application relates to the field of communication technologies, and in particular, to a method, an apparatus, and an electronic device for determining a network energy consumption control strategy. Background Art

[0002] With the rapid development of new network services such as 5G, the Internet of Things, and the Internet of Vehicles, increasingly severe real-world problems such as tight power supply and increasing environmental protection pressure force operators to efficiently utilize and dynamically manage resources such as spectrum, power, and network energy consumption, so as to reduce communication costs, improve user experience, and extend battery life.

[0003] The network energy consumption control strategies adopted by related technologies mainly focus on static optimization based on small-scale models. The energy consumption control strategy is obtained by solving the energy consumption minimization problem using a static optimization model. However, it ignores the interaction between base stations in the network and is difficult to accurately control the energy consumption of network devices in the face of a complex and changing network environment, resulting in poor optimization effects.

[0004] In response to the above problems, no effective solution has been proposed yet. Summary of the Invention

[0005] Embodiments of the present application provide a method, an apparatus, and an electronic device for determining a network energy consumption control strategy, so as to at least solve the technical problem that the network energy consumption optimization method adopted by related technologies is difficult to accurately control the energy consumption of network devices in the face of a complex and changing network environment, resulting in poor optimization effects.

[0006] According to one aspect of the embodiments of the present application, a method for determining a network energy consumption control strategy is provided, including: obtaining characteristic data of multiple base stations in a communication network, where the characteristic data is used to reflect the operating state of the base stations; analyzing the characteristic data using a network energy consumption model to obtain an initial strategy, where the network energy consumption model is used to predict the energy consumption of multiple base stations under different network configurations, and the initial strategy is used to control the energy consumption of the communication network; iteratively optimizing the initial strategy in a multi-dimensional collaborative optimization framework to obtain a target strategy, where the multi-dimensional collaborative optimization framework is used to represent the constraint relationship between the interaction between multiple base stations and the energy consumption; determining network configuration parameters corresponding to the base stations according to the target strategy, and configuring the communication network using the network configuration parameters.

[0007] In some embodiments of the present application, the initial policy is iteratively optimized under a multi-dimensional collaborative optimization framework to obtain a target policy, including: determining a first parameter under the initial policy, where the first parameter is used to quantitatively represent the influence degree of the network topologies corresponding to multiple base stations on energy consumption; determining a first matrix according to the first parameter and a second parameter, where the second parameter is used to quantitatively represent the intensity of energy consumption optimization, and the first matrix is used to reflect the sensitivity of network performance to energy consumption changes; predicting feature data by using the first matrix and a second matrix to obtain a prediction result, where the second matrix is a transition probability matrix between different time points of the network state, and the prediction result is used to reflect the change relationship of the feature data corresponding to multiple base stations over time within a preset time period; adjusting the control parameters in the initial policy according to the prediction result to obtain the target policy, where the control parameters are used to adjust the network configuration of the communication network.

[0008] In some embodiments of the present application, determining the first parameter under the initial policy includes: obtaining the geographical location information of multiple base stations and determining a topology matrix corresponding to the geographical location information, where the elements in the topology matrix are used to reflect the mutual influence degree between base stations located at different geographical locations; determining the initial network configuration parameters corresponding to multiple base stations in the initial policy; and determining the first parameter according to the topology matrix and the initial network configuration parameters.

[0009] In some embodiments of the present application, the topology matrix is a weighted adjacency matrix, where each element in the weighted adjacency matrix is used to reflect the mutual influence degree between the corresponding two base stations, and the weight corresponding to each element is determined according to the physical distance and signal strength between the two base stations.

[0010] In some embodiments of the present application, the second parameter is determined by the following method: determining a network load index corresponding to the feature data, where the network load index is used to quantitatively represent the difference degree between the current network load state and the average load state within a preset statistical period; increasing the initial setting value of the second parameter when the network load index is in a first threshold interval; and decreasing the initial setting value of the second parameter when the network load index is in a second threshold interval, where the end value of the second threshold interval is less than the start value of the first threshold interval.

[0011] In some embodiments of the present application, it further includes: determining the total number of base stations in the communication network; determining the number of regional divisions according to the total number of base stations when the total number of base stations exceeds a preset numerical control threshold; dividing the communication network into the number of regional division sub-regions, and respectively determining the target policy corresponding to each sub-region.

[0012] In some embodiments of the present application, it further includes: obtaining the computing resource utilization rate of a first sub-region, where the first sub-region is any one of the sub-regions divided by the number of regions; determining a second sub-region from the sub-regions divided by the number of regions when the computing resource utilization rate meets a preset condition, where the computing resource utilization rate of the second sub-region meets the computing task of the first sub-region; and migrating the computing task of the first sub-region from the computing node of the first sub-region to the computing node of the second sub-region.

[0013] According to another aspect of the embodiments of the present application, there is also provided a device for determining a network energy consumption control strategy, including: an acquisition module, configured to acquire characteristic data of multiple base stations in a communication network; an analysis module, configured to analyze the characteristic data by using a network energy consumption model to obtain an initial strategy; an optimization module, configured to iteratively optimize the initial strategy under a multi-dimensional collaborative optimization framework to obtain a target strategy; and an execution module, configured to determine network configuration parameters corresponding to the base stations according to the target strategy and configure the communication network by using the network configuration parameters.

[0014] According to still another aspect of the embodiments of the present application, there is also provided an electronic device, including: a memory and a processor, where the memory is used to store program instructions; the processor is connected to the memory and is configured to execute to implement the method for determining the above-mentioned network energy consumption control strategy.

[0015] According to still another aspect of the embodiments of the present application, there is also provided a non-volatile storage medium, which includes a stored computer program, where the device where the non-volatile storage medium is located executes the method for determining the above-mentioned network energy consumption control strategy by running the computer program.

[0016] According to still another aspect of the embodiments of the present application, there is also provided a computer program product, including computer instructions, where the computer instructions implement the method for determining the above-mentioned network energy consumption control strategy when executed by a processor.

[0017] In the embodiments of the present application, by adopting the method of combining a dynamic multi-dimensional collaborative optimization framework with real-time characteristic data and an intelligent network energy consumption model, through continuously iteratively optimizing the initial strategy and adaptively adjusting network configuration parameters, the purpose of balancing the accurate control of the energy consumption of the communication network and maintaining high performance is achieved, and the technical effect of significantly improving the energy consumption optimization efficiency and effect is realized. Furthermore, it solves the technical problem that the network energy consumption optimization method adopted in the related art is difficult to accurately control the energy consumption of network devices and results in poor optimization effects when facing a complex and changeable network environment. Description of the Drawings

[0018] The accompanying drawings described herein are used to provide a further understanding of the present application and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation of the present application. In the drawings:

[0019] Figure 1 is a hardware structure block diagram of a computer terminal for a method of determining a network energy consumption control strategy according to an embodiment of the present application;

[0020] Figure 2 is a flowchart of a method of determining a network energy consumption control strategy according to an embodiment of the present application;

[0021] Figure 3 is a flowchart of training a network energy consumption model for a method of determining a network energy consumption control strategy according to an embodiment of the present application;

[0022] Figure 4 is an iterative optimization flowchart of a multi-dimensional collaborative optimization framework for a method of determining a network energy consumption control strategy according to an embodiment of the present application;

[0023] Figure 5 is a schematic structural diagram of a device for determining a network energy consumption control strategy according to an embodiment of the present application. Detailed Embodiments

[0024] In order to enable those skilled in the art to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0025] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned accompanying drawings are used to distinguish similar objects and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily need to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0026] In order to better understand the embodiments of the present application, the technical terms involved in the embodiments of the present application are explained as follows:

[0027] Multi-Dimensional Collaborative Optimization Framework: An optimization technique that comprehensively considers multiple optimization objectives and constraints, and seeks the global optimal solution by coordinating the interactions between variables in different dimensions (such as space, time, resource allocation, etc.).

[0028] Network Topology-Energy Consumption Correlation Model (abbreviated as NTC): A mathematical model used to describe the relationship between network structure characteristics (such as connectivity, distance, signal strength, etc.) and system energy consumption, and can quantify the impact degree of the topological structure between base stations on energy consumption. In this application, the NTC model accurately captures the complex interactions between base stations by introducing the theory of complex variable functions and the concept of entanglement in topology, provides key topological information for formulating energy consumption optimization strategies, is the basis for constructing the first parameter (topological entanglement energy consumption metric), and ensures the accuracy and efficiency of the optimization algorithm.

[0029] Network Performance-Energy Consumption Collaborative Impact Matrix Model (abbreviated as NP-CIM): An analysis tool used to evaluate the sensitivity of network performance metrics (such as latency, throughput, quality of service, etc.) to energy consumption changes, and measures the coupling effect between performance and energy consumption under different network configurations by establishing a matrix. In this application, the NP-CIM model is used to associate the connection between network performance and energy consumption optimization objectives during the optimization process, helps the system effectively control energy consumption while improving performance, and avoids sacrificing user experience by simply pursuing energy conservation.

[0030] Network State Evolution Dynamics Model (abbreviated as NSED): A model constructed based on the principles of dynamics, used to predict the evolution law of network states over time, including the change trends of key parameters such as network traffic, base station load, and user behavior. In this application, the NSED model can accurately predict future network states through state transition dynamics equations. For example, when dealing with sudden traffic peaks or network failures, this model can quickly adjust optimization strategies to maintain the stable operation of the network.

[0031] Multi-Objective Optimization-Based Control Strategy Model (MOOCSM): A control theory method used to find the optimal control strategy among multiple conflicting or complementary objectives (such as minimizing energy consumption, maximizing performance, maintaining fairness, etc.), ensuring the overall improvement of system performance. In this application, MOOCSM achieves a dynamic balance between energy consumption and network performance by iteratively optimizing control parameters. In actual deployment, this model can quickly respond to network changes, generate control strategies that meet the requirements of multi-objective optimization, and ensure the efficient operation of the network and the realization of green communication.

[0032] With the full deployment of 5G networks and the continuous evolution of 6G technologies, the energy consumption problem of communication infrastructure has become increasingly prominent, and there are many problems with the energy consumption optimization methods adopted by related technologies. First of all, most methods adopt a single optimization objective, only focusing on reducing energy consumption while ignoring network performance and user experience. This one-sided pursuit of energy conservation often leads to a decline in service quality in actual applications. Secondly, most related technologies use simple neural networks or reinforcement learning algorithms, which are difficult to accurately capture the complex interactions between base stations and network topology characteristics. This algorithmic limitation makes the optimization results often locally optimal and unable to achieve global collaborative optimization of the network. In addition, related technologies adopt a centralized architecture, where the optimization decisions of all base stations are uniformly calculated and issued by the central server. This architecture has serious computational latency and communication overhead problems when facing large-scale networks. Especially when the network state changes rapidly, centralized optimization often cannot keep up with the network dynamics, resulting in poor optimization effects.

[0033] To solve the above technical problems, the embodiments of this application provide corresponding solutions, which are described in detail below.

[0034] The method embodiment for determining the network energy consumption control strategy provided by the embodiments of this application can be executed in a mobile terminal, a computer terminal, or a similar computing device. Figure 1 The hardware structure block diagram of a computer terminal for implementing the method for determining the network energy consumption control strategy is shown. As Figure 1As shown, the computer terminal 10 may include one or more processors (illustrated as 102a, 102b, ……, 102n in the figure) (the processor may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission module 106 for communication functions connected via wired and / or wireless networks. In addition, it may further include: a display, a keyboard, a cursor control device, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, and a BUS bus. Those of ordinary skill in the art can understand that Figure 1 the structure shown is only schematic and does not limit the structure of the above-mentioned electronic device. For example, the computer terminal 10 may further include more or fewer components than Figure 1 shown in, or have a different configuration from Figure 1 that shown.

[0035] It should be noted that the above one or more processors and / or other data processing circuits are generally referred to as "data processing circuits" herein. The data processing circuit may be embodied in whole or in part as software, hardware, firmware, or any combination thereof. In addition, the data processing circuit may be a single independent processing module, or be incorporated in whole or in part into any one of the other elements in the computer terminal 10. As involved in the embodiments of the present application, the data processing circuit is a kind of processor control (such as the selection of a variable resistance terminal path connected to an interface).

[0036] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the method for determining the network energy consumption control strategy in the embodiments of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implements the method for determining the network energy consumption control strategy described above. The memory 104 may include a high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memories. In some instances, the memory 104 may further include a memory remotely set relative to the processor, and these remote memories can be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0037] The transmission module 106 is used to receive or send data via a network. Specific examples of the above-mentioned network may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission module 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission module 106 can be a Radio Frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0038] The display can be, for example, a touch-screen liquid crystal display (LCD), which enables users to interact with the user interface of the computer terminal 10.

[0039] It should be noted here that in some alternative embodiments, the above Figure 1 illustrated computer terminal may include hardware elements (including circuits), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware elements and software elements. It should be pointed out that Figure 1 is only an example of a specific specific instance and is intended to show the types of components that may exist in the above computer terminal.

[0040] Under the above operating environment, an embodiment of a method for determining a network energy consumption control strategy is provided in the embodiments of the present application. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0041] Figure 2 is a flowchart of a method for determining a network energy consumption control strategy according to an embodiment of the present application. As Figure 2 shown, the method includes the following steps:

[0042] Step S202, obtaining characteristic data of multiple base stations in a communication network, where the characteristic data is used to reflect the operating status of the base stations.

[0043] In the above step S202, the characteristic data includes a series of indicators that can comprehensively reflect the operating status of the base stations, such as the real-time traffic, signal strength, power consumption level, number of user connections, device health status, etc. of the base stations, reflecting the current working mode, device performance parameters, and communication quality of the users connected thereto. The characteristic data constitutes a real-time snapshot of the network status, enabling the system to accurately perceive changes in the network environment.

[0044] In some embodiments of the present application, the operating data of the base station can be collected in real time through sensors installed on the base station and the built-in monitoring system of the communication device. It should be noted that the data collection frequency can be dynamically adjusted according to network requirements, generally set between 1 and 10 minutes to ensure the real-time nature of the data and the feasibility of processing. For example, when the system detects a sudden increase in traffic in a certain area, it will automatically increase the monitoring frequency of the base stations in that area to once every 1 minute to capture network status changes in a timely manner. This method solves the problem of optimization delay caused by data lag in traditional optimization methods, ensuring that the system can quickly respond to network changes and adjust the energy consumption strategy in a timely manner.

[0045] In a large data environment, the high dimensionality of feature data easily leads to low computing efficiency. To solve this problem, the collected original feature data is cleaned and normalized to exclude invalid or redundant information, and then data mining techniques such as principal component analysis (PCA) and clustering analysis are used to screen out the key feature set that has the greatest impact on energy consumption. For example, through the analysis of historical data, it is found that the correlation between the base station power consumption and the number of user connections and signal strength is the highest, so these two features are preferentially retained.

[0046] Step S204: Analyze the feature data using the network energy consumption model to obtain an initial strategy, where the network energy consumption model is used to predict the energy consumption of multiple base stations under different network configurations, and the initial strategy is used to control the energy consumption of the communication network.

[0047] In the above step S204, the network energy consumption model is used to quantify and predict the energy consumption of multiple base stations in the communication network under given network configuration conditions. For example, through methods such as deep neural networks and complex variable function theory, the feature data of the base station (such as traffic, power, renewable energy supply situation, etc.) is integrated to predict the energy consumption under different configurations. In some embodiments of the present application, a deep neural network (Deep Neural Network, abbreviated as DNN) can be used to analyze the feature data of the base station and train a model that can predict the energy consumption under different network configurations. The model contains multiple hidden layers to enhance its learning ability for the complex relationships of the feature data. For example, the input layer receives information such as the traffic and power usage of the base station, the intermediate layer performs feature transformation through complex variable function theory, and the output layer gives the energy consumption prediction value.

[0048] Figure 3 is a flowchart of the training of the network energy consumption model of a method for determining a network energy consumption control strategy according to an embodiment of the present application, as Figure 3 shown. In some embodiments of the present application, the training process may include the following steps:

[0049] Step S302: Start, indicating the starting point of the network energy consumption model training process.

[0050] Step S304: Input the base station feature data, and import the key operating status information from the real-time or historical data stream of the base station, including but not limited to data such as the power consumption, working mode, signal strength, ambient temperature, number of user connections, traffic type and size of the base station. These data are transmitted to the model training system through a network interface or a data pipeline as the input source for model training.

[0051] Step S306: Data preprocessing, perform cleaning, normalization and standardization on the input base station feature data, remove outliers and missing data to ensure the quality and consistency of the data. At the same time, convert non-numeric features into numeric features for easy model processing.

[0052] Step S308: Feature selection, use methods such as statistical analysis, principal component analysis (PCA), and correlation analysis to screen out the features that have a greater impact on network energy consumption. For example, by calculating the correlation coefficients between features, select the top N% (such as the top 20%) of the features most relevant to energy consumption, or use algorithms such as recursive feature elimination (RFE) for feature selection.

[0053] Step S310: Build a deep neural network, design a deep neural network architecture according to the dimension of the feature data and the optimization goal, including an input layer, multiple hidden layers and an output layer. The hidden layer can adopt convolutional layers, fully connected layers, LSTM layers, etc., combined with advanced techniques such as attention mechanisms or residual connections to enhance the model's expressive ability and learning efficiency.

[0054] Step S312: Model training, input the preprocessed base station feature data into the deep neural network, and use the backpropagation algorithm for model training to minimize the error between the predicted energy consumption and the actual energy consumption. During the training process, cross-validation or an early stopping mechanism can be used to avoid overfitting and ensure the robustness and generalization ability of the model.

[0055] Step S314: Judge whether the performance meets the requirements. After the model training is completed, evaluate the prediction performance of the model through a test data set, such as calculating indicators such as the mean absolute error (MAE), mean square error (MSE), or correlation coefficient between the predicted energy consumption and the true energy consumption, and judge whether the model reaches the predetermined performance threshold.

[0056] Step S316: In the case where the performance does not meet the requirements, adjust the model parameters. If it is found in step S314 that the model performance does not meet the standard, the system will automatically or manually adjust the model parameters, such as increasing the number of network layers, changing the learning rate, adjusting the regularization term, or replacing the optimization algorithm, and then return to step S312 for retraining.

[0057] Step S318: Save the model when the performance meets the requirements. Once the performance of the model passes the evaluation in Step S314, the system will save all the parameters of the model, including weights, biases, network structure, etc., for use in subsequent steps.

[0058] Step S320: End, which represents the end point of the network energy consumption model training process. The successfully trained model will be used in the subsequent network energy consumption control strategy generation phase.

[0059] The initial strategy is the first-round control strategy derived by the network energy consumption model based on feature data analysis, and is used to guide the initial energy consumption management of the base station, such as adjusting the base station working mode, power distribution, etc. In some embodiments of the present application, the preprocessed feature data is used as input and fed into the trained network energy consumption model for prediction. The model will calculate the expected energy consumption value under the current network configuration according to the input feature data, and convert the abstract network operation state into a specific energy consumption result; according to the energy consumption result predicted by the model, a set of control parameters are generated to guide how the base station adjusts the working mode, power distribution, resource scheduling, etc. to achieve the purpose of energy saving or energy efficiency optimization. For example, if the model predicts that the energy consumption of a base station in a certain area is high, the generated control parameters include reducing the transmission power of the base station in that area, increasing the enabling frequency of the intelligent sleep mode, optimizing traffic distribution, etc.; based on the control parameters, an initial energy consumption control strategy (i.e., the initial strategy) is formed.

[0060] Step S206: Iteratively optimize the initial strategy under the multi-dimensional collaborative optimization framework to obtain the target strategy, where the multi-dimensional collaborative optimization framework is used to characterize the constraint relationship between the interactions among multiple base stations and energy consumption.

[0061] In the above Step S206, the multi-dimensional collaborative optimization framework is a set of mathematical models and algorithms that comprehensively consider the interaction among multiple base stations in the network, the constraint relationship between energy consumption and network performance.

[0062] Figure 4 It is an iterative optimization flowchart of the multi-dimensional collaborative optimization framework for a method for determining a network energy consumption control strategy according to an embodiment of the present application. As Figure 4 shown, in some embodiments of the present application, the initial strategy can be iteratively optimized through the following steps to obtain the target strategy. Specifically:

[0063] Step S402: Start, which is the starting point of the optimization process.

[0064] Step S404: Construct a topological entanglement energy consumption metric (corresponding to the network topology and energy consumption correlation model). Based on the physical locations of base stations in the network, signal coverage ranges, communication link states, etc., construct a topological entanglement metric under the theory of complex variable functions to quantify the impact of the interaction between base stations on the overall network energy consumption. This metric is achieved by calculating the Riemann-Zeta function values under the complex variable parameters between base stations, reflecting the potential impact of the dynamic changes in network topological characteristics on energy consumption.

[0065] Step S406: Calculate the performance-energy consumption coupling matrix (corresponding to the network performance and energy consumption co-influence matrix model). Establish a matrix model, where the elements of the matrix describe the coupling relationship between network performance indicators (such as signal quality, number of user connections, data transmission rate, etc.) and energy consumption. By analyzing historical data, it is possible to determine how the energy consumption level of each base station is affected by changes in network performance, and how network performance changes with energy consumption adjustment.

[0066] Step S408: Construct the state transition dynamics equation (corresponding to the network state evolution dynamics model). Design a set of dynamics equations to describe the law of network state change over time, taking into account factors such as control strategies, topological entanglement, and performance-energy consumption coupling. These equations form a closed loop, capable of reflecting how the network dynamically adjusts under different control strategies, and the immediate and long-term impacts of such adjustments on energy consumption and performance.

[0067] Step S410: Optimize the control strategy (corresponding to the control strategy model based on multi-objective optimization). Based on the constructed topological entanglement energy consumption metric, performance-energy consumption coupling matrix, and state transition dynamics equation, adopt multi-objective optimization algorithms (such as Pareto optimization, genetic algorithm, particle swarm optimization, etc.) to find the control strategy solutions that can simultaneously meet multiple objectives such as the lowest energy consumption, the best network performance, and the highest utilization of renewable energy, so as to adjust the initial strategy to achieve the optimal state under the multi-dimensional collaborative optimization framework and realize the unity of energy consumption, performance, and green communication.

[0068] Step S412: Judge whether it converges. Check whether the optimization process has reached the predetermined convergence criteria. For example, whether the adjustment amount of the control strategy is less than the preset threshold, whether the improvement rate of the optimization objective function has decreased to a certain level, etc.

[0069] Step S414: Update the parameters. If it is judged in Step S412 that the optimization process has not converged, the system will adjust the parameters of the optimization algorithm, such as the learning rate, search range, regularization term, etc., and then return to Step S406 to perform the optimization calculation again.

[0070] Step S416: Output the optimal control strategy. When it is determined in step S412 that the optimization process converges, the system will output the final control strategy, that is, the target strategy, for the energy consumption management of the actual network.

[0071] Step S418: End, which is the termination point of the optimization process. At this time, the iterative optimization process is completed, and the target strategy is generated and available for execution.

[0072] In some embodiments of the present application, the target strategy is determined through the following steps: determining a first parameter under the initial strategy, where the first parameter is used to quantitatively represent the influence degree of the network topology structure corresponding to multiple base stations on the energy consumption; determining a first matrix according to the first parameter and a second parameter, where the second parameter is used to quantitatively represent the intensity of energy consumption optimization, and the first matrix is used to reflect the sensitivity of network performance to energy consumption changes; using the first matrix and a second matrix to predict the feature data to obtain a prediction result, where the second matrix is the transition probability matrix between different time points of the network state, and the prediction result is used to reflect the change relationship of the feature data corresponding to multiple base stations over time within a preset time period; adjusting the control parameters in the initial strategy according to the prediction result to obtain the target strategy, where the control parameters are used to adjust the network configuration of the communication network.

[0073] Specifically, the training of the network energy consumption strategy can be achieved progressively through four mathematical models, which are the topological entanglement energy consumption metric ε(s), the performance-energy consumption coupling matrix C(ω), the state transition dynamics equation and the optimal control strategy U * , and this progressive model design ensures the comprehensiveness and accuracy of the optimization process.

[0074] (1) The calculation formula of the topological entanglement energy consumption metric is:

[0075]

[0076] where ε(s) is the topological entanglement energy consumption metric (the first parameter); ζ(s) is the Riemann-Zeta function, which is used to characterize the complexity of the network topology structure; N is the number of base stations; s is a complex variable parameter, which is used to control the convergence of the Riemann-Zeta function and the sensitivity to topological features; T is the network topology matrix, which represents the weighted adjacency matrix of the physical connection relationship and signal strength between base stations; E is the network energy consumption matrix, which records the energy consumption situation of each base station in the network.

[0077] The first parameter can be determined in the following manner: Obtain the geographical location information of multiple base stations, and determine the topology matrix corresponding to the geographical location information, where the elements in the topology matrix are used to reflect the degree of mutual influence between base stations located at different geographical locations; determine the initial network configuration parameters corresponding to the multiple base stations in the initial policy; and determine the first parameter based on the topology matrix and the initial network configuration parameters.

[0078] It should be noted that the topology matrix can be a weighted adjacency matrix, where each element in the weighted adjacency matrix is used to reflect the degree of mutual influence between the corresponding two base stations, and the weight corresponding to each element is determined based on the physical distance and signal strength between the two base stations.

[0079] In some embodiments of the present application, s = 2 + i can be selected as the initial value of the complex variable parameter (indicating that s is a complex number with a real part of 2 and an imaginary part of 1), which is an empirical value obtained based on a large number of experiments. At this point, the Riemann-Zeta function has good convergence. The network topology matrix T can be in the form of a weighted adjacency matrix, where the weight is calculated based on the physical distance and signal strength between the base stations. For example, for two base stations 100 meters apart, their weight can be set to 0.8, while the weight corresponding to base stations 500 meters apart can be reduced to 0.3. This way of setting weights can truly reflect the degree of mutual influence between the base stations.

[0080] (2) The calculation formula for the performance - energy consumption coupling matrix is:

[0081]

[0082] Among them, C(ω) is the performance - energy consumption coupling matrix (the first matrix); ω is the entanglement degree parameter (the second parameter); λ N is the eigenvalue, used to reflect the inherent characteristics of each base station; ε(s) ij (taking ij as an example) is the specific value of the energy consumption metric generated by topological entanglement between the i-th base station and the j-th base station.

[0083] The second parameter can be determined in the following manner: Determine the network load index corresponding to the characteristic data, where the network load index is used to quantitatively represent the difference degree between the current network load status and the average load status within the preset statistical period; when the network load index is in the first threshold interval, increase the initial setting value of the second parameter; when the network load index is in the second threshold interval, decrease the initial setting value of the second parameter, where the end value of the second threshold interval is less than the start value of the first threshold interval.

[0084] In practical applications, the selection of the entanglement degree parameter ω is crucial. ω can be initially set to 0.5 and then dynamically adjusted according to the network state. For example, when the network load is heavy, the value of ω can be appropriately increased to strengthen the intensity of energy consumption optimization; when the network load is light, the value of ω can be decreased to ensure communication quality.

[0085] The eigenvalue λ in the matrix N reflects the inherent characteristics of each base station and can be extracted from historical data through the principal component analysis (PCA) method. For example, the first 20% of the principal components are taken.

[0086] By introducing complex variable functions and topological entanglement theory, this application has successfully solved the problem that traditional methods are difficult to accurately describe the complex interactions between base stations. For example, in a 5G network test with 100 base stations, the method of this application captures approximately 30% more implicit associations between base stations than traditional methods, and these associations play a key role in energy consumption optimization.

[0087] It should be noted that during the actual deployment process, the algorithm parameters can be dynamically adjusted according to the network scale. For example, for a small network (not exceeding the first preset number of base stations, such as 50), the number of eigenvalues can be set to 1 / 2 of the total number of base stations; for a large network (exceeding the second preset number of base stations, such as 200), this ratio can be reduced to 1 / 4 to balance the calculation efficiency and optimization effect. In this way, the accuracy of energy consumption optimization is improved, and a rapid response to the dynamic changes of the network is achieved. In practical applications, this application can reduce energy consumption by 15 - 20% while maintaining network performance, which is significantly better than related technologies. Moreover, this application has strong scalability and can adapt to communication networks of different scales and types.

[0088] (3) The calculation formula for the state transition dynamics equation is:

[0089]

[0090] where is the rate of change of the state transition matrix with time (prediction result), indicating how the network state changes with time under the adjustment of the network control strategy; Φ is the state transition matrix (the second matrix); K is the upper limit of the derivative order; is the partial derivative term, reflecting the sensitivity of the model to changes in the entanglement parameter ω; T is the transpose.

[0091] By combining the state transition matrix Φ and the coupling matrix C(ω), the dynamics equation forms a closed-loop evolution system. In practical applications, since the contribution of the high-order derivative term decreases rapidly, the influence after more than 3 orders can be ignored, and the upper limit of the derivative order K can be set to 3.

[0092] In some embodiments of the present application, the initial value of the state transition matrix Φ can be obtained through statistical analysis of historical data. For example, network status data for the past 24 hours can be selected, the transition probabilities between adjacent time instants can be calculated, and an initial state transition matrix can be formed. This initialization method based on actual data can accelerate the convergence speed of the system.

[0093] In the calculation formula of the above state transfer dynamics equation, the partial derivative term reflects the sensitivity of the system to changes in entanglement parameters. When the network load undergoes a sudden change, these high-order derivative terms can help the system quickly adjust to a new equilibrium state. For example, during a large-scale sports event, when the base station traffic suddenly increases by 300%, the method adopted in the present application can complete the re-optimization of network configuration within 5 minutes, while the traditional method takes 15 - 20 minutes.

[0094] (4) Solving the optimal control strategy, whose calculation formula is:

[0095]

[0096] where, U * is the optimal control strategy (i.e., the target strategy); represents finding the strategy that minimizes the objective function among all possible control strategy sets U; T is the time period; γ is the regularization parameter, used to control the penalty degree of the optimization process for the change amplitude of the control strategy; U is the control strategy; is the square of the Frobenius norm, measuring the strength of the control strategy U, that is, the degree to which it changes the network operation state.

[0097] In some embodiments of the present application, the time period T can be set to 30 minutes. This value is an empirical value obtained based on a large number of experiments and can achieve a good balance between response real-time performance and optimization stability. The selection of the regularization parameter γ can adopt an adaptive adjustment mechanism. For example, when the system is in a stable state, the value of γ is appropriately increased (between 0.1 and 0.5) to prevent over-optimization; when the system needs to respond quickly, the value of γ is decreased (can be reduced to about 0.01) to improve the flexibility of control.

[0098] Step S208, according to the target strategy, determine the network configuration parameters corresponding to the base station, and configure the communication network using the network configuration parameters.

[0099] In the above step S208, the target strategy refers to the optimal control strategy obtained through the multi-dimensional collaborative optimization framework. The network configuration parameters include but are not limited to the transmission power of the base station, frequency allocation, energy consumption mode (such as sleep, energy saving, high performance, etc.), access and utilization methods of renewable energy, working state of the heat dissipation system, etc. These parameters are the specific operation points for implementing the target strategy.

[0100] In some embodiments of the present application, the following steps may also be performed: determining the total number of base stations in the communication network; in the case where the total number of base stations exceeds a preset numerical control threshold, determining the number of area divisions according to the total number of base stations; dividing the communication network into the number of area divisions of sub-areas, and respectively determining the target strategy corresponding to each sub-area.

[0101] Specifically, by counting the total number of base stations in the communication network and comparing this number with the preset numerical control threshold, it is determined whether area division is required. For example, the preset numerical control threshold is 200 base stations. When the counted number of base stations exceeds 200, the system starts the area division program. Once it is determined that area division is required, the system determines the number of area divisions according to the total number of base stations and the ideal number of base stations in each sub-area (for example, 20 - 30). According to the network status and resource characteristics of the sub-areas, the target strategy of each sub-area is determined separately.

[0102] The present application also introduces a dynamic load balancing mechanism. Specifically: obtaining the computing resource utilization rate of the first sub-area, where the first sub-area is any one of the number of area divisions of sub-areas; in the case where the computing resource utilization rate meets the preset conditions, determining the second sub-area from the number of area divisions of sub-areas, where the computing resource utilization rate of the second sub-area meets the computing tasks of the first sub-area; migrating the computing tasks of the first sub-area from the computing nodes of the first sub-area to the computing nodes of the second sub-area.

[0103] In actual deployment, a distributed computing architecture can be adopted to decompose complex optimization problems into multiple sub-problems for parallel solution. When the computing load of a certain sub-area is too heavy, some computing tasks can be automatically migrated to a sub-area with a lighter load. Practice shows that this mechanism can increase the utilization rate of computing resources by about 25% and reduce the optimization delay by 40% at the same time.

[0104] It should be noted that the present application may also include a fault recovery mechanism. When a base station fault is detected, the topological entanglement energy consumption metric can be recalculated immediately and the control strategy can be adjusted accordingly. Specifically, a three-level fault response mechanism is set: for minor faults (such as the influence range is less than 3 base stations), the system adopts a local adjustment strategy; for medium faults (such as the influence range is 3 - 10 base stations), area re-optimization is started; for severe faults (such as the influence range exceeds 10 base stations), full network optimization is performed.

[0105] In some embodiments of the present application, it further includes an intelligent energy-saving suggestion generator, which includes two key neural network modules, respectively for device recognition mapping and control strategy generation. Preferably, the first neural network adopts a deep convolutional structure, including 5 convolutional layers and 3 fully connected layers. The input layer receives device description information, which includes features such as device type, model, operating parameters, etc., and encodes these features into a 768-dimensional vector. For example, for a 5G base station cooling system, its feature vector will include key parameters such as ambient temperature, device power consumption, and cooling efficiency. The second neural network is responsible for generating specific control parameters. By adopting a recurrent neural network (LSTM) structure enhanced by an attention mechanism, it can effectively capture the temporal dependence relationship between device control parameters. For example, specifically, when dealing with the control strategy of an air conditioning system, the network will automatically learn the optimal coordination between the temperature setting value, wind speed adjustment, and start-stop timing, so as to achieve refined management of energy consumption.

[0106] In some embodiments of the present application, it further includes a data format converter, which includes three cascaded neural networks. The first network is responsible for policy conversion, converting high-level energy consumption optimization goals into specific control parameters. For example, a residual structure can be adopted, including 15 residual blocks, and each residual block includes two convolutional layers and a short-circuit connection. This structural design significantly improves the convergence speed and optimization effect of the network. In actual tests, the residual structure shortens the training time by about 40% compared with a ordinary feedforward network. The second network focuses on instruction generation, converting control parameters into an instruction sequence that can be directly executed by the device. By introducing an instruction optimization module based on a graph attention mechanism, it can automatically learn the dependence relationship between different control instructions and avoid instruction conflicts. For example, when dealing with the power supply system of a base station, this module can ensure that the switching instruction of the backup power supply perfectly matches the control instruction of the main power supply, effectively avoiding the risk of power supply interruption. The third network is responsible for hardware compatibility inspection. This network can adopt a method of modeling hardware characteristics based on a graph neural network, which can accurately evaluate the matching degree between control instructions and actual hardware. For example, a compatibility scoring mechanism between 0 and 1 can be designed, and when the score is lower than a preset threshold (such as 0.8), the instruction adjustment process is triggered.

[0107] Through the above steps S202 to S208, by adopting a dynamic multi-dimensional collaborative optimization framework to combine real-time feature data with an intelligent network energy consumption model, continuously iterating and optimizing the initial strategy and adaptively adjusting network configuration parameters, the purpose of balancing the precise control of communication network energy consumption and maintaining high performance is achieved, and the technical effect of significantly improving the energy consumption optimization efficiency and effect is realized. Furthermore, it solves the technical problem that the network energy consumption optimization method adopted in the related technology is difficult to achieve precise control of network device energy consumption and results in poor optimization effect when facing a complex and changeable network environment.

[0108] Figure 5 It is a structural diagram of a device for determining a network energy consumption control strategy according to an embodiment of the present application. As Figure 5 shown, the device includes:

[0109] An acquisition module 502, configured to acquire characteristic data of multiple base stations in a communication network;

[0110] An analysis module 504, configured to analyze the characteristic data by using a network energy consumption model to obtain an initial strategy;

[0111] An optimization module 506, configured to iteratively optimize the initial strategy under a multi-dimensional collaborative optimization framework to obtain a target strategy;

[0112] An execution module 508, configured to determine network configuration parameters corresponding to the base stations according to the target strategy, and configure the communication network by using the network configuration parameters.

[0113] It should be noted that Figure 5 the device for determining the network energy consumption control strategy shown is used to execute Figure 2 the method for determining the network energy consumption control strategy shown. Therefore Figure 2 the relevant explanations in the method for determining the network energy consumption control strategy in Figure 5 also apply to the device for determining the network energy consumption control strategy shown, which will not be elaborated here.

[0114] An embodiment of the present application further provides an electronic device, which includes a memory and a processor. The memory is used to store program instructions, and the processor is connected to the memory and is configured to execute the steps of implementing the method for determining the network energy consumption control strategy in each embodiment of the present application.

[0115] For example, the processor executes the following functions by executing the program instructions stored in the memory: acquiring characteristic data of multiple base stations in a communication network, where the characteristic data is used to reflect the operating status of the base stations; analyzing the characteristic data by using a network energy consumption model to obtain an initial strategy, where the network energy consumption model is used to predict the energy consumption of multiple base stations under different network configurations, and the initial strategy is used to control the energy consumption of the communication network; iteratively optimizing the initial strategy under a multi-dimensional collaborative optimization framework to obtain a target strategy, where the multi-dimensional collaborative optimization framework is used to characterize the constraint relationship between the interaction between multiple base stations and the energy consumption; determining network configuration parameters corresponding to the base stations according to the target strategy, and configuring the communication network by using the network configuration parameters.

[0116] An embodiment of the present application also provides a non-volatile storage medium, which includes a stored computer program. Wherein, the device where the non-volatile storage medium is located executes the steps of the method for determining the network energy consumption control strategy in each embodiment of the present application by running the computer program.

[0117] An embodiment of the present application also provides a computer program product, including computer instructions, which implement the steps of the method for determining the network energy consumption control strategy in each embodiment of the present application when executed by a processor.

[0118] An embodiment of the present application also provides a computer program, which implements the steps of the method for determining the network energy consumption control strategy in each embodiment of the present application when executed by a processor.

[0119] The serial numbers of the above embodiments of the present application are only for description and do not represent the advantages or disadvantages of the embodiments.

[0120] In the above embodiments of the present application, the descriptions of each embodiment have their own focuses. For the parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0121] In several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the division of the units can be a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of the units or modules can be in electrical or other forms.

[0122] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0123] In addition, the functional units in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0124] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), mobile hard disks, magnetic disks, or optical discs.

[0125] The above are only the preferred embodiments of this application. It should be noted that for those of ordinary skill in the art, without departing from the principle of this application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this application.

Claims

1. A method for determining a network energy consumption control strategy, characterized in that Including: Obtain the characteristic data of multiple base stations in a communication network, where the characteristic data is used to reflect the operating status of the base stations; Analyze the characteristic data by using a network energy consumption model to obtain an initial strategy, where the network energy consumption model is used to predict the energy consumption of the multiple base stations under different network configurations, and the initial strategy is used to control the energy consumption of the communication network; Iteratively optimize the initial strategy in a multi-dimensional collaborative optimization framework to obtain a target strategy, where the multi-dimensional collaborative optimization framework is used to characterize the constraint relationship between the interaction among the multiple base stations and the energy consumption; According to the target strategy, determine the network configuration parameters corresponding to the base stations, and configure the communication network by using the network configuration parameters.

2. The method according to claim 1, wherein Iteratively optimize the initial strategy in a multi-dimensional collaborative optimization framework to obtain a target strategy, including: Determine a first parameter under the initial strategy, where the first parameter is used to quantitatively represent the influence degree of the network topology structure corresponding to the multiple base stations on the energy consumption; Determine a first matrix according to the first parameter and a second parameter, where the second parameter is used to quantitatively represent the intensity of energy consumption optimization, and the first matrix is used to reflect the sensitivity degree of network performance to energy consumption changes; Use the first matrix and the second matrix to predict the characteristic data to obtain a prediction result, where the second matrix is a transition probability matrix between different time points of the network state, and the prediction result is used to reflect the change relationship of the characteristic data corresponding to the multiple base stations over time within a preset time period; Adjust the control parameters in the initial strategy according to the prediction result to obtain the target strategy, where the control parameters are used to adjust the network configuration of the communication network.

3. The method according to claim 2, wherein Determine the first parameter under the initial strategy, including: Obtain the geographical location information of the multiple base stations, and determine a topology matrix corresponding to the geographical location information, where the elements in the topology matrix are used to reflect the mutual influence degree between the base stations located at different geographical locations; Determine the initial network configuration parameters corresponding to the multiple base stations in the initial strategy; Determine the first parameter according to the topology matrix and the initial network configuration parameters.

4. The method according to claim 3, characterized in that, The topology matrix is a weighted adjacency matrix, where each element in the weighted adjacency matrix is used to reflect the mutual influence degree between the corresponding two base stations, and the weight corresponding to each element is determined according to the physical distance and signal strength of the two base stations.

5. The method according to claim 2, wherein The second parameter is determined by the following method: Determine a network load index corresponding to the characteristic data, where the network load index is used to quantitatively represent the difference degree between the current network load status and the average load status within a preset statistical period; When the network load index is in a first threshold interval, increase the initial setting value of the second parameter; When the network load index is in a second threshold interval, decrease the initial setting value of the second parameter, where the end value of the second threshold interval is less than the start value of the first threshold interval.

6. The method according to claim 1, wherein The method further includes: Determine the total number of base stations in the communication network; In the case that the total number of the base stations exceeds a preset numerical control threshold, determine the number of regional divisions according to the total number of the base stations; Divide the communication network into the number of sub-regions of the regional divisions, and respectively determine the sub-target strategies corresponding to each sub-region.

7. The method according to claim 6, wherein The method further includes: Obtain the computing resource utilization rate of a first sub-region, where the first sub-region is any one of the number of sub-regions of the regional divisions; In the case that the computing resource utilization rate meets a preset condition, determine a second sub-region from the number of sub-regions of the regional divisions, where the computing resource utilization rate of the second sub-region meets the computing tasks of the first sub-region; Migrate the computing tasks of the first sub-region from the computing nodes of the first sub-region to the computing nodes of the second sub-region.

8. An apparatus for determining a network energy consumption control strategy, characterized in that, Includes: An acquisition module, configured to acquire characteristic data of multiple base stations in a communication network, where the characteristic data is used to reflect the operating state of the base stations; An analysis module, configured to analyze the characteristic data by using a network energy consumption model to obtain an initial strategy, where the network energy consumption model is used to predict the energy consumption of the multiple base stations under different network configurations, and the initial strategy is used to control the energy consumption of the communication network; An optimization module, configured to iteratively optimize the initial strategy in a multi-dimensional collaborative optimization framework to obtain a target strategy, where the multi-dimensional collaborative optimization framework is used to characterize the constraint relationship between the interaction between the multiple base stations and the energy consumption; An execution module, configured to determine the network configuration parameters corresponding to the base stations according to the target strategy, and configure the communication network by using the network configuration parameters.

9. An electronic device, characterized in that, Includes: A memory and a processor, where the memory is used to store program instructions; the processor is connected to the memory and is configured to execute the method for determining the network energy consumption control strategy according to any one of claims 1 to 7.

10. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored computer program, where the device where the non-volatile storage medium is located executes the method for determining the network energy consumption control strategy according to any one of claims 1 to 7 by running the computer program.

11. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, the method for determining the network energy consumption control strategy according to any one of claims 1 to 7 is implemented.

Images