Base station communication resource scheduling method and system based on electric power big data

Abstract

The invention provides a base station communication resource scheduling method and system based on electric power big data, and is applied to the technical field of electric power and communication control. The method comprises the following steps: fitting an association relationship between power consumption and communication load based on historical communication load data, historical power consumption data and historical meteorological data of each base station in a target area; performing power consumption prediction based on the historical power consumption data of the base station and the time sequence prediction model; based on the association relationship between the power consumption and the communication load and the power consumption data in the prediction period, obtaining the communication load in the prediction period; based on the historical user behavior data and the historical communication demand data of the base station, performing communication demand prediction of the base station by using a neural network model; and scheduling communication resources in the prediction period target area based on the prediction period communication load and the prediction period communication demand. The problems that the communication base station is large in energy consumption and poor in stability when serving as a power load, and adverse effects are brought to power grid dispatching and energy management are solved.

Description

Technical Field

[0001] The present invention relates to the field of power and communication control technology, and in particular to a base station communication resource scheduling method and system based on power big data. Background Art

[0002] With the vigorous development of the fifth generation mobile communication technology (5G), the number and coverage of 5G base stations, as the core component of the new generation of communication infrastructure, have expanded rapidly. However, the high energy consumption characteristics of 5G base stations have brought unprecedented pressure to the power grid, and also posed new challenges to energy management and scheduling.

[0003] Current communication resource scheduling schemes often only focus on optimizing communication power, resulting in high energy consumption of communication base stations and poor stability as power loads, which has an adverse impact on power grid scheduling and energy management. Summary of the invention

[0004] In order to overcome the problem that the above-mentioned communication base stations have high energy consumption and poor stability as power loads, which brings adverse effects on power grid scheduling and energy management, the present invention provides a base station communication resource scheduling method and system based on power big data.

[0005] On the one hand, the present invention provides a base station communication resource scheduling method based on power big data, comprising:

[0006] Based on the historical communication load data, historical power consumption data and historical meteorological data of each base station in the target area, the correlation between the power consumption and communication load of each base station in the target area is fitted;

[0007] Power consumption is predicted based on the historical power consumption data of each base station and the time series prediction model to obtain the power consumption data of each base station during the forecast period;

[0008] Based on the correlation between the power consumption of each base station and the communication load and the power consumption data of the forecast period, the communication load of each base station in the forecast period is obtained;

[0009] Based on the historical user behavior data and historical communication demand data of each base station, the communication demand of each base station in the target area is predicted using a neural network model to obtain the predicted communication demand of each base station;

[0010] The communication resources in the target area in the forecast period are scheduled based on the forecast period communication load and forecast period communication demand of each base station.

[0011] Optionally, fitting the correlation between the power consumption and the communication load of each base station in the target area based on the historical communication load data, the historical power consumption data and the historical meteorological data of each base station in the target area includes:

[0012] Obtain historical communication load data, historical power consumption data, historical meteorological data and base station locations of each base station in the target area;

[0013] Taking the historical power consumption data as the target variable, and the historical communication load data, historical meteorological data and base station location as the influencing factors of the target variable, a linear regression model is used to fit the correlation between the power consumption and communication load of each base station in the target area.

[0014] Optionally, the correlation between the power consumption and the communication load is:

[0015] P=a0+a1L+a2T+a3H+a4Lon+a5Lat+∈

[0016] Among them, P represents the power consumption of the base station, L represents the communication load of the base station, T represents the weather temperature at the base station location, H represents the air humidity at the base station location, Lon and Lat are the longitude and latitude of the base station location, ∈ is the error term, and a0, a1, a2, a3, a4, and a5 are fitting coefficients respectively.

[0017] Optionally, the power consumption forecasting based on the historical power consumption data of each base station and the time series forecasting model to obtain the power consumption data of each base station in the forecast period includes:

[0018] The historical power consumption data of each base station in the target area is subjected to CEEMDAN-SSA double decomposition to obtain multiple eigenmode sequences;

[0019] Power consumption is predicted based on multiple intrinsic mode sequences and time series prediction models to obtain the prediction sequence of each intrinsic mode;

[0020] The prediction sequence of each eigenmode is reconstructed based on the time sequence to obtain the power consumption data of each base station in the target area during the forecast period.

[0021] Optionally, the historical power consumption data of each base station in the target area is subjected to complete set empirical mode decomposition-singular spectrum analysis double decomposition to obtain multiple eigenmode sequences, including:

[0022] Based on the complete set empirical mode decomposition, the historical power consumption data of each base station in the target area is decomposed once to obtain multiple decomposed mode subsequences;

[0023] For each decomposed modal subsequence, the multi-scale entropy of each decomposed modal subsequence is determined by scale transformation and local entropy calculation, and the reconstructed aggregation of similar subsequences in the decomposed modal subsequence is performed based on the multi-scale entropy of each decomposed modal subsequence;

[0024] The reconstructed aggregation results are decomposed twice using singular spectrum analysis to extract multiple eigenmode sequences.

[0025] Optionally, after performing power consumption prediction based on the historical power consumption data of each base station and the time series prediction model to obtain the power consumption data of each base station in the predicted period, the method further includes:

[0026] Inputting the sample prediction error of the time series prediction model and the meteorological data corresponding to the sample into the error prediction model to obtain the corresponding predicted power consumption error;

[0027] Performing error correction on the power consumption data for the forecast period based on the forecast power consumption error;

[0028] The prediction error sequence is a data sequence formed by the difference between the sample prediction result and the sample label based on the time series prediction model, and the error prediction model is a trained bidirectional long short-term memory network.

[0029] Optionally, the neural network model includes a plurality of hidden layers, and based on the historical user behavior data and historical communication demand data of each base station, the communication demand of each base station in the target area is predicted by using the neural network model to obtain the predicted communication demand of each base station, including:

[0030] Taking historical user behavior data and historical communication demand data in the target area as input data;

[0031] Performing layer-by-layer feature mapping and nonlinear processing on the input data through multiple hidden layers to obtain the predicted communication demand of the target area;

[0032] Wherein, the neural network model is a trained multi-layer perceptron.

[0033] Optionally, the scheduling of communication resources in the target area in the forecast period based on the forecast period communication load and the forecast period communication demand of each base station includes:

[0034] Clustering the base stations in the target area based on the positions of the base stations in the target area to obtain multiple base station clusters;

[0035] For each base station cluster, based on the predicted communication load of the base stations in the cluster and the predicted communication demand of the target area, the particle swarm optimization algorithm is used to optimize the scheduling of the communication resources of the cluster in the predicted period to obtain the communication resource scheduling strategy of the cluster in the predicted period.

[0036] Optionally, based on the predicted period communication load of the base stations in the cluster and the predicted period communication demand of the target area, the communication resources of the cluster in the predicted period are optimized and scheduled using a particle swarm optimization algorithm to obtain a communication resource scheduling strategy for the cluster in the predicted period, including:

[0037] Taking each base station in the cluster as a particle, taking the multi-dimensional communication resource allocation strategy of each base station as the position of the corresponding particle, and taking the adjustment amount of the multi-dimensional communication resources of each base station as the speed of the corresponding particle;

[0038] Based on each resource allocation strategy of the cluster, the user communication demand satisfaction, power load balancing and energy efficiency of the corresponding cluster are calculated, and based on the weighted fusion of the user communication demand satisfaction, power load balancing and energy efficiency of the cluster, the fitness value corresponding to the current resource allocation strategy is obtained;

[0039] Iteratively updating the speed and position of each particle based on the fitness value until a termination condition is reached, and obtaining a global optimal position of each particle as a communication resource scheduling strategy for the cluster in the prediction period;

[0040] Among them, the multi-dimensional communication resources include power resources, spectrum resources and communication user resources.

[0041] On the other hand, the present invention also provides a base station communication resource scheduling system based on power big data, comprising:

[0042] A fitting module, used to fit the correlation between the power consumption and the communication load of each base station in the target area based on the historical communication load data, historical power consumption data and historical meteorological data of each base station in the target area;

[0043] The power load forecasting module is used to forecast power consumption based on the historical power consumption data of each base station and the time series forecasting model, and obtain the power consumption data of each base station in the forecast period;

[0044] A communication load determination module, configured to obtain the communication load of each base station in the forecast period based on the correlation between the power consumption of each base station and the communication load and the power consumption data in the forecast period;

[0045] The demand prediction module is used to predict the communication demand of each base station in the target area based on the historical user behavior data and historical communication demand data of each base station using a neural network model to obtain the predicted communication demand of each base station;

[0046] The scheduling module is used to schedule communication resources in the target area during the forecast period based on the forecast period communication load and forecast period communication demand of each base station.

[0047] On the other hand, the present invention also provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0048] The memory is used to store one or more programs;

[0049] When the one or more programs are executed by the at least one processor, any one of the above-mentioned base station communication resource scheduling methods based on power big data is implemented.

[0050] On the other hand, the present invention also provides a readable storage medium having an execution program stored thereon, and when the execution program is executed, any one of the above-mentioned base station communication resource scheduling methods based on power big data is implemented.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] The present invention provides a method and system for dispatching communication resources of a base station based on electric power big data. The method and system construct a correlation between electric power consumption and communication load, and mine the influence of communication load on electric power consumption data. Then, electric power consumption is predicted based on a time series prediction model, and the communication load in a prediction period is calculated based on the correlation between electric power consumption and communication load and the predicted electric power consumption data in the prediction period, so as to realize the prediction of communication load considering the coupling relationship between electric power consumption and communication load. Finally, communication resources in the prediction period are dispatched based on the predicted communication demand and communication load, so that the corresponding electric power consumption data is taken into account in the communication resource dispatching process, thereby avoiding the problem of power grid control caused by the large energy consumption of individual communication base stations and the poor stability of electric power load caused by the large energy consumption of the individual communication base stations, and realizing the coordinated control of electric power and communication resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A flow chart of a method for scheduling base station communication resources based on power big data according to the present invention;

[0054] Figure 2 It is a schematic structural diagram of the electronic device of the present invention. DETAILED DESCRIPTION

[0055] The specific implementation modes of the present invention are further described in detail below with reference to the accompanying drawings.

[0056] Example 1

[0057] The present invention provides a method for dispatching base station communication resources based on power big data, as shown in the schematic diagram Figure 1 As shown, including:

[0058] Step S110, fitting the correlation between the power consumption and the communication load of each base station in the target area based on the historical communication load data, historical power consumption data and historical meteorological data of each base station in the target area;

[0059] Step S120, performing power consumption forecasting based on the historical power consumption data of each base station and the time series forecasting model to obtain the power consumption data of each base station in the forecast period;

[0060] Step S130, based on the correlation between the power consumption of each base station and the communication load and the power consumption data of the predicted period, obtaining the communication load of each base station in the predicted period;

[0061] Step S140, based on the historical user behavior data and historical communication demand data of each base station, the communication demand of each base station in the target area is predicted by using a neural network model to obtain the predicted communication demand of each base station;

[0062] Step S150 , scheduling communication resources within the target area in the forecast period based on the forecast period communication load and forecast period communication demand of each base station.

[0063] In this example implementation, the target area may be an administrative area of ​​a city or a certain area within a city. The base station may be various types of communication base stations, such as a 5G base station. The historical power consumption data may include historical power loads, and the historical meteorological data refers to the meteorological data of the past period of time (such as the past year or several years) at the location of the base station, and the meteorological data may include temperature, humidity, wind speed, etc. The location of the base station may include the latitude and longitude of the base station. The communication load, regional meteorology, and base station location may be used as factors affecting regional power consumption, and the historical data may be used to fit the relationship between power consumption and communication load, thereby mining the coupling relationship between power load and communication load, and predicting the communication load of the target area based on the coupling relationship, thereby scheduling the communication resources of the target area based on the predicted communication load and communication demand, realizing coordinated control of power and communication power, optimizing power consumption during communication, and avoiding the problems of excessive energy consumption and poor load stability in the communication resource scheduling process. By utilizing power big data, the present invention can perceive the operating status of the base station in real time, and dynamically adjust the communication resource scheduling strategy according to actual needs, realize the shared utilization of communication resources in the region, improve resource utilization, reduce energy consumption, enhance base station stability, and effectively improve the operational efficiency and economic benefits of the 5G network.

[0064] In some embodiments, the step S110 of fitting the correlation between the power consumption and the communication load of each base station in the target area based on the historical communication load data, the historical power consumption data and the historical meteorological data of each base station in the target area includes:

[0065] Obtain historical communication load data, historical power consumption data, historical meteorological data and base station locations of each base station in the target area;

[0066] Taking the historical power consumption data as the target variable, and the historical communication load data, historical meteorological data and base station location as the influencing factors of the target variable, a linear regression model is used to fit the correlation between the power consumption and communication load of each base station in the target area.

[0067] In this example implementation, a communication base station sample library is established by collecting power consumption data, communication load, base station location information, and meteorological data (including temperature, humidity, wind speed) of each base station in the target area (such as a city administrative area) in the past few years. Specifically, the communication base station sample library Y is as follows:

[0068]

[0069] Among them, P i , P n represent the power consumption data sequence of base station i and base station n respectively, L i , L n They represent the communication load sequence of base station i and base station n respectively, T i 、T n Respectively represent the temperature series of base station i and base station n, H i , H n Respectively represent the humidity sequence of base station i and base station n, W i , W n Respectively represent the surrounding wind speed sequence of base station i and base station n, Lon i , Lon n Respectively represent the longitude of base station i and base station n, Lat i , Lat n Represent the latitude of base station i and base station n respectively, and n is the number of base stations in the target area. The power consumption data of the base station is used as the target variable, and the communication load, base station temperature, base station humidity and base station location information are used as input features. Based on the sample library, a linear regression model is used to fit the relationship between P and L, and the correlation between base station power consumption and communication load is extracted. The model can be expressed as:

[0070] P=a0+a1L+a2T+a3H+a4Lon+a5Lat+∈

[0071] Among them, P represents the power consumption of the base station, L represents the communication load of the base station, T represents the weather temperature at the base station location, H represents the air humidity at the base station location, Lon and Lat are the longitude and latitude of the base station location, ∈ is the error term, and a0, a1, a2, a3, a4, and a5 are fitting coefficients respectively.

[0072] In an example implementation, the step S120 of performing power consumption prediction based on the historical power consumption data of each base station and the time series prediction model to obtain the power consumption data of each base station during the prediction period includes:

[0073] The historical power consumption data of each base station in the target area is subjected to CEEMDAN-SSA double decomposition to obtain multiple eigenmode sequences;

[0074] Power consumption is predicted based on multiple intrinsic mode sequences and time series prediction models to obtain the prediction sequence of each intrinsic mode;

[0075] The prediction sequence of each eigenmode is reconstructed based on the time sequence to obtain the power consumption data of each base station in the target area during the forecast period.

[0076] In this example implementation, CEEMDAN-SSA dual decomposition is an algorithm for processing time series data that combines two methods, namely, fully adaptive noise ensemble empirical mode decomposition (CEEMDAN) and singular spectrum analysis (SSA), and is mainly used to remove noise in the sequence and extract the main rules to obtain multiple eigenmode sequences. The time series prediction model can be a trained BiLSTM model, and the extracted multiple eigenmode sequences are input into the trained BiLSTM model to capture the rules of each component and output the prediction sequence of each eigenmode; the multiple eigenmode sequences can be sequences with time sequence labels, and the prediction sequence of each eigenmode is reconstructed based on the time sequence label to obtain the predicted period power consumption data of each base station in the target area.

[0077] Exemplarily, the historical power consumption data of each base station in the target area is subjected to complete set empirical mode decomposition-singular spectrum analysis double decomposition to obtain multiple eigenmode sequences, including:

[0078] Based on the complete set empirical mode decomposition, the historical power consumption data of each base station in the target area is decomposed once to obtain multiple decomposed mode subsequences;

[0079] For each decomposed modal subsequence, the multi-scale entropy of each decomposed modal subsequence is determined by scale transformation and local entropy calculation, and the reconstructed aggregation of similar subsequences in the decomposed modal subsequence is performed based on the multi-scale entropy of each decomposed modal subsequence;

[0080] The reconstructed aggregation results are decomposed twice using singular spectrum analysis to extract multiple eigenmode sequences.

[0081] In this example implementation, the complete ensemble empirical mode decomposition (CEEMDAN) is used to perform a primary decomposition of historical power consumption data (such as base station power load) to obtain multiple decomposition modal subsequences, and then the multi-scale entropy is used to calculate the complexity of each decomposition modal subsequence, and similar decomposition modal sub-sequences are reconstructed and aggregated according to the multi-scale entropy; the denoising function of the singular spectrum analysis (SSA) is then used to perform a secondary decomposition on the new sequence obtained by the reconstruction and aggregation, remove the noise in the sequence, extract a series of trend and periodic components, and obtain multiple eigenmode sequences, thereby further improving the prediction accuracy.

[0082] In an example implementation, after performing power consumption prediction based on the historical power consumption data of each base station and the time series prediction model to obtain the power consumption data of each base station during the predicted period, the method further includes:

[0083] Inputting the sample prediction error of the time series prediction model and the meteorological data corresponding to the sample into the error prediction model to obtain the corresponding predicted power consumption error;

[0084] Performing error correction on the power consumption data for the forecast period based on the forecast power consumption error;

[0085] In this example implementation, considering the impact of meteorological factors on the prediction performance, the power consumption data in the prediction period is error corrected. Specifically, the prediction error sequence is a data sequence formed by the difference between the sample prediction result and the sample label based on the time series prediction model, and the error prediction model is a trained bidirectional long short-term memory network. The sample is the base station power consumption sequence, and the meteorological data corresponding to the sample refers to the meteorological data sequence with the same time and place as the sample. A certain amount of predicted power consumption load can be predicted as the test set data according to the original power consumption load (sample) by the BiLSTM model, and the test set data can be divided into an error training set and an error test set in proportion, and the error training set and the sample label of the corresponding prediction data are used to form a prediction error sequence. The prediction error sequence and the weather data associated with the corresponding sample constitute a multidimensional sequence and are input into the error prediction model for error prediction. The obtained prediction error is used to correct the error of the test set data to obtain the final power consumption data of the base station in the prediction period. The input data can be normalized before entering the error prediction model to speed up the processing speed. Accordingly, when the prediction error is obtained, the prediction error is denormalized to achieve numerical recovery.

[0086] In an example implementation, the step S130 uses a neural network model to predict the communication demand of each base station in the target area based on the historical user behavior data and the historical communication demand data of each base station to obtain the predicted communication demand of each base station, including:

[0087] Taking historical user behavior data and historical communication demand data in the target area as input data;

[0088] The input data is subjected to layer-by-layer feature mapping and nonlinear processing through multiple hidden layers to obtain the predicted communication demand of the target area.

[0089] In this example implementation, historical user behavior data refers to user behavior data of users within the coverage area of ​​the base station over a period of time in the past (such as the past year), and the user behavior data includes the average number of daily trips of users, the frequency of trips during peak hours, the types of activities during holidays (such as staying at home, shopping in shopping malls, etc.), and the length of time different apps are used on the user's device. Historical communication demand data refers to the communication demand data of users within the coverage area of ​​the base station over a period of time in the past (such as the past year), and the communication demand data includes call volume, data traffic, number of connections, etc. The neural network model is a trained multilayer perceptron, and the training process of the multilayer perceptron is as follows: divide the training data set into a training set (80%), a validation set (10%), and a test set (10%), use the training set to train the multilayer perceptron (MPL), input a vector consisting of user behavior features and communication demand features, output future user communication demands, complete model training through the back-propagation algorithm and the adjustment of hyperparameters such as hidden layer size and learning rate, use the validation set and the test set to verify and test the trained model, and based on the test results, use the mean square error MSE and R 2 The performance of the model is evaluated by indicators such as . The multilayer perceptron includes an input layer, multiple hidden layers and an output layer. The input layer and the output layer are used to receive input data and output results. The input data is subjected to layer-by-layer feature mapping and nonlinear processing through multiple hidden layers to obtain the communication demand of the target area in the forecast period. The multilayer perceptron mines the rules of historical user behavior and historical communication demand of users, thereby predicting the communication demand of users in the future time period.

[0090] In an example implementation, the scheduling of communication resources in the target area in the forecast period based on the forecast period communication load and the forecast period communication demand of each base station in step S140 includes:

[0091] Clustering the base stations in the target area based on the positions of the base stations in the target area to obtain multiple base station clusters;

[0092] For each base station cluster, based on the predicted communication load of the base stations in the cluster and the predicted communication demand of the target area, the particle swarm optimization algorithm is used to optimize the scheduling of the communication resources of the cluster in the predicted period to obtain the communication resource scheduling strategy of the cluster in the predicted period.

[0093] In this example implementation, a clustering algorithm (such as K-means) can be used to divide the base stations in the target area into multiple base station clusters according to the location of each base station, so that base stations with closer distances can be used as a base station cluster, and a cluster can be used as a super base station for resource scheduling between clusters. Specifically, first determine the number of clusters to be divided (such as 3), randomly select k from each base station as the initial cluster center, calculate the distance from each base station to each cluster center, and assign the base station to the cluster to which the cluster center closest to it belongs, recalculate the cluster center of each cluster, repeat the iteration until the algorithm converges, complete the clustering of each base station, and obtain multiple base station clusters. Within each base station cluster, the particle swarm optimization algorithm can be used to optimize the scheduling of the communication resources of the cluster during the prediction period to obtain the communication resource scheduling strategy of the cluster during the prediction period. This example can realize base station resource sharing within the cluster and improve the utilization rate of communication power sources. It can also realize multi-base station complementarity and integrate the communication power resources of multiple base stations to form a virtual "super base station" to improve the overall communication capability. It can also realize cross-network collaboration and coordinate scheduling with the networks of other operators to make full use of cross-network resources and improve the overall energy efficiency level of regional communication infrastructure.

[0094] Exemplarily, optimizing and scheduling the communication resources of the cluster within the prediction period by using a particle swarm optimization algorithm includes:

[0095] Taking each base station in the cluster as a particle, taking the multi-dimensional communication resource allocation strategy of each base station as the position of the corresponding particle, and taking the adjustment amount of the multi-dimensional communication resources of each base station as the speed of the corresponding particle;

[0096] Based on each resource allocation strategy of the cluster, the user communication demand satisfaction, power load balancing and energy efficiency of the corresponding cluster are calculated, and based on the weighted fusion of the user communication demand satisfaction, power load balancing and energy efficiency of the cluster, the fitness value corresponding to the current resource allocation strategy is obtained;

[0097] Iteratively updating the speed and position of each particle based on the fitness value until a termination condition is reached, and obtaining a global optimal position of each particle as a communication resource scheduling strategy for the cluster in the prediction period;

[0098] In this example implementation, the multi-dimensional communication resources include power resources, spectrum resources, and communication user resources. Define each base station in the cluster as a particle, and the dimensions of the particle include the power allocation, spectrum allocation, user allocation, etc. of the base station. Assuming there are s base stations, each base station has m resource parameters (such as power, spectrum, etc.), then the dimension of each particle is s×m; initialize the position and speed of each particle, such as randomly generating a set of initial particle positions X j and speed V j, where X j represents the position of the jth particle, X j Corresponding to a resource allocation strategy of the base station, speed V j represents the moving direction and speed of the jth particle in the solution space, V j The adjustment direction and regulation of the multi-dimensional communication resources of the corresponding base station; the objective function of the optimization problem is used as the fitness function, and the fitness function is defined to include three indicators, namely user demand satisfaction, power load balance and energy efficiency. The fitness function is expressed as follows:

[0099] f(X j )=ω1QoS(X j )+ω2PB(X j )+ω3EE(X j )

[0100] Among them, QoS represents user demand satisfaction, PB represents power load balancing, EE represents energy efficiency, and ω1, ω2, and ω3 are the weights of the corresponding indicators. By continuously iteratively updating the speed and position of the particles, the fitness value of each particle is calculated. When the algorithm converges, the global optimal position of each particle is output to obtain the optimal base station communication resource allocation plan within the cluster. This example can realize the dynamic adjustment and allocation of communication resources according to communication needs, thereby realizing the flexible use of communication resources.

[0101] Traditional energy management methods are often based on static and local data, which makes it difficult to fully and accurately reflect the real-time energy consumption and communication power status of communication base stations. However, the current analysis methods and systems for power consumption data of communication base stations are still imperfect. On the one hand, the lack of efficient data processing and analysis algorithms makes it difficult to extract valuable information from massive data; on the other hand, existing scheduling strategies often focus on the optimization of a single goal, ignoring the complex interactive relationship between the communication power of communication base stations and power grid scheduling. Therefore, developing a method and system for analyzing the communication power scheduling potential of communication base stations based on power big data is of great significance for achieving efficient utilization and sustainable development of regional communication infrastructure. In this context, the present invention uses big data technology to deeply mine and analyze the power consumption data of communication base stations, which becomes a key way to improve the efficient utilization of regional communication infrastructure. As an important information source reflecting the operating status of 5G base stations, power consumption data contains rich communication scheduling potential. By real-time monitoring and analyzing these data, the energy consumption characteristics, communication load changes and resource utilization efficiency of communication base stations can be accurately grasped, thereby providing strong support for the formulation of scientific and reasonable scheduling strategies. The present invention aims to deeply mine power consumption data, accurately evaluate the utilization and shared utilization value of communication base stations, and provide a scientific basis for formulating scientific and reasonable scheduling strategies, thereby promoting the deep integration and coordinated development of 5G technology and energy management. The present invention aims to solve the problem of the lack of efficient scheduling methods for communication base stations in the prior art and realize the efficient utilization of regional communication infrastructure.

[0102] Example 2

[0103] Based on the same inventive concept, the present invention also provides a base station communication resource scheduling system based on power big data, the system comprising:

[0104] A fitting module, used to fit the correlation between the power consumption and the communication load of each base station in the target area based on the historical communication load data, historical power consumption data and historical meteorological data of each base station in the target area;

[0105] The power load forecasting module is used to forecast power consumption based on the historical power consumption data of each base station and the time series forecasting model, and obtain the power consumption data of each base station in the forecast period;

[0106] A communication load determination module, configured to obtain the communication load of each base station in the forecast period based on the correlation between the power consumption of each base station and the communication load and the power consumption data in the forecast period;

[0107] The demand prediction module is used to predict the communication demand of each base station in the target area based on the historical user behavior data and historical communication demand data of each base station using a neural network model to obtain the predicted communication demand of each base station;

[0108] The scheduling module is used to schedule communication resources in the target area during the forecast period based on the forecast period communication load and forecast period communication demand of each base station.

[0109] In a possible implementation, the fitting module includes:

[0110] The data acquisition submodule is used to obtain the historical communication load data, historical power consumption data, historical meteorological data and base station location of each base station in the target area;

[0111] The fitting submodule is used to use the historical power consumption data as the target variable, the historical communication load data, historical meteorological data and base station location as the influencing factors of the target variable, and use a linear regression model to fit the correlation between the power consumption and communication load of each base station in the target area.

[0112] In a possible implementation manner, the correlation relationship between the power consumption and the communication load is:

[0113] P=a0+a1L+a2T+a3H+a4Lon+a5Lat+∈

[0114] Among them, P represents the power consumption of the base station, L represents the communication load of the base station, T represents the weather temperature at the base station location, H represents the air humidity at the base station location, Lon and Lat are the longitude and latitude of the base station location, ∈ is the error term, and a0, a1, a2, a3, a4, and a5 are fitting coefficients respectively.

[0115] In a possible implementation, the power load prediction module includes:

[0116] The decomposition submodule is used to perform CEEMDAN-SSA double decomposition on the historical power consumption data of each base station in the target area to obtain multiple eigenmode sequences;

[0117] A prediction submodule is used to predict power consumption based on multiple intrinsic mode sequences and time series prediction models to obtain a prediction sequence for each intrinsic mode;

[0118] The reconstruction submodule is used to reconstruct the prediction sequence of each eigenmode based on the time sequence to obtain the power consumption data of each base station in the target area during the prediction period.

[0119] In a possible implementation manner, the decomposition submodule is specifically used for:

[0120] Based on the complete set empirical mode decomposition, the historical power consumption data of each base station in the target area is decomposed once to obtain multiple decomposed mode subsequences;

[0121] For each decomposed modal subsequence, the multi-scale entropy of each decomposed modal subsequence is determined by scale transformation and local entropy calculation, and the reconstructed aggregation of similar subsequences in the decomposed modal subsequence is performed based on the multi-scale entropy of each decomposed modal subsequence;

[0122] The reconstructed aggregation results are decomposed twice using singular spectrum analysis to extract multiple eigenmode sequences.

[0123] In a possible implementation manner, the power load prediction module further includes an error correction submodule, and the error correction submodule is used to:

[0124] Inputting the sample prediction error of the time series prediction model and the meteorological data corresponding to the sample into the error prediction model to obtain the corresponding predicted power consumption error;

[0125] Performing error correction on the power consumption data for the forecast period based on the forecast power consumption error;

[0126] Among them, the prediction error sequence is a data sequence formed by the difference between the sample prediction result and the sample label based on the time series prediction model, and the error prediction model is a trained bidirectional long short-term memory network.

[0127] In a possible implementation, the neural network model includes multiple hidden layers, and the demand forecasting module is specifically used to:

[0128] Taking historical user behavior data and historical communication demand data in the target area as input data;

[0129] Performing layer-by-layer feature mapping and nonlinear processing on the input data through multiple hidden layers to obtain the predicted communication demand of the target area;

[0130] Wherein, the neural network model is a trained multi-layer perceptron.

[0131] In a possible implementation, the scheduling module includes:

[0132] A clustering submodule, used to cluster the base stations in the target area based on the positions of the base stations in the target area to obtain multiple base station clusters;

[0133] The optimization submodule is used to optimize the scheduling of the communication resources of each base station cluster during the prediction period by using a particle swarm optimization algorithm based on the predicted communication load of the base stations in the cluster and the predicted communication demand of the target area during the prediction period, so as to obtain a communication resource scheduling strategy for the cluster during the prediction period.

[0134] In a possible implementation, the optimization submodule is specifically used for:

[0135] Taking each base station in the cluster as a particle, taking the multi-dimensional communication resource allocation strategy of each base station as the position of the corresponding particle, and taking the adjustment amount of the multi-dimensional communication resources of each base station as the speed of the corresponding particle;

[0136] Based on each resource allocation strategy of the cluster, the user communication demand satisfaction, power load balancing and energy efficiency of the corresponding cluster are calculated, and based on the weighted fusion of the user communication demand satisfaction, power load balancing and energy efficiency of the cluster, the fitness value corresponding to the current resource allocation strategy is obtained;

[0137] Iteratively updating the speed and position of each particle based on the fitness value until a termination condition is reached, and obtaining a global optimal position of each particle as a communication resource scheduling strategy for the cluster in the prediction period;

[0138] Among them, the multi-dimensional communication resources include power resources, spectrum resources and communication user resources.

[0139] Example 3

[0140] like Figure 2 As shown, the present invention also provides an electronic device, which may be a computer device, a single-chip device, an intelligent mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, the processor, and the transceiver component are connected via a bus; the memory may be used to store an execution program, and an exemplary execution program may include instructions; the processor is used to execute the instructions stored in the memory. The memory may also be used to store data, which may be called and / or modified when the instructions are executed.

[0141] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in a storage medium to implement the corresponding method flow or corresponding functions, so as to implement the steps of a base station communication resource scheduling method based on power big data in the above-mentioned embodiment.

[0142] Example 4

[0143] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It can be understood that the storage medium here can include both built-in storage media in electronic devices and, of course, extended storage media supported by electronic devices. The storage medium provides a storage space that stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a base station communication resource scheduling method based on power big data in the above embodiment.

[0144] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0145] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0146] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0147] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0148] 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 its protection scope. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that after reading the present invention, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the application, but these changes, modifications or equivalent substitutions are all within the protection scope of the claims to be approved.

Claims

1. A method for scheduling base station communication resources based on power big data, characterized in that: include: Based on the historical communication load data, historical power consumption data and historical meteorological data of each base station in the target area, the correlation between the power consumption and communication load of each base station in the target area is fitted; Power consumption is predicted based on the historical power consumption data of each base station and the time series prediction model to obtain the power consumption data of each base station during the forecast period; Based on the correlation between the power consumption of each base station and the communication load and the power consumption data of the predicted period, the communication load of each base station in the predicted period is obtained; Based on the historical user behavior data and historical communication demand data of each base station, the communication demand of each base station in the target area is predicted using a neural network model to obtain the predicted communication demand of each base station; The communication resources in the target area in the forecast period are scheduled based on the forecast period communication load and forecast period communication demand of each base station.

2. The method according to claim 1, characterized in that The method of fitting the correlation between the power consumption and the communication load of each base station in the target area based on the historical communication load data, the historical power consumption data and the historical meteorological data of each base station in the target area includes: Obtain historical communication load data, historical power consumption data, historical meteorological data and base station locations of each base station in the target area; Taking the historical power consumption data as the target variable, and the historical communication load data, historical meteorological data and base station location as the influencing factors of the target variable, a linear regression model is used to fit the correlation between the power consumption and communication load of each base station in the target area.

3. The method according to claim 2, characterized in that The correlation between the power consumption of each base station in the target area and the communication load is: P=a0+a1L+a2T+a3H+a4Lon+a5Lat+∈ Among them, P represents the power consumption of the base station, L represents the communication load of the base station, T represents the weather temperature at the base station location, H represents the air humidity at the base station location, Lon and Lat are the longitude and latitude of the base station location, ∈ is the error term, and a0, a1, a2, a3, a4, and a5 are the fitting coefficients respectively.

4. The method according to claim 1, characterized in that The power consumption forecasting is performed based on the historical power consumption data of each base station and the time series forecasting model to obtain the power consumption data of each base station during the forecast period, including: The historical power consumption data of each base station in the target area is subjected to CEEMDAN-SSA double decomposition to obtain multiple eigenmode sequences; Power consumption is predicted based on multiple intrinsic mode sequences and time series prediction models to obtain the prediction sequence of each intrinsic mode; The prediction sequence of each eigenmode is reconstructed based on the time sequence to obtain the power consumption data of each base station in the target area during the forecast period.

5. The method according to claim 4, characterized in that The historical power consumption data of each base station in the target area is subjected to complete set empirical mode decomposition-singular spectrum analysis double decomposition to obtain multiple eigenmode sequences, including: Based on the complete set empirical mode decomposition, the historical power consumption data of each base station in the target area is decomposed once to obtain multiple decomposed mode subsequences; For each decomposed modal subsequence, the multi-scale entropy of each decomposed modal subsequence is determined by scale transformation and local entropy calculation, and the reconstructed aggregation of similar subsequences in the decomposed modal subsequence is performed based on the multi-scale entropy of each decomposed modal subsequence; The reconstructed aggregation results are decomposed twice using singular spectrum analysis to extract multiple eigenmode sequences.

6. The method according to claim 4, characterized in that After the power consumption data of each base station is obtained by performing power consumption forecasting based on the historical power consumption data of each base station and the time series forecasting model, the following is also included: Inputting the sample prediction error of the time series prediction model and the meteorological data corresponding to the sample into the error prediction model to obtain the corresponding predicted power consumption error; Performing error correction on the power consumption data for the forecast period based on the forecast power consumption error; The prediction error sequence is a data sequence formed by the difference between the sample prediction result and the sample label based on the time series prediction model, and the error prediction model is a trained bidirectional long short-term memory network.

7. The method according to claim 1, characterized in that The neural network model includes a plurality of hidden layers. Based on the historical user behavior data and historical communication demand data of each base station, the communication demand of each base station in the target area is predicted by using the neural network model to obtain the predicted communication demand of each base station, including: Taking historical user behavior data and historical communication demand data in the target area as input data; Performing layer-by-layer feature mapping and nonlinear processing on the input data through multiple hidden layers to obtain the predicted communication demand of the target area; Wherein, the neural network model is a trained multi-layer perceptron.

8. The method according to claim 1, characterized in that: The scheduling of communication resources in the target area in the forecast period based on the forecast period communication load and the forecast period communication demand of each base station includes: Clustering the base stations in the target area based on the positions of the base stations in the target area to obtain multiple base station clusters; For each base station cluster, based on the predicted communication load of the base stations in the cluster and the predicted communication demand of the target area, the particle swarm optimization algorithm is used to optimize the scheduling of the communication resources of the cluster in the predicted period to obtain the communication resource scheduling strategy of the cluster in the predicted period.

9. The method according to claim 8, characterized in that Based on the predicted period communication load of the base stations in the cluster and the predicted period communication demand of the target area, the communication resources of the cluster in the predicted period are optimized and scheduled by using a particle swarm optimization algorithm to obtain a communication resource scheduling strategy of the cluster in the predicted period, including: Taking each base station in the cluster as a particle, taking the multi-dimensional communication resource allocation strategy of each base station as the position of the corresponding particle, and taking the adjustment amount of the multi-dimensional communication resources of each base station as the speed of the corresponding particle; Based on each resource allocation strategy of the cluster, the user communication demand satisfaction, power load balancing and energy efficiency of the corresponding cluster are calculated, and based on the weighted fusion of the user communication demand satisfaction, power load balancing and energy efficiency of the cluster, the fitness value corresponding to the current resource allocation strategy is obtained; Iteratively updating the speed and position of each particle based on the fitness value until a termination condition is reached, and obtaining a global optimal position of each particle as a communication resource scheduling strategy for the cluster in the prediction period; Among them, the multi-dimensional communication resources include power resources, spectrum resources and communication user resources.

10. A base station communication resource scheduling system based on power big data, characterized in that: include: A fitting module, used to fit the correlation between the power consumption and the communication load of each base station in the target area based on the historical communication load data, historical power consumption data and historical meteorological data of each base station in the target area; The power load forecasting module is used to forecast power consumption based on the historical power consumption data of each base station and the time series forecasting model, and obtain the power consumption data of each base station in the forecast period; A communication load determination module, configured to obtain the communication load of each base station in the forecast period based on the correlation between the power consumption of each base station and the communication load and the power consumption data in the forecast period; The demand prediction module is used to predict the communication demand of each base station in the target area based on the historical user behavior data and historical communication demand data of each base station using a neural network model to obtain the predicted communication demand of each base station; The scheduling module is used to schedule communication resources in the target area during the forecast period based on the forecast period communication load and forecast period communication demand of each base station.

Images