An energy saving control method and device of a base station, electronic equipment and storage medium

Abstract

The application discloses an energy-saving control method and device of a base station, electronic equipment and a storage medium. The application identifies different load operation days in each week and low load time periods in different operation days in each week based on historical daily operation load data of the base station and by means of clustering analysis and linear regression. In this way, the working scenes of each base station can be identified, and an energy-saving scheme of each base station in the corresponding working scene is generated. Based on this, the energy-saving method of the base station can be matched with the actual business situation of the base station and the network demand of users, so that the energy consumption of the base station is reduced while meeting the use demand of users and the business demand of the base station. Therefore, the application realizes intelligent management of the energy consumption of the base station and is suitable for large-scale application and popularization in the field of base station energy saving.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of base station energy saving, and particularly relates to a base station energy saving control method and device, an electronic device and a storage medium. BACKGROUND

[0002] In order to ensure that mobile users can use mobile services anytime and anywhere, and meet the communication needs of users in mobile terminal number intensive areas and busy time periods, a large number of base station devices are arranged in the existing network to realize complete coverage of mobile signals and multi-layer coverage in areas with high traffic volume. However, it is found through statistics that the utilization rate of base station devices is not high, and the average utilization rate is less than 50%. The instantaneous utilization rate is even less than 10% in some time periods. In this case, the continuous operation of the base station causes the problem of energy waste. At the same time, with the development of communication technology, the distribution of fifth-generation mobile communication technology base stations is also more and more extensive. While providing people with higher quality network services, it also further increases the energy consumption of base stations. Therefore, how to reduce the energy waste of base stations has become the focus of attention of all parties.

[0003] At present, in order to reduce the energy consumption of base stations, artificial experience is mostly used to set the base station to an energy saving state in a fixed time period (for example, the base station can be set to an energy saving state at 00:00-5:00 at night). Although this method can achieve a certain energy saving effect, since the network demand of users served by the base station is dynamically changing, there may be a situation that there are more active users in the fixed time period and fewer active users outside the fixed time period. It can be seen that the current energy saving mode of the base station has the problem that it does not match the actual business situation of the base station and the network demand of users, and the energy saving effect is not good. Therefore, how to provide a method for controlling the energy saving of the base station based on the network demand of users has become a problem to be solved. SUMMARY

[0004] The purpose of the present application is to provide a base station energy saving control method, device, electronic device and storage medium, to solve the problem that the energy saving mode of the base station in the prior art does not match the actual business situation of the base station and the network demand of users, thereby causing the problem of poor energy saving effect.

[0005] In order to achieve the above-mentioned purpose, the present application adopts the following technical scheme:

[0006] In a first aspect, a base station energy saving control method is provided, comprising:

[0007] obtaining a running load data set of a target base station in a historical time period, wherein the running load data set comprises a plurality of time-continuous running load data sequences, and the time length of any running load data sequence is 24 hours;

[0008] performing clustering analysis on the several operation load data sequences in the operation load data set to obtain k operation load clusters, and performing linear regression processing on the several operation load data sequences to obtain a set of daily operation load trend functions of the target base station, wherein the set of daily operation load trend functions comprises a daily operation load trend function corresponding to each day in a week of the target base station, and k is a positive integer;

[0009] selecting an optimal operation load cluster from the k operation load clusters, and determining week operation load change trend information of the target base station based on the optimal operation load cluster;

[0010] determining a daily operation energy-saving time period of the target base station in a week according to the set of daily operation load trend functions;

[0011] generating an energy-saving control instruction of the target base station according to the week operation load change trend information and the daily operation energy-saving time period, wherein the energy-saving control instruction comprises energy-saving operation days of the target base station in a week, a running start time and a running end time of each energy-saving operation day, and an energy-saving strategy of each energy-saving operation day, and the energy-saving strategy comprises symbol-off, time-slot-off, channel-off or device hibernation;

[0012] sending the energy-saving control instruction to the target base station, so that the target base station executes the energy-saving control instruction, so as to complete energy-saving control of the target base station after executing the energy-saving control instruction.

[0013] Based on the above disclosure, the application first obtains continuous operation load data sequences in a historical time period of a target base station in units of 24 hours; then, the corresponding operation service load rule of the target base station is analyzed based on the foregoing data, that is, clustering analysis is performed on the obtained several operation load data sequences, and linear regression processing is performed, to obtain an operation load cluster to which each operation load data sequence belongs and a daily operation load trend function corresponding to each day of the target base station in a week; then, the application determines an optimal operation load cluster according to the cluster to which the foregoing data belongs, and determines a week operation load change trend of the target base station according to the optimal operation load cluster, which is equivalent to the load change of the target base station in a week, such as high-load operation days from Monday to Friday, and low-load operation days on Saturday and Sunday; based on this, this step is equivalent to dividing different operation load days of the target base station in a week, and can provide data basis for generating a corresponding energy-saving scheme of the target base station.

[0014] Meanwhile, the application also determines the daily energy-saving operation time period of the target base station in a week based on the daily operation load trend functions of each base station; finally, according to the aforementioned weekly operation load change trend information and the daily energy-saving operation time period, the energy-saving operation day of the target base station in a week, the start time and end time of each energy-saving operation day and the corresponding energy-saving strategy are determined, and the energy-saving control instruction is generated and sent to the target base station, so that the target base station works according to the instruction, and the energy-saving operation of the target base station is finally realized.

[0015] Through the above design, the application identifies the different load operation days in a week and the low load time period of each day in a week of the base station based on the historical daily operation load data of the base station and by means of clustering analysis and linear regression; in this way, the working scene of each base station is identified, and the energy-saving scheme of each base station under the corresponding working scene is generated; based on this, the energy-saving method of the base station provided by the application can match the actual business situation of the base station and the network demand of users, so that the energy consumption of the base station is reduced while meeting the use demand of users and the business demand of the base station, thereby the application realizes the intelligent management of the energy consumption of the base station, and is suitable for large-scale application and promotion in the field of base station energy saving.

[0016] In one possible design, the clustering analysis is performed on a plurality of operation load data sequences in the operation load data set to obtain k operation load clusters, including:

[0017] The first clustering density of each operation load data sequence in the operation load data set is calculated, and the first average clustering density of the operation load data set is calculated, and based on the first average clustering density and the first clustering density of each operation load data sequence, k operation load data sequences are determined as initial clustering centers;

[0018] The distance between the pth target operation load data sequence in the target operation load data set and each initial clustering center is calculated, and based on the distance between the pth target operation load data sequence and each initial clustering center, the pth target operation load data sequence is divided into a target initial clustering center, wherein the target operation load data set is the operation load data set after deleting all initial clustering centers, and the target initial clustering center is the initial clustering center with the minimum distance from the pth target operation load data sequence;

[0019] p is increased by 1, and the distance between the pth target operation load data sequence and each initial clustering center is recalculated, until p is equal to P, and k initial operation load clusters are obtained, wherein the initial value of p is 1, and P is the total number of target operation load data sequences;

[0020] calculating a cluster center of each initial operation load cluster, and determining a cluster criterion function value based on the cluster center of each initial operation load cluster and the operation load data set;

[0021] determining whether the cluster criterion function value is less than or equal to a preset cluster threshold value;

[0022] If not, updating the initial cluster center to the cluster center of each initial operation load cluster, and recalculating the distance between the pth target operation load data sequence in the target operation load data set and each initial cluster center until the cluster criterion function value is less than or equal to the preset cluster threshold value, so as to take the k initial operation load clusters corresponding to the cluster criterion function value less than or equal to the preset cluster threshold value as the k operation load clusters.

[0023] In one possible design, calculating a first cluster density of each operation load data sequence in the operation load data set comprises:

[0024] calculating the Euclidean distance between each pair of operation load data sequences in the operation load data set, and summing all the Euclidean distances to obtain a Euclidean distance summary value;

[0025] determining an average Euclidean distance of the operation load data set by using the Euclidean distance summary value;

[0026] for the ith operation load data sequence in the operation load data set, calculating a first cluster density of the ith operation load data sequence based on the average Euclidean distance and the Euclidean distance between the ith operation load data sequence and each operation load data sequence;

[0027] incrementing i by 1, and recalculating the first cluster density of the ith operation load data sequence based on the average Euclidean distance and the Euclidean distance between the ith operation load data sequence and each operation load data sequence until i is equal to n, where the initial value of i is 1 and n is the total number of operation load data sequences.

[0028] In one possible design, determining the average Euclidean distance of the operation load data set by using the Euclidean distance summary value comprises:

[0029] the average Euclidean distance of the operation load data set is calculated by using the following formula (1);

[0030]

[0031] In the above formula (1), denotes the average Euclidean distance, s denotes the Euclidean distance aggregate value, denotes the number of sets composed of any two operation load data sequences selected from n operation load data sequences, and n denotes the total number of operation load data sequences;

[0032] Correspondingly, based on the average Euclidean distance and the Euclidean distance between the i-th operation load data sequence and each operation load data sequence, the first clustering density of the i-th operation load data sequence is calculated, comprising:

[0033] The first clustering density of the i-th operation load data sequence is calculated by using the following formula (2);

[0034]

[0035] In the above formula (2), w i denotes the i-th operation load data sequence, m(w i ) denotes the first clustering density of the i-th operation load data sequence, s(w i , w j ) denotes the Euclidean distance between the i-th operation load data sequence and the j-th operation load data sequence, denotes the density function, and

[0036] In one possible design, based on the first average clustering density and the first clustering density of each operation load data sequence, k operation load data sequences are determined as initial clustering centers, comprising:

[0037] Based on the first average clustering density and the first clustering density of each operation load data sequence, the density attribute of each operation load data sequence is determined, wherein the density attribute includes a density sparse point or a density dense point;

[0038] From the operation load data set, the operation load data sequence with the density attribute of the density sparse point is extracted to form a sparse data set;

[0039] The second clustering density of each operation load data sequence in the sparse data set and the second average clustering density of the sparse data set are calculated, and from a plurality of second clustering densities, the second clustering density greater than the second average clustering density is extracted to form a density set using the extracted second clustering density;

[0040] It is judged whether the number of the maximum second clustering density in the density set is equal to 1;

[0041] If yes, the operation load data sequence corresponding to the maximum second clustering density is taken as an initial clustering center;

[0042] deleting a target second clustering density from the density set to obtain an updated density set, wherein the target second clustering density corresponds to a running load data sequence whose Euclidean distance with specified data is less than an average Euclidean distance of the running load data set, and the specified data is a remaining running load data sequence after the running load data sequence corresponding to the target second clustering density is deleted from the running load data set;

[0043] calculating a second average clustering density of the updated density set, and extracting a second clustering density greater than the second average clustering density from the second clustering densities of the density set until k initial clustering centers are obtained.

[0044] In a possible design, if the number of the maximum second clustering densities in the density set is greater than 1, the method further includes:

[0045] composing a candidate clustering center set by using the running load data sequence corresponding to the maximum second clustering density;

[0046] for the hth running load data sequence in the candidate clustering center set, calculating the Euclidean distance between the hth running load data sequence and each running load data sequence in the candidate clustering center set, and summing the Euclidean distances between the hth running load data sequence and each running load data sequence in the candidate clustering center set to obtain a clustering center value of the hth running load data sequence in the candidate clustering center set;

[0047] increasing h by 1 until h is equal to H, to obtain the clustering center value of each running load data sequence in the candidate clustering center set, wherein the initial value of h is 1, and H is the total number of data sequences in the candidate clustering center set;

[0048] taking the running load data sequence corresponding to the minimum clustering center value in the candidate clustering center set as an initial clustering center.

[0049] In a possible design, selecting an optimal running load cluster from the k running load clusters includes:

[0050] calculating a silhouette coefficient of each running load cluster, and selecting a running load cluster with the maximum silhouette coefficient as the optimal running load cluster;

[0051] Correspondingly, determining the weekly running load change trend information of the target base station based on the optimal running load cluster includes:

[0052] determining whether a profile coefficient of the optimal operation load cluster is greater than or equal to a preset threshold value;

[0053] if yes, generating a weekly operation load change trend clustering graph of the target base station based on the k operation load clusters;

[0054] determining high-load operation days and low-load operation days of the target base station in a week according to the weekly operation load change trend clustering graph;

[0055] composing the weekly operation load change trend information by using the high-load operation days and the low-load operation days.

[0056] In a second aspect, an energy-saving control device of a base station is provided, which comprises:

[0057] a data acquisition unit configured to acquire a set of operation load data of a target base station in a historical time period, wherein the set of operation load data comprises a plurality of time-continuous operation load data sequences, and any operation load data sequence corresponds to a time length of 24 hours;

[0058] a data analysis unit configured to perform clustering analysis on the plurality of operation load data sequences in the set of operation load data to obtain k operation load clusters, and perform linear regression processing on the plurality of operation load data sequences to obtain a set of daily operation load trend functions of the target base station, wherein the set of daily operation load trend functions comprises a daily operation load trend function corresponding to each day in a week of the target base station, and k is a positive integer;

[0059] the data analysis unit is further configured to select an optimal operation load cluster from the k operation load clusters, and determine a weekly operation load change trend information of the target base station based on the optimal operation load cluster;

[0060] the data analysis unit is further configured to determine a daily energy-saving time period of the target base station in a week according to the set of daily operation load trend functions;

[0061] an energy-saving strategy generation unit configured to generate an energy-saving control instruction of the target base station according to the weekly operation load change trend information and the daily energy-saving time period of the target base station, wherein the energy-saving control instruction comprises energy-saving operation days of the target base station in a week, a start time and an end time of each energy-saving operation day, and an energy-saving strategy of each energy-saving operation day, and the energy-saving strategy comprises symbol shutdown, time slot shutdown, channel shutdown or device hibernation;

[0062] The sending unit is configured to send the energy-saving control instruction to the target base station, so that the target base station executes the energy-saving control instruction, so as to complete the energy-saving control of the target base station after executing the energy-saving control instruction.

[0063] In a third aspect, another energy-saving control device of a base station is provided, which is taken as an electronic device and includes a memory, a processor and a transceiver connected in sequence, wherein the memory is configured to store a computer program, the transceiver is configured to transceive messages, and the processor is configured to read the computer program and execute the energy-saving control method of the base station as in the first aspect or any possible design of the first aspect.

[0064] In a fourth aspect, a storage medium is provided, and the storage medium stores instructions, which, when executed on a computer, execute the energy-saving control method of the base station as in the first aspect or any possible design of the first aspect.

[0065] In a fifth aspect, a computer program product is provided, and the computer program product includes instructions, which, when executed on a computer, cause the computer to execute the energy-saving control method of the base station as in the first aspect or any possible design of the first aspect.

[0066] Advantages:

[0067] (1) The energy-saving method of the base station provided by the present application can match the actual business situation of the base station and the network demand of the user, so as to reduce the energy consumption of the base station while meeting the use demand of the user and the business demand of the base station, thereby realizing the intelligent management of the energy consumption of the base station and being suitable for large-scale application and promotion in the field of base station energy saving. BRIEF DESCRIPTION OF DRAWINGS

[0068] Figure 1 A step flowchart of the energy-saving control method of the base station provided by the embodiment of the present application is shown in the figure.

[0069] Figure 2 A structure diagram of the energy-saving control device of the base station provided by the embodiment of the present application is shown in the figure.

[0070] Figure 3 A structure diagram of the electronic device provided by the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0071] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the present application will be briefly introduced below in connection with the drawings and the description of the embodiments or the prior art. Obviously, the following description of the drawings only constitutes some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings. It should be noted that the description of these embodiments is used to help understand the present application, but does not constitute a limitation on the present application.

[0072] It should be understood that although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element can be called a second element, and similarly, a second element can be called a first element without departing from the scope of the example embodiments of the present application.

[0073] It should be understood that for the term "and / or" that may appear herein, it only describes the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which means that there are three cases of A alone, B alone, and A and B together. For the term "and" that may appear herein, it describes another association object relationship, which means that there can be two relationships, for example, A and B, which means that there are two cases of A alone and A and B together. In addition, for the character " / " that may appear herein, it generally means that the associated objects before and after are an "or" relationship.

[0074] Embodiment:

[0075] Referring to Figure 1 As shown in the figure, the energy saving control method of the base station provided by the embodiment can identify the different load operation days in a week and the low load period of each operation day in a week according to the historical business operation load data of the base station, and generate an energy saving control scheme of the base station based on the same. In this way, the method can identify the working scenario of the base station, and realize customized energy saving control of the base station based on the working scenario of the base station, so as to reduce the energy consumption of the base station while meeting the user use demand and the base station business demand, and further realize intelligent management of the energy consumption of the base station. Optionally, the method can be but is not limited to running at the base station management end side, and the base station management end can be but is not limited to a personal computer (PC), a tablet computer or a smart phone. It can be understood that the foregoing execution subject does not constitute a limitation on the embodiments of the present application, and accordingly, the running steps of the method can be but are not limited to the steps S1-S6 shown below.

[0076] S1. Obtain a set of operation load data of a target base station in a historical time period, wherein the set of operation load data comprises a plurality of time-continuous operation load data sequences, and any operation load data sequence corresponds to a time length of 24 hours; in a specific application, the target base station is a base station that needs to implement energy consumption management, and each operation load data sequence is essentially a service load data record of the target base station in 24 hours, and the service data of the target base station can include but is not limited to: active user service, uplink and downlink average traffic, packet data convergence protocol data, network data (such as uplink and downlink carrier utilization rate and various signal occupation rate, etc.), etc. Of course, the aforementioned historical time period can be data of 10 or 20 consecutive weeks in a year, which can be set according to actual use, and is not limited here. In addition, in this embodiment, after obtaining the set of operation load data, data preprocessing is performed, such as deleting error values, supplementing missing data, etc. After processing, the target base station corresponding service operation load rule can be analyzed based on the data set, as shown in the following step S2.

[0077] S2. Cluster analysis is performed on a plurality of operation load data sequences in the set of operation load data to obtain k operation load clusters, and linear regression processing is performed on the plurality of operation load data sequences to obtain a set of daily operation load trend functions of the target base station, wherein the set of daily operation load trend functions comprises a daily operation load trend function corresponding to each day of the target base station in a week, and k is a positive integer. In this embodiment, the cluster analysis of the data sequences in the set of operation load data is essentially to divide each operation load data sequence into a plurality of clusters, so as to determine the weekly operation load change mode of the target base station according to the divided clusters. In this embodiment, the weekly operation load change mode of the target base station can be an operation load change trend rule (such as the operation load of Monday to Friday is at a peak value, and the operation load of Saturday and Sunday is at a valley value; or such as the operation load of Monday to Friday is at a valley value, and the operation load of Saturday and Sunday is at a peak value), an operation load change trend consistent (the operation load has no obvious change in a week, and the load is approximately the same), or an operation load change without a rule. In this way, the aforementioned operation load change mode can provide a data basis for the generation of a subsequent energy saving scheme.

[0078] Meanwhile, by linearly regressing the aforementioned several operation load data sequences, a daily operation load trend function of the target base station in a week can be obtained; then, by means of the function, the low load period of the target base station in each day of a week can be determined; in this way, in combination with the aforementioned weekly operation load change mode, the energy-saving scheme of the target base station can be determined; in this embodiment, the operation load data sequences belonging to the same operation day in different weeks can be but not limited to linearly regressed, so as to regress the daily operation load trend function of the target base station in a week; for example, the operation load data sequences belonging to Monday in different weeks are linearly regressed, so as to obtain the daily operation load trend function of the target base station corresponding to Monday; of course, the generation principles of the daily operation load trend functions of the remaining days are the same as the aforementioned example, and will not be described herein.

[0079] In this embodiment, the following provides one of the methods for performing clustering analysis on the aforementioned operation load data set, which can but not limited to the steps S21-S26.

[0080] S21. Calculate the first clustering density of each operation load data sequence in the operation load data set, and the first average clustering density of the operation load data set, and determine k operation load data sequences as initial clustering centers based on the first average clustering density and the first clustering density of each operation load data sequence; in this embodiment, step S21 is the process of determining initial clustering centers, that is, before performing clustering analysis, k operation load data sequences need to be selected from the aforementioned operation load data set as initial clustering centers; then, iteration clustering is continuously performed until the end condition is met; in specific application, this embodiment determines the density sparse data sequences in each operation load data sequence by using the clustering density between each operation load data sequence and the average clustering density of all operation load data sequences; then, the initial clustering centers can be determined based on the density sparse data sequences; optionally, the calculation process of the first clustering density of each operation load data sequence can but not limited to the following S21a-S21d.

[0081] S21a. Calculate the Euclidean distance between each two operation load data sequences in the operation load data set, and sum all the Euclidean distances to obtain a Euclidean distance summary value; in this embodiment, the Euclidean distance between each two operation load data sequences includes itself, for example, assuming that there are operation load data sequences t1, t2 and t3, then for operation load data sequence t1, the Euclidean distance between t1 and t1, the Euclidean distance between t1 and t2, and the Euclidean distance between t1 and t3 are calculated; of course, the calculation of the Euclidean distance of the remaining operation load data sequences is the same as the aforementioned example, and will not be described herein.

[0082] After obtaining the Euclidean distance aggregate value of the running load data set, the average Euclidean distance of the running load data set is calculated, and then based on the average Euclidean distance and the Euclidean distance between each running load data sequence, the first clustering density of each running load data sequence is calculated.

[0083] S21b. The average Euclidean distance of the running load data set is determined by using the Euclidean distance aggregate value. In this embodiment, the average Euclidean distance of the running load data set can be calculated by using the following formula (1), but is not limited thereto.

[0084]

[0085] In the above formula (1), s represents the average Euclidean distance, and s represents the Euclidean distance aggregate value. The number of sets composed of any two running load data sequences from n running load data sequences is represented by s, and n represents the total number of running load data sequences.

[0086] Thus, after the average Euclidean distance is calculated, the first clustering density of each running load data sequence can be calculated in combination with the Euclidean distance between each running load data sequence, as shown in the following step S21c.

[0087] S21c. For the i-th running load data sequence in the running load data set, the first clustering density of the i-th running load data sequence is calculated based on the average Euclidean distance and the Euclidean distance between the i-th running load data sequence and each running load data sequence. In specific implementation, the first clustering density of the i-th running load data sequence can be calculated by using the following formula (2), but is not limited thereto.

[0088]

[0089] In the above formula (2), w i represents the i-th running load data sequence, m(w i ) represents the first clustering density of the i-th running load data sequence, s(w i ,w j ) represents the Euclidean distance between the i-th running load data sequence and the j-th running load data sequence, represents the density function, and

[0090] The above formula (2) is described on the basis of the foregoing example.

[0091] ​Assuming that the first running load data sequence t1 is taken as an example, then the average Euclidean distance is calculated first, and the difference between the Euclidean distance of t1 and itself is calculated, and then the value of the density function (denoted as q1) is calculated based on the difference. Then, the average Euclidean distance is calculated again, and the difference between the Euclidean distance of t1 and the second running load data sequence (t2) is calculated, and then the value of the density function in the second operation (denoted as q2) is calculated based on the difference. Similarly, the average Euclidean distance is calculated, and the difference between the Euclidean distance of t1 and the third running load data sequence (t3) is calculated, and then the value of the density function in the third operation (denoted as q3) is calculated. Finally, the sum of q1, q2 and q3 is calculated, and the first clustering density of t1 is obtained. Thus, i is continuously incremented by 1, and the first clustering density is calculated by using the above formula (2), until i is equal to n, and then the first clustering density of each running load data sequence is obtained, wherein the loop process is shown in the following step S21d.

[0092] S21d. i is incremented by 1, and the first clustering density of the ith running load data sequence is calculated again based on the average Euclidean distance and the Euclidean distance between the ith running load data sequence and each running load data sequence, until i is equal to n, and the first clustering density of each running load data sequence is obtained, wherein the initial value of i is 1, and n is the total number of running load data sequences.

[0093] Thus, after the first clustering density of each running load data sequence is calculated by the above steps S21a-S21d, the sum of all first clustering densities is calculated, and the average is taken, and then the first average clustering density is obtained, and then the k initial clustering centers are determined based on the first average clustering density and each first clustering density. Optionally, the selection process of the initial clustering centers can be but is not limited to the following steps S21e-S21k.

[0094] S21e. The density attribute of each running load data sequence is determined based on the first average clustering density and the first clustering density of each running load data sequence, wherein the density attribute includes a density sparse point or a density dense point. In this embodiment, for any running load data sequence, the product of the first average clustering density and the parameter b can be but is not limited to calculated, and then it is judged whether the product is greater than the first clustering density of the any running load data sequence. If yes, it is determined that the density attribute of the any running load data sequence is a density sparse point. If the product is less than the first clustering density of the any running load data sequence, it is determined that the density attribute of the any running load data sequence is a density dense point. Optionally, 0<b<1, and preferably 0.2.

[0095] After the density attribute of each operation load data sequence is determined, the density sparse sequence is selected based on the density attribute, so that the initial clustering center is determined based on the selected sequence; wherein the specific process is shown in the following steps S21f-S21k.

[0096] S21f. The operation load data sequence with the density attribute of density sparse point is extracted from the operation load data set to constitute the sparse data set; in this embodiment, step S21f is equivalent to screening the density sparse data sequence from the operation load data set, and then determining the initial clustering center based on the clustering density of the density sparse data sequence (i.e. the operation load data sequence with the density attribute of density sparse point) and the average clustering density of all the density sparse data sequences again, as shown in the following steps S21g-S21k.

[0097] S21g. The second clustering density of each operation load data sequence in the sparse data set and the second average clustering density of the sparse data set are calculated, and the second clustering density greater than the second average clustering density is extracted from the second clustering densities to constitute the density set; in this embodiment, the calculation process of the second clustering density of each operation load data sequence in the sparse data set and the second average clustering density of the set can be referred to the foregoing steps S21a-S21d, and the principle is not described again; and after the second clustering density of each operation load data sequence in the sparse data set and the second average clustering density of the sparse data set are obtained, the second clustering density greater than the second average clustering density is screened out to constitute the density set; finally, the initial clustering center can be determined according to the density set.

[0098] The foregoing step S21g is described with an example as follows:

[0099] Suppose that the sparse data set includes operation load data sequences t2, t3, t4, t5 and t8, wherein the second clustering density of t2 is 1.3, the second clustering density of t3 is 1.6, the second clustering density of t4 is 2, the second clustering density of t5 is 1.9, and the second clustering density of t8 is 2.5; and the second average clustering of the sparse data set is 1.86; then the density set includes 2, 1.9 and 2.5; of course, in the case of different number of operation load data sequences, the confirmation principle of the density set is the same as the foregoing example, which is not described again.

[0100] After the density set is obtained, the initial clustering center is determined based on the number of the maximum second clustering density in the density set, as shown in the following steps S21h-S21k.

[0101] S21h. determining whether the number of the maximum second cluster density in the density set is equal to 1.

[0102] S21i. If yes, the operation load data sequence corresponding to the maximum second cluster density is taken as an initial cluster center; in the embodiment, based on the foregoing example, the maximum second cluster density in the density set is 2.5, and the number thereof is only 1, thus the operation load data sequence corresponding to the second cluster density 2.5 can be taken as an initial cluster center, that is, the operation load data sequence t8 is taken as an initial cluster center.

[0103] After an initial cluster center is determined, the second cluster density in the density set needs to be deleted to obtain an updated density set, and then the second average cluster density of the updated density set (for the sake of distinction, hereinafter referred to as the updated second average cluster density) is recalculated; at this time, the foregoing step S21g is repeated, that is, the second cluster density greater than the updated second average cluster density is extracted from the updated density set to serve as the density set; finally, an initial cluster center can be determined again based on the same principle of the foregoing step S21h and step S21i; based on this, the foregoing operation is repeatedly performed until k initial cluster centers are determined; in the embodiment, the loop process is shown in the following steps S21j and S21k.

[0104] S21j. deleting the target second cluster density from the density set to obtain an updated density set, wherein the Euclidean distance between the operation load data sequence corresponding to the target second cluster density and the specified data is less than the average Euclidean distance of the operation load data set, and the specified data is the remaining operation load data sequence after the operation load data sequence corresponding to the target second cluster density is deleted from the operation load data set; in the embodiment, the maximum second cluster density in the foregoing step S21i is also deleted when the target second cluster density is deleted; at the same time, it has been described that the specified data is the remaining data after the operation load data sequence corresponding to the target second cluster density is deleted from the operation load data set; thus, as long as the Euclidean distance between the operation load data sequence corresponding to the target second cluster density and any specified data is less than the average Euclidean distance, the target second cluster density can be deleted.

[0105] As, after determining an initial clustering center, the density set includes 2 and 1.9, at this time, the Euclidean distance between the running load data sequence t4 and each specified data (assuming that the specified data has the running load data sequences t1, t2, t3, t5-t8) is calculated, thus, the Euclidean distance between t4 and t1, t4 and t2, t4 and t3, t4 and t5, t4 and t6, t4 and t7, and t4 and t8 is calculated, at this time, as long as any one of the aforementioned 7 distances is less than the average Euclidean distance, then the second clustering density corresponding to t4 (i.e. the second clustering density 2) is deleted from the density set; of course, the updating principle of the remaining different density sets is the same as the aforementioned example, and will not be described here.

[0106] After obtaining the updated density set, the second average clustering density thereof is recalculated; then, the size relationship between the second average clustering density of the updated density set and each second clustering density in the updated density set is used to determine a new density set; finally, the aforementioned steps S21i and S21j are repeated to determine an initial clustering center again; in this way, the k initial clustering centers are determined from the running load data set by continuously cycling according to the principle, as shown in the following step S21k.

[0107] S21k. Calculate the second average clustering density of the updated density set, and extract the second clustering density greater than the second average clustering density from the second clustering densities of the density set until k initial clustering centers are obtained; in the present embodiment, if the number of the maximum second clustering density in the density set in the aforementioned step S21i is greater than 1, then an initial clustering center is determined by using the following steps:

[0108] First, the candidate clustering center set is composed of the running load data sequence corresponding to the maximum second clustering density.

[0109] Then, for the hth running load data sequence in the candidate clustering center set, the Euclidean distance between the hth running load data sequence in the candidate clustering center set and each running load data sequence in the candidate clustering center set is calculated, and the sum of the Euclidean distance between the hth running load data sequence in the candidate clustering center set and each running load data sequence in the candidate clustering center set is obtained to obtain the clustering center value of the hth running load data sequence in the candidate clustering center set.

[0110] Next, h is incremented by 1 until h is equal to H to obtain the clustering center value of each running load data sequence in the candidate clustering center set, wherein the initial value of h is 1 and H is the total number of data sequences in the candidate clustering center set.

[0111] Finally, the running load data sequence corresponding to the minimum cluster center value in the candidate cluster center set is taken as an initial cluster center; of course, after the initial cluster center is determined, the foregoing step S21j is executed, and then the loop is continuously performed until the k initial cluster centers are determined.

[0112] The following describes the operation steps when the number of the maximum second cluster density in the density set is greater than 1 by taking an example:

[0113] Suppose that there are three maximum second cluster densities, and the corresponding running load data sequences are t3, t5 and t7; at this time, the candidate cluster center set is {t3, t5, t7}, wherein for the running load data sequence t3, the Euclidean distances between t3 and t3, t3 and t5 and t3 and t7 are calculated, and then the sum of the three Euclidean distances is obtained to obtain the cluster center value (supposed to be 3) of the running load data sequence t3; then, the cluster center values of the running load data sequences t5 and t7 are calculated (supposed to be 3.4 and 2.9 respectively) by using the foregoing same principle; finally, the running load data sequence corresponding to the minimum cluster center value, that is, t7, is taken as the initial cluster center; of course, the determination principle of the initial cluster center is the same as the foregoing example in the case of different numbers of the maximum second cluster density, and thus is not described herein.

[0114] Thus, by the foregoing steps S21a-S21k, the cluster density and the average cluster density of each running load data sequence are calculated by using the Euclidean distances between the running load data sequences and the average Euclidean distance, and the initial cluster center is determined based on the two indexes; in this way, the problem that the traditional clustering method is easily affected by the selection of the initial cluster center and falls into a local optimal solution and the problem that an effective cluster cannot be obtained are avoided; thus, a more reasonable initial cluster center can be provided for the subsequent clustering analysis of the running load data sequences, thereby improving the effectiveness and accuracy of the clustering.

[0115] After the k initial cluster centers are determined, the clustering analysis of a plurality of running load data sequences can be performed, as shown in the following steps S22-S26.

[0116] S22. Calculate the distance between the pth target running load data sequence in a target running load data set and each initial cluster center, and divide the pth target running load data sequence to a target initial cluster center based on the distance between the pth target running load data sequence and each initial cluster center, wherein the target running load data set is the running load data set after all the initial cluster centers are deleted, and the target initial cluster center is the initial cluster center with the minimum distance to the pth target running load data sequence.

[0117] S23. Increment p by 1, and re-calculate the distance between the pth target operation load data sequence and each initial clustering center until p equals P, to obtain k initial operation load clusters, wherein the initial value of p is 1, and P is the total number of target operation load data sequences.

[0118] S24. Calculate the clustering center of each initial operation load cluster, and determine the clustering criterion function value based on the clustering center of each initial operation load cluster and the operation load data set; in this embodiment, for any initial operation load cluster, the mean of the sum of each data sequence in the any initial operation load cluster is taken as the clustering center of the any initial operation load cluster; meanwhile, the expression of the clustering criterion function is shown in the following formula (3).

[0119]

[0120] In the above formula (3), E represents the clustering criterion function, w i represents the ith operation load data sequence, C v represents the vth clustering center (i.e. each initial operation load cluster is determined at each iteration when the step S23 is reached, and thus k clustering centers are obtained, and the C v is the vth clustering center in the k clustering centers).

[0121] After the clustering criterion function value at this iteration is calculated based on the above formula (3), it can be compared with the preset clustering threshold value to determine whether the clustering is ended, as shown in the following steps S25 and S26.

[0122] S25. Determine whether the clustering criterion function value is less than or equal to the preset clustering threshold value; in this embodiment, the preset clustering threshold value can be specifically set according to actual use, which is not specifically limited here; meanwhile, if the clustering criterion function is less than or equal to the preset clustering threshold value, it indicates that the clustering is optimal, at this time, the above k initial operation load clusters can be taken as the operation load clusters and output; otherwise, it indicates that the clustering is not optimal, at this time, the following step S26 needs to be performed.

[0123] S26. If not, update the initial clustering center to the clustering center of each initial operation load cluster, and re-calculate the distance between the pth target operation load data sequence and each initial clustering center in the target operation load data set until the clustering criterion function value is less than or equal to the preset clustering threshold value, so as to take the k initial operation load clusters corresponding to the clustering criterion function value less than or equal to the preset clustering threshold value as the k operation load clusters.

[0124] Thus, through the aforementioned steps S21 to S26, several operating load data sequences can be divided into several clusters, so as to determine the weekly operating load change pattern of the target base station within a week based on the divided clusters, thereby providing data basis for the generation of subsequent energy-saving schemes for the target base station.

[0125] In this embodiment, after determining the clusters to which several operating load data sequences belong, it is necessary to select the optimal cluster from the divided clusters so as to determine the weekly operating load change pattern of the target base station within a week based on the optimal cluster and obtain the corresponding weekly operating load change trend information; the specific determination process is shown in step S3 below.

[0126] S3. From the k operational load clusters, select the optimal operational load cluster, and based on the optimal operational load cluster, determine the weekly operational load change trend information of the target base station. In this embodiment, for example, but not limited to, first calculating the profile coefficient of each operational load cluster, and then selecting the operational load cluster with the largest profile coefficient as the optimal operational load cluster; then, based on the profile coefficient of the optimal operational load cluster, determine the weekly operational load change pattern of the target base station; finally, based on the weekly operational load change pattern, select different methods to determine the weekly operational load change trend information.

[0127] In this embodiment, the process of calculating the profile coefficient of any operating load cluster is as follows: First, calculate the sample profile coefficient of each operating load data sequence in the any operating load cluster; then, sum all the sample profile coefficients and take the mean of the summation results to obtain the profile coefficient of the any operating load cluster.

[0128] Optionally, the formula for calculating the sample profile coefficient of the u-th operating load data sequence in any operating load cluster is shown in Equation (4) below.

[0129]

[0130] In equation (4) above, g(u) represents the sample profile coefficient of the u-th operating load data sequence in any operating load cluster, b(u) represents the minimum value of the average distance between the u-th operating load data sequence and the other operating load clusters, and a(u) is the average distance between the u-th operating load data sequence and the other operating load data sequences in any operating load cluster; wherein, the average distance between the u-th operating load data sequence and any other operating load cluster is the average distance between the u-th operating load data sequence and each data sequence in any other operating load cluster, and u=1,2,3,..,U, where U is the total number of data in any operating load cluster.

[0131] In the embodiment, after the profile coefficient of the optimal operation load cluster is determined, the weekly operation load change mode of the target base station and the weekly operation load change trend of the target base station can be identified according to the profile coefficient, as shown in the following steps S31-S34.

[0132] S31. Determine whether the profile coefficient of the optimal operation load cluster is greater than or equal to a preset threshold value; in the embodiment, the preset threshold value can be but is not limited to 0.65; wherein, if the profile coefficient of the optimal operation load cluster is greater than 0.65, it is determined that the weekly operation load change mode of the target base station is a regular operation load change trend (for example, the operation load is at a peak value from Monday to Friday, and is at a valley value on Saturday and Sunday); if the profile coefficient of the optimal operation load cluster is less than 0.65, and the average distance of the data in the optimal operation load cluster is less than or equal to 0.1, it is determined that the weekly operation load change mode of the target base station is a consistent operation load change trend (the operation load does not change significantly within a week, and the load is approximately the same); and if the profile coefficient does not satisfy the above two conditions, it is determined that the weekly operation load change mode of the target base station is an irregular operation load change trend.

[0133] When it is determined that the weekly operation load change mode of the target base station is a regular operation load change trend, the load change rule of the target base station can be analyzed according to the divided clusters, as shown in the following step S32.

[0134] S32. If yes, generate a weekly operation load change trend clustering diagram of the target base station based on the k operation load clusters; in the embodiment, the clustering diagram can be used to determine the operation load clustering of the target base station within a week, that is, the high-load operation day and the low-load operation day of the target base station within a week; and then, the high-load operation day and the low-load operation day obtained by the foregoing analysis can be used to generate the weekly operation load change trend information of the target base station, as shown in the following steps S33 and S34.

[0135] S33. Determine the high-load operation day and the low-load operation day of the target base station within a week according to the weekly operation load change trend clustering diagram.

[0136] S34. Use the high-load operation day and the low-load operation day to constitute the weekly operation load change trend information; in the embodiment, if the weekly operation load change mode of the target base station is a consistent operation load change trend, it is determined that the operation load of the target base station is consistent every day within a week, so that the weekly operation load change trend information of the target base station is determined to be that every day within a week is a high-load operation day; and if the weekly operation load change mode of the target base station is an irregular operation load change trend, no energy-saving operation is performed, and the process is ended.

[0137] Thus, by the aforementioned step S3 and its sub-steps, the application can determine the high-load operation day and the low-load operation day of the target base station in a week, and then, in combination with the daily operation load trend function set of the target base station, the operation day and the working time period of each operation day in a week, on which the target base station needs to perform energy-saving operation, can be determined; specifically, as shown in the following steps S4 and S5.

[0138] S4. According to the daily operation load trend function set, the daily energy-saving operation time period of the target base station in a week is determined; in the embodiment, for any daily operation load trend function, the load variation coefficient of the base station operation day corresponding to the daily operation load trend function can be calculated based on the daily operation load trend function; and then, based on the load variation coefficient, the daily energy-saving operation time period of the base station operation day is determined; optionally, the calculation formula of the load variation coefficient is as follows:

[0139] R = |f max (t) - |f min (t) | (5)

[0140] In the above formula (5), R is the load variation coefficient, f max (t) represents the maximum value of any daily operation load trend function, and f min (t) represents the minimum value of any daily operation load trend function.

[0141] At the same time, the continuous time period less than R / 4 in any daily operation load trend function is taken as the daily energy-saving operation time period of the target base station in the base station operation day; for example, assuming that any daily operation load trend function corresponds to Monday, and the continuous time period less than R / 4 in the daily operation load trend function is from 23:00 to 7:00; then, the daily energy-saving operation time period of the target base station corresponding to Monday is from 23:00 to 7:00; of course, the determination principle of the daily energy-saving operation time period of each base station operation day is the same as the aforementioned example, and will not be described here.

[0142] After obtaining the daily energy-saving operation time period of the target base station in a week, the energy-saving control instruction of the target base station can be generated in combination with the corresponding weekly operation load variation trend information, as shown in the following step S5.

[0143] S5. generating the energy-saving control instruction of the target base station according to the weekly operation load change trend information of the target base station and the daily energy-saving operation time period, wherein the energy-saving control instruction comprises the energy-saving operation day of the target base station in a week, the operation start time and the operation end time of each energy-saving operation day, and the energy-saving strategy of each energy-saving operation day, and the energy-saving strategy comprises symbol off, time slot off, channel off or device hibernation; in the embodiment, since the weekly operation load change trend information contains the high-load operation day or the high-load operation day and the low-load operation day of the target base station; then, different energy-saving strategies are provided for the target base station according to the type of the load operation day of the target base station.

[0144] In a specific application, if the weekly operation load change trend information contains the high-load operation day and the low-load operation day, two kinds of energy-saving strategies are set in a week, wherein the high-load operation day corresponds to one kind of energy-saving strategy, the low-load operation day corresponds to one kind of energy-saving strategy, and the energy-saving operation start time and the energy-saving operation end time of the target base station in the high-load operation day are the start time and the end time of the daily energy-saving operation time period corresponding to the high-load operation day; similarly, the energy-saving operation start time and the energy-saving operation end time of the target base station in the low-load operation day are the start time and the end time of the daily energy-saving operation time period corresponding to the low-load operation day.

[0145] Similarly, if the weekly operation load change trend information contains only the high-load operation day, the corresponding energy-saving strategy is only one kind, that is, the energy-saving operation day is Monday to Sunday, and the operation time period is the daily energy-saving operation time period of the high-load operation day.

[0146] For example, the weekly operation load change trend information of the target base station contains the high-load operation day and the low-load operation day, wherein the high-load operation day is Monday to Friday, the low-load operation day is Saturday to Sunday (the corresponding scene is an office building), the daily energy-saving operation time period of the high-load operation day is 0 to 7, and the daily energy-saving operation time period of the low-load operation day is 0 to 24; then the energy-saving control instruction of the target base station is: the energy-saving operation day 1: Monday to Friday, the operation start time of the energy-saving operation day is 0, the operation end time is 7; the energy-saving strategy is: strategy one (including symbol off, time slot off, channel off or device hibernation); the energy-saving operation day 2: Saturday to Sunday, the operation start time of the energy-saving operation day is 0, the operation end time is 24; the energy-saving strategy is: strategy two (symbol off, time slot off, channel off or device hibernation).

[0147] For example, the weekly operation load change trend information of the target base station includes high-load operation days (Monday to Sunday, i.e., the change trend is consistent), and the daily energy-saving operation time period of the high-load operation days is from 23:00 to 5:00; then, the energy-saving control instruction of the target base station is: the energy-saving operation days are 3: Monday to Sunday, the operation start time of the energy-saving operation days is 23:00, and the operation end time is 5:00; the energy-saving strategy is: strategy three (including symbol off, time slot off, channel off, or device hibernation).

[0148] For another example, the weekly operation load change trend information of the target base station includes high-load operation days and low-load operation days, wherein the high-load operation days are Saturday to Sunday, the low-load operation days are Monday to Friday (which corresponds to the scenario of a residence), and the daily energy-saving operation time period of the high-load operation days is from 1:00 to 8:00, and the daily energy-saving operation time period of the low-load operation days is from 9:00 to 6:00; then, the energy-saving control instruction of the target base station is: the energy-saving operation days are 4: Monday to Friday, the operation start time of the energy-saving operation days is 9:00, and the operation end time is 6:00; the energy-saving strategy is: strategy four (including symbol off, time slot off, channel off, or device hibernation); the energy-saving operation days are 5: Saturday to Sunday, the operation start time of the energy-saving operation days is 1:00, and the operation end time is 8:00; the energy-saving strategy is: strategy six (symbol off, time slot off, channel off, or device hibernation).

[0149] Of course, the generation process of the energy-saving control instruction of the target base station with different weekly operation load change trend information and daily energy-saving operation time periods is the same as the aforementioned example principle, and details are not repeated here.

[0150] In this embodiment, symbol off is two-dimensional scheduling, including time domain and frequency domain (such as resource block RB, bandwidth part (BWP), and serving cell), so it includes both scheduling symbol number and RB data, and the devices to be turned off are the reference PA (power amplifier) and part of the transceiver, which has no impact on the key performance of the network; time slot off is mainly to converge services to certain time slots, and the remaining time slots are not scheduled to achieve the purpose of turning off devices for energy saving; channel off means turning off part of the channels (channel value PA and the PA corresponding TRX (transceiver) and DPD (Digital Pre-Distortion device)) under a certain load to achieve the purpose of energy saving.

[0151] In addition, in this embodiment, any one of the aforementioned symbol off, time slot off, channel off, and device hibernation can be selected as the energy-saving strategy; of course, strategy combination can also be performed; specifically, the selection of the aforementioned energy-saving strategy is a commonly used technology for reference energy-saving strategy control, and the principle is not repeated here.

[0152] After the energy-saving control instruction of the target base station is generated, it can be sent to the target base station to realize the energy-saving operation of the base station, as shown in the following step S6.

[0153] S6. The energy-saving control instruction is sent to the target base station to make the target base station execute the energy-saving control instruction so as to complete the energy-saving control of the target base station after executing the energy-saving control instruction. In this embodiment, the target base station can operate according to the instruction to execute the corresponding energy-saving strategy in the determined energy-saving operation day and the corresponding time period, so as to finally achieve the purpose of reducing the energy consumption of the base station.

[0154] Therefore, through the energy-saving control method of the base station described in detail in the foregoing steps S1-S6, the application can identify the different load operation days in a week and the low load time period of each operation day in a week of the base station according to the historical business operation load data of the base station, and generate the energy-saving control scheme of the base station based on the same. In this way, the method can identify the working scenario of the base station and realize the customized energy-saving control of the base station based on the working scenario of the base station, so as to reduce the energy consumption of the base station while meeting the use demand of users and the business demand of the base station, and further realize the intelligent management of the energy consumption of the base station.

[0155] As shown in Figure 2 the second aspect of the embodiment provides a hardware device for implementing the energy-saving control method of the base station in the first aspect of the embodiment, which comprises:

[0156] a data acquisition unit configured to acquire a set of operation load data of a target base station in a historical time period, wherein the set of operation load data comprises a plurality of time-continuous operation load data sequences, and the time length of any operation load data sequence is 24 hours.

[0157] a data analysis unit configured to perform cluster analysis on the plurality of operation load data sequences in the set of operation load data to obtain k operation load clusters, and perform linear regression processing on the plurality of operation load data sequences to obtain a set of daily operation load trend functions of the target base station, wherein the set of daily operation load trend functions comprises a daily operation load trend function corresponding to each day in a week of the target base station, and k is a positive integer.

[0158] the data analysis unit is configured to select an optimal operation load cluster from the k operation load clusters, and determine the weekly operation load change trend information of the target base station based on the optimal operation load cluster.

[0159] the data analysis unit is configured to determine the daily operation energy-saving time period of the target base station in a week according to the set of daily operation load trend functions.

[0160] The energy-saving strategy generation unit is configured to generate an energy-saving control instruction of the target base station according to the weekly operation load change trend information of the target base station and daily operation energy-saving time periods, wherein the energy-saving control instruction comprises energy-saving operation days of the target base station in a week, an operation start time and an operation end time of each energy-saving operation day, and an energy-saving strategy of each energy-saving operation day, and the energy-saving strategy comprises symbol-off, time-slot-off, channel-off or device hibernation.

[0161] The sending unit is configured to send the energy-saving control instruction to the target base station, so that the target base station executes the energy-saving control instruction, so as to complete energy-saving control of the target base station after executing the energy-saving control instruction.

[0162] The working process, working details and technical effects of the device provided by the embodiment can be referred to the first aspect of the embodiment, and will not be repeated here.

[0163] As shown in Figure 3 The third aspect of the embodiment provides another energy-saving control device of a base station, taking an electronic device as an example, which comprises a memory, a processor and a transceiver connected in sequence, wherein the memory is configured to store a computer program, the transceiver is configured to transceive messages, and the processor is configured to read the computer program and execute the energy-saving control method of a base station as described in the first aspect of the embodiment.

[0164] For example, the memory can include, but is not limited to, a random access memory (RAM), a read-only memory (ROM), a flash memory, a first-in first-out memory (FIFO) and / or a first-in last-out memory (FILO) and the like; specifically, the processor can include one or more processing cores, such as a 4-core processor, an 8-core processor and the like. The processor can be implemented in at least one of a hardware form of a DSP (Digital Signal Processing), a FPGA (Field-Programmable Gate Array) and a PLA (Programmable Logic Array), and the processor can also include a main processor and a coprocessor. The main processor is a processor for processing data in an awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in a standby state.

[0165] In some embodiments, the processor can be integrated with a GPU (Graphics Processing Unit) that is responsible for rendering and drawing the content required to be displayed on the display screen, for example, the processor can not be limited to a microprocessor of STM32F105 series, a RISC (reduced instruction set computer) microprocessor, an X86 architecture processor, or a processor integrated with an embedded NPU (neural-network processing units); the transceiver can be but not limited to a WIFI transceiver, a Bluetooth transceiver, a GPRS transceiver, a ZigBee transceiver, a 3G transceiver, a 4G transceiver, and / or a 5G transceiver, etc. In addition, the device can further include but not limited to a power module, a display screen, and other necessary components.

[0166] The working process, working details and technical effects of the electronic device provided in the embodiment can be referred to the first aspect of the embodiment, and will not be repeated here.

[0167] The fourth aspect of the embodiment provides a storage medium storing instructions of the energy saving control method of the base station in the first aspect of the embodiment, that is, the storage medium stores instructions, and when the instructions run on a computer, the energy saving control method of the base station in the first aspect of the embodiment is executed.

[0168] The storage medium refers to a carrier for storing data, which can include but is not limited to floppy disks, optical disks, hard disks, flash memories, USB flash disks, and / or Memory Sticks, etc., and the computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0169] The working process, working details and technical effects of the storage medium provided in the embodiment can be referred to the first aspect of the embodiment, and will not be repeated here.

[0170] The fifth aspect of the embodiment provides a computer program product containing instructions, which, when running on a computer, causes the computer to execute the energy saving control method of the base station in the first aspect of the embodiment, wherein the computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0171] Finally, it should be noted that the above description is only the preferred embodiment of the present application, and is not intended to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A power saving control method of a base station, characterized by, The method comprises the following steps: acquiring a running load data set of a target base station in a historical time period, wherein the running load data set comprises a plurality of time-continuous running load data sequences, and any running load data sequence corresponds to a time length of 24 hours; performing cluster analysis on the plurality of running load data sequences in the running load data set to obtain k running load clusters, and performing linear regression processing on the plurality of running load data sequences to obtain a daily running load trend function set of the target base station, wherein the daily running load trend function set comprises a daily running load trend function corresponding to each day in a week of the target base station, and k is a positive integer; selecting an optimal running load cluster from the k running load clusters, and determining a weekly running load change trend information of the target base station based on the optimal running load cluster; determining daily running energy-saving time periods of the target base station in a week according to the daily running load trend function set; generating an energy-saving control instruction of the target base station according to the weekly running load change trend information and the daily running energy-saving time periods of the target base station, wherein the energy-saving control instruction comprises energy-saving running days of the target base station in a week, a running start time and a running end time of each energy-saving running day, and an energy-saving strategy of each energy-saving running day, and the energy-saving strategy comprises symbol-off, time-slot-off, channel-off or device hibernation; sending the energy-saving control instruction to the target base station, so that the target base station executes the energy-saving control instruction to complete energy-saving control of the target base station after executing the energy-saving control instruction; performing cluster analysis on the plurality of running load data sequences in the running load data set to obtain k running load clusters, comprising: calculating a first cluster density of each running load data sequence in the running load data set and a first average cluster density of the running load data set, and determining k running load data sequences as initial cluster centers based on the first average cluster density and the first cluster density of each running load data sequence; calculating distances between a pth target running load data sequence in a target running load data set and each initial cluster center, and dividing the pth target running load data sequence to a target initial cluster center based on the distances between the pth target running load data sequence and each initial cluster center, wherein the target running load data set is the running load data set after deleting all initial cluster centers, and the target initial cluster center is an initial cluster center with the minimum distance to the pth target running load data sequence; increasing p by 1 and recalculating the distances between the pth target running load data sequence and each initial cluster center until p is equal to P to obtain k initial running load clusters, wherein the initial value of p is 1, and P is the total number of target running load data sequences; calculating a cluster center of each initial operation load cluster, and determining a cluster criterion function value based on the cluster center of each initial operation load cluster and the operation load data set; determining whether the cluster criterion function value is less than or equal to a preset cluster threshold value; if not, updating the initial cluster center to the cluster center of each initial operation load cluster, and recalculating the distance between the pth target operation load data sequence in the target operation load data set and each initial cluster center until the cluster criterion function value is less than or equal to the preset cluster threshold value, so that the k initial operation load clusters corresponding to the cluster criterion function value less than or equal to the preset cluster threshold value are taken as the k operation load clusters; calculating a first cluster density of each operation load data sequence in the operation load data set, comprising: calculating the Euclidean distance between each pair of operation load data sequences in the operation load data set, and summing all Euclidean distances to obtain a Euclidean distance summary value; determining an average Euclidean distance of the operation load data set using the Euclidean distance summary value; for the ith operation load data sequence in the operation load data set, calculating a first cluster density of the ith operation load data sequence based on the average Euclidean distance and the Euclidean distance between the ith operation load data sequence and each operation load data sequence; increasing i by 1 and recalculating the first cluster density of the ith operation load data sequence based on the average Euclidean distance and the Euclidean distance between the ith operation load data sequence and each operation load data sequence until i is equal to n, wherein the initial value of i is 1 and n is the total number of operation load data sequences; determining an average Euclidean distance of the operation load data set using the Euclidean distance summary value, comprising: using the following formula (1) to calculate the average Euclidean distance of the operation load data set; In the above formula (1), represents the average Euclidean distance, s represents the Euclidean distance aggregate value, represents the number of sets composed of any two of the n operation load data sequences, and n represents the total number of operation load data sequences; correspondingly, calculating the first cluster density of the ith operation load data sequence based on the average Euclidean distance and the Euclidean distance between the ith operation load data sequence and each operation load data sequence, comprising: using the following formula (2) to calculate the first cluster density of the ith operation load data sequence; In the above formula (2), w i represents the i-th running load data sequence, m(w i ) represents the first clustering density of the i-th running load data sequence, s(w i , w j ) represents the Euclidean distance between the i-th running load data sequence and the j-th running load data sequence, represents the density function, and determining k operation load data sequences as initial cluster centers based on the first average cluster density and the first cluster density of each operation load data sequence, comprising: determining a density attribute of each operation load data sequence based on the first average cluster density and the first cluster density of each operation load data sequence, wherein the density attribute includes a density sparse point or a density dense point; extracting operation load data sequences with the density attribute of the density sparse point from the operation load data set to form a sparse data set; calculating a second clustering density of each of the running load data sequences in the sparse data set and a second average clustering density of the sparse data set, and extracting a second clustering density greater than the second average clustering density from the second clustering densities to form a density set using the extracted second clustering densities; determining whether the number of the maximum second clustering densities in the density set is equal to 1; if yes, taking the running load data sequence corresponding to the maximum second clustering density as an initial clustering center; deleting a target second clustering density from the density set to obtain an updated density set, wherein the Euclidean distance between the running load data sequence corresponding to the target second clustering density and specified data is less than the average Euclidean distance of the running load data set, and the specified data is the remaining running load data sequence after deleting the running load data sequence corresponding to the target second clustering density from the running load data set; calculating the second average clustering density of the updated density set, and reextracting a second clustering density greater than the second average clustering density from the second clustering densities of the density set until k initial clustering centers are obtained.

2. The method of claim 1, wherein, if the number of the maximum second clustering densities in the density set is greater than 1, the method further comprises: forming a candidate clustering center set using the running load data sequences corresponding to the maximum second clustering densities; for the hth running load data sequence in the candidate clustering center set, calculating the Euclidean distance between the hth running load data sequence in the candidate clustering center set and each running load data sequence in the candidate clustering center set, and summing the Euclidean distances between the hth running load data sequence in the candidate clustering center set and each running load data sequence in the candidate clustering center set to obtain a clustering center value of the hth running load data sequence in the candidate clustering center set; increasing h by 1 until h is equal to H to obtain the clustering center value of each running load data sequence in the candidate clustering center set, wherein the initial value of h is 1 and H is the total number of data sequences in the candidate clustering center set; taking the running load data sequence corresponding to the minimum clustering center value in the candidate clustering center set as an initial clustering center.

3. The method of claim 1, wherein, selecting an optimal running load cluster from the k running load clusters, comprising: calculating the silhouette coefficient of each running load cluster, and selecting the running load cluster with the maximum silhouette coefficient as the optimal running load cluster; correspondingly, determining the weekly running load change trend information of the target base station based on the optimal running load cluster, comprising: determining whether the silhouette coefficient of the optimal running load cluster is greater than or equal to a preset threshold value; if yes, generating a weekly running load change trend clustering map of the target base station based on the k running load clusters; determining the high-load running day and the low-load running day of the target base station within a week according to the weekly running load change trend clustering map; forming the weekly running load change trend information using the high-load running day and the low-load running day.

4. An energy saving control device of a base station, characterized by comprising: The energy-saving control method for the base station of any one of claims 1-3, and the device comprises: a data acquisition unit configured to acquire a set of operation load data of the target base station in a historical time period, wherein the set of operation load data comprises a plurality of time-continuous operation load data sequences, and any operation load data sequence corresponds to a time length of 24 hours; a data analysis unit configured to perform cluster analysis on the plurality of operation load data sequences in the set of operation load data to obtain k operation load clusters, and perform linear regression processing on the plurality of operation load data sequences to obtain a set of daily operation load trend functions of the target base station, wherein the set of daily operation load trend functions comprises a daily operation load trend function corresponding to each day of the target base station in a week, and k is a positive integer; the data analysis unit is configured to select an optimal operation load cluster from the k operation load clusters, and determine a weekly operation load change trend information of the target base station based on the optimal operation load cluster; the data analysis unit is configured to determine a daily energy-saving time period of the target base station in a week according to the set of daily operation load trend functions; an energy-saving strategy generation unit configured to generate an energy-saving control instruction of the target base station according to the weekly operation load change trend information and the daily energy-saving time period of the target base station, wherein the energy-saving control instruction comprises energy-saving operation days of the target base station in a week, a start time and an end time of each energy-saving operation day, and an energy-saving strategy of each energy-saving operation day, and the energy-saving strategy comprises symbol-off, time-slot-off, channel-off or device hibernation; a sending unit configured to send the energy-saving control instruction to the target base station, so that the target base station executes the energy-saving control instruction to complete energy-saving control of the target base station after executing the energy-saving control instruction.

5. An electronic device, comprising: comprises: a memory, a processor and a transceiver connected in sequence, wherein the memory is configured to store a computer program, the transceiver is configured to transceive messages, and the processor is configured to read the computer program and execute the energy-saving control method of the base station of any one of claims 1-3.

6. A storage medium, characterized by The storage medium has instructions stored thereon, and when the instructions run on a computer, the energy-saving control method of the base station of any one of claims 1-3 is executed.

Images