CN116996926B - Data-driven base station power consumption modeling method - Google Patents
Data-driven base station power consumption modeling method Download PDFInfo
- Publication number
- CN116996926B CN116996926B CN202311247898.8A CN202311247898A CN116996926B CN 116996926 B CN116996926 B CN 116996926B CN 202311247898 A CN202311247898 A CN 202311247898A CN 116996926 B CN116996926 B CN 116996926B
- Authority
- CN
- China
- Prior art keywords
- base station
- power consumption
- data
- model
- training
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000012549 training Methods 0.000 claims abstract description 39
- 238000013508 migration Methods 0.000 claims abstract description 14
- 230000005012 migration Effects 0.000 claims abstract description 14
- 238000007781 pre-processing Methods 0.000 claims abstract description 9
- 238000010801 machine learning Methods 0.000 claims abstract description 6
- 238000013526 transfer learning Methods 0.000 claims description 15
- 238000012935 Averaging Methods 0.000 claims description 7
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 2
- 238000013527 convolutional neural network Methods 0.000 claims description 2
- 230000007704 transition Effects 0.000 claims 1
- 230000006870 function Effects 0.000 description 9
- 238000003066 decision tree Methods 0.000 description 3
- 238000012937 correction Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000005192 partition Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/06—Testing, supervising or monitoring using simulated traffic
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/096—Transfer learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W88/00—Devices specially adapted for wireless communication networks, e.g. terminals, base stations or access point devices
- H04W88/08—Access point devices
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Hardware Design (AREA)
- Medical Informatics (AREA)
- Geometry (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The application discloses a data-driven base station power consumption modeling method, which comprises the following steps: s1, taking a base station of any category as an initial base station, collecting data of the initial base station, preprocessing the collected data, and constructing an original sample set; s2, constructing a base station power consumption model through a machine learning algorithm, and training the base station power consumption model through an original sample set to obtain an initial base station-like power consumption model; s3, performing migration learning on the basis of the power consumption model of the initial base station to obtain base station power consumption models of different categories. According to the application, the base station data is collected to model the power consumption of the base station, and then the migration learning method is utilized to finish the modeling of the power consumption of different types of base stations, so that the base station power consumption modeling method has good accuracy and generalization.
Description
Technical Field
The application relates to base station power consumption modeling, in particular to a data-driven base station power consumption modeling method.
Background
Existing base station power consumption modeling methods are generally based on physical models or empirical rules by analyzing the power consumption of each component of the base station and then modeling. Such modeling methods have a number of drawbacks: first, when performing careful power consumption analysis and modeling on each component of the base station, a great deal of expertise and time-consuming work are required, the process is complicated, and the modeling efficiency is low. Secondly, the existing method can only model a base station with a specific model or a specific configuration at a time, and the model is weak in generalization due to the immobilization of parameters, so that the existing method is difficult to adapt to different base stations, different network scenes and different deployment conditions. Thirdly, the existing method only considers a few key factors, ignores the influence of factors such as different service type flows and the like on power consumption, and the defects cause insufficient generalization and accuracy of a finally constructed base station power consumption model.
Disclosure of Invention
The application aims to overcome the defects of the prior art and provides a data-driven base station power consumption modeling method which is used for modeling the base station power consumption by collecting base station data and then completing the modeling of different types of base station power consumption by utilizing a migration learning method, thereby having good accuracy and generalization.
The aim of the application is realized by the following technical scheme: a data-driven base station power consumption modeling method comprises the following steps:
s1, taking a base station of any category as an initial base station, collecting data of the initial base station, preprocessing the collected data, and constructing an original sample set;
s2, constructing a base station power consumption model through a machine learning algorithm, and training the base station power consumption model through an original sample set to obtain an initial base station-like power consumption model;
s3, performing migration learning on the basis of the power consumption model of the initial base station, and obtaining base station power consumption models of different categories:
firstly, removing an output layer of a power consumption model of an initial base station, and then connecting a layer of transfer learning model to form a base station power consumption model based on transfer learning;
pre-training a base station power consumption model based on transfer learning by using an original sample set:
the power consumption model part of the initial base station is kept unchanged during training, only the transfer learning model is updated, and a pre-training model is obtained after training is finished;
for any base station with different types from the initial base station, collecting data of the base station for preprocessing to obtain a sample set of the base station, and then training a pre-training model by utilizing the sample set of the base station:
in the training process, in the pre-training model, the power consumption model part of the initial base station is kept unchanged, only the migration learning model is updated, and the power consumption model of the current base station is obtained after the training is finished.
The beneficial effects of the application are as follows: according to the application, detailed component power consumption analysis is not needed, only base station power consumption data are required to be collected, and then modeling is directly carried out, so that the modeling efficiency is improved; a large amount of data is adopted to train the model, and a migration learning method is adopted, so that generalization of the model is improved, and the model can be suitable for base stations of different types or configurations; the nonlinear relation between the power consumption and different service flows can be automatically learned, and the accuracy and the stability of the model are improved.
Drawings
FIG. 1 is a flow chart of the method of the present application.
Detailed Description
The technical solution of the present application will be described in further detail with reference to the accompanying drawings, but the scope of the present application is not limited to the following description.
As shown in fig. 1, a data-driven base station power consumption modeling method includes the following steps: the method comprises the following steps:
s1, taking a base station of any category as an initial base station, collecting data of the initial base station, preprocessing the collected data, and constructing an original sample set;
s101, for any base station under an initial base station, collecting geographic position information, coverage scenes, power consumption, antenna hanging height, coverage radius and uplink and downlink traffic of different service types of the base station in a plurality of moments, and forming a data sample:
in each moment, geographic position information, coverage scene, power consumption, antenna hanging height, coverage radius and uplink and downlink flow of different service types of the base station are taken as sample characteristics of the moment, and the power consumption of the base station is taken as a sample label;
setting that data at T moments are collected, and obtaining T data samples;
it should be noted that, although data of T times is collected, when the base station determines, its coverage scene, geographic location, antenna hanging height, and coverage radius are determined, and at different times, uplink and downlink traffic and power consumption of different service types are different.
S102, repeatedly executing the step S101 for a plurality of base stations under the initial base station, and obtaining a data sample of each base station; providing a total of M base stations, and obtaining M x T data samples;
in the embodiment of the present application, the plurality of base stations may be all base stations under the initial base station, or may be M base stations selected from the base stations when the number of base stations is greater.
Judging whether the power consumption information contained in each obtained data sample is missing or not, and filling the power consumption information if the power consumption information is missing;
the process for filling the power consumption information comprises the following steps:
set the firstiThe first base stationtThe data at the individual moments in time are missing,，/>the filling mode includes
From the firstiOutside of individual base stationsM-1 base station, selectAnd the first and secondiThe base station is located at a distance from the nearest base station,will choose +.>The first base stationtAveraging the power consumption data at each moment to obtain a first average value;
then from the firstiThe remainder of the individual base stationsT-From the power consumption data of 1 moment, select andtthe closest time of dayTaking out the power consumption values at each moment, and then averaging to obtain a second average value, < >>；
After averaging the first average value and the second average value, filling the missing power consumption value;
compared with the traditional time sequence filling mode, the filling mode effectively improves the filling accuracy.
In the embodiment of the application, padding can be performed from two directions of time sequence and characteristics, and the accuracy of padding can be improved, for example:
from the firstiOutside of individual base stationsM-1 base station, select the firstiThe coverage scenes of the base stations are the same, and then the base station closest to the antenna hanging height and the coverage radius of the ith base station is selected from the base stations:
for each base station, the antenna hanging height of the base station is different from the antenna hanging height of the ith base station, then an absolute value is obtained, the coverage radius of the base station is different from the antenna radius of the ith base station, then an absolute value is obtained, and then the two absolute values are summed; then the base station with the smallest summation result is taken and the base station with the smallest summation result is processedtTaking the power consumption value at each moment as a filled reference value;
then from the firstiThe remainder of the individual base stationsT-From the power consumption data of 1 moment, select andtthe closest time of dayThe power consumption values at each instant are taken out and then averaged to obtain a correction value +.>；
And averaging the reference value and the correction value, and filling the missing power consumption value.
S103, repeatedly executing the step S102 for N times to obtain N.times.M.T data samples, adding the data samples into an original sample set, and constructing the finished original sample set.
S2, constructing a base station power consumption model through a machine learning algorithm, and training the base station power consumption model through an original sample set to obtain an initial base station-like power consumption model;
the machine learning algorithm includes an XGBoost algorithm, a deep neural network algorithm, or a convolutional neural network algorithm.
Taking an XGBoost regression algorithm as an example, taking the power consumption of a base station as a label, and taking geographic position, coverage scene, antenna hanging height, coverage, uplink and downlink flows of different services and the like as characteristics. Assuming a total of n samples of the collected dataset, each sample has m features, denoted by set D:
wherein,is a feature vector +_>Is a label.
The algorithm takes a decision tree as a base learner, and the conversion function is as follows:
where q represents the tree structure and T represents the number of leaves of the tree. In each iteration, a decision tree is generated, and the model sums the trees to obtain a prediction function during predictionTo express:
here, theRepresenting the space of the generated decision tree.
For loss functions of modelsThe representation, then, can yield the objective function of the model:
to improve model robustness, regular terms are added into the objective functionObtaining a corrected objective function:
the goal of the model is to minimize the modified objective function. Because the model adopts the algorithm of the addition model, the time complexity of the prior spanning tree is generated when calculating the objective functionIs known, the objective function to be minimized can be further simplified:
then the second-order Taylor expansion method is used for the re-alignmentApproximation is performed:
wherein the method comprises the steps ofAnd->Are respectively->For->First-order and second-order partial derivatives are obtained. Wherein the method comprises the steps ofIs a constant which can be ignored when the maximum value is found:
substituting the expression of the time complexity of the tree into the following expression can be obtained, whereinRepresenting the set of samples in the j-th leaf:
so for a tree of defined structure, the optimal weight for each leaf can be obtained by solving the quadratic equationAnd optimum value of the loss function:
/>
the model is growingWhen in tree, a complete greedy method is adopted, all possible situations are listed at each partition point, and then objective function values before and after partition are calculatedAnd selecting the situation that the information gain is maximum for segmentation. And growing the tree in the mode during each iteration until a preset termination condition is met, and building the power consumption model of the target type base station.
When the power consumption model is constructed by using a deep neural network algorithm or a convolution neural network algorithm, the characteristics of each sample in the sample set are only needed to be used as model input, the labels are used as expected input, and then the neural network training is carried out in a back propagation mode.
For the built base station power consumption model, as the model only aims at a certain specific type of base station, in order to build the whole network base station power consumption model, the model is subjected to migration learning, so that the whole network base station power consumption model is obtained. By using the migration learning method, the knowledge migration of the model can be realized, the requirement on training data can be reduced, and the efficiency, accuracy and generalization capability of model construction are higher.
S3, performing migration learning on the basis of the power consumption model of the initial base station, and obtaining base station power consumption models of different categories: in the process, firstly, based on an established power consumption model, a pre-training and fine-tuning process is added to realize migration learning.
Firstly, removing an output layer of a power consumption model of an initial base station, and then connecting a layer of transfer learning model to form a base station power consumption model based on transfer learning; in an embodiment of the application, the transfer learning model includes ResNet and VGGNet.
The pre-training process comprises the following steps: pre-training a base station power consumption model based on transfer learning by using an original sample set: features of the data samples in the original sample set are used as inputs in the pre-training process, and labels are used as expected outputs of the model. The power consumption model part of the initial base station is kept unchanged during training, only the transfer learning model is updated, and a pre-training model is obtained after training is finished; this process is implemented using optimization algorithms, including but not limited to random gradient descent.
Fine tuning: for any base station with different type from the initial base station, data of the base station is collected for preprocessing (the process of collecting the data of the base station for preprocessing is performed according to step S102), a sample set of the base station is obtained, and then the sample set of the base station is utilized to train a pre-training model: during training, taking the characteristics of the base station sample set as input and taking the label as expected output; when training is performed, the power consumption model part of the initial class base station in the pre-training model is kept unchanged, only the transfer learning model is updated, and the power consumption model of the current class base station is obtained after the training is finished. This process is implemented using optimization algorithms, including but not limited to random gradient descent.
And (3) performing migration learning on each base station of different types in the whole network according to the step (S3) to obtain power consumption models of the base stations of different types.
In the embodiment of the present application, the modeling of the whole network base station may be performed for a certain target area (city, a certain area of city, etc., villages and towns, etc.), or may be performed for a whole range formed by a plurality of target areas in a larger range.
In an embodiment of the present application, the types of base stations are categorized according to the POIs in which the base stations are located, including but not limited to, business centers, highways, metropolitan villages, residential communities, suburban roads, industrial parks, schools, subways, and the like. Of course, other classification methods are also possible, and the setting can be performed according to the needs of the user.
In the application, the existing base station power consumption modeling method is considered to model by analyzing the power consumption of the base station antenna unit and the baseband unit, so that the geographic position of the base station and the traffic of different types of services are not considered. The difference in geographical location may have an effect on the propagation of the signal and thus on the power consumption of the antenna unit. For example, in urban dense areas, signal propagation can be affected by blocking and multipath effects, resulting in antenna elements requiring more power to maintain coverage and signal quality. And the traffic (such as voice communication, video stream, etc.) of different types of services has different load requirements on the baseband unit, so that the power consumption of the baseband unit can be influenced. Therefore, in order to improve the accuracy of the model, the application selects the geographic position of the base station and the traffic of different types of services as parameters. Meanwhile, in the transfer learning process, the pre-training and the fine-tuning are adopted, and as the pre-training process adopts the original data set, only the current base station data set with smaller data size is adopted in the fine-tuning process, so that the training difficulty and complexity are greatly reduced. And the modeling efficiency of the base station power consumption model is improved.
The above embodiments are only for illustrating the technical solution of the present application, and are not limiting; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims (7)
1. A data-driven base station power consumption modeling method is characterized in that: the method comprises the following steps:
s1, taking a base station of any category as an initial base station, collecting data of the initial base station, preprocessing the collected data, and constructing an original sample set;
s2, constructing a base station power consumption model through a machine learning algorithm, and training the base station power consumption model through an original sample set to obtain an initial base station-like power consumption model;
s3, performing migration learning on the basis of the power consumption model of the initial base station, and obtaining base station power consumption models of different categories:
firstly, removing an output layer of a power consumption model of an initial base station, and then connecting a layer of transfer learning model to form a base station power consumption model based on transfer learning;
pre-training a base station power consumption model based on transfer learning by using an original sample set:
the power consumption model part of the initial base station is kept unchanged during training, only the transfer learning model is updated, and a pre-training model is obtained after training is finished;
for any base station with different types from the initial base station, collecting data of the base station for preprocessing to obtain a sample set of the base station, and then training a pre-training model by utilizing the sample set of the base station:
in the training process, in the pre-training model, the power consumption model part of the initial base station is kept unchanged, only the migration learning model is updated, and the power consumption model of the current base station is obtained after the training is finished.
2. A method of modeling data driven base station power consumption as claimed in claim 1, wherein: the step S1 includes:
s101, for any base station under an initial base station, collecting geographic position information, coverage scenes, power consumption, antenna hanging height, coverage radius and uplink and downlink traffic of different service types of the base station in a plurality of moments, and forming a data sample:
in each moment, geographic position information, coverage scene, power consumption, antenna hanging height, coverage radius and uplink and downlink flow of different service types of the base station are taken as sample characteristics of the moment, and the power consumption of the base station is taken as a sample label;
setting that data at T moments are collected, and obtaining T data samples;
s102, repeatedly executing the step S101 for a plurality of base stations under the initial base station, and obtaining a data sample of each base station; providing a total of M base stations, and obtaining M x T data samples;
judging whether the power consumption information contained in each obtained data sample is missing or not, and filling the power consumption information if the power consumption information is missing;
s103, repeatedly executing the step S102 for N times to obtain N.M.T data samples, and adding the N.M.T data samples into an original sample set to complete the construction of the original sample set.
3. A method of modeling data driven base station power consumption as claimed in claim 2, wherein: the process for filling the power consumption information comprises the following steps:
set the firstiThe first base stationtThe data at the individual moments in time are missing,，/>the filling mode includes:
from the firstiOutside of individual base stationsM-1 base station, selectAnd the first and secondiThe base station closest to the individual base station +.>Will choose +.>The first base stationtAveraging the power consumption data at each moment to obtain a first average value;
then from the firstiThe remainder of the individual base stationsT-From the power consumption data of 1 moment, select andtthe closest time of dayTaking out the power consumption values at each moment, and then averaging to obtain a second average value, < >>；
And after averaging the first average value and the second average value, filling the missing power consumption value.
4. A method of modeling data driven base station power consumption as claimed in claim 2, wherein: the geographic position information comprises longitude and latitude coordinates of a base station, and the coverage scene comprises an indoor scene and an outdoor scene; the service type includes a video service, a voice service, or a game service.
5. A method of modeling data driven base station power consumption as claimed in claim 1, wherein: the machine learning algorithm includes an XGBoost algorithm, a deep neural network algorithm, or a convolutional neural network algorithm.
6. A method of modeling data driven base station power consumption as claimed in claim 1, wherein: the transition learning model includes ResNet or VGGNet.
7. A method of modeling data driven base station power consumption as claimed in claim 2, wherein: for any base station of a type different from the initial base station, the process of acquiring the data of the base station and preprocessing is performed according to step S101.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311247898.8A CN116996926B (en) | 2023-09-26 | 2023-09-26 | Data-driven base station power consumption modeling method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311247898.8A CN116996926B (en) | 2023-09-26 | 2023-09-26 | Data-driven base station power consumption modeling method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116996926A CN116996926A (en) | 2023-11-03 |
| CN116996926B true CN116996926B (en) | 2023-11-28 |
Family
ID=88526945
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311247898.8A Active CN116996926B (en) | 2023-09-26 | 2023-09-26 | Data-driven base station power consumption modeling method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116996926B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106714223A (en) * | 2015-11-16 | 2017-05-24 | 中国移动通信集团公司 | Method and device for building base station energy consumption model |
| CN110880020A (en) * | 2019-10-30 | 2020-03-13 | 西安交通大学 | Self-adaptive trans-regional base station energy consumption model migration and compensation method |
| CN113987698A (en) * | 2021-10-11 | 2022-01-28 | 常州工学院 | Machine tool power consumption model modeling method based on data driving |
| WO2022036867A1 (en) * | 2020-08-21 | 2022-02-24 | 广东电网有限责任公司清远供电局 | Method for reconstructing three-dimensional model of transformer substation |
| CN114885340A (en) * | 2022-05-07 | 2022-08-09 | 东南大学 | Ultra-dense wireless network power distribution method based on deep transfer learning |
-
2023
- 2023-09-26 CN CN202311247898.8A patent/CN116996926B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106714223A (en) * | 2015-11-16 | 2017-05-24 | 中国移动通信集团公司 | Method and device for building base station energy consumption model |
| CN110880020A (en) * | 2019-10-30 | 2020-03-13 | 西安交通大学 | Self-adaptive trans-regional base station energy consumption model migration and compensation method |
| WO2022036867A1 (en) * | 2020-08-21 | 2022-02-24 | 广东电网有限责任公司清远供电局 | Method for reconstructing three-dimensional model of transformer substation |
| CN113987698A (en) * | 2021-10-11 | 2022-01-28 | 常州工学院 | Machine tool power consumption model modeling method based on data driving |
| CN114885340A (en) * | 2022-05-07 | 2022-08-09 | 东南大学 | Ultra-dense wireless network power distribution method based on deep transfer learning |
Non-Patent Citations (1)
| Title |
|---|
| 基于深度学习的自编码器端到端物理层优化方案;叶佩文;贾向东;杨小蓉;胡海霞;;计算机工程(第12期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116996926A (en) | 2023-11-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109743683B (en) | Method for determining position of mobile phone user by adopting deep learning fusion network model | |
| CN111210082B (en) | Precipitation prediction method based on optimized BP neural network algorithm | |
| CN106912015B (en) | Personnel trip chain identification method based on mobile network data | |
| CN109063938B (en) | Air quality prediction method based on PSODE-BP neural network | |
| CN112902969B (en) | Path planning method of unmanned aerial vehicle in data collection process | |
| CN106780089B (en) | Permanent basic farmland planning method based on neural network cellular automaton model | |
| CN109635511A (en) | A kind of high-rise residential areas forced-ventilated schemes generation design method generating confrontation network based on condition | |
| CN109635748B (en) | Method for extracting road characteristics in high-resolution image | |
| CN103778477B (en) | Monitor station cloth station method and apparatus | |
| CN114092769A (en) | Transformer substation multi-scene inspection analysis method based on federal learning | |
| CN113505842B (en) | Automatic urban building extraction method suitable for large-scale regional remote sensing image | |
| CN112685504B (en) | Production process-oriented distributed migration chart learning method | |
| CN108616892B (en) | 4G base station site selection method based on hybrid immune algorithm | |
| CN112004233B (en) | Network planning method based on big data mining | |
| CN115641718B (en) | Short-time traffic flow prediction method based on bayonet flow similarity and semantic association | |
| CN111405586A (en) | Wireless intelligent propagation method based on Cost231-Hata model | |
| Hu et al. | An intelligent UAV deployment scheme for load balance in small cell networks using machine learning | |
| CN110378527A (en) | A kind of water station site selecting method | |
| CN115965171A (en) | Micro-park site selection method based on ant colony optimization algorithm | |
| CN113592132A (en) | Rainfall objective forecasting method based on numerical weather forecast and artificial intelligence | |
| CN114492941A (en) | Whole-county photovoltaic prediction method based on cluster division and data enhancement | |
| CN117374956A (en) | Short-term prediction method for photovoltaic power generation of comprehensive energy station | |
| CN116996926B (en) | Data-driven base station power consumption modeling method | |
| CN115051925A (en) | Time-space sequence prediction method based on transfer learning | |
| CN113342873B (en) | Population analysis unit division method based on city morphology and convergence mode |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |