Energy-saving control system for temporarily adjusting power of base station

Abstract

The invention provides an energy-saving control system for temporarily adjusting the power of a base station, which relates to the technical field of base station optimization, and comprises a data caching module, a base station simulation model, a user feedback model, an initial solution generation module for generating an optimal initial solution, and a multi-target annealing optimization algorithm based on the optimal initial solution. The multi-target annealing optimization module is used for optimizing the optimal initial solution by combining a base station simulation model and a user feedback model so as to obtain an optimal solution, and the base station adjusting module is used for adjusting state parameters of all signal transmitting units in the base station based on the optimal solution. According to the invention, when the state parameters of each signal transmitting unit in the base station are optimized by using the multi-target annealing algorithm, not only are user demand parameters and base station output power considered, but also communication quality parameters are considered, so that the energy-saving optimization of the base station is ensured, the signal quality is ensured, and adverse use experience caused to users is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of base station optimization, and specifically provides an energy-saving control system for temporarily adjusting the power of a base station. Background Art

[0002] The power consumption of 5G base stations has increased by nearly 3-4 times compared to that of 4G base stations. Moreover, due to the limitation of the coverage range, the number of 5G base stations has also increased significantly, resulting in a sharp rise in the total energy consumption and operating costs. While 5G base stations bring convenience to people's work and life, the soaring electricity bills of 5G base stations also make operators overwhelmed.

[0003] In the prior art, a method and system for controlling the output power of a base station transmitter (classified as H04W52) with the authorization announcement number of "CN119255354B" includes: Step S1: Set a signal strength threshold and determine the minimum output power; obtain the minimum power data of the base station under different user density loads and generate a minimum power curve; Step S2: Set a collection period, monitor the user density load of the base station within each period, and obtain the daily user density load curve; determine the rising interval and the falling interval according to the maximum point and the minimum point, and determine the adjustment time point; and obtain the adjustment interval according to the adjustment time point to obtain the target power; Step S3: Real-time obtain the user density load of the current base station and generate the current load curve; obtain the offset and the rising and falling amounts in real time by comparing the corresponding parts of the current load curve and the load standard curve; correct the next adjustment time point to obtain the predicted adjustment time point and the corresponding predicted target power, so as to adaptively adjust the base station power according to user needs, thereby achieving the effect of saving the energy consumption of the base station.

[0004] However, the prior art still has relatively large defects. For example, when controlling the power of the base station, the above prior art only considers the factor of user density load and does not consider the influencing factor of signal quality, resulting in the problem that the signal quality deteriorates after the base station power is adjusted, affecting the user experience. Moreover, the above prior art adjusts the overall output power of the base station, but there are multiple adjustment schemes for adjusting the base station output power to the target value, and the advantages and disadvantages of different adjustment schemes are also inconsistent, and it is difficult for the above prior art to determine the optimal adjustment scheme.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The purpose of the present invention is to provide an energy-saving control system for temporarily adjusting the power of a base station to solve the problems raised in the above background art.

[0007] To achieve the above object, the present invention provides the following technical solutions: An energy-saving control system for temporarily adjusting the power of a base station, comprising: A data caching module, configured to cache the operation parameters and performance parameters of the base station within a historical monitoring time period, as well as the user demand parameters and communication quality parameters of each sector within the communication coverage area of the base station. The operation parameters include the output power of the base station and the state parameters of each signal transmitting unit within the base station, which include output power, amplifier gain, antenna orientation angle, and antenna tilt angle. The performance parameters include the effective radiation power and power density of each sector. The user demand parameters include the number of users and network traffic. The communication quality parameters include reference signal received power and signal-to-noise ratio; A base station simulation model, the base station simulation model constructed based on a deep neural network has the state parameters of each signal transmitting unit as input, and the output is the output power and performance parameters of the base station, and the model is trained based on the data in the data caching module; A user feedback model, the user feedback model constructed based on a random forest network has the performance parameters of the base station and the user demand parameters of each sector as input, and the output is the communication quality parameters of each sector, and the model is trained based on the data in the data caching module; An initial solution generation module, configured to obtain the user demand parameters of each sector at the current moment, determine an initial solution set in combination with the data in the data caching module, calculate the evaluation index of each initial solution in the initial solution set based on the output power of the base station and the communication quality parameters, and screen out the optimal initial solution based on the evaluation index. The solution is the state parameters of each signal transmitting unit within the base station; A multi-objective annealing optimization module, configured to minimize the output power of the base station and optimize the communication quality parameters as optimization objectives, based on the multi-objective annealing optimization algorithm, and optimize the optimal initial solution in combination with the base station simulation model and the user feedback model to obtain the optimal solution; A base station adjustment module, configured to adjust the state parameters of each signal transmitting unit within the base station at the current moment to be consistent with the state parameters of the optimal solution.

[0008] Preferably, the selection criterion for the historical monitoring time period is: if within a historical time period, the ratio of the standard deviation to the mean of the number of users in each sector is not greater than a first preset threshold, and the ratio of the standard deviation to the mean of the network traffic in each sector is not greater than a second preset threshold, and the duration of this historical time period is not less than a third preset threshold, then this historical time period is defined as the historical monitoring time period.

[0009] Preferably, the method for determining the initial solution set is as follows: If within a historical monitoring time period, the relative error of the number of users in each sector with respect to the number of users in the corresponding sector at the current moment is not higher than the fourth preset threshold, and the relative error of the network traffic in each sector with respect to the network traffic in the corresponding sector at the current moment is not higher than the fifth preset threshold, then the state parameters of each signal transmitting unit in the base station corresponding to this historical monitoring time period are used as the initial solution, and all the initial solutions are aggregated to form an initial solution set; Among them, the calculation formulas for the relative error of the number of users and the relative error of user traffic are as follows:

[0010] In the formula, is the number of users in the o-th sector at the current moment, is the number of users in the o-th sector within the historical monitoring time period, is the relative error of the number of users in the o-th sector, is the network traffic in the o-th sector at the current moment, is the network traffic in the o-th sector within the historical monitoring time period, is the relative error of the network traffic in the o-th sector, o is the index of the sector, and , is the number of sectors in the communication coverage area of the base station, and the relative errors of other parameters are the same.

[0011] Preferably, the method for obtaining the optimal initial solution is as follows: Based on the base station output power and communication quality parameters, calculate the evaluation index of each initial solution in the initial solution set, and use the initial solution corresponding to the minimum evaluation index as the optimal initial solution. The calculation formula for the evaluation index is as follows:

[0012] In the formula, is the base station output power of the initial solution, is the maximum value of the base station output power in the initial solution set, is the base station output power evaluation coefficient of the initial solution; In the formula, is the reference signal received power of the initial solution in the o-th sector, is the maximum value of the reference signal received power in the initial solution set in the o-th sector, is the network traffic of the initial solution in the o-th sector, is the cumulative value of the network traffic of the initial solution in all sectors, is the reference signal received power evaluation coefficient of the initial solution; In the formula, is the signal-to-noise ratio of the initial solution in the o-th sector, is the maximum value of the signal-to-noise ratio in the initial solution set in the o-th sector, is the signal-to-noise ratio evaluation coefficient of the initial solution; In the formula, is the evaluation index of the initial solution, , , are respectively the weights of the base station output power evaluation coefficient, the reference signal received power evaluation coefficient, and the signal-to-noise ratio evaluation coefficient in the calculation of the evaluation index, , , The specific values of are determined by the analytic hierarchy process.

[0013] Preferably, the process of obtaining the optimal solution is as follows: 1) Set the initial temperature and the cooling rate; 2) Take the optimal initial solution as the current solution, use the safe operation of the signal transmission unit as a constraint condition, generate a random perturbation amount based on the initial solution set to perturb the optimal initial solution, and obtain multiple neighborhood solutions; 3) For each neighborhood solution, first input it into the base station simulation model to obtain the corresponding base station output power and performance parameters, and input the obtained performance parameters and the user demand parameters at the current moment into the user feedback model to obtain the corresponding communication quality parameters; 4) Based on the Euclidean distance of the base station output power, the communication quality parameters, the neighborhood solution, and the state parameters at the current moment, compare and analyze each neighborhood solution with the current solution one by one to determine the accepted new solution. The specific logic is as follows: If the base station output power of the k-th neighborhood solution is less than the base station output power of the current solution, and all communication quality parameters of the k-th neighborhood solution are better than the communication quality parameters of the current solution, then accept the k-th neighborhood solution as the new solution; If the base station output power of the k-th neighborhood solution is not less than the base station output power of the current solution, and all communication quality parameters of the k-th neighborhood solution are not better than the communication quality parameters of the current solution, then do not accept the k-th neighborhood solution as the new solution; If there exists a k-th neighborhood solution whose base station output power is less than the base station output power of the current solution, or at least one communication quality parameter of the k-th neighborhood solution is better than the communication quality parameter of the current solution, then accept the new solution with a certain probability. The formula for the probability of accepting the new solution is as follows:

[0014] In the formula, is the probability of accepting the k-th neighborhood solution as the new solution, is the evaluation index of the k-th neighborhood solution, is the evaluation index of the current solution, T is the current iteration temperature, is the Euclidean distance between the k-th neighborhood solution and the state parameters at the current moment, is the mean Euclidean distance between all neighborhood solutions corresponding to the current solution and the state parameters at the current moment, is the index of the neighborhood solution, and , is the number of neighborhood solutions; 5) Each new solution obtained is taken as the current solution, and iterative optimization is performed based on the same method until the predetermined number of iterations is reached, and all the new solutions obtained are summarized to form a new solution set; 6) Calculate the evaluation index of each new solution in the new solution set, and select the new solution corresponding to the minimum evaluation index as the optimal solution.

[0015] Preferably, the step 2) comprises the following steps: 2.1) Based on the first The output power of the jth signal transmitting unit is calculated, and the output power variance of the jth signal transmitting unit in the initial solution set is calculated. The function expression is as follows: ; ;

[0016] In the formula, is the initial solution set, The initial solution The output power of each signal transmitting unit, is the first The average output power of the signal transmission unit, is the first The output power variance of the signal transmission unit is is the index of the initial solution in the initial solution set, and , is the number of initial solutions in the initial solution set, is the index of the signal transmission unit in the base station, and , is the number of signal transmission units in the base station; 2.2) Based on the initial solution set The output power variance of the signal transmitting unit is used to construct a normal distribution for generating random disturbances, thereby generating the kth neighborhood solution. The output power of a signal transmitting unit, the function expression is as follows:

[0017] In the formula, The current solution The output power of each signal transmitting unit, For the The neighborhood solution The output power of a signal transmitting unit To generate the th neighborhood solution, the th signal transmitting unit's output power uses a random perturbation amount The random perturbation amount is a normal distribution with a mean of 0 and a variance of Indicates a random perturbation amount generated based on the normal distribution ; 2.3) Based on the same method as in steps 2.1) and 2.2), generate the other parameter values in each neighborhood solution

[0018] Preferably, the base station simulation model selects the deep learning framework of TensorFlow, and it includes an input layer for receiving the state parameters of each signal transmitting unit, multiple hidden layers for data processing of the input data, and ReLU is selected as the activation function on the hidden layer, and an output layer for outputting the base station output power and performance parameters

[0019] Preferably, the user feedback model includes an input layer for receiving performance parameters and the user demand parameters of each sector, a random tree construction module for tree construction based on the training set, an integration module for calculating the final output based on the prediction results of each tree, an output layer for outputting the communication quality parameters of each sector, and an evaluation and tuning module for performance evaluation and hyperparameter tuning

[0020] Compared with the prior art, the beneficial effects of the present invention are The energy-saving control system for temporarily adjusting the base station power of the present invention performs simulation through the base station simulation model and the user feedback model to simulate the base station output power and communication quality parameters under different combinations of state parameters, laying a foundation for in-depth optimization and adjustment. When optimizing the state parameters of each signal transmitting unit in the base station using the multi-objective annealing algorithm, not only the user demand parameters and the base station output power are considered, but also the communication quality parameters are considered, so as to ensure that while optimizing the energy consumption of the base station, the signal quality is guaranteed and an adverse user experience is avoided BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic diagram of the module structure of the overall system of the present invention DETAILED DESCRIPTION OF THE INVENTION

[0022] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to specific embodiments

[0023] ​​It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention shall have the ordinary meanings understood by those of ordinary skill in the field to which the present invention pertains. The "first", "second" and similar terms used in the present invention do not denote any order, quantity or importance, but are only used to distinguish different components. The terms such as "including" or "comprising" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left" and "right" are only used to represent relative position relationships, and when the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0024] Embodiment: Please refer to Figure 1 , the present invention provides an energy-saving control system for temporarily adjusting the power of a base station, including: A data caching module, configured to cache the operating parameters, performance parameters of the base station within a historical monitoring time period, as well as the user demand parameters and communication quality parameters of each sector within the communication coverage area of the base station. The operating parameters include the output power of the base station and the status parameters of each signal transmitting unit within the base station, which include output power, amplifier gain, antenna orientation angle and antenna tilt angle. The performance parameters include the effective radiated power and power density of each sector. The user demand parameters include the number of users and network traffic. The communication quality parameters include reference signal received power and signal-to-noise ratio; It should be noted that within the historical monitoring time period, the collection frequencies of each parameter are the same to ensure timestamp alignment, and the collection frequency can be once every 1 second, once every 2 seconds, once every 5 seconds, etc. In this way, the time-series data of each parameter is obtained, and the time-series data of each parameter is respectively subjected to mean value processing, and the processed result is used as the parameter value under this historical monitoring time period. For example, the multiple base station output powers collected within the historical monitoring time period are subjected to mean value processing, and the obtained base station output power mean value is used as the base station output power within this historical monitoring time period. In addition, the collection frequency can be set by the staff according to the actual situation on the basis of meeting subsequent simulation training, and no limitation is made here; It should be noted that a sector is a specific part within the coverage area of a base station, generally referring to an area within a certain angular range. The communication coverage area of a base station can be divided into single-sector, dual-sector, triple-sector or six-sector, which is specifically set according to the actual usage of the base station. For example, a common triple-sector divides the communication coverage range of the base station into three sectors with an angular range of 120 degrees each. The signal transmitting unit includes a T / R component and an antenna. The output power is the output power of the transmitter in the T / R component, and the amplifier gain is the amplification gain of the signal amplifier in the T / R component. The output power of the base station, as well as the output power, amplifier gain, antenna orientation angle and antenna tilt angle of each signal transmitting unit in the base station, can be obtained through the base station management system BMS. The effective radiated power, power density, number of users, network traffic, reference signal received power, and signal-to-noise ratio of each sector can be obtained through the base station network quality monitoring device. This is the prior art and will not be elaborated here; Among them, the selection criteria for the historical monitoring time period are as follows: If, within a historical time period, the ratio of the standard deviation to the mean of the number of users in each sector is not greater than the first preset threshold, and the ratio of the standard deviation to the mean of the network traffic in each sector is not greater than the second preset threshold, and the duration of this historical time period is not less than the third preset threshold, then this historical time period is defined as the historical monitoring time period; It should be noted that the setting of the first preset threshold is used to screen out the time periods with stable users, and the setting of the second preset threshold is used to screen out the time periods with stable network traffic. The two are combined to screen out the time periods that take into account both stable users and stable network traffic. On this basis, the setting of the third preset threshold is used to screen out the time periods with strong stability and persistence. Therefore, through the setting of the first preset threshold, the second preset threshold and the third preset threshold, the time periods with stable and persistent user requirements are screened out, so as to facilitate the subsequent analysis of the logical relationship between user requirements and other parameters of the base station. Compared with analyzing the logical relationship between user requirements and other parameters of the base station based on the data at a single moment, the influence of accidental errors is reduced and the reliability of the results is ensured; In general applications, if the ratio of the standard deviation to the mean of a set of data is less than 30%, the fluctuations of this set of data can be considered acceptable. Therefore, the first preset threshold and the second preset threshold are set below 0.3. Also, since the degree of network traffic fluctuations affected by the usage scenario is greater compared to the number of users, the value range of the first preset threshold is adaptively set between 0.1 and 0.15, and the value range of the second preset threshold is between 0.1 and 0.3. To ensure the richness of data samples, the third preset threshold is set to not less than 15 minutes. Of course, the specific values of the first preset threshold, the second preset threshold, and the third preset threshold can also be set by the staff according to the actual situation. For example, if the operating time of this base station is short, the first preset threshold and the second preset threshold can be appropriately increased, and the third preset threshold can be decreased to obtain sufficient sample data to meet the subsequent simulation training. On the contrary, if the operating time of this base station is long and there are sufficient sample data, the first preset threshold and the second preset threshold can be appropriately decreased, and the third preset threshold can be increased to obtain more stable sample data to ensure the accuracy of the subsequent simulation training. There is no limitation here; The base station simulation model. The input of the base station simulation model constructed based on the deep neural network is the state parameters of each signal transmitting unit, and the output is the base station output power and performance parameters, and the model is trained based on the data in the data cache module; As an implementation, the base station simulation model selects the deep learning framework of TensorFlow, and it includes an input layer for receiving the state parameters of each signal transmitting unit, multiple hidden layers for processing the input data, and ReLU is selected as the activation function on the hidden layers, and an output layer for outputting the base station output power and performance parameters. The process of training the base station simulation model is as follows: After establishing the mapping relationship between the state parameters of each signal transmitting unit, the base station output power, and the performance parameters at each historical monitoring moment and summarizing them to form the first sample set, the first sample set is divided into a training set, a test set, and a validation set according to the ratio of 70:15:15. And the batch size and the number of training epochs are set. The state parameters of each signal transmitting unit in the training set are used as the input, and the corresponding base station output power and performance parameters are used as the output labels to train the base station simulation model. The root mean square error is used as the loss function. During the training process, the model parameters are updated through backpropagation to minimize the loss function value. Specifically, the model parameters can be updated through optimization algorithms such as Adam and SGD. And during the training process, the model hyperparameters (such as the learning rate, batch size, the number of hidden layers, etc.) are adjusted through the validation set to optimize the model performance. After reaching the predetermined number of training epochs, the test set is input into the base station simulation model for performance testing. If the deviation between the model prediction value and the true value is between 2% and 5%, it is considered that the training is completed; otherwise, the training is restarted; It should be noted that the base station output power is comprehensively determined by the states of various signal transmitting units of the base station. Increasing the output power or gain of a certain transmitting unit usually increases the overall base station output power. The amplifier gain, antenna azimuth angle, and tilt angle affect the signal transmission efficiency and propagation characteristics, thereby directly affecting the effective radiation power and power density of the base station. Moreover, there is a relatively complex non-linear relationship between the input and output. Therefore, this base station simulation model uses a deep neural network to capture the complex non-linear relationship between the input (i.e., the state parameters of each signal transmitting unit in the base station) and the output (i.e., the base station output power and performance parameters), so as to achieve the purpose of predicting the base station output power and performance parameters based on the state parameters of each signal transmitting unit in the base station.

[0025] The user feedback model, the user feedback model constructed based on the random forest network has the performance parameters of the base station and the user demand parameters of each sector as inputs, and the communication quality parameters of each sector as outputs, and the model is trained based on the data in the data cache module; As an implementation, the user feedback model includes an input layer for receiving performance parameters and user demand parameters of each sector, a random tree construction module for constructing trees based on the training set, an integration module for calculating the final output based on the prediction results of each tree, an output layer for outputting the communication quality parameters of each sector, and an evaluation and tuning module for performance evaluation and hyperparameter tuning. Among them, the random tree construction module constructs random trees based on the sample selection (i.e., randomly extracting samples from the training set using the bootstrap method to establish multiple independent decision trees), feature selection (i.e., randomly selecting features at the node splitting of each tree, and then selecting the best splitting point based on these features), and decision tree construction (i.e., constructing each tree through a recursive method until the stopping condition is reached) steps in the conventional technology. The features are the input parameters of the model. The integration module obtains the final output by averaging the prediction results of each tree. The specific training process is as follows: After establishing the mapping relationship for the performance parameters of the base station, the user demand parameters of each sector, and the communication quality parameters of each sector at each historical monitoring moment, summarize them to form a second sample set. Divide the second sample set into a training set, a test set, and a validation set according to the ratio of 70:15:15, and set the hyperparameters of the random forest (such as the number of trees, the maximum depth, the maximum number of sample splits). Use the performance parameters of the base station and the user demand parameters of each sector in the training set as inputs, and use the corresponding communication quality parameters of each sector as outputs to train the user feedback model. During the training process, the random tree construction module will generate multiple decision trees and integrate them to form the final model. Monitor the performance metrics of the user feedback model (such as mean absolute error, mean squared error, etc.) during the training process to ensure the effectiveness of model training, and adjust the hyperparameters based on the validation results of the validation set (such as coefficient of determination, root mean square error) during the training process to avoid overfitting or underfitting problems. Finally, input the test set into the user feedback model for performance testing. If the deviation between the model prediction value and the true value is between 2% and 5%, it is considered that the training is completed; otherwise, retrain; It should be noted that the effective radiated power and power density directly reflect the signal coverage ability and intensity of the base station. These parameters will affect the signal quality received by users during communication. The number of users and network traffic reveal the network load situation during this time period. Changes in user demand may directly affect communication quality, especially at high loads, which may lead to signal congestion or quality degradation. Moreover, both the input and output are high-dimensional features involving all sectors. Therefore, this user feedback model uses a random forest model that can effectively process high-dimensional features to capture the non-linear function relationship between the input (i.e., the effective radiated power, power density, number of users, and network traffic of each sector) and the output (i.e., the reference signal received power and signal-to-noise ratio of each sector), so as to achieve the purpose of predicting communication quality parameters based on performance parameters and user demand parameters.

[0026] The initial solution generation module is used to obtain the user demand parameters of each sector at the current moment, determine the initial solution set in combination with the data in the data cache module, calculate the evaluation index of each initial solution in the initial solution set based on the base station output power and communication quality parameters, and screen out the optimal initial solution based on the evaluation index. The solution is the state parameters of each signal transmitting unit in the base station; Among them, the method for determining the initial solution set is as follows: If the relative error between the number of users in each sector and the number of users in the corresponding sector at the current moment is not higher than the fourth preset threshold during a historical monitoring time period, and the relative error between the network traffic in each sector and the network traffic in the corresponding sector at the current moment is not higher than the fifth preset threshold, then use the state parameters of each signal transmitting unit in the base station corresponding to this historical monitoring time period as the initial solution, and summarize all the initial solutions to form the initial solution set; Among them, the calculation formulas for the relative error of the number of users and the relative error of user traffic are as follows:

[0027] In the formula, is the number of users in the o-th sector at the current moment, is the number of users in the o-th sector during the historical monitoring time period, is the relative error of the number of users in the o-th sector, is the network traffic of the o-th sector at the current moment, is the network traffic of the o-th sector during the historical monitoring time period, is the relative error of the network traffic of the o-th sector. o is the index of the sector, and , is the number of sectors within the communication coverage area of the base station. The relative errors of other parameters are the same and will not be elaborated here; It should be noted that the settings of the fourth preset threshold and the fifth preset threshold are both for finding an initial solution similar to the user demand parameters at the current moment. Their values are set by the staff according to actual needs and are not restricted here. Generally, the fourth preset threshold should not be higher than 5%, and the fifth preset threshold should not be higher than 10% to ensure the richness and quality of the initial solution set; Among them, the method for obtaining the optimal initial solution is as follows: Based on the base station output power and communication quality parameters, calculate the evaluation index of each initial solution in the initial solution set, and take the initial solution corresponding to the minimum evaluation index as the optimal initial solution. The calculation formula of the evaluation index is as follows:

[0028] In the formula, is the base station output power of the initial solution, is the maximum value of the base station output power in the initial solution set, is the evaluation coefficient of the base station output power of the initial solution, which is used to evaluate the quality of this initial solution in the initial solution set from the perspective of base station energy saving effect. The larger the evaluation coefficient of the base station output power, the better this initial solution is in the evaluation of base station energy saving effect; It should be noted that The setting of introduces a reference object (i.e., In the formula, is the reference signal receiving power of the initial solution in the o-th sector, is the maximum value of the reference signal receiving power of the initial solution set in the o-th sector, is the network traffic of the initial solution in the o-th sector, is the cumulative value of network traffic in all sectors for the initial solution, is the reference signal received power evaluation coefficient of the initial solution, which is used to evaluate the quality of the initial solution in the initial solution set from the signal reception strength. The smaller the reference signal received power evaluation coefficient, the better the initial solution in terms of signal reception strength evaluation; It should be noted that, the setting of introduces a reference object (i.e., ) to evaluate the signal reception strength of the initial solution in the o-th sector. The larger the reference signal received power of the initial solution in the o-th sector, the better the initial solution performs in terms of signal reception strength in the o-th sector compared to other initial solutions in the initial solution set, the setting of is used to perform a weighted comprehensive evaluation of the signal reception strengths of all sectors of the initial solution, the larger, the greater the proportion of the network traffic in the o-th sector in all sectors, and thus the greater the importance of the o-th sector. Therefore, through the setting of, an adaptive weighted assignment is performed for the reference signal received power evaluation results of each sector, the larger, the better the initial solution performs in terms of signal reception strength in the o-th sector, and the smaller the reference signal received power evaluation coefficient; In the formula, is the signal-to-noise ratio of the initial solution in the o-th sector, is the maximum value of the signal-to-noise ratio of the initial solution set in the o-th sector, is the signal-to-noise ratio evaluation coefficient of the initial solution, which is used to evaluate the quality of the initial solution in the initial solution set from the signal quality. The smaller the signal-to-noise ratio evaluation coefficient, the better the initial solution in terms of signal reception strength evaluation; It should be noted that, the setting of introduces a reference object (i.e., ) to evaluate the signal quality of the initial solution in the o-th sector. The larger the signal-to-noise ratio of the initial solution in the o-th sector, the better the initial solution performs in terms of signal quality in the o-th sector compared to other initial solutions in the initial solution set, the setting of is used to perform a weighted comprehensive evaluation of the signal qualities of all sectors of the initial solution, the larger, the greater the proportion of the network traffic in the o-th sector in all sectors, and thus the greater the importance of the o-th sector. Therefore, through the setting of, an adaptive weighted assignment is performed for the signal-to-noise ratio evaluation results of each sector, the larger, the better the initial solution performs in terms of signal quality in the o-th sector, and the smaller the signal-to-noise ratio evaluation coefficient; In the formula, is the evaluation index of the initial solution, which is used to comprehensively evaluate the initial solution from three aspects: the energy-saving effect of the base station, the signal reception strength, and the signal quality. The smaller the evaluation index, the better the comprehensive performance of the initial solution, and the greater the evaluation coefficient. , , are the weights of the base station output power evaluation coefficient, the reference signal received power evaluation coefficient, and the signal-to-noise ratio evaluation coefficient in the calculation of the evaluation index respectively. , , The specific values of are determined by the analytic hierarchy process. The analytic hierarchy process is an existing technology and will not be elaborated here. It should be noted that in the previous description, it has been discussed that the larger the base station output power evaluation coefficient, the smaller the reference signal received power evaluation coefficient and the signal-to-noise ratio evaluation coefficient, and the better the performance of the initial solution. Also, since both the reference signal received power evaluation coefficient and the signal-to-noise ratio evaluation coefficient evaluate the performance of the initial solution from the communication quality, while the base station output power evaluation coefficient evaluates the performance of the initial solution from the energy-saving effect, the base station output power evaluation coefficient is used as the denominator, and the reference signal received power evaluation coefficient and the signal-to-noise ratio evaluation coefficient are accumulated to form the numerator, in order to form to construct the functional relationship of the evaluation index. And the larger the base station output power evaluation coefficient, the smaller the reference signal received power evaluation coefficient and the signal-to-noise ratio evaluation coefficient, the better the performance of the initial solution, and the smaller the evaluation index. The setting of adding 1 to the denominator avoids the situation where the denominator is equal to 0.

[0029] The multi-objective annealing optimization module is used to optimize the optimal initial solution based on the multi-objective annealing optimization algorithm, combined with the base station simulation model and the user feedback model, with the goal of minimizing the base station output power and optimizing the communication quality parameters to obtain the optimal solution. The specific process is as follows: 1) Set the initial temperature and the cooling rate. The value range of the initial temperature can be set between 10 and 100 to facilitate solving the complex optimization problem in this application. Of course, it can also start from a smaller value, monitor the running effect and convergence speed of the optimization algorithm, and then gradually increase the initial temperature until a suitable value is found. The value range of the cooling rate can be set between 0.8 - 0.99, and can be dynamically adjusted according to the optimization effect. For example, in continuous multi-round optimization processes, if the number of times of accepting new solutions is less than the expected value, then reduce the cooling rate to slow down the temperature drop speed and increase the possibility of exploration. 2) Take the optimal initial solution as the current solution, use the safe operation of the signal transmission unit as a constraint condition, generate a random perturbation amount based on the initial solution set to perturb the optimal initial solution, and obtain multiple neighborhood solutions. The specific process is as follows: 2.1) Based on the The output power of the j-th signal transmitting unit in the initial solution set is calculated, and the variance of the output power of the j-th signal transmitting unit in the initial solution set is as follows:

[0030]

[0031]

[0032] In the formula, is the output power of the j-th signal transmitting unit in the i-th initial solution in the initial solution set, is the index of the i-th initial solution in the initial solution set, is the index of the j-th signal transmitting unit, is the average value of the output power of the j-th signal transmitting unit in the initial solution set, is the index of the j-th signal transmitting unit in the initial solution set, is the variance of the output power of the j-th signal transmitting unit in the initial solution set, which is used to measure the output power fluctuation range of the j-th signal transmitting unit in the initial solution set, is the index of the j-th signal transmitting unit, is the number of signal transmitting units in the base station, is the index of the initial solution in the initial solution set, and , is the number of initial solutions in the initial solution set, is the index of the signal transmitting unit in the base station, and , is the number of signal transmitting units in the base station; 2.2) Based on the variance of the output power of the j-th signal transmitting unit in the initial solution set, a normal distribution for generating a random perturbation amount is constructed, and the output power of the j-th signal transmitting unit in the k-th neighborhood solution is generated as follows: The function expression is as follows: In the formula,

[0033] is the output power of the j-th signal transmitting unit in the current solution, is the output power of the j-th signal transmitting unit in the k-th neighborhood solution, is the index of the j-th signal transmitting unit, is the index of the k-th neighborhood solution, is the index of the j-th signal transmitting unit, is the random perturbation amount used to generate the output power of the j-th signal transmitting unit in the k-th neighborhood solution, is a normal distribution with a mean of 0 and a variance of is the index of the k-th neighborhood solution, and is the index of the j-th signal transmitting unit, is the variance of the output power of the j-th signal transmitting unit in the initial solution set, , represents the random perturbation amount generated based on the normal distribution , is the index of the neighborhood solution, and , K is the number of neighborhood solutions, which is a positive integer not less than 2. The larger the value of K, the more neighborhood solutions are generated based on the current solution, the stronger the optimization exploration ability, but the optimization rate will also decrease. Its specific value is set by the staff according to the actual situation and is not restricted here; It should be noted that setting the mean value in the normal distribution used to generate the random perturbation amount to 0 enables the generated multiple neighborhood solutions to be evenly distributed around the current solution, thereby achieving the purpose of comprehensively and widely exploring neighborhood solutions based on the high-quality current solution, which helps to discover potential high-quality solutions not covered by the initial solution set and avoid the problem of falling into local optima. The variance of the initial solution set can reflect the true dispersion degree of high-quality solutions. Using it as the setting of the variance in the normal distribution for generating the random perturbation amount can prevent the algorithm from over-exploring in unnecessary regions, especially for the high-dimensional features of this application, greatly improving the optimization pertinence and optimization efficiency; 2.3) Based on the same method as in steps 2.1) and 2.2), generate the other parameter values in each neighborhood solution; Among them, the constraint conditions include: the output power of any signal transmitting unit is not higher than the rated maximum output power of this signal transmitting unit, that is, not higher than the rated maximum output power of the transmitter, so as to avoid the problem of transmitter damage. The rated maximum output power of the transmitter can be obtained based on the product manual. The amplifier gain of any signal transmitting unit is not higher than the rated maximum gain of this signal transmitting unit, that is, not higher than the rated maximum gain of the signal amplifier, so as to avoid the problem of nonlinear distortion caused by the signal amplifier being overloaded. The rated maximum gain of the signal amplifier can be obtained based on the product manual. The antenna orientation angle and antenna tilt angle of any signal transmitting unit are both within the adjustment range of the antenna, so as to avoid the problem that the optimized parameters are difficult to implement. The adjustment range of the antenna can be obtained based on the product specification table of the antenna, or set by the staff according to the actual situation; 3) For each neighborhood solution, first input it into the base station simulation model to obtain the corresponding base station output power and performance parameters, and input the obtained performance parameters and the user demand parameters at the current moment into the user feedback model to obtain the corresponding communication quality parameters; 4) Based on the Euclidean distance of the base station output power, communication quality parameters, neighborhood solution, and state parameters at the current moment, compare and analyze each neighborhood solution with the current solution one by one to determine the accepted new solution. The specific logic is as follows: If the base station output power of the k-th neighborhood solution is less than the base station output power of the current solution, and each communication quality parameter of the k-th neighborhood solution is better than the communication quality parameter of the current solution, then accept the k-th neighborhood solution as the new solution; If the base station output power of the k-th neighborhood solution is not less than that of the current solution, and all communication quality parameters of the k-th neighborhood solution are not better than those of the current solution, then the k-th neighborhood solution is not accepted as a new solution; If there exists a k-th neighborhood solution with a base station output power less than that of the current solution, or at least one communication quality parameter of the k-th neighborhood solution is better than that of the current solution, then the new solution is accepted with a certain probability. The formula for the acceptance probability of the new solution is as follows:

[0034] In the formula, is the probability of accepting the k-th neighborhood solution as a new solution, is the evaluation index of the k-th neighborhood solution, is the evaluation index of the current solution. The methods for obtaining the evaluation indices of the neighborhood solution and the current solution are as follows: The neighborhood solution and the current solution are aggregated to form a solution set, and then calculated using the same method as the evaluation index of the above initial solution. Moreover, when calculating the evaluation indices of the neighborhood solution and the current solution, the network traffic of each sector at the current moment is used. The specific calculation process is not elaborated here; In the formula, T is the current iteration temperature, which is specifically determined according to the initial temperature, the cooling rate, and the current iteration round number. This is prior art and is not elaborated here; It should be noted that The larger the difference between and , the worse the k-th neighborhood solution is compared to the current solution. Therefore, the conventional is used to construct the main part of the probability. When the difference between and In the formula, is the Euclidean distance between the state parameters of the k-th neighborhood solution and the current moment, which reflects the degree of difference between the state parameters of the k-th neighborhood solution and the state parameters at the current moment. The larger In the formula, is the average Euclidean distance between all neighborhood solutions corresponding to the current solution and the state parameters at the current moment. The setting of is used to introduce a reference object The larger the value, the higher the difficulty and energy consumed in adjusting the state parameters at the current moment to the state parameters of the kth neighborhood solution, and the worse the applicability of the kth neighborhood solution. The smaller it is, the smaller the probability of accepting the kth neighborhood solution as the new solution should also decrease. Therefore, The probability of accepting the kth neighborhood solution is adjusted in the form of, so as to ensure the reliability of the quality of the new solution. The advantage of this is that the difficulty and workload of adjustment are difficult to be accurately quantified as optimization targets intuitively, but they actually affect the adjustment work of the base station. Therefore, this application innovatively calculates the Euclidean distance and considers it in the step of acceptance probability, so that the optimization consideration of the state parameters is more comprehensive and specific, and the quality and applicability of the new solution are guaranteed; 5) Each new solution obtained is taken as the current solution, and iterative optimization is performed based on the same method until the predetermined number of iterations is reached, and all the new solutions obtained are summarized to form a new solution set; It should be noted that when the new solution is used as the current solution for perturbation to generate a neighborhood solution, the variance used is the variance of the corresponding parameters of all the accepted new solutions. The calculation method is the same as above and will not be repeated here. 6) Calculate the evaluation index of each new solution in the new solution set, and select the new solution corresponding to the minimum evaluation index as the optimal solution. The calculation method of the evaluation index of each new solution in the new solution set is consistent with the above, and when calculating the evaluation index of the new solution, the network traffic of each sector at the current moment is used. The specific calculation process is not repeated here.

[0035] It should be noted that the meaning of each communication quality parameter of the neighborhood solution being better than each communication quality parameter of the current solution is: the reference signal received power of the neighborhood solution is higher than the reference signal received power corresponding to the current solution, and the signal-to-noise ratio of the neighborhood solution is higher than the signal-to-noise ratio corresponding to the current solution. This is common sense and will not be elaborated here. The base station adjustment module is used to adjust the state parameters of each signal transmission unit in the base station at the current moment to be consistent with the state parameters of the optimal solution. The specific adjustment method is a known constant and will not be described here.

[0036] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0037] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will realize that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0038] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units. They may be located in one place or distributed over multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0039] As described above, the above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all of them should be covered by the protection scope of this application.

Claims

1. An energy-saving control system for temporarily adjusting base station power, characterized in that: include: A data cache module is used to cache the operating parameters and performance parameters of the base station within the historical monitoring time period, as well as the user demand parameters and communication quality parameters of each sector within the communication coverage area of ​​the base station, wherein the operating parameters include the output power of the base station and the state parameters of each signal transmitting unit in the base station, including the output power, amplifier gain, antenna orientation angle and antenna tilt angle, the performance parameters include the effective radiated power and power density of each sector, the user demand parameters include the number of users and network traffic, and the communication quality parameters include the reference signal received power and signal-to-noise ratio; A base station simulation model, wherein the base station simulation model constructed based on a deep neural network has state parameters of each signal transmission unit as input, outputs base station output power and performance parameters, and performs model training based on data in a data cache module; A user feedback model, wherein the user feedback model constructed based on a random forest network has performance parameters of the base station and user demand parameters of each sector as input, communication quality parameters of each sector as output, and performs model training based on data in a data cache module; The initial solution generation module is used to obtain the user demand parameters of each sector at the current moment, and determine the initial solution set in combination with the data in the data cache module, and calculate the evaluation index of each initial solution in the initial solution set based on the base station output power and communication quality parameters, and screen out the optimal initial solution based on the evaluation index, which is the state parameter of each signal transmission unit in the base station; The multi-objective annealing optimization module is used to optimize the optimal initial solution based on the multi-objective annealing optimization algorithm and in combination with the base station simulation model and the user feedback model, with the optimization objectives of minimizing the base station output power and optimizing the communication quality parameters, so as to obtain the optimal solution; The base station adjustment module is used to adjust the state parameters of each signal transmission unit in the base station at the current moment to be consistent with the state parameters of the optimal solution.

2. The energy-saving control system for temporarily adjusting base station power according to claim 1, characterized in that: The selection criteria for the historical monitoring time period are: if within a historical time period, the standard deviation of the number of users in each sector to the mean ratio is not greater than the first preset threshold, and the standard deviation of the network traffic in each sector to the mean ratio is not greater than the second preset threshold, and the duration of this historical time period is not less than the third preset threshold, then this historical time period is defined as the historical monitoring time period.

3. The energy-saving control system for temporarily adjusting base station power according to claim 1, characterized in that: The method for determining the initial solution set is: if within a historical monitoring time period, the relative error between the number of users in each sector and the number of users in the corresponding sector at the current moment is not higher than the fourth preset threshold, and the relative error between the network traffic in each sector and the network traffic in the corresponding sector at the current moment is not higher than the fifth preset threshold, then the state parameters of each signal transmitting unit in the base station corresponding to this historical monitoring time period are used as the initial solution, and all the initial solutions are summarized to form an initial solution set.

4. The energy-saving control system for temporarily adjusting base station power according to claim 3, characterized in that: The relative error calculation formula of the number of users and user traffic is as follows: ; In the formula, is the number of users in the oth sector at the current moment, is the number of users in the oth sector during the historical monitoring period, is the relative error of the number of users in the oth sector, is the network traffic of the oth sector at the current moment, is the network traffic of the oth sector during the historical monitoring period, is the relative error of network traffic in the oth sector, o is the index of the sector, and , It is the number of sectors within the communication coverage area of ​​the base station.

5. The energy-saving control system for temporarily adjusting base station power according to claim 4, characterized in that: The method for obtaining the optimal initial solution is: based on the base station output power and the communication quality parameter, the evaluation index of each initial solution in the initial solution set is calculated, and the initial solution corresponding to the minimum evaluation index is used as the optimal initial solution. The calculation formula of the evaluation index is as follows: ; In the formula, is the base station output power of the initial solution, is the maximum output power of the base station in the initial solution set, is the base station output power evaluation coefficient of the initial solution; In the formula, is the reference signal received power of the initial solution in the oth sector, is the maximum value of the reference signal received power in the oth sector of the initial solution set, is the network traffic of the initial solution in the oth sector, is the cumulative value of network traffic in all sectors of the initial solution, is the reference signal received power evaluation coefficient of the initial solution; In the formula, is the signal-to-noise ratio of the initial solution in the oth sector, is the maximum signal-to-noise ratio of the initial solution set in the oth sector, is the signal-to-noise ratio evaluation coefficient of the initial solution; In the formula, is the evaluation index of the initial solution, , , are the weights of the base station output power evaluation coefficient, the reference signal received power evaluation coefficient, and the signal-to-noise ratio evaluation coefficient in the evaluation index calculation. , , The specific value of is selected and determined through the hierarchical analysis method.

6. The energy-saving control system for temporarily adjusting base station power according to claim 1, characterized in that: The process of obtaining the optimal solution is as follows: 1) Set the initial temperature and cooling rate; 2) Taking the optimal initial solution as the current solution and the safe operation of the signal transmission unit as the constraint condition, a random perturbation is generated based on the initial solution set to perturb the optimal initial solution and obtain multiple neighborhood solutions; 3) For each neighborhood solution, first input it into the base station simulation model to obtain the corresponding base station output power and performance parameters, and then input the obtained performance parameters and the user demand parameters at the current moment into the user feedback model to obtain the corresponding communication quality parameters; 4) Based on the Euclidean distance of the base station output power, communication quality parameters, neighborhood solutions and the current state parameters, each neighborhood solution is compared and analyzed with the current solution to determine the accepted new solution; 5) Each new solution is taken as the current solution, and iterative optimization is performed based on the same method until the predetermined number of iterations is reached, and all the new solutions are summarized to form a new solution set; 6) Calculate the evaluation index of each new solution in the new solution set, and select the new solution corresponding to the minimum evaluation index as the optimal solution.

7. The energy-saving control system for temporarily adjusting base station power according to claim 6, characterized in that: The logic for accepting the new solution is: If the base station output power of the kth neighborhood solution is less than the base station output power of the current solution, and the communication quality parameters of the kth neighborhood solution are better than the communication quality parameters of the current solution, then the kth neighborhood solution is accepted as the new solution; If the base station output power of the kth neighborhood solution is not less than the base station output power of the current solution, and the communication quality parameters of the kth neighborhood solution are not better than the communication quality parameters of the current solution, the kth neighborhood solution is not accepted as a new solution; If there is a k-th neighborhood solution whose base station output power is less than the base station output power of the current solution, or if there is at least one communication quality parameter in the k-th neighborhood solution that is better than the communication quality parameter of the current solution, then the new solution is accepted with a certain probability. The formula for the probability of accepting the new solution is as follows: ; In the formula, is the probability of accepting the kth neighborhood solution as the new solution, is the evaluation index of the kth neighborhood solution, is the evaluation index of the current solution, T is the current iteration temperature, is the Euclidean distance between the kth neighborhood solution and the state parameter at the current moment, is the mean Euclidean distance between all neighborhood solutions corresponding to the current solution and the state parameters at the current moment, is the index of the neighborhood solution, and , is the number of neighborhood solutions.

8. The energy-saving control system for temporarily adjusting base station power according to claim 6, characterized in that: The step 2) comprises the following steps: 2.1) Based on the first The output power of the jth signal transmitting unit is calculated, and the output power variance of the jth signal transmitting unit in the initial solution set is calculated. The function expression is as follows: ; ; In the formula, is the initial solution set, The initial solution The output power of each signal transmitting unit is is the first The average output power of the signal transmission unit, is the first The output power variance of the signal transmission unit is is the index of the initial solution in the initial solution set, and , is the number of initial solutions in the initial solution set, is the index of the signal transmission unit in the base station, and , is the number of signal transmission units in the base station; 2.2) Based on the initial solution set The output power variance of the signal transmitting unit is used to construct a normal distribution for generating random disturbances, thereby generating the kth neighborhood solution. The output power of a signal transmitting unit, the function expression is as follows: ; In the formula, The current solution The output power of each signal transmitting unit is For the The first The output power of each signal transmitting unit is To generate the The neighborhood solution The amount of random perturbation used by the output power of each signal transmitting unit, The mean is 0 and the variance is The normal distribution of Represents a normal distribution The amount of random perturbation generated ; 2.3) Based on the same method as steps 2.1) and 2.2), generate other parameter values ​​in each neighborhood solution.

9. The energy-saving control system for temporarily adjusting base station power according to claim 1, characterized in that: The base station simulation model uses the deep learning framework of TensorFlow.

10. The energy-saving control system for temporarily adjusting base station power according to claim 9, characterized in that: The base station simulation model includes an input layer for receiving state parameters of each signal transmission unit, multiple hidden layers for performing data processing on the input data, and ReLU is selected as the activation function on the hidden layer, and an output layer for outputting the output power and performance parameters of the base station.

11. The energy-saving control system for temporarily adjusting base station power according to claim 1, characterized in that: The user feedback model includes an input layer for receiving performance parameters and user demand parameters of each sector, a random tree construction module for tree construction based on a training set, an integration module for calculating the final output based on the prediction results of each tree, an output layer for outputting communication quality parameters of each sector, and an evaluation and tuning module for performance evaluation and hyperparameter tuning.

Images