CN114339967A - Method and device for predicting base station traffic - Google Patents
Method and device for predicting base station traffic Download PDFInfo
- Publication number
- CN114339967A CN114339967A CN202111605305.1A CN202111605305A CN114339967A CN 114339967 A CN114339967 A CN 114339967A CN 202111605305 A CN202111605305 A CN 202111605305A CN 114339967 A CN114339967 A CN 114339967A
- Authority
- CN
- China
- Prior art keywords
- energy
- saving
- traffic
- period
- base station
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 62
- 238000009826 distribution Methods 0.000 claims abstract description 85
- 238000012549 training Methods 0.000 claims description 25
- 238000004590 computer program Methods 0.000 claims description 16
- 230000001186 cumulative effect Effects 0.000 claims description 15
- 239000011159 matrix material Substances 0.000 claims description 10
- 101150071746 Pbsn gene Proteins 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 5
- 230000000694 effects Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 7
- 230000005540 biological transmission Effects 0.000 description 6
- 230000002159 abnormal effect Effects 0.000 description 3
- 238000007418 data mining Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Abstract
The embodiment of the invention provides a method and a device for predicting base station traffic. The method comprises the following steps: predicting the accumulated value of the traffic of all energy-saving periods and the probability distribution of the traffic of all energy-saving periods in a target energy-saving period of the base station; and predicting the traffic of each energy-saving period in the target period of the base station according to the accumulated value of the traffic of all the energy-saving periods in the target energy-saving period and the probability distribution of the traffic. The invention can solve the problem that the prediction accuracy and the robustness can not be ensured simultaneously when the multi-step time sequence is adopted to predict the base station traffic in the related technology, and achieves the effect of ensuring the prediction accuracy and the robustness simultaneously.
Description
Technical Field
The embodiment of the invention relates to the technical field of data mining, in particular to a method and a device for predicting base station traffic.
Background
In the initial development stage of the 5G service, the network load is low, the proportion of light-load or no-load base stations is very high, and the 5G base stations adopting a large-scale Multiple Input Multiple Output (Massive MIMO) technology further increase the power consumption and bring higher electricity cost pressure. At present, implementing deep sleep aiming at light load or no-load time periods becomes a main strategy for saving energy of a base station, and a data mining algorithm is required to accurately predict the service volume index of the base station in continuous time periods.
The traffic prediction of the energy-saving period of the current base station mainly adopts two schemes: 1. selecting an energy-saving time interval of the next energy-saving period by adopting a statistical analysis method based on a Key Performance Indicator (KPI) of a low-delay historical base station, and predicting the traffic volume of the energy-saving time interval; 2. and (3) adopting prediction algorithms such as data mining or deep learning, counting the external influence data (such as weather, holiday information, geographical position information and the like) superimposed by the low/medium/low/high time delay related KPI indexes, and adopting a single-step time sequence method to predict the traffic of a plurality of continuous time intervals in the next energy-saving period.
The first method has high requirement on time delay, cannot consider index change trend and amplitude by means of a traditional statistical method, and often has large deviation from an actual result; the second method can only carry out multiple predictions on multiple time intervals in a period, wastes a large amount of computing resources and increases prediction time, and moreover, index prediction results of the multiple time intervals are not associated and restricted, so that the overall trend is greatly different from the actual trend.
In the related art, the multi-step timing prediction problem needs to predict indexes in a plurality of future time periods, so two schemes of setting a time sliding window equal to step modeling and adding a prediction result into a historical sequence to be fitted again are generally adopted. The first scheme abandons the characteristics of the latest t-1 period, possibly reduces the accuracy of the model but can enhance the robustness, and the second scheme increases the accuracy of the period at a close distance and introduces the deviation of the prediction result, so that the model robustness is reduced.
Disclosure of Invention
The embodiment of the invention provides a method and a device for predicting base station traffic, which are used for at least solving the problem that the prediction accuracy and the robustness cannot be simultaneously ensured when the base station traffic is predicted by adopting a multi-step time sequence in the related technology.
According to an embodiment of the present invention, a method for predicting traffic of a base station is provided, including: predicting the accumulated value of the traffic of all energy-saving periods and the probability distribution of the traffic of all energy-saving periods in a target energy-saving period of the base station; and predicting the traffic of each energy-saving period in the target period of the base station according to the accumulated value of the traffic of all the energy-saving periods in the target energy-saving period and the probability distribution of the traffic.
In an exemplary embodiment, before predicting an accumulated value of traffic volumes of all energy saving periods and a probability distribution of traffic volumes of all energy saving periods within a target energy saving cycle of a base station, the method further includes: and determining a target energy-saving period to be predicted, an energy-saving time period and a traffic index of the base station, wherein the traffic is the utilization rate of the PRB.
In an exemplary embodiment, after determining the target energy saving cycle, the energy saving period and the traffic index to be predicted by the base station, the method further includes: determining a model input characteristic data set according to a target energy-saving cycle, an energy-saving time period and a traffic index, wherein the model input characteristic data set comprises: traffic timing trend indicator data, base station KPI trend indicator data, and external indicator data.
In one exemplary embodiment, the traffic timing trend indicator data includes: the average PRB utilization rate, the lowest PRB utilization rate and the highest PRB utilization rate of each energy-saving time interval in a plurality of energy-saving periods, and the highest PRB utilization rate, the average PRB utilization rate and the lowest PRB utilization rate of each energy-saving period; the base station KPI trend index data comprises: the average number of RRC users, the minimum number of RRC users and the maximum number of RRC users in each energy-saving period in a plurality of energy-saving periods, and the maximum number of RRC users, the average number of RRC users and the minimum number of RRC users in each energy-saving period; the external index data includes: time characteristics and regional characteristics, wherein the time characteristics at least comprise one of the following: the year, month and date of the energy-saving time period, whether the week is weekend or not, whether the holiday is festival or not and whether the hour is hour or not; the geographic features include at least one of: the province where the base station is located, the city of land, longitude and latitude information, the ID of town, suburb and base station cell.
In an exemplary embodiment, after determining the model input feature data set according to the target energy saving cycle, the energy saving period and the traffic index, the method further includes: selecting traffic time sequence trend index data, base station KPI trend index data and other external index data of the base station as a basic feature data set by taking an energy-saving period as a unit, wherein a field set of the basic feature data set is [ X1 ]; adding a predicted value Y1 in the basic characteristic data set to generate an accumulated quantity model data set, wherein the predicted value Y1 is the sum of the utilization rates of PRBs in each energy-saving period of the next energy-saving period, and the field set of the accumulated quantity model data set is [ X1, Y1 ]; adding a probability distribution W1 of PRB utilization rate of each energy-saving period of an energy-saving cycle and the number Y2 of energy-saving periods to be predicted in the basic characteristic data set to generate a probability distribution model data set, wherein the field set of the probability distribution model data set is [ X1, W1, Y2 ].
In an exemplary embodiment, the cumulative model data set and the probability distribution model data set are input into a LightGBM model for training, and a cumulative model for predicting PRB utilization and a probability distribution model for PRB utilization are obtained respectively.
In an exemplary embodiment, inputting the cumulative model dataset into a LightGBM model for training to obtain a cumulative model for predicting PRB utilization includes: and inputting the accumulative quantity model data set into a lightGBM model by adopting a regression mode to train so as to obtain an accumulative quantity model for predicting the PRB utilization rate, wherein the field set X1 of the basic characteristic data set is used as a characteristic, and the sum Y1 of the PRB utilization rates of each energy-saving period in the next energy-saving period is used as a predicted value.
In an exemplary embodiment, inputting the probability distribution model dataset into a LightGBM model for training to obtain a probability distribution model for predicting PRB utilization includes: and inputting the probability distribution model data set into a lightGBM model for training by adopting a multi-classification mode by taking a field set X1 of a basic characteristic data set as a characteristic, taking the probability distribution W1 of the PRB utilization rate of each energy-saving time interval as a weight, and taking the number Y2 of the energy-saving time intervals to be predicted as a predicted value to obtain the probability distribution model for predicting the PRB utilization rate.
In one exemplary embodiment, further comprising: predicting and obtaining an accumulated value of PRB utilization rates of all energy-saving time periods in a target energy-saving period based on a trained accumulated model; and predicting and obtaining the probability distribution of the PRB utilization rate of all energy-saving time periods in the target energy-saving period based on the trained probability distribution model.
In one exemplary embodiment, further comprising: inputting the accumulated value of the PRB utilization rates of all energy-saving periods in the target energy-saving period into the probability distribution model by taking 1/N as weight, and acquiring a PRB probability distribution matrix of each energy-saving period in the target energy-saving period, wherein N is the number of the energy-saving periods; and factoring the accumulated value of the PRB utilization rates of all the energy-saving periods in the target energy-saving period with the probability distribution of the PRB utilization rate of each energy-saving period to determine the PRB utilization rate of each energy-saving period in the target energy-saving period.
According to another embodiment of the present invention, there is provided an apparatus for predicting traffic of a base station, including: the first prediction module is used for predicting the accumulated value of the traffic of all energy-saving time intervals and the probability distribution of the traffic of all energy-saving time intervals in the target energy-saving period of the base station; and the second prediction module is used for predicting the traffic of each energy-saving time period in the target period of the base station according to the accumulated value of the traffic of all the energy-saving time periods in the target energy-saving period and the probability distribution of the traffic.
In an exemplary embodiment, the determining module is configured to determine a target energy saving period, an energy saving period, and a traffic indicator to be predicted by the base station, where the traffic is a PRB utilization.
In one exemplary embodiment, the determining module includes: a determining unit, configured to determine a model input feature data set according to a target energy saving cycle, an energy saving time period, and a traffic index, where the model input feature data set includes: traffic timing trend indicator data, base station KPI trend indicator data, and external indicator data.
According to a further embodiment of the present invention, there is also provided a computer-readable storage medium having a computer program stored thereon, wherein the computer program is arranged to perform the steps of any of the above method embodiments when executed.
According to yet another embodiment of the present invention, there is also provided an electronic device, including a memory in which a computer program is stored and a processor configured to execute the computer program to perform the steps in any of the above method embodiments.
According to the invention, the traffic of each energy-saving time period can be predicted more accurately based on the accumulated value of the traffic of all energy-saving time periods and the probability distribution of the traffic in the target energy-saving period, and meanwhile, the result prediction is carried out based on the probability distribution of the traffic indexes of each time period, so that the probability distribution of each time period has constraint, the robustness is higher, the abnormal value of a certain time period is not easy to appear, and the accuracy of the prediction result is improved. Therefore, the problem that the prediction accuracy and the robustness cannot be ensured simultaneously when the multi-step time sequence is adopted to predict the base station traffic in the related technology can be solved, and the effect of ensuring the prediction accuracy and the robustness simultaneously can be achieved.
Drawings
Fig. 1 is a block diagram of a hardware configuration of a computer terminal that operates a base station traffic prediction method;
fig. 2 is a flowchart of a base station traffic prediction method according to an embodiment of the present invention;
fig. 3 is a block diagram of a base station traffic prediction apparatus according to an embodiment of the present invention;
fig. 4 is a block diagram of a base station traffic prediction apparatus according to another embodiment of the present invention;
FIG. 5 is a flow diagram of a method for predicting base station traffic based on multi-step timing probability distributions in accordance with an embodiment of the present invention;
fig. 6 is a flow chart of a method of base station traffic incorporating embodiments of the present invention.
Detailed Description
Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings in conjunction with the embodiments.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
The method embodiments provided in the embodiments of the present application may be executed in a mobile terminal, a computer terminal, or a similar computing device. Taking the operation on a computer terminal as an example, fig. 1 is a hardware structure block diagram of a computer terminal operating the base station traffic prediction method according to the embodiment of the present invention. As shown in fig. 1, the computer terminal may include one or more processors 102 (only one is shown in fig. 1) (the processor 102 may include, but is not limited to, a Processing device such as a Microprocessor (MCU) or a Programmable logic device (FPGA)) and a memory 104 for storing data, wherein the computer terminal may further include a transmission device 106 for communication function and an input/output device 108. It will be understood by those skilled in the art that the structure shown in fig. 1 is only an illustration and is not intended to limit the structure of the computer terminal. For example, the computer terminal may also include more or fewer components than shown in FIG. 1, or have a different configuration than shown in FIG. 1.
The memory 104 may be used to store a computer program, for example, a software program and a module of an application software, such as a computer program corresponding to the method for predicting the traffic of the base station in the embodiment of the present invention, and the processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, so as to implement the method described above. The memory 104 may include high speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include memory located remotely from the processor 102, which may be connected to a computer terminal over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The transmission device 106 is used for receiving or transmitting data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider of the computer terminal. In one example, the transmission device 106 includes a Network adapter (NIC), which can be connected to other Network devices through a base station so as to communicate with the internet. In one example, the transmission device 106 may be a Radio Frequency (RF) module, which is used for communicating with the internet in a wireless manner.
In this embodiment, a method for predicting traffic of a base station operating on the computer terminal is provided, and fig. 2 is a flowchart of the method for predicting traffic of a base station according to the embodiment of the present invention, as shown in fig. 2, the flowchart includes the following steps:
step S202, predicting the accumulated value of the traffic of all energy-saving time periods and the probability distribution of the traffic of all energy-saving time periods in the target energy-saving period of the base station;
step S204, predicting the traffic volume of each energy-saving time interval in the target period of the base station according to the accumulated value of the traffic volumes of all the energy-saving time intervals in the target energy-saving period and the probability distribution of the traffic volumes.
Before step S202 in this embodiment, the method may further include: and determining a target energy-saving period to be predicted, an energy-saving time period and a traffic index of the base station, wherein the traffic is the utilization rate of the PRB.
In an embodiment, after determining the target energy saving cycle, the energy saving period, and the traffic index to be predicted by the base station, the method may further include: determining a model input characteristic data set according to a target energy-saving cycle, an energy-saving time period and a traffic index, wherein the model input characteristic data set comprises: traffic timing trend indicator data, base station KPI trend indicator data, and external indicator data.
In this embodiment, the traffic time series trend index data includes: the average PRB utilization rate, the lowest PRB utilization rate and the highest PRB utilization rate of each energy-saving time interval in a plurality of energy-saving periods, and the highest PRB utilization rate, the average PRB utilization rate and the lowest PRB utilization rate of each energy-saving period; the base station KPI trend index data comprises: the average number of RRC users, the minimum number of RRC users and the maximum number of RRC users in each energy-saving period in a plurality of energy-saving periods, and the maximum number of RRC users, the average number of RRC users and the minimum number of RRC users in each energy-saving period; the external index data includes: time characteristics and regional characteristics, wherein the time characteristics at least comprise one of the following: the year, month and date of the energy-saving time period, whether the week is weekend or not, whether the holiday is festival or not and whether the hour is hour or not; the geographic features include at least one of: the province where the base station is located, the city of land, longitude and latitude information, the ID of town, suburb and base station cell.
In an exemplary embodiment, after determining the model input feature data set according to the target energy saving cycle, the energy saving period, and the traffic index, the method may further include: selecting traffic time sequence trend index data, base station KPI trend index data and other external index data of the base station as a basic feature data set by taking an energy-saving period as a unit, wherein a field set of the basic feature data set is [ X1 ]; adding a predicted value Y1 in the basic characteristic data set to generate an accumulated quantity model data set, wherein the predicted value Y1 is the sum of the utilization rates of PRBs in each energy-saving period of the next energy-saving period, and the field set of the accumulated quantity model data set is [ X1, Y1 ]; adding a probability distribution W1 of PRB utilization rate of each energy-saving period of an energy-saving cycle and the number Y2 of energy-saving periods to be predicted in the basic characteristic data set to generate a probability distribution model data set, wherein the field set of the probability distribution model data set is [ X1, W1, Y2 ].
In an exemplary embodiment, the method may further include: and inputting the accumulative quantity model data set and the probability distribution model data set into a LightGBM model for training to respectively obtain an accumulative quantity model for predicting the PRB utilization rate and a probability distribution model for predicting the PRB utilization rate.
In an exemplary embodiment, inputting the cumulative model dataset into a LightGBM model for training to obtain a cumulative model for predicting PRB utilization may further include: and inputting the accumulative quantity model data set into a lightGBM model by adopting a regression mode to train so as to obtain an accumulative quantity model for predicting the PRB utilization rate, wherein the field set X1 of the basic characteristic data set is used as a characteristic, and the sum Y1 of the PRB utilization rates of each energy-saving period in the next energy-saving period is used as a predicted value.
In an exemplary embodiment, inputting the probability distribution model dataset into the LightGBM model for training to obtain the probability distribution model for predicting the PRB utilization may further include: and inputting the probability distribution model data set into a lightGBM model for training by adopting a multi-classification mode by taking a field set X1 of a basic characteristic data set as a characteristic, taking the probability distribution W1 of the PRB utilization rate of each energy-saving time interval as a weight, and taking the number Y2 of the energy-saving time intervals to be predicted as a predicted value to obtain the probability distribution model for predicting the PRB utilization rate.
In an exemplary embodiment, the method may further include: predicting and obtaining an accumulated value of PRB utilization rates of all energy-saving time periods in a target energy-saving period based on a trained accumulated model; and predicting and obtaining the probability distribution of the PRB utilization rate of all energy-saving time periods in the target energy-saving period based on the trained probability distribution model.
In an exemplary embodiment, the method may further include: inputting the accumulated value of the PRB utilization rates of all energy-saving periods in the target energy-saving period into the probability distribution model by taking 1/N as weight, and acquiring a PRB probability distribution matrix of each energy-saving period in the target energy-saving period, wherein N is the number of the energy-saving periods; and factoring the accumulated value of the PRB utilization rates of all the energy-saving periods in the target energy-saving period with the probability distribution of the PRB utilization rate of each energy-saving period to determine the PRB utilization rate of each energy-saving period in the target energy-saving period.
Through the steps, the same data set can be adopted as that of the traditional single-step time-interval model in the embodiment, only two model training processes are needed, and the existing prediction method needs to train for 24 times respectively in different time intervals, so that the resource of model calculation is greatly reduced. Meanwhile, probability distribution restriction is added among the prediction results in a plurality of time periods, robustness is improved, and abnormal values in a certain time period are not easy to occur. Therefore, the problem that the prediction accuracy and the robustness cannot be ensured simultaneously when the multi-step time sequence is adopted to predict the base station traffic in the related technology can be solved, and the effect of ensuring the prediction accuracy and the robustness simultaneously can be achieved.
Through the above description of the embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but the former is a better implementation mode in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., a Read-Only Memory/Random Access Memory (ROM/RAM), a magnetic disk, an optical disk) and includes instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, or a network device) to execute the method according to the embodiments of the present invention.
In this embodiment, a device for predicting traffic volume of a base station is also provided, and the device is used to implement the foregoing embodiments and preferred embodiments, and details are not repeated for what has been described. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.
Fig. 3 is a block diagram of a prediction apparatus of base station traffic according to an embodiment of the present invention, as shown in fig. 3, the apparatus includes: a first prediction module 10 and a second prediction module 20.
A first prediction module 10, configured to predict an accumulated value of traffic volumes in all energy-saving periods and a probability distribution of traffic volumes in all energy-saving periods within a target energy-saving cycle of a base station;
a second predicting module 20, configured to predict the traffic volume of each energy-saving period in the target period of the base station according to the aggregate value of the traffic volumes of all energy-saving periods in the target energy-saving period and the probability distribution of the traffic volumes.
Fig. 4 is a block diagram of a prediction apparatus of base station traffic according to an embodiment of the present invention, and as shown in fig. 4, the apparatus includes a determination module 30 in addition to all modules shown in fig. 3.
A determining module 30, configured to determine a target energy saving period, an energy saving time period, and a traffic indicator to be predicted by the base station, where the traffic is a PRB utilization rate.
In this embodiment, the determining module 30 further comprises a determining unit 31.
A determining unit 31, configured to determine a model input feature data set according to a target energy saving cycle, an energy saving time period, and a traffic index, where the model input feature data set includes: traffic timing trend indicator data, base station KPI trend indicator data, and external indicator data.
It should be noted that, the above modules may be implemented by software or hardware, and for the latter, the following may be implemented, but not limited to: the modules are all positioned in the same processor; alternatively, the modules are respectively located in different processors in any combination.
In order to facilitate understanding of the technical solutions provided by the present invention, the following detailed description will be made with reference to embodiments of specific scenarios.
The embodiment of the invention provides a method for predicting base station traffic based on multi-step time sequence probability distribution. The method comprises the following steps: and converting the multi-step time sequence prediction of all the periods of the traffic into two combined predictions, and predicting the accumulated value of the traffic indexes of all the periods in the energy-saving period and the traffic distribution probability of all the periods in the period.
By the method provided by the embodiment of the invention, the calculation resources of the traditional multi-step time sequence prediction model in the aspects of model characteristic generation and model training prediction can be saved, the model operation time is shortened, and the correlation constraint among multi-period prediction results is promoted so as to improve the accuracy. Meanwhile, due to total amount constraint among a plurality of time periods, the prediction abnormal value is reduced, and the robustness of the model is improved.
Fig. 5 is a flowchart of a method for predicting base station traffic based on multi-step timing probability distribution according to an embodiment of the present invention, as shown in fig. 5, the method includes the following steps:
step S501, service quantity indexes, energy-saving cycles and energy-saving time periods predicted in the energy-saving time periods of the base stations are determined.
Step S502, determining indexes of different input characteristics of the model, and generating a data set of the input characteristics of the model.
In particular, the feature data may contain three types of data: generating a traffic time sequence trend index, generating other KPI trend indexes related to the base station and other external data indexes:
in step S503, a cumulative quantity model data set and a probability distribution model data set are generated.
Specifically, a base signature DATASET is generated, a cumulative model DATASET DATASET1 is generated based on the base signature DATASET, and the field sets are [ X1, Y1]
And generating a probability distribution model data set, and generating a data set DATASET2 by recording the number of times of repeating the basic characteristic data set [ X1] to a time period needing prediction every line. The database 2 adds the time period to be predicted Y2 corresponding to each row record as the prediction index, and generates a probability distribution model data set database 3 with field sets of [ X1, W1, Y2 ].
And step S504, training a model by using LIGHT TGBM.
Specifically, training the cumulant model by using the light tgbm includes: the data set DATASET1 is characterized by X1 and Y1 as a predicted value, and the training set is input into a LIGHT TGBM model for training in a regression mode.
Training a probability distribution model by using LIGHT TGBM, comprising: the data set DATASET2 is characterized by X1, W1 as weight, Y2 as predicted value, and the training set is input into LIGHT TGBM model for training, and is trained by multi-classification mode, and the probability matrix with the output result as class is set.
Step S505, integrate the prediction results.
Specifically, firstly, predicting a business index value cumulant P _ sum through a cumulant model; setting weight, and generating a predicted value probability distribution matrix P _ prb of each time interval by using a probability distribution model; and finally, factoring the accumulated quantity of the predicted value with the PRB probability distribution matrix P _ PRB of each time interval to obtain the predicted value of all time intervals in one day.
An embodiment of the present invention further provides a method for base station traffic in combination with the specific example, and fig. 6 is a flowchart of the method for base station traffic according to the embodiment of the present invention, as shown in fig. 6, the method includes the following steps:
step S601, determining an energy-saving cycle, an energy-saving time period and a traffic index.
Specifically, a traffic index, an energy-saving cycle and an energy-saving time period predicted in an energy-saving time period of a base station are determined; a data set of model input features is determined.
In the embodiment, the utilization rate of the PRB is taken as a target traffic index, the energy-saving cycle is one day, and the energy-saving time interval is 24 hours; the input feature data contains three types of data: traffic timing trend indicators, base station related KPI trend indicators, and other external data indicators;
the traffic time sequence trend index comprises the following steps: average PRB utilization, lowest PRB utilization, highest PRB utilization per hour in approximately 7 days; the highest PRB utilization rate, the average PRB utilization rate and the lowest PRB utilization rate in each day of 7 days;
the base station related KPI trend indexes comprise: average number of RRC users, minimum number of RRC users, maximum number of RRC users per hour in 7 days; the highest RRC user number, the average RRC user number and the lowest RRC user number in each day of 7 days;
other external data metrics include: time characteristics: the year, month, date, whether weekend, whether holiday, hour of the period; regional characteristics: the province, city, longitude and latitude information, town, suburb, and ID characteristics of the base station cell.
In step S602, a basic feature data set DATASET0 is generated.
Specifically, the traffic time sequence trend index, the base station related KPI trend index and other external data indexes of each base station cell in the current day are calculated as a basic feature set DATASET0 in units of days, and the field set of the basic feature is [ X1 ].
In this embodiment, the method further includes: generating a cumulative model DATASET DATASET 1:
specifically, the sum of the PRB utilization rates in the next 24-hour period added after each row of the basic feature data set is recorded as Y1, and a cumulative model data set DATASET1 is generated, wherein the field set is [ X1, Y1 ].
In this embodiment, the method further includes: generating a probability distribution model DATASET DATASET 3:
specifically, the base signature DATASET [ X1] was repeated 24 times per row record, each row record corresponding to a 24 hour period of the day to generate the DATASET DATASET 2. The sum of the PRB utilization per time period/current day PRB utilization aggregate is taken as the PRB probability distribution per time period, which is W1. Each row records the corresponding number of hours of the day as Y2. DATASET2 adds PRB probability distribution W1 of each row record as weight, DATASET2 adds the number of hours of the day Y2 corresponding to each row record as predictor, and generates probability distribution model DATASET DATASET3 with field sets of [ X1, W1, Y2 ].
Step S603, train an accumulative quantity model.
Specifically, the DATASET DATASET1 is characterized by X1 and Y1 as a predicted value, and the training set is input into the LIGHT TGBM model for training in a regression mode.
Step S604, a probability distribution model is trained.
Specifically, the DATASET2 is characterized by X1, W1 as a weight, and Y2 as a predicted value, and the training set is input into the light tgbm model for training, and is trained in a multi-classification manner, and a probability matrix with the output result as a class is set.
In step S605, the cumulative model predicts PRB utilization cumulative quantity P _ sum.
Step S606, a PRB probability distribution matrix P _ PRB for each energy-saving period is predicted.
Specifically, a PRB probability distribution matrix P _ PRB for each period is generated using a probability distribution model with 1/24 as a weight.
Step S607, factoring the PRB utilization cumulative quantity P _ sum with the PRB probability distribution matrix P _ PRB for each time interval, to obtain the PRB utilization of all time intervals in one day.
The original data of the embodiment of the invention is from the load data of the Chinese telecommunication 4G base station, the base station kpi data in one year is used, compared with the time-sharing prediction method realized by the prior art, the resource occupancy rate of the scheme is reduced by 23%, the wap index is reduced from 912.5 to 613.4, the 32% is improved, and the reliability and the practicability of the model are checked.
Embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program is arranged to perform the steps of any of the above-mentioned method embodiments when executed.
In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to: various media capable of storing computer programs, such as a usb disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic disk, or an optical disk.
Embodiments of the present invention also provide an electronic device comprising a memory having a computer program stored therein and a processor arranged to run the computer program to perform the steps of any of the above method embodiments.
In an exemplary embodiment, the electronic apparatus may further include a transmission device and an input/output device, wherein the transmission device is connected to the processor, and the input/output device is connected to the processor.
For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary embodiments, and details of this embodiment are not repeated herein.
It will be apparent to those skilled in the art that the various modules or steps of the invention described above may be implemented using a general purpose computing device, they may be centralized on a single computing device or distributed across a network of computing devices, and they may be implemented using program code executable by the computing devices, such that they may be stored in a memory device and executed by the computing device, and in some cases, the steps shown or described may be performed in an order different than that described herein, or they may be separately fabricated into various integrated circuit modules, or multiple ones of them may be fabricated into a single integrated circuit module. Thus, the present invention is not limited to any specific combination of hardware and software.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the principle of the present invention should be included in the protection scope of the present invention.
Claims (15)
1. A method for predicting traffic of a base station, comprising:
predicting the accumulated value of the traffic of all energy-saving periods and the probability distribution of the traffic of all energy-saving periods in a target energy-saving period of the base station;
and predicting the traffic of each energy-saving period in the target period of the base station according to the accumulated value of the traffic of all the energy-saving periods in the target energy-saving period and the probability distribution of the traffic.
2. The method of claim 1, wherein before predicting the aggregate value of the traffic volumes of all energy-saving periods and the probability distribution of the traffic volumes of all energy-saving periods within the target energy-saving period of the base station, further comprising:
and determining a target energy-saving period to be predicted, an energy-saving time period and a traffic index of the base station, wherein the traffic is the utilization rate of the PRB.
3. The method of claim 2, wherein after determining the target energy-saving period, the energy-saving period and the traffic index to be predicted by the base station, the method further comprises:
determining a model input characteristic data set according to a target energy-saving cycle, an energy-saving time period and a traffic index, wherein the model input characteristic data set comprises: traffic timing trend indicator data, base station KPI trend indicator data, and external indicator data.
4. The method of claim 3, wherein,
the traffic time series trend index data includes: the average PRB utilization rate, the lowest PRB utilization rate and the highest PRB utilization rate of each energy-saving time interval in a plurality of energy-saving periods, and the highest PRB utilization rate, the average PRB utilization rate and the lowest PRB utilization rate of each energy-saving period;
the base station KPI trend index data comprises: the average number of RRC users, the minimum number of RRC users and the maximum number of RRC users in each energy-saving period in a plurality of energy-saving periods, and the maximum number of RRC users, the average number of RRC users and the minimum number of RRC users in each energy-saving period;
the external index data includes: time characteristics and regional characteristics, wherein the time characteristics at least comprise one of the following: the year, month and date of the energy-saving time period, whether the week is weekend or not, whether the holiday is festival or not and whether the hour is hour or not; the geographic features include at least one of: the province where the base station is located, the city of land, longitude and latitude information, the ID of town, suburb and base station cell.
5. The method of claim 3, wherein after determining the model input feature data set according to the target energy saving cycle, the energy saving period, and the traffic indicator, further comprising:
selecting traffic time sequence trend index data, base station KPI trend index data and other external index data of the base station as a basic feature data set by taking an energy-saving period as a unit, wherein a field set of the basic feature data set is [ X1 ];
adding a predicted value Y1 in the basic characteristic data set to generate an accumulated quantity model data set, wherein the predicted value Y1 is the sum of the utilization rates of PRBs in each energy-saving period of the next energy-saving period, and the field set of the accumulated quantity model data set is [ X1, Y1 ];
adding a probability distribution W1 of PRB utilization rate of each energy-saving period of an energy-saving cycle and the number Y2 of energy-saving periods to be predicted in the basic characteristic data set to generate a probability distribution model data set, wherein the field set of the probability distribution model data set is [ X1, W1, Y2 ].
6. The method of claim 5, further comprising:
and inputting the accumulative quantity model data set and the probability distribution model data set into a LightGBM model for training to respectively obtain an accumulative quantity model for predicting the PRB utilization rate and a probability distribution model for predicting the PRB utilization rate.
7. The method of claim 6, wherein inputting the cumulative model dataset into a LightGBM model for training to obtain a cumulative model for predicting PRB utilization comprises:
and inputting the accumulative quantity model data set into a lightGBM model by adopting a regression mode to train so as to obtain an accumulative quantity model for predicting the PRB utilization rate, wherein the field set X1 of the basic characteristic data set is used as a characteristic, and the sum Y1 of the PRB utilization rates of each energy-saving period in the next energy-saving period is used as a predicted value.
8. The method of claim 6, wherein inputting the probability distribution model dataset into a LightGBM model for training to obtain a probability distribution model for predicting PRB utilization comprises:
and inputting the probability distribution model data set into a lightGBM model for training by adopting a multi-classification mode by taking a field set X1 of a basic characteristic data set as a characteristic, taking the probability distribution W1 of the PRB utilization rate of each energy-saving time interval as a weight, and taking the number Y2 of the energy-saving time intervals to be predicted as a predicted value to obtain the probability distribution model for predicting the PRB utilization rate.
9. The method of claim 8, further comprising:
predicting and obtaining an accumulated value of PRB utilization rates of all energy-saving time periods in a target energy-saving period based on a trained accumulated model;
and predicting and obtaining the probability distribution of the PRB utilization rate of all energy-saving time periods in the target energy-saving period based on the trained probability distribution model.
10. The method of claim 9, further comprising:
inputting the accumulated value of the PRB utilization rates of all energy-saving periods in the target energy-saving period into the probability distribution model by taking 1/N as weight, and acquiring a PRB probability distribution matrix of each energy-saving period in the target energy-saving period, wherein N is the number of the energy-saving periods;
and factoring the accumulated value of the PRB utilization rates of all the energy-saving periods in the target energy-saving period with the probability distribution of the PRB utilization rate of each energy-saving period to determine the PRB utilization rate of each energy-saving period in the target energy-saving period.
11. An apparatus for predicting traffic of a base station, comprising:
the first prediction module is used for predicting the accumulated value of the traffic of all energy-saving time intervals and the probability distribution of the traffic of all energy-saving time intervals in the target energy-saving period of the base station;
and the second prediction module is used for predicting the traffic of each energy-saving period in the target period of the base station according to the accumulated value of the traffic of all the energy-saving periods in the target energy-saving period and the probability distribution of the traffic.
12. The apparatus of claim 11, further comprising:
and the determining module is used for determining a target energy-saving period to be predicted, an energy-saving time period and a traffic index of the base station, wherein the traffic is the utilization rate of the PRB.
13. The apparatus of claim 12, wherein the determining module comprises:
a determining unit, configured to determine a model input feature data set according to a target energy saving cycle, an energy saving time period, and a traffic index, where the model input feature data set includes: traffic timing trend indicator data, base station KPI trend indicator data, and external indicator data.
14. A computer-readable storage medium, in which a computer program is stored, wherein the computer program is arranged to perform the method of any of claims 1 to 10 when executed.
15. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, and wherein the processor is arranged to execute the computer program to perform the method of any of claims 1 to 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111605305.1A CN114339967A (en) | 2021-12-24 | 2021-12-24 | Method and device for predicting base station traffic |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111605305.1A CN114339967A (en) | 2021-12-24 | 2021-12-24 | Method and device for predicting base station traffic |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114339967A true CN114339967A (en) | 2022-04-12 |
Family
ID=81012074
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111605305.1A Pending CN114339967A (en) | 2021-12-24 | 2021-12-24 | Method and device for predicting base station traffic |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114339967A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115426030A (en) * | 2022-09-06 | 2022-12-02 | 广州爱浦路网络技术有限公司 | Satellite energy-saving method and device based on big data |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015207795A (en) * | 2014-04-17 | 2015-11-19 | 日本電信電話株式会社 | Communication traffic prediction device, method and program |
| CN110766232A (en) * | 2019-10-30 | 2020-02-07 | 支付宝(杭州)信息技术有限公司 | Dynamic prediction method and system thereof |
| CN111328084A (en) * | 2018-12-14 | 2020-06-23 | 中国移动通信集团广东有限公司 | Method and device for evaluating cell capacity |
| CN112434885A (en) * | 2020-12-15 | 2021-03-02 | 中国联合网络通信集团有限公司 | Service prediction method and device for energy-saving cell |
| CN112835771A (en) * | 2019-11-22 | 2021-05-25 | 中国移动通信有限公司研究院 | Traffic prediction method, device, equipment and storage medium |
| CN113207162A (en) * | 2021-04-14 | 2021-08-03 | 浪潮天元通信信息系统有限公司 | Base station energy consumption intelligent control method based on service prediction |
-
2021
- 2021-12-24 CN CN202111605305.1A patent/CN114339967A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015207795A (en) * | 2014-04-17 | 2015-11-19 | 日本電信電話株式会社 | Communication traffic prediction device, method and program |
| CN111328084A (en) * | 2018-12-14 | 2020-06-23 | 中国移动通信集团广东有限公司 | Method and device for evaluating cell capacity |
| CN110766232A (en) * | 2019-10-30 | 2020-02-07 | 支付宝(杭州)信息技术有限公司 | Dynamic prediction method and system thereof |
| CN112835771A (en) * | 2019-11-22 | 2021-05-25 | 中国移动通信有限公司研究院 | Traffic prediction method, device, equipment and storage medium |
| CN112434885A (en) * | 2020-12-15 | 2021-03-02 | 中国联合网络通信集团有限公司 | Service prediction method and device for energy-saving cell |
| CN113207162A (en) * | 2021-04-14 | 2021-08-03 | 浪潮天元通信信息系统有限公司 | Base station energy consumption intelligent control method based on service prediction |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115426030A (en) * | 2022-09-06 | 2022-12-02 | 广州爱浦路网络技术有限公司 | Satellite energy-saving method and device based on big data |
| CN115426030B (en) * | 2022-09-06 | 2023-05-05 | 广州爱浦路网络技术有限公司 | Satellite energy saving method and satellite energy saving device based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107333232B (en) | Terminal positioning method and network equipment | |
| KR101976189B1 (en) | Method of providing analysis service of floating population | |
| CN101442762B (en) | Method and apparatus for analyzing network performance and locating network fault | |
| CN103747523A (en) | User position predicating system and method based on wireless network | |
| Mehmeti et al. | Performance modeling, analysis, and optimization of delayed mobile data offloading for mobile users | |
| CN105531746A (en) | Management of data collected for traffic analysis | |
| Rodoplu et al. | A multiscale algorithm for joint forecasting–scheduling to solve the massive access problem of IoT | |
| CN103731916A (en) | Wireless-network-based user position predicting system and method | |
| US20190287391A1 (en) | Traffic demand prediction system and traffic demand prediction apparatus | |
| CN116167581A (en) | Battery demand estimation method and device, scheduling method and computer equipment | |
| CN114339967A (en) | Method and device for predicting base station traffic | |
| CN114462722B (en) | New energy power generation light-weight high-precision cloud prediction system, method and device | |
| CN111800807A (en) | Method and device for alarming number of base station users | |
| CN112287503B (en) | Dynamic space network construction method for traffic demand prediction | |
| CN116756522B (en) | Probability forecasting method and device, storage medium and electronic equipment | |
| Aldhyani et al. | An integrated model for prediction of loading packets in network traffic | |
| CN111311014B (en) | Service data processing method, device, computer equipment and storage medium | |
| US11671874B2 (en) | Communication band calculation device, communication band calculation method, and medium | |
| Li et al. | A meta-learning based framework for cell-level mobile network traffic prediction | |
| Tang et al. | Forecasting Network Traffic at Large Time Scales by Using Dual-Related Method | |
| Chen | Human Habits Investigation: from mobility reconstruction to mobile traffic prediction | |
| CN113260064B (en) | Access method and system of large-scale internet of things equipment | |
| CN117528552B (en) | Mobile phone data transmission method and system based on 5G | |
| US11962475B2 (en) | Estimating properties of units using system state graph models | |
| Zhu et al. | Prediction of cellular network channel utilization based on graph convolutional networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |