Base station power distribution method and device, equipment and storage medium

Abstract

The invention provides a base station power distribution method and device, equipment and a storage medium, and the method comprises the steps: obtaining target data corresponding to each base station in a target region; determining traffic change information based on the target data; predicting a base station power consumption demand of the base station based on the traffic change information, and predicting a base station power demand of the base station based on the base station power consumption demand and the real-time traffic of the base station; determining a power distribution strategy based on the power demand of each base station; and carrying out dynamic power distribution on each base station based on the power distribution strategy. The embodiment of the invention can dynamically adjust the power distribution of the base station and reduce the power consumption of the base station.

Description

Technical Field

[0001] The present invention relates to the field of mobile communication technology, and in particular to a base station power allocation method, device, equipment and storage medium. Background Art

[0002] With the rapid development of mobile communication technology, the power consumption of base stations, as the core component of mobile communication networks, has received increasing attention. In order to reduce the power consumption of base stations, it is very important to allocate appropriate power to the base stations.

[0003] In the prior art, a linear regression model or a traditional machine learning algorithm is usually used to determine the current power consumption of a base station, and a power allocation strategy is determined based on the current power consumption of the base station to reduce the power consumption of the base station.

[0004] However, in the above method, since the user's behavior habits and traffic usage are complex and changeable, when the power allocation strategy is determined only based on the current power consumption of the base station, the determined power allocation strategy is inaccurate and cannot achieve the purpose of reducing the power consumption of the base station. Summary of the invention

[0005] The present invention provides a base station power allocation method, device, equipment and storage medium, which are used to solve the defects of inaccurate power allocation strategy determined in the prior art and high power consumption of base stations, and realize the purpose of dynamically adjusting the power allocation of base stations and reducing the power consumption of base stations.

[0006] The present invention provides a base station power allocation method, comprising: For each base station in the target area, obtaining target data corresponding to the base station; Based on the target data, determining flow change information; Predicting a base station power consumption requirement of the base station based on the traffic change information, and predicting a base station power demand of the base station based on the base station power consumption requirement and the real-time traffic of the base station; Determining a power allocation strategy based on the power requirements of each of the base stations; Based on the power allocation strategy, dynamic power allocation is performed on each of the base stations.

[0007] According to a base station power allocation method provided by the present invention, the determining of flow change information based on the target data includes: Based on the target data, determining a base station power consumption data set; Inputting the base station power consumption data set into an aggregate decision tree model to obtain traffic change information output by the aggregate decision tree model; The aggregate decision tree model is obtained by training an initial aggregate decision tree model based on sample base station power consumption data of the sample base stations and corresponding traffic change information labels.

[0008] According to a base station power allocation method provided by the present invention, the predicting of the base station power consumption requirement of the base station based on the traffic change information includes: Based on the flow change information, generating a feature extraction model; Inputting the base station power consumption data set into the feature extraction model to obtain a feature data set output by the feature extraction model; generating a power consumption demand curve based on the characteristic data set; Based on the generated power consumption demand curve, a temporal convolutional network model is constructed; The base station power consumption requirement of the base station is predicted based on the temporal convolutional network model.

[0009] According to a base station power allocation method provided by the present invention, the power allocation strategy is determined based on the power requirements of each base station, including: For each of the base stations, obtaining the static power consumption and dynamic power consumption of the base station; Based on the static power consumption and the dynamic power consumption of each of the base stations, constructing an objective function with the goal of minimizing the total power consumption of all base stations; The objective function is solved to obtain the power allocation strategy.

[0010] According to a base station power allocation method provided by the present invention, the dynamic power allocation is performed on each base station based on the power allocation strategy, including: Based on the power allocation strategy, determining the initial allocated power of each of the base stations; For each of the base stations, obtaining the execution power consumption of the base station when the base station is running based on the initial allocated power; Based on the execution power consumption of the base station and the initial allocation power of the base station, a target allocation power of the base station is determined.

[0011] According to a base station power allocation method provided by the present invention, the method further includes: Based on the traffic data and traffic demand data of each base station, determine a low traffic period and a low demand area, wherein the low traffic period is a period corresponding to traffic data less than a traffic threshold, and the low demand area is an area corresponding to traffic demand data less than a demand threshold; Determine the impact value of the base station corresponding to the low traffic period and the low demand area on the network performance of the target area after shutting down; Based on the impact value, a target base station shutdown strategy is determined, where the target base station shutdown strategy is used to characterize a target base station that is shut down within a target time period.

[0012] According to a base station power allocation method provided by the present invention, determining a target base station shutdown strategy based on the impact value includes: Based on the impact values ​​of each base station corresponding to the low traffic period and the low demand area, determining an initial base station shutdown strategy from a shutdown strategy set; each shutdown strategy in the shutdown strategy set is used to characterize the base stations that need to be shut down in each time period; Execute a shutdown operation based on the initial base station shutdown strategy, and monitor the network performance and user experience value of the target area in real time; The target base station shutdown strategy is determined based on the network performance and the user experience value.

[0013] According to a base station power allocation method provided by the present invention, the base station power consumption data set is determined based on the target data, including: Performing data cleaning on the target data to obtain cleaned target data; Normalizing the cleaned target data to obtain normalized target data; Extracting key feature variable data from the normalized target data; The key characteristic variable data is standardized to obtain the base station power consumption data set.

[0014] The present invention also provides a base station power distribution device, comprising: An acquisition module, used for acquiring target data corresponding to each base station in a target area; A determination module, used to determine flow change information based on the target data; A prediction module, configured to predict a base station power consumption requirement of the base station based on the traffic change information, and predict a base station power demand of the base station based on the base station power consumption requirement and the real-time traffic of the base station; The determining module is further configured to determine a power allocation strategy based on the power requirements of each base station; The allocation module is used to dynamically allocate power to each base station based on the power allocation strategy.

[0015] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, any of the above-mentioned base station power allocation methods is implemented.

[0016] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the base station power allocation method described in any one of the above is implemented.

[0017] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the base station power allocation method described above is implemented.

[0018] The base station power allocation method, device, equipment and storage medium provided by the present invention can predict the base station power consumption demand of the base station in each future time period by acquiring target data corresponding to the base station and determining the traffic change information based on the target data, and predict the base station power demand of the base station based on the base station power consumption demand and the real-time traffic, and determine the power allocation strategy based on the base station power demand, and implement dynamic power allocation for each base station according to the power allocation strategy. Since the base station power consumption demand of the base station in each future time period can be predicted based on the traffic change information, the base station power consumption demand of the base station in the future time period can be further predicted in combination with the base station power consumption demand and the real-time traffic of the base station, and the power allocation strategy determined based on the base station power demand can also adapt to the traffic changes in each future time period, so that the power allocation of each base station can be dynamically adjusted according to the traffic changes based on the above-mentioned power allocation strategy, so as to achieve the purpose of minimizing the power consumption of the base station. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present invention or 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 A schematic diagram of a flow chart of a base station power allocation method provided in an embodiment of the present invention.

[0021] Figure 2 A schematic diagram of the structure of a base station power allocation device provided in an embodiment of the present invention.

[0022] Figure 3 A schematic diagram of the physical structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] In the prior art, the power consumption management of base stations mainly relies on traditional static power allocation strategies. These strategies usually allocate power based on experience or simple linear regression models, and lack the ability to predict and adjust the dynamic changes of base station power consumption in real time.

[0025] In view of the above problems, an embodiment of the present invention provides a base station power allocation method, in which the base station power consumption demand of the base station in each future time period can be predicted by acquiring the user equipment access data, traffic data and environmental data corresponding to the base station, and determining the traffic change information based on these data, and the base station power consumption demand of the base station can be predicted based on the base station power consumption demand and the real-time power consumption, and the power allocation strategy is determined based on the base station power demand, and according to the power allocation strategy, dynamic power allocation is implemented for each base station. Since the base station power consumption demand of the base station in each future time period can be predicted based on the traffic change information, the base station power consumption demand is related to the traffic change information, and it changes with the change of the traffic change information. When the base station power consumption demand of the base station in the future time period is predicted in combination with the base station power consumption demand and the real-time power consumption of the base station, the power allocation strategy determined based on the base station power demand can also adapt to the traffic changes in each future time period, so that the power allocation of each base station can be dynamically adjusted based on the above power allocation strategy as the traffic changes, so as to achieve the purpose of reducing the power consumption of the base station to the greatest extent.

[0026] Combine the following Figure 1 The base station power allocation method provided by an embodiment of the present invention is described. The embodiment of the present invention can be applied to communication scenarios, especially to scenarios of how to dynamically predict the power consumption of base stations and dynamically adjust the power allocation of base stations. The execution subject of this method can be an electronic device such as a terminal device, a computer, a server, a server cluster, or a specially designed base station power allocation device, or a base station power allocation device set in the electronic device, and the base station power allocation device can be implemented by software, hardware, or a combination of both.

[0027] Figure 1 A schematic diagram of a flow chart of a base station power allocation method provided by an embodiment of the present invention, such as Figure 1 As shown, the method includes: Step 101: For each base station in a target area, obtain target data corresponding to the base station.

[0028] In this step, the target area is an area where base station power consumption monitoring or prediction is required. At least one base station is deployed in the target area, and the target data includes at least one of user equipment access data, flow data, and environmental data. Through the distributed data collection module, multiple high-precision user access sensors, flow sensors, and at least one type of environmental sensors are widely deployed in the coverage area of ​​each base station to monitor at least one of user equipment access data, flow data, and environmental data in real time, which can ensure the comprehensiveness and real-time nature of data collection.

[0029] In the embodiments of the present invention, the target data includes user device access data, traffic data and environmental data as an example for explanation. For the case where the target data includes other data, the implementation method is similar to that including user device access data, traffic data and environmental data, and will not be repeated here.

[0030] In addition, the collected user device access data, traffic data, and environmental data can be transmitted to the cloud server through the wireless communication network. Wireless transmission technology not only improves the speed and efficiency of data transmission, but also ensures the real-time and integrity of the data. This real-time transmission mechanism enables the system to obtain the latest data in a timely manner, thereby monitoring and analyzing the power consumption of the base station in real time, and quickly responding to changing network environments and user needs. Compared with traditional methods that rely on wired networks or manual monitoring, the efficiency and accuracy of data collection are improved in the embodiments of the present invention, providing a basis for subsequent data analysis and power consumption optimization.

[0031] Among them, the user access sensor collects user equipment access data in real time. The user equipment access data can also be understood as the data of the user equipment accessing the base station, for example, it may include the user equipment identification, access time and access frequency. After the collection, the user equipment access data will be further converted into a structured format, as shown in the following formula (1): (1)

[0032] in, Indicates that the user equipment accesses data. represents the identifier of the user equipment u, Indicates the access time. Indicates the access frequency, Represents the set of all base stations in the target area.

[0033] The flow sensor monitors the real-time flow of each channel of the base station. The real-time flow of each channel includes uplink flow and downlink flow. After collecting the real-time flow, the real-time flow will be recorded in time series, as shown in the following formula (2): (2)

[0034] in, Indicates flow data, Indicates the upstream traffic. represents the downstream traffic, t represents the time series, and T represents the preset time period in the past.

[0035] The environmental sensor monitors the environmental data of the base station coverage area. The environmental data can also be understood as environmental parameters, including temperature, humidity and wind speed. After the environmental data is collected, the environmental data can also be transmitted to the data collection module in real time. The environmental data can be, for example, as shown in the following formula (3): (3)

[0036] in, Represents environmental data, represents the temperature of the e-th data point, represents the humidity of the e-th data point, represents the wind speed of the e-th data point, Represents a collection of data points.

[0037] Furthermore, the collected user device access data, traffic data, and environment data can be transmitted to the cloud server through the wireless communication network, as shown in the following formula (4): (4)

[0038] in, Represents the dataset transferred to the cloud server.

[0039] Step 102: Determine flow change information based on target data.

[0040] In this step, the traffic change information may include the change information of the number of terminals accessing the base station and / or the data traffic change information, and the traffic change information is related to the user usage pattern. The user usage pattern can be understood as the user's behavior habits or data usage methods in a certain period of time, which may include the frequency of the user's access to the network in a certain period of time, the use of data traffic and other related behaviors.

[0041] By acquiring the target data, we can analyze the terminal access situation and user usage patterns, and further obtain traffic change information. The traffic change information may include the changing trend of data traffic, which is related to the user usage pattern. If the user usage pattern changes, the traffic change information may also change accordingly. It may also include change information on the number of terminals accessing the base station. The larger the number of terminals accessing the base station, the greater the traffic change information.

[0042] Step 103: predicting the base station power consumption requirement of the base station based on the traffic change information, and predicting the base station power requirement of the base station based on the base station power consumption requirement and the real-time traffic.

[0043] In this step, after the traffic change information is obtained, a base station power consumption prediction model can be constructed based on the traffic change information, so as to predict the base station power consumption requirements in various future time periods according to the power consumption prediction model.

[0044] In addition, real-time traffic is collected in real time by the monitoring system of the base station, which can represent the number of terminals currently connected to the base station and the data traffic currently consumed. Combining the predicted base station power consumption requirements in future time periods with the current real-time traffic, the base station power requirements in future time periods can be more accurately evaluated. For example, if the real-time traffic represents that there are currently more terminals accessing the base station or the current data traffic is large, the base station power requirements in future time periods will increase.

[0045] Since the predicted power consumption demand of the base station and the current real-time traffic can be combined, the power demand of the base station in the future time period can be determined by considering the power consumption of the base station. This can ensure that the power required by the base station is provided while reducing the power consumption of the base station, thereby improving network performance and service quality.

[0046] Step 104: Determine a power allocation strategy based on the power requirements of each base station.

[0047] In this step, after determining the base station power requirements of each base station, in order to ensure that the power requirements of each base station are met and to ensure the overall network performance, the power consumption of each base station also needs to be considered. Therefore, by establishing an optimization function that minimizes total power consumption, a genetic algorithm is used to solve the objective function to determine the optimal power allocation strategy.

[0048] It should be understood that the above optimal power allocation strategy can not only ensure the power allocation of each base station, but also minimize the overall power consumption of the base station.

[0049] Step 105: Based on the power allocation strategy, dynamically allocate power to each base station.

[0050] In this step, after the power allocation strategy is determined, the corresponding power is allocated to each base station according to the power allocation strategy. When the base station is running according to the power, the target data can be continuously monitored, so as to re-determine the base station power demand and power consumption of each base station according to the above steps based on the newly collected target data. In the case of deviation, the above power allocation strategy will be dynamically adjusted.

[0051] The base station power allocation method provided by the embodiment of the present invention obtains target data including user equipment access data, traffic data and environmental data corresponding to the base station, and determines traffic change information based on these target data, so as to predict the base station power consumption demand of the base station in each future time period, and predict the base station power demand of the base station based on the base station power consumption demand and real-time power consumption, and determine the power allocation strategy based on the base station power demand, and implement dynamic power allocation for each base station according to the power allocation strategy. Since the base station power consumption demand of the base station in each future time period can be predicted based on the traffic change information, the base station power consumption demand of the base station in the future time period can be further predicted in combination with the base station power consumption demand and the real-time traffic of the base station, and the power allocation strategy determined based on the base station power demand can also adapt to the traffic changes in each future time period, so that the power allocation of each base station can be dynamically adjusted according to the traffic changes based on the above-mentioned power allocation strategy, so as to achieve the purpose of minimizing the power consumption of the base station.

[0052] Exemplarily, based on the above embodiment, when determining the flow change information based on the target data, it can be done in the following manner: Based on the target data, the base station power consumption data set is determined, and the base station power consumption data set is input into the aggregation decision tree model to obtain the traffic change information output by the aggregation decision tree model, wherein the aggregation decision tree model is obtained by training the initial aggregation decision tree model based on the sample base station power consumption data of the sample base station and the corresponding traffic change information labels.

[0053] Specifically, when determining the base station power consumption data set based on the collected target data, the target data may be preprocessed first, and a feature selection algorithm may be used to extract key feature variables from the preprocessed target data, thereby generating a standardized base station power consumption data set. Exemplarily, when determining the base station power consumption data set, the target data may be cleaned first to obtain cleaned target data, and then the cleaned target data may be normalized to obtain the normalized target data, and then key feature variable data may be extracted from the normalized target data, thereby normalizing the key feature variable data to obtain the base station power consumption data set.

[0054] Specifically, the collected user access data, traffic data, and environment data can be cleaned, denoised, and completed based on the following equations (5) and (6) to generate a clean data set: (5)

[0055] (6)

[0056] Here, n represents the number of data groups, wherein multiple data collected at the same time can be understood as a group of data. represents the mean of the i-th data set transmitted to the cloud server, represents the standard deviation of the i-th data set transmitted to the cloud server, represents the i-th set of data transmitted to the cloud server, represents the i-1th set of data transmitted to the cloud server, represents the i+1th set of data transmitted to the cloud server, represents the indicator function, R represents the reasonable range of the data, Represents the interpolation function, which is used to fill in the missing data.

[0057] After obtaining the cleaned target data, the cleaned target data can be calculated according to the following formula (7): Normalization is performed to convert data of different data types into values ​​within the same scale range: (7)

[0058] in, represents the normalized target data, represents the minimum value in the cleaned target data, represents the maximum value in the cleaned target data, and a and b represent the normalized target ranges, respectively.

[0059] Afterwards, the mutual information method (MI) can be applied to extract key feature variable data from the normalized target data according to the following formula (8): : (8)

[0060] Among them, MI (,) represents the mutual information function, Y represents the target variable, represents the mutual information between the i-th normalized target data and the target variable, and m represents the number of normalized target data. The mutual information function MI measures the dependency or correlation between the normalized target data and the target variable. Based on this metric, the normalized target data with a high correlation with the target variable can be selected as the input for subsequent model training.

[0061] Furthermore, the extracted key feature variable data can be expressed as follows: Perform standardization to generate a standardized base station power consumption data set : (9)

[0062] in, represents the mean of the key feature variable data, Represents the standard deviation of the key feature variable data.

[0063] In this embodiment, by performing data cleaning on the target data and normalizing the cleaned target data, key feature variable data are extracted from the normalized target data, and these key feature variable data can accurately reflect the factors affecting the power consumption of the base station. In addition, the extracted key feature variable data are standardized to obtain the power consumption data of the base station. The above-mentioned data preprocessing can ensure the quality and consistency of the data, and the data preprocessing and feature extraction process are completely independent of manual operation, which can improve the efficiency and accuracy of determining the power consumption data of the base station. Furthermore, the above-mentioned feature selection process can not only improve the accuracy of data analysis, but also make the power consumption data of the base station more standardized and structured, which is convenient for subsequent modeling and prediction.

[0064] After determining the base station power consumption data set Afterwards, the base station power consumption dataset will be Input the trained aggregate decision tree model Perform periodic learning and determine the aggregate decision tree model based on the following formula (10): Output user usage patterns and traffic change information : (10)

[0065] Where t represents the timestamp, Represents the base station power consumption dataset The state at time t can also be understood as the base station power consumption data set A collection of feature values ​​or data points at time t. These features can include indicators in multiple dimensions, such as user access data, traffic data (uplink traffic, downlink traffic), and environmental data (temperature, humidity, wind speed, etc.). represents the weight of the jth node in the i-th decision tree model, Represents an indicator function, indicating whether the input belongs to the region , Represents the output value of the node, represents the weight of the i-th decision tree model, and k represents the number of decision tree models.

[0066] The above-mentioned aggregate decision tree model can be trained in the following way: after collecting sample target data of sample base stations, cleaning, normalizing, feature extracting and standardizing the sample target data, the sample base station power consumption data of the sample base stations can be obtained. , based on sample base station power consumption data The initial training set and validation set are constructed according to the following formula (11), which are used for model training and performance evaluation respectively: (11)

[0067] The above formula (11) represents the power consumption data of the sample base station Twenty percent of the data is used as the test set, and the remaining eighty percent of the data is used as the training set. represents the training set, represents the test set, where and They can be expressed by the following equations (12) and (13) respectively: (12)

[0068] (13)

[0069] Where N represents the total number of sample base station power consumption data, X represents the sample base station power consumption data, represents the power consumption data of the i-th sample base station, Y represents the output label, that is, the traffic change information label, represents the i-th output label.

[0070] After obtaining the training set, the adaptive weighted aggregate decision tree algorithm can be used for model training. The weight of each aggregate decision tree is dynamically adjusted according to its performance according to the adaptive weighted method based on the following formula (14): (14)

[0071] in, represents the aggregated decision tree model, Represents the i-th decision tree model In the training set The loss on Represents the jth decision tree model In the training set The loss on represents the adjustment parameter, represents the weight of the i-th decision tree model, and k represents the number of decision tree models.

[0072] After the aggregate decision tree model is trained, it can also be based on the test set The cross-validation method is used to evaluate the performance of the aggregated decision tree model, as shown in the following formula (15) to ensure the generalization ability of the aggregated decision tree model: (15)

[0073] in, Indicates the test score, represents the number of samples in the i-th test set, (x, y) represents the power consumption data of the test sample base station and the corresponding actual traffic change information, Represents the evaluation result of the i-th cross-validation data.

[0074] when If it is less than the preset value, it means that the generalization performance of the aggregated decision tree model is poor, and it is necessary to return to fine-tune the model parameters of the aggregated decision tree model. In this way, the generalization ability of the aggregated decision tree model can be improved.

[0075] In this embodiment, after determining the base station power consumption data set based on the target data, the base station power consumption data set is input into the aggregated decision tree model for periodic learning, and the traffic change information output by the aggregated decision tree model can be obtained. By integrating multiple decision tree models, the accuracy and stability of the traffic change information prediction can be improved. In addition, through the mechanism of periodic learning, the aggregated decision tree model can be regularly updated and adjusted to adapt to the ever-changing user behavior and network environment.

[0076] Exemplarily, based on the above embodiments, when predicting the base station power consumption demand of the base station based on traffic change information, a feature extraction model can be generated based on the traffic change information, and the base station power consumption data set can be input into the feature extraction model to obtain a feature data set output by the feature extraction model, and a power consumption demand curve is generated based on the feature data set. After a timing convolutional network model is constructed based on the generated power consumption demand curve, the base station power consumption demand of the base station is predicted based on the timing convolutional network model.

[0077] Specifically, after determining the flow change information, the feature extraction model can be generated according to the following formula (16): : (16)

[0078] in, represents a constant term, Representation feature extraction model The i-th model parameter of Indicates the status of the identified user's usage pattern and traffic change information in the past i-th time step, represents the error term and m represents the number of model parameters.

[0079] Base station power consumption dataset Input to feature extraction model In the above example, key features are extracted to obtain feature data sets. , as shown in the following formula (17): (17)

[0080] in, represents the power consumption data set from the i-th base station, X represents the power consumption data set from all base stations, represents the feature data extracted from the i-th base station power consumption dataset, and Y represents the features extracted from all base station power consumption datasets.

[0081] After determining the feature data set After that, by using the feature dataset Perform time series analysis to identify the traffic change trend of users in the feature data set, and analyze the seasonal components and periodic changes of traffic changes. Based on the identified information, the power consumption demand curve can be generated according to the following formula (18): (18)

[0082] Among them, p represents the order of the autoregressive term, which represents the relationship between the current value and the value of the previous p time steps, that is, the impact of the historical data of p time steps on the current time step. q represents the order of the moving average term, which represents the relationship between the error term at the current moment and the error term at the previous q moments, reflecting the impact of the random disturbance of the previous q time steps on the current time step. Represents the power consumption requirement at time t, or other variable values ​​that you want to predict. It is a constant term, which indicates the basic value of power consumption demand at the current moment without being affected by historical data. It can be understood as an offset of the time series model. represents the autoregressive coefficient, represents the moving average coefficient, represents the error term, represents the seasonal component, capturing periodic changes, Indicates frequency, Indicates phase offset.

[0083] After generating the power consumption demand curve, the temporal convolutional network model is constructed based on the power consumption demand curve according to the following formula (19): : (19)

[0084] Where L is the number of layers of TCN, represents the convolution kernel size of the lth layer, represents the convolution kernel weight of the lth layer, represents the sub-convolution kernel weight of the lth layer, and They represent the bias terms, represents the activation function, represents the activation output of the l-1th layer, represents the subconvolution kernel size of the lth layer, represents the power consumption requirement at the tijth moment determined based on the power consumption requirement curve.

[0085] Furthermore, we can build a temporal convolutional network model based on , the power consumption demand of the base station in each period in the future is predicted by the following formula (20): (20)

[0086] in, represents the power consumption demand predicted in the h period after time t, M is the number of nodes in the TCN output layer, represents the weight of the mth node, represents the weight of the input layer, represents the magnitude of the seasonal component, represents the hyperbolic tangent activation function, represents the seasonal component, capturing periodic changes, Indicates frequency, Indicates phase offset.

[0087] Based on the above prediction results, the final base station power consumption prediction model can be determined according to the following formula (21): , to predict the power consumption requirements of base stations in future time periods: (twenty one)

[0088] in, Represents the base station power consumption prediction model The model parameters are n and n represents the step size in the time series. The base station power consumption prediction model can be further adjusted to make the final base station power consumption prediction model better fit the actual situation. It is used to correct the power consumption requirements of the base station in different time periods, so that a more accurate prediction value of the power consumption requirements of the base station can be obtained based on the above base station power consumption prediction model.

[0089] In this embodiment, after generating a feature extraction model through traffic change information, a feature data set is extracted from the base station power consumption data set based on the feature extraction model, and after generating a power consumption demand curve according to the feature data set, a model based on time series analysis and timing convolutional network constructed according to the power consumption demand curve can accurately predict the power consumption demand of the base station in each time period in the future, thereby improving the accuracy and effectiveness of power consumption management.

[0090] Exemplarily, based on the above-mentioned embodiments, when determining the power allocation strategy based on the power requirements of each base station, it is possible to obtain the static power consumption and dynamic power consumption of each base station, and based on the static power consumption and dynamic power consumption of each base station, construct an objective function with the goal of minimizing the total power consumption of all base stations, solve the objective function, and obtain the power allocation strategy.

[0091] Specifically, for each base station in the target area, the static power consumption and dynamic power consumption of the base station can be obtained, so that the objective function with the goal of minimizing the total power consumption of all base stations can be constructed according to the following formula (22): (twenty two)

[0092] in, represents the total power consumption of the jth base station, represents the static power consumption of the jth base station, represents the dynamic power consumption of the jth base station, and m represents the number of base stations.

[0093] After constructing the objective function, the genetic algorithm can be used to solve the objective function, as shown in the following equation (23), and the optimal power allocation strategy can be obtained: (twenty three)

[0094] in, It represents the optimal power allocation of the jth base station. The genetic algorithm is used to iteratively solve the problem through selection, crossover and mutation operations, so that the optimal power allocation of each base station can be obtained. By selecting and adjusting the power allocation of each base station, the total power consumption can be minimized.

[0095] In this embodiment, based on the static power consumption and dynamic power consumption of each base station, an objective function with the goal of minimizing the total power consumption of all base stations is constructed, and the objective function is solved. The resulting power allocation strategy can ensure that the power allocation of each base station can both meet real-time needs and minimize total power consumption, thereby improving overall energy efficiency.

[0096] Exemplarily, based on the above-mentioned embodiments, when dynamic power allocation is performed on each base station based on a power allocation strategy, the initial allocated power of each base station can be determined based on the power allocation strategy, and for each base station, the execution power consumption of the base station when it operates based on the initial allocated power is obtained, and the target allocated power of the base station is determined based on the execution power consumption of the base station and the initial allocated power of the base station.

[0097] Specifically, the initial power allocation strategy of each base station can be determined according to the optimal power allocation strategy by using the following formula (24), that is, the power allocation of each base station can be dynamically adjusted: (twenty four)

[0098] in, represents the initial allocated power of the jth base station at time t, that is, the adjusted power allocation, Represents the adjustment factor. represents the power demand at time t, which can be determined by the following equation (25): (25)

[0099] in, Represents real-time data at time ti, that is, real-time traffic. Indicates the adjustment parameter.

[0100] For each base station, the execution effect of the power allocation strategy can be monitored in real time, that is, the execution power consumption of the base station when it is running based on the initial allocated power can be obtained. When a deviation occurs, it is iteratively adjusted according to the actual situation based on the execution power consumption of the base station and the initial allocated power of the base station according to the following formula (26): (26)

[0101] in, It represents the target allocated power of the jth base station at time t, that is, the final power allocation. represents the adjustment factor, It represents the actual power consumption of the base station at time t, which reflects the actual power consumption of the base station at time t.

[0102] In this embodiment, after the initial allocated power of each base station is determined based on the power allocation strategy, the execution power consumption of the base station when it is running based on the initial allocated power is obtained, and the target allocated power of the base station is determined based on the execution power consumption of the base station and the initial allocated power of the base station. In this way, the power allocation of the base station can be adjusted in real time during the execution of the base station, so that the base station allocation can be optimized and the overall energy consumption can be reduced.

[0103] Exemplarily, on the basis of the above embodiments, in order to further reduce power consumption, the shutdown strategy of the base station can also be determined based on the optimal power allocation strategy. Exemplarily, based on the traffic data and traffic demand data of each base station, a low traffic period and a low demand area can be determined, the low traffic period is a period corresponding to the traffic data less than the traffic threshold, and the low demand area is an area corresponding to the traffic demand data less than the demand threshold. The impact value of the base station corresponding to the low traffic period and the low demand area on the network performance of the target area after shutting down is determined, and based on the impact value, the shutdown strategy of the target base station is determined. The shutdown strategy of the target base station is used to characterize the target base station that is shut down within the target time period.

[0104] Specifically, the low traffic period and low demand area can be determined according to the following formula (27): (27)

[0105] in, represents the set of low traffic periods and low demand areas at time t, represents the traffic data of the jth base station at time t, represents the traffic demand data of the jth base station at time t, Indicates the flow threshold. Indicates the demand threshold.

[0106] For base stations in low traffic periods and low demand areas, the impact of shutting down the base station on the network performance of the target area can be determined according to the following formula (28): (28)

[0107] in, represents the impact on network performance when the jth base station is turned off in the low traffic period and low demand area at time t, represents the total power consumption of the jth base station, represents the set of neighboring base stations of the jth base station, represents the power consumption of the i-th neighbor base station after traffic redirection.

[0108] After calculating the impact value of the base station shutdown on the network performance of the target area, based on the impact value, the target base station shutdown strategy is determined according to the following formula (29): (29)

[0109] in, represents the target base station shutdown strategy at time t, Represents a shutdown strategy set.

[0110] The target base station shutdown strategy can be understood as the best way to minimize the impact of power consumption on overall network performance by selecting certain base stations to shut down at time t. By calculating and selecting the shutdown solution that can maximize power savings and minimize the impact on network operation, unnecessary energy consumption can be effectively reduced during periods of low demand or low traffic. represents all possible combinations of shutdown schemes, including different base station combinations and their corresponding shutdown times. Each scheme in the shutdown strategy set involves shutting down different base stations. For example, one scheme may choose to shut down base station j1, while another scheme may choose to shut down base stations j2 or j3, or even a combination of multiple base stations. Each shutdown scheme also includes the choice of shutdown time, that is, the duration of the base station being shut down.

[0111] After the shutdown strategy of the target base station is determined, the shutdown operation of the corresponding base station will be automatically executed.

[0112] In this embodiment, after determining the low-flow period and low-demand area, the impact value of shutting down the base station corresponding to the low-flow period and low-demand area on the network performance of the target area will be determined. Based on the impact value, the target base station shutdown strategy is determined. The corresponding target base station can be shut down within the target time period based on the target base station shutdown strategy, thereby effectively reducing unnecessary energy consumption during periods of low demand or low flow.

[0113] Exemplarily, based on the above-mentioned embodiments, when determining the shutdown strategy of the target base station based on the impact value, the initial base station shutdown strategy can be determined from the shutdown strategy set based on the impact value of each base station corresponding to the low-traffic period and the low-demand area; each shutdown strategy in the shutdown strategy set is used to characterize the base stations that need to be shut down in each time period, and the shutdown operation is performed based on the initial base station shutdown strategy, and after monitoring the network performance and user experience value of the target area in real time, the target base station shutdown strategy is determined based on the network performance and user experience value.

[0114] Specifically, after the initial base station shutdown strategy is determined from the shutdown strategy set based on the impact value of each base station, the shutdown operation of the corresponding base station will be performed based on the initial base station shutdown strategy, and the network performance and user experience value in the target area after the base station is shut down will be monitored in real time.

[0115] When the network performance and user experience values ​​are abnormal, the base station shutdown strategy will be dynamically adjusted according to the following formula (30) to obtain the target base station shutdown strategy: (30)

[0116] in, Indicates the power consumption distribution after adjusting the base station shutdown strategy. represents the target network performance indicator, Indicates the actual network performance indicator, Represents the adjustment factor.

[0117] In addition, it is necessary to regularly evaluate the effect of the target base station shutdown strategy and iteratively optimize the shutdown strategy according to the evaluation results according to the following formula (31) to ensure the best balance between network performance and energy saving effect: (31)

[0118] in, represents the final target base station shutdown strategy, represents the weight parameter, A cost function representing the user experience, which is related to the user experience value.

[0119] In this embodiment, after the initial base station shutdown strategy is determined and the shutdown operation is executed, the network performance and user experience value of the target area will be monitored in real time, and the final target base station shutdown strategy will be determined based on the network performance and user experience value. The target base station shutdown strategy can be adjusted in real time according to the actual network environment, ensuring the accuracy of the target base station shutdown strategy, reducing unnecessary power consumption, and achieving significant energy-saving effects.

[0120] In addition, in a dual-mode 4G+5G environment, the best network mode can be dynamically selected based on artificial intelligence algorithms, including monitoring user needs and energy consumption, dynamically adjusting the network mode, and selecting single 4G or dual 4G+5G, taking into account both network performance and energy consumption optimization. Cloud-enhanced technology can also be used to dynamically schedule and manage base station resources through the K8S system, including deploying the K8S system to achieve containerized management of base station resources, and dynamically scheduling base station computing and storage resources according to real-time needs.

[0121] The base station power allocation device provided by an embodiment of the present invention is described below. The base station power allocation device described below and the base station power allocation method described above can be referred to each other.

[0122] Figure 2 A schematic diagram of a base station power distribution device according to an embodiment of the present invention is provided. Figure 2 As shown, the base station power allocation device 200 includes: The acquisition module 11 is used to acquire target data corresponding to each base station in the target area; A determination module 12, configured to determine flow change information based on the target data; A prediction module 13, configured to predict a base station power consumption requirement of the base station based on the traffic change information, and predict a base station power demand of the base station based on the base station power consumption requirement and the real-time traffic of the base station; The determination module 12 is further configured to determine a power allocation strategy based on the power requirements of each base station; The allocation module 14 is configured to perform dynamic power allocation on each of the base stations based on the power allocation strategy.

[0123] In an exemplary embodiment, the determining module 12 is specifically configured to: Based on the target data, determining a base station power consumption data set; Inputting the base station power consumption data set into an aggregate decision tree model to obtain traffic change information output by the aggregate decision tree model; The aggregate decision tree model is obtained by training an initial aggregate decision tree model based on sample base station power consumption data of the sample base stations and corresponding traffic change information labels.

[0124] In an exemplary embodiment, the prediction module 13 is specifically configured to: Based on the flow change information, generating a feature extraction model; Inputting the base station power consumption data set into the feature extraction model to obtain a feature data set output by the feature extraction model; generating a power consumption demand curve based on the characteristic data set; Based on the generated power consumption demand curve, a temporal convolutional network model is constructed; The base station power consumption requirement of the base station is predicted based on the temporal convolutional network model.

[0125] In an exemplary embodiment, the determining module 12 is specifically configured to: For each of the base stations, obtaining the static power consumption and dynamic power consumption of the base station; Based on the static power consumption and the dynamic power consumption of each of the base stations, constructing an objective function with the goal of minimizing the total power consumption of all base stations; The objective function is solved to obtain the power allocation strategy.

[0126] In an exemplary embodiment, the allocation module 14 is specifically configured to: Based on the power allocation strategy, determining the initial allocated power of each of the base stations; For each of the base stations, obtaining the execution power consumption of the base station when the base station is running based on the initial allocated power; Based on the execution power consumption of the base station and the initial allocation power of the base station, a target allocation power of the base station is determined.

[0127] In an exemplary embodiment, the determining module 12 is further configured to: Based on the traffic data and traffic demand data of each base station, determine a low traffic period and a low demand area, wherein the low traffic period is a period corresponding to the traffic data less than the traffic threshold, and the low demand area is an area corresponding to the traffic demand data less than the demand threshold; Determine the impact value of the base station corresponding to the low traffic period and the low demand area on the network performance of the target area after shutting down; Based on the impact value, a target base station shutdown strategy is determined, where the target base station shutdown strategy is used to characterize a target base station that is shut down within a target time period.

[0128] In an exemplary embodiment, the determining module 12 is specifically configured to: Based on the impact values ​​of each base station corresponding to the low traffic period and the low demand area, determining an initial base station shutdown strategy from a shutdown strategy set; each shutdown strategy in the shutdown strategy set is used to characterize the base stations that need to be shut down in each time period; Execute a shutdown operation based on the initial base station shutdown strategy, and monitor the network performance and user experience value of the target area in real time; The target base station shutdown strategy is determined based on the network performance and the user experience value.

[0129] In an exemplary embodiment, the determining module 12 is specifically configured to: Performing data cleaning on the target data to obtain cleaned target data; Normalizing the cleaned target data to obtain normalized target data; Extracting key feature variable data from the normalized target data; The key characteristic variable data is standardized to obtain the base station power consumption data set.

[0130] The device of this embodiment can be used to execute the method of any embodiment in the base station power allocation method side embodiment. Its specific implementation process and technical effects are similar to those in the base station power allocation method side embodiment. For details, please refer to the detailed description in the base station power allocation method side embodiment, which will not be repeated here.

[0131] Figure 3 A schematic diagram of the physical structure of an electronic device provided by an embodiment of the present invention, such as Figure 3 As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330 and a communication bus 340, wherein the processor 310, the communication interface 320 and the memory 330 complete mutual communication through the communication bus 340. The processor 310 may call the logic instructions in the memory 330 to execute the base station power allocation method, the method comprising: for each base station in the target area, obtaining the target data corresponding to the base station; based on the target data, determining the flow change information; predicting the base station power consumption demand of the base station based on the flow change information, and predicting the base station power demand of the base station based on the base station power consumption demand and the real-time flow of the base station; determining the power allocation strategy based on the power demand of each base station; and dynamically allocating power to each base station based on the power allocation strategy.

[0132] In addition, the logic instructions in the above-mentioned memory 330 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0133] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the base station power allocation method provided by the above methods, and the method includes: for each base station in the target area, obtaining target data corresponding to the base station; based on the target data, determining traffic change information; predicting the base station power consumption requirement of the base station based on the traffic change information, and predicting the base station power requirement of the base station based on the base station power consumption requirement and the real-time traffic of the base station; determining a power allocation strategy based on the power requirement of each base station; and dynamically allocating power to each base station based on the power allocation strategy.

[0134] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the base station power allocation method provided by the above-mentioned methods, the method comprising: acquiring target data corresponding to each base station in a target area; determining traffic change information based on the target data; predicting the base station power consumption requirement of the base station based on the traffic change information, and predicting the base station power demand of the base station based on the base station power consumption requirement and the real-time traffic of the base station; determining a power allocation strategy based on the power demand of each base station; and dynamically allocating power to each base station based on the power allocation strategy.

[0135] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0136] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0137] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A base station power allocation method, characterized in that: include: For each base station in the target area, obtaining target data corresponding to the base station; Based on the target data, determining flow change information; Predicting a base station power consumption requirement of the base station based on the traffic change information, and predicting a base station power requirement of the base station based on the base station power consumption requirement and real-time traffic; Determining a power allocation strategy based on the power requirements of each of the base stations; Based on the power allocation strategy, dynamic power allocation is performed on each of the base stations.

2. The base station power allocation method according to claim 1, characterized in that: The determining of flow change information based on the target data includes: Based on the target data, determining a base station power consumption data set; Inputting the base station power consumption data set into an aggregate decision tree model to obtain traffic change information output by the aggregate decision tree model; The aggregate decision tree model is obtained by training an initial aggregate decision tree model based on sample base station power consumption data of the sample base stations and corresponding traffic change information labels.

3. The base station power allocation method according to claim 2, characterized in that: The predicting the base station power consumption requirement of the base station based on the traffic change information includes: Based on the flow change information, generating a feature extraction model; Inputting the base station power consumption data set into the feature extraction model to obtain a feature data set output by the feature extraction model; generating a power consumption demand curve based on the characteristic data set; Based on the generated power consumption demand curve, a temporal convolutional network model is constructed; The base station power consumption requirement of the base station is predicted based on the temporal convolutional network model.

4. The base station power allocation method according to claim 1, characterized in that: The determining of a power allocation strategy based on the power requirements of each base station includes: For each of the base stations, obtaining the static power consumption and dynamic power consumption of the base station; Based on the static power consumption and the dynamic power consumption of each of the base stations, constructing an objective function with the goal of minimizing the total power consumption of all base stations; The objective function is solved to obtain the power allocation strategy.

5. The base station power allocation method according to claim 4, characterized in that: The dynamically allocating power to each of the base stations based on the power allocation strategy includes: Based on the power allocation strategy, determining the initial allocated power of each of the base stations; For each of the base stations, obtaining the execution power consumption of the base station when the base station is running based on the initial allocated power; Based on the execution power consumption of the base station and the initial allocation power of the base station, a target allocation power of the base station is determined.

6. The base station power allocation method according to any one of claims 1 to 5, characterized in that: The method further comprises: Based on the traffic data and traffic demand data of each base station, determine a low traffic period and a low demand area, wherein the low traffic period is a period corresponding to traffic data less than a traffic threshold, and the low demand area is an area corresponding to traffic demand data less than a demand threshold; Determine the impact value of the base station corresponding to the low traffic period and the low demand area on the network performance of the target area after shutting down; Based on the impact value, a target base station shutdown strategy is determined, where the target base station shutdown strategy is used to characterize a target base station that is shut down within a target time period.

7. The base station power allocation method according to claim 6, characterized in that: The determining, based on the impact value, a target base station shutdown strategy includes: Based on the impact values ​​of each base station corresponding to the low traffic period and the low demand area, determining an initial base station shutdown strategy from a shutdown strategy set; each shutdown strategy in the shutdown strategy set is used to characterize the base stations that need to be shut down in each time period; Execute a shutdown operation based on the initial base station shutdown strategy, and monitor the network performance and user experience value of the target area in real time; The target base station shutdown strategy is determined based on the network performance and the user experience value.

8. The base station power allocation method according to claim 2, characterized in that: The determining, based on the target data, a base station power consumption data set includes: Performing data cleaning on the target data to obtain cleaned target data; Normalizing the cleaned target data to obtain normalized target data; Extracting key feature variable data from the normalized target data; The key characteristic variable data is standardized to obtain the base station power consumption data set.

9. A base station power distribution device, characterized in that: include: An acquisition module, used for acquiring target data corresponding to each base station in a target area; A determination module, used to determine flow change information based on the target data; A prediction module, configured to predict a base station power consumption requirement of the base station based on the traffic change information, and predict a base station power requirement of the base station based on the base station power consumption requirement and the real-time traffic of the base station; The determining module is further configured to determine a power allocation strategy based on the power requirements of each base station; An allocation module is used to dynamically allocate power to each base station based on the power allocation strategy.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the base station power allocation method according to any one of claims 1 to 8 is implemented.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the base station power allocation method according to any one of claims 1 to 8 is implemented.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the base station power allocation method according to any one of claims 1 to 8 is implemented.

Images