CN111405583B - Tidal effect avoidance method, tidal effect avoidance device and computer-readable storage medium - Google Patents
Tidal effect avoidance method, tidal effect avoidance device and computer-readable storage medium Download PDFInfo
- Publication number
- CN111405583B CN111405583B CN201910000461.1A CN201910000461A CN111405583B CN 111405583 B CN111405583 B CN 111405583B CN 201910000461 A CN201910000461 A CN 201910000461A CN 111405583 B CN111405583 B CN 111405583B
- Authority
- CN
- China
- Prior art keywords
- service
- success rate
- prediction model
- rate prediction
- internet
- 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
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/22—Traffic simulation tools or models
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/147—Network analysis or design for predicting network behaviour
-
- 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
-
- 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
Embodiments of the present invention provide a tidal effect avoidance method, apparatus, and computer-readable storage medium, the method comprising: when the service runs, determining configuration parameters which enable functions of the service success rate prediction model to obtain maximum values based on the service success rate prediction model obtained through training and feature vectors reported by the terminal of the Internet of things; and transmitting the configuration parameters to the Internet of things terminal, and adjusting a service model and sending service data in a peak shifting mode by the Internet of things terminal.
Description
Technical Field
The present invention relates to the field of mobile communications technologies, and in particular, to a tidal effect avoidance method, apparatus, and computer readable storage medium.
Background
With the rapid development of mobile communication, the number of mobile subscribers and the traffic volume have increased. With the rapid increase of the number of terminals of the internet of things and the continuous increase of service types, the tidal effect in the wireless communication field is more obvious, and the terminal shows obvious regional characteristics and time characteristics. For example: during the working period, people gather in a large amount in the CBD area, and migrate to the residential area in a large amount after going out of work. The phenomenon causes traffic flow of terminal equipment of the Internet of things, so that burst large traffic occurs in a hot spot area at a specific moment, network congestion is caused, even access is not possible, and service failure occurs, namely tidal effect. When tidal effects occur, for terminal services of the internet of things, the conditions of service interaction failure and frequent retries can occur, and for network equipment, high-load operation and even service rejection can occur, so that the service quality is greatly influenced, and the common problem puzzling operators and service manufacturers is solved.
At present, two processing modes aiming at tidal effects are mainly adopted, one is that a terminal side actively avoids concurrent access conditions, and a typical method is that random time delay is added when service data is transmitted, so that instantaneous concurrency of the terminal can be reduced when actual service is transmitted, and when data of an access terminal is increased, the probability of collision conflict is increased, and the suppression effect of the tidal effects is weakened; another typical method is to optimize on the network device side, including adding a buffer queue mechanism or an elastic capacity expansion mechanism, aiming at increasing network access capability, where the network device introduces a buffer queue mode, and when the concurrency is large, there is a problem of access rejection due to data discarding, and the elastic capacity expansion mode has weak practical operability.
Disclosure of Invention
In view of the foregoing, it is desirable for embodiments of the present invention to provide a tidal effect avoidance method, apparatus, and computer readable storage medium.
In order to achieve the above object, the technical solution of the embodiment of the present invention is as follows:
the embodiment of the invention provides a tidal effect avoidance method, which comprises the following steps:
when the service runs, determining configuration parameters which enable functions of the service success rate prediction model to obtain maximum values based on the service success rate prediction model obtained through training and feature vectors reported by the terminal of the Internet of things;
and transmitting the configuration parameters to the Internet of things terminal, and adjusting a service model and sending service data in a peak shifting mode by the Internet of things terminal.
Optionally, before the service is run, the method further includes:
based on the feature vector reported by the Internet of things terminal during test operation, training by combining with a logistic regression LR method to obtain a service success rate prediction model; different service success rate prediction models correspond to different base stations.
The feature vector reported by the terminal of the internet of things based on the test operation is trained by combining an LR method to obtain a service success rate prediction model, and the method comprises the following steps:
dividing a set of feature vectors reported by an Internet of things terminal in at least one test run period to obtain a data set S for training and a data set T for testing;
training a preset business success rate prediction model by utilizing the data set S for training to obtain a group of fitted parameters;
and generalizing the service success rate prediction model obtained through training by utilizing the data set T for testing to obtain the service success rate prediction model.
The feature vector reported by the terminal of the Internet of things comprises the following parameters:
x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, …, xn, and y; wherein, the liquid crystal display device comprises a liquid crystal display device,
the method comprises the steps that x1 is data packet length, x2 is QoS grade, x3 is transmission time, x4 is transmission delay offset, x5 is concurrent terminal number, x6 is end-to-end service delay, x7 is retransmission times, x8 is signal-to-interference-plus-noise ratio (SINR), x9 is Reference Signal Received Power (RSRP), and x10 to xn are extensible fields; and y is the success rate of the service.
The determining, based on the service success rate prediction model obtained by training and the feature vector reported by the internet of things terminal, a configuration parameter for enabling a function of the service success rate prediction model to obtain a maximum value includes:
classifying the feature vector data reported by the Internet of things terminal according to the base station identifier, and inquiring a corresponding service success rate prediction model based on the classification result; each type of feature vector data corresponds to one base station and corresponds to one business success rate prediction model;
and determining configuration parameters which enable functions of the business success rate prediction model to obtain maximum values based on the feature vector data of each type and the corresponding business success rate prediction model.
The configuration parameters include the following parameters:
e 0 ’，e 1 ’，e 2 ’，e3，…e n 'A'; wherein, the liquid crystal display device comprises a liquid crystal display device,
said e 0 ' to maximize the transmission delay offset of the function of the traffic success rate prediction model,
said e 1 ' to maximize the number of traffic retransmissions as a function of the traffic success rate prediction model,
said e 2 ' to maximize the function of the traffic success rate prediction model,
said e 3 ' to e n ' an extensible influence factor that maximizes the function of the business success rate prediction model.
Optionally, the method further comprises:
and when the service runs, updating the trained service success rate prediction model according to a preset period based on the feature vector reported by the Internet of things terminal in real time.
The embodiment of the invention also provides a tide effect evasion device, which comprises:
the determining module is used for determining configuration parameters which enable functions of the service success rate prediction model to obtain maximum values based on the service success rate prediction model obtained through training and the feature vectors reported by the Internet of things terminal during service operation;
and the sending module is used for sending the configuration parameters to the Internet of things terminal, and is used for the Internet of things terminal to adjust the service model and send service data in a peak shifting mode.
The embodiment of the invention also provides a tide effect evasion device, which comprises: a processor and a memory for storing a computer program capable of running on the processor,
wherein the processor is configured to execute the steps of the above method when running the computer program.
The embodiment of the invention also provides a computer readable storage medium, on which a computer program is stored, which when being executed by a processor, implements the steps of the above method.
The tidal effect avoidance method, the tidal effect avoidance device and the computer readable storage medium provided by the embodiment of the invention are characterized in that when in service operation, configuration parameters which enable functions of a service success rate prediction model to obtain maximum values are determined based on a service success rate prediction model obtained through training and a feature vector reported by an Internet of things terminal; and transmitting the configuration parameters to the Internet of things terminal, and adjusting a service model and sending service data in a peak shifting mode by the Internet of things terminal. The embodiment of the invention determines the configuration parameters for finally adjusting the service model by the terminal of the Internet of things based on the service success rate prediction model obtained by training the machine learning method and the feature vector reported by the terminal of the Internet of things so as to realize the optimal configuration of the peak-shifting transmission data mode among the terminals.
In addition, the embodiment of the invention can also update the model in time based on the latest reported feature vector to form a complete closed-loop service self-adaptive optimization flow, can better adapt to continuous changes of service scenes, can adaptively correct the prediction model when the conditions of expanding service scale, updating service flow, adding other Internet of things products newly and the like occur, and can be rapidly applied to the existing service flow through a closed-loop adjustment mechanism, thereby ensuring that the service can operate under the optimal configuration parameters.
In addition, the embodiment of the invention provides a prediction model for the service success rate by taking the service success rate as a main decision basis, a simplified function of the configuration parameter and the service success rate is obtained based on the model, and the optimal value of the configuration parameter is obtained by solving the maximum value of the function under the specified condition. The configuration parameters obtained based on the decision mechanism are the schemes with highest efficiency and lowest terminal energy consumption on the basis of ensuring the success rate of the service, can maximally utilize the network bearing capacity, and simultaneously save the terminal energy consumption.
Drawings
FIG. 1 is a schematic flow chart of a tidal effect avoidance method according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of a tidal effect avoidance method according to an embodiment of the present invention;
FIG. 3 is a schematic view of a tidal effect avoidance apparatus according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a tidal effect avoidance apparatus according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a tidal effect avoidance apparatus according to an embodiment of the present invention;
fig. 6 is a schematic diagram of a tidal effect avoidance system according to an embodiment of the present invention.
Detailed Description
The invention is described below with reference to the drawings and examples.
The embodiment of the invention provides a tidal effect avoidance method, as shown in fig. 1, which comprises the following steps:
step 101: when the service runs, determining configuration parameters which enable functions of the service success rate prediction model to obtain maximum values based on the service success rate prediction model obtained through training and feature vectors reported by the terminal of the Internet of things;
step 102: and transmitting the configuration parameters to the Internet of things terminal, and adjusting a service model and sending service data in a peak shifting mode by the Internet of things terminal.
The embodiment of the invention determines the configuration parameters for finally adjusting the service model by the terminal of the Internet of things based on the service success rate prediction model obtained by training the machine learning method and the feature vector reported by the terminal of the Internet of things so as to realize the optimal configuration of the peak-shifting transmission data mode among the terminals.
In one embodiment, as shown in fig. 2, before the service is run, the method further includes:
step 100: based on the feature vector reported by the Internet of things terminal during test operation, training by combining with a logistic regression LR method to obtain a service success rate prediction model; different service success rate prediction models correspond to different base stations.
In the embodiment of the invention, the feature vector reported by the terminal of the internet of things based on the test run is trained by combining with an LR method to obtain a service success rate prediction model, which comprises the following steps:
dividing a set of feature vectors reported by an Internet of things terminal in at least one test run period to obtain a data set S for training and a data set T for testing;
training a preset business success rate prediction model by utilizing the data set S for training to obtain a group of fitted parameters;
and generalizing the service success rate prediction model obtained through training by utilizing the data set T for testing to obtain the service success rate prediction model.
In the embodiment of the invention, the feature vector reported by the terminal of the internet of things comprises the following parameters:
x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, …, xn, and y; wherein, the liquid crystal display device comprises a liquid crystal display device,
the method comprises the steps that x1 is data packet length, x2 is QoS grade, x3 is transmission time, x4 is transmission delay offset, x5 is concurrent terminal number, x6 is end-to-end service delay, x7 is retransmission times, x8 is signal-to-interference-plus-noise ratio (SINR), x9 is Reference Signal Received Power (RSRP), and x10 to xn are extensible fields; and y is the success rate of the service.
In the embodiment of the invention, the service success rate prediction model obtained based on training and the feature vector reported by the terminal of the internet of things determine the configuration parameters for enabling the function of the service success rate prediction model to obtain the maximum value, and the configuration parameters comprise:
classifying the feature vector data reported by the Internet of things terminal according to the base station identifier, and inquiring a corresponding service success rate prediction model based on the classification result; each type of feature vector data corresponds to one base station and corresponds to one business success rate prediction model;
and determining configuration parameters which enable functions of the business success rate prediction model to obtain maximum values based on the feature vector data of each type and the corresponding business success rate prediction model.
In the embodiment of the present invention, the configuration parameters include the following parameters:
e 0 ’，e 1 ’，e 2 ’，e3，…e n 'A'; wherein, the liquid crystal display device comprises a liquid crystal display device,
said e 0 ' to maximize the transmission delay offset of the function of the traffic success rate prediction model,
said e 1 ' to maximize the number of traffic retransmissions as a function of the traffic success rate prediction model,
said e 2 ' to maximize the function of the traffic success rate prediction model,
said e 3 ' to e n ' an extensible influence factor that maximizes the function of the business success rate prediction model.
In one embodiment, the method further comprises:
and when the service runs, updating the trained service success rate prediction model according to a preset period based on the feature vector reported by the Internet of things terminal in real time.
In order to implement the above method, an embodiment of the present invention further provides a tidal effect avoidance apparatus, as shown in fig. 3, including:
the determining module 301 is configured to determine, when a service runs, a configuration parameter that makes a function of the service success rate prediction model obtain a maximum value based on a service success rate prediction model obtained by training and a feature vector reported by an internet of things terminal;
and the sending module 302 is configured to send the configuration parameters to the terminal of the internet of things, and is used for the terminal of the internet of things to adjust a service model and send service data in a peak shifting manner.
In one embodiment, as shown in fig. 4, the apparatus further comprises: the training module 300 is used for obtaining a service success rate prediction model based on the feature vector reported by the terminal of the internet of things in test operation by combining with a logistic regression LR method; different service success rate prediction models correspond to different base stations.
In the embodiment of the present invention, the training module 300 is based on the feature vector reported by the terminal of the internet of things during test operation, and combines the LR method to train to obtain the service success rate prediction model, and includes:
dividing a set of feature vectors reported by an Internet of things terminal in at least one test run period to obtain a data set S for training and a data set T for testing;
training a preset business success rate prediction model by utilizing the data set S for training to obtain a group of fitted parameters;
and generalizing the service success rate prediction model obtained through training by utilizing the data set T for testing to obtain the service success rate prediction model.
In the embodiment of the invention, the feature vector reported by the terminal of the internet of things comprises the following parameters:
x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, …, xn; wherein, the liquid crystal display device comprises a liquid crystal display device,
the data packet length of x1, the QoS class of x2, the transmission time of x3, the transmission delay offset of x4, the number of concurrent terminals of x5, the end-to-end service delay of x6, the retransmission times of x7, the signal-to-interference-plus-noise ratio SINR of x8, the reference signal received power RSRP of x9, and the expandable fields of x10 to xn.
In the embodiment of the present invention, the determining module 301 determines, based on the service success rate prediction model obtained by training and the feature vector reported by the internet of things terminal, a configuration parameter that makes a function of the service success rate prediction model obtain a maximum value, including:
classifying the feature vector data reported by the Internet of things terminal according to the base station identifier, and inquiring a corresponding service success rate prediction model based on the classification result; each type of feature vector data corresponds to one base station and corresponds to one business success rate prediction model;
and determining configuration parameters which enable functions of the business success rate prediction model to obtain maximum values based on the feature vector data of each type and the corresponding business success rate prediction model.
In the embodiment of the present invention, the configuration parameters include the following parameters:
e 0 ’，e 1 ’，e 2 ’，e3，…e n 'A'; wherein, the liquid crystal display device comprises a liquid crystal display device,
said e 0 ' to maximize the transmission delay offset of the function of the traffic success rate prediction model,
said e 1 ' to maximize the number of traffic retransmissions as a function of the traffic success rate prediction model,
said e 2 ' to maximize the function of the traffic success rate prediction model,
said e 3 ' to e n ' an extensible influence factor that maximizes the function of the business success rate prediction model.
In one embodiment, as shown in fig. 5, the apparatus further comprises: and the updating module 303 is used for updating the trained service success rate prediction model according to a preset period based on the feature vector reported by the internet of things terminal in real time when the service runs.
The embodiment of the invention also provides a tide effect evasion device, which comprises: a processor and a memory for storing a computer program capable of running on the processor,
wherein the processor, when executing the computer program, performs:
when the service runs, determining configuration parameters which enable functions of the service success rate prediction model to obtain maximum values based on the service success rate prediction model obtained through training and feature vectors reported by the terminal of the Internet of things;
and transmitting the configuration parameters to the Internet of things terminal, and adjusting a service model and sending service data in a peak shifting mode by the Internet of things terminal.
Before the service is run, the processor is further configured to execute, when running the computer program:
based on the feature vector reported by the Internet of things terminal during test operation, training by combining with a logistic regression LR method to obtain a service success rate prediction model; different service success rate prediction models correspond to different base stations.
When the feature vector reported by the internet of things terminal during test operation and the LR method is combined for training to obtain a service success rate prediction model, the processor is further used for executing the following steps when the computer program is run:
dividing a set of feature vectors reported by an Internet of things terminal in at least one test run period to obtain a data set S for training and a data set T for testing;
training a preset business success rate prediction model by utilizing the data set S for training to obtain a group of fitted parameters;
and generalizing the service success rate prediction model obtained through training by utilizing the data set T for testing to obtain the service success rate prediction model.
The feature vector reported by the terminal of the Internet of things comprises the following parameters:
x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, …, xn, and y; wherein, the liquid crystal display device comprises a liquid crystal display device,
the method comprises the steps that x1 is data packet length, x2 is QoS grade, x3 is transmission time, x4 is transmission delay offset, x5 is concurrent terminal number, x6 is end-to-end service delay, x7 is retransmission times, x8 is signal-to-interference-plus-noise ratio (SINR), x9 is Reference Signal Received Power (RSRP), and x10 to xn are extensible fields; and y is the success rate of the service.
When the service success rate prediction model obtained based on training and the feature vector reported by the internet of things terminal determine the configuration parameters which enable the function of the service success rate prediction model to obtain the maximum value, the processor is further used for executing the computer program when running:
classifying the feature vector data reported by the Internet of things terminal according to the base station identifier, and inquiring a corresponding service success rate prediction model based on the classification result; each type of feature vector data corresponds to one base station and corresponds to one business success rate prediction model;
and determining configuration parameters which enable functions of the business success rate prediction model to obtain maximum values based on the feature vector data of each type and the corresponding business success rate prediction model.
Wherein the configuration parameters include the following parameters:
e 0 ’，e 1 ’，e 2 ’，e3，…e n 'A'; wherein, the liquid crystal display device comprises a liquid crystal display device,
said e 0 ' to maximize the transmission delay offset of the function of the traffic success rate prediction model,
said e 1 ' to maximize the number of traffic retransmissions as a function of the traffic success rate prediction model,
said e 2 ' to maximize the function of the traffic success rate prediction model,
said e 3 ' to e n ' an extensible influence factor that maximizes the function of the business success rate prediction model.
The processor is further configured to execute, when the computer program is executed:
and when the service runs, updating the trained service success rate prediction model according to a preset period based on the feature vector reported by the Internet of things terminal in real time.
It should be noted that: in the device provided in the above embodiment, when the tidal effect avoidance process is performed, only the division of each program module is used for illustration, in practical application, the process allocation may be performed by different program modules according to needs, that is, the internal structure of the device is divided into different program modules, so as to complete all or part of the processes described above. In addition, the apparatus provided in the foregoing embodiments and the corresponding method embodiments belong to the same concept, and specific implementation processes of the apparatus and the corresponding method embodiments are detailed in the method embodiments, which are not described herein again.
In an exemplary embodiment, the present invention further provides a computer readable storage medium, which may be FRAM, ROM, PROM, EPROM, EEPROM, flash Memory, magnetic surface Memory, optical disk, or CD-ROM; but may be a variety of devices including one or any combination of the above-described memories, such as a mobile phone, computer, tablet device, personal digital assistant, or the like.
The embodiment of the present invention also provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs:
when the service runs, determining configuration parameters which enable functions of the service success rate prediction model to obtain maximum values based on the service success rate prediction model obtained through training and feature vectors reported by the terminal of the Internet of things;
and transmitting the configuration parameters to the Internet of things terminal, and adjusting a service model and sending service data in a peak shifting mode by the Internet of things terminal.
Before the service is run, the computer program, when run by the processor, further performs:
based on the feature vector reported by the Internet of things terminal during test operation, training by combining with a logistic regression LR method to obtain a service success rate prediction model; different service success rate prediction models correspond to different base stations.
When the feature vector reported by the internet of things terminal based on test operation and the LR method are combined for training to obtain a service success rate prediction model, the computer program is executed by the processor and further executes:
dividing a set of feature vectors reported by an Internet of things terminal in at least one test run period to obtain a data set S for training and a data set T for testing;
training a preset business success rate prediction model by utilizing the data set S for training to obtain a group of fitted parameters;
and generalizing the service success rate prediction model obtained through training by utilizing the data set T for testing to obtain the service success rate prediction model.
The feature vector reported by the terminal of the Internet of things comprises the following parameters:
x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, …, xn, and y; wherein, the liquid crystal display device comprises a liquid crystal display device,
the method comprises the steps that x1 is data packet length, x2 is QoS grade, x3 is transmission time, x4 is transmission delay offset, x5 is concurrent terminal number, x6 is end-to-end service delay, x7 is retransmission times, x8 is signal-to-interference-plus-noise ratio (SINR), x9 is Reference Signal Received Power (RSRP), and x10 to xn are extensible fields; and y is the success rate of the service.
When the service success rate prediction model obtained based on training and the feature vector reported by the internet of things terminal determine the configuration parameters which enable the function of the service success rate prediction model to obtain the maximum value, the computer program is executed by the processor and further executes the following steps:
classifying the feature vector data reported by the Internet of things terminal according to the base station identifier, and inquiring a corresponding service success rate prediction model based on the classification result; each type of feature vector data corresponds to one base station and corresponds to one business success rate prediction model;
and determining configuration parameters which enable functions of the business success rate prediction model to obtain maximum values based on the feature vector data of each type and the corresponding business success rate prediction model.
Wherein the configuration parameters include the following parameters:
e 0 ’，e 1 ’，e 2 ’，e3，…e n 'A'; wherein, the liquid crystal display device comprises a liquid crystal display device,
said e 0 ' to maximize the transmission delay offset of the function of the traffic success rate prediction model,
said e 1 ' to maximize the number of traffic retransmissions as a function of the traffic success rate prediction model,
said e 2 ' to maximize the function of the traffic success rate prediction model,
said e 3 ' to e n ' an extensible influence factor that maximizes the function of the business success rate prediction model.
The computer program, when executed by the processor, further performs:
and when the service runs, updating the trained service success rate prediction model according to a preset period based on the feature vector reported by the Internet of things terminal in real time.
The invention is described below in connection with scene embodiments.
The embodiment relates to two types of fields in application, one type is a key field carried in a data packet in a service interaction flow of a terminal, and the other type is a field value calculated by a background server according to historical data. These fields are all influencing factors that influence the success rate of service interaction, and will be used as feature vectors in the machine learning model input space, as shown in the following table 1:
TABLE 1
It should be noted that, the descriptive field in the above fields is an immutable field, which is used in the data processing process, but does not belong to the feature vector of the input space of the prediction model. Feature vectors are variables that affect the predictive model.
When the analysis finds that the client (the terminal of the Internet of things) performs service interaction with the background, the service success rate (y) is commonly influenced by multidimensional factors, wherein the factors influencing the success rate are expressed as feature vectors (x 1, x2, x3, x4, x5, x6, x7, x8, x9, x10, … and xn), wherein x1 is the data packet length, x2 is the QoS grade, x3 is the transmission time, x4 is the transmission delay offset, x5 is the concurrent terminal number, x6 is the end-to-end service delay, x7 is the retransmission times, x8 is the SINR, x9 is the RSRP, and x10 to xn are the extensible fields (which can be self-determined in combination with specific services).
And predicting the service success rate y of the whole system by adopting an LR method, and then firstly obtaining a hypothesis function model:
here y=hθ (x), θ being a variable parameter of the model.
When the system is on line, the system is firstly operated in an existing random delay time mode, and training data is accumulated through a test operation period (generally more than 2 weeks). The resulting set of feature vectors M (x 1, x2, x3, x4, x5, x6, …, xn, y) is divided into a set of data S for training, a set of data T for testing.
And training the machine learning model by using the training set S to obtain a group of parameters theta, and realizing the generalization process of the model by using the testing set T to finally obtain the applicable business success rate prediction model.
A group of configuration parameter vectors (e 0, e1, e2, e3 and … en) which affect the success rate of the service at the terminal side are set, wherein e0 is a transmission delay offset, e1 is the service retransmission times, e2 is a service platform address, and e3 to en are extensible influencing factors according to specific service types. Based on the function of the service success rate prediction model, a group of configuration parameters (e 0', e1', e2', e3', …, en ') which can enable the function to obtain maximum value points are obtained, the parameter combination is issued to the terminal side of the Internet of things, and the terminal of the Internet of things adjusts the service receiving and dispatching model according to the parameter combination, so that service concurrency is reduced to the greatest extent on the premise of meeting service quality, and service quality is improved.
In the application process, the machine learning model can be continuously adaptively optimized due to the continuous input data set, so that the parameter prediction effect is ensured not to deviate greatly along with the change of the service scale and the network configuration.
The overall system composition and business flow of this embodiment are shown in fig. 6, and include: the machine learning prediction service module, the parameter decision module, the service platform and the client (the terminal of the Internet of things) are characterized in that the specific service execution steps are as follows:
the whole business is divided into two stages, wherein the first stage is a test operation modeling stage, and the second stage is an online use stage; wherein, the liquid crystal display device comprises a liquid crystal display device,
the test run modeling phase flow comprises the following steps:
step one: the client presets the existing random time delay tidal prevention method;
step two: the client runs with the service for at least 2 weeks;
step three: the service platform reports the feature vector reported by the client side in the test run to the machine learning prediction service module through an interface;
step four: the machine learning prediction service module adopts an LR method to fit the service success rate, and trains by utilizing a data set S for training in the feature vector to obtain a group of fitted parameters theta;
step five: and testing and verifying the obtained prediction model function through a data set T used for testing in the feature vector, completing generalization to enable the prediction model function to have universal applicability, and completing preliminary establishment of the business success rate prediction model.
The online operation phase flow comprises the following steps:
step one: the client carries the necessary feature vectors and description fields in the table 1 when the service interaction is performed normally;
step two: the service platform receives the data, processes and packages the data and sends the data to the parameter decision module;
step three: the parameter decision module classifies the reported data (the data corresponding to the same base station is divided into one type) according to the description variable base station ECI, and queries the corresponding function of the latest business success rate prediction model from the machine learning prediction service module for each type of data;
step four: the parameter decision module uses the function of the service success rate prediction model, takes the variable in the preset configuration parameter vector as an adjustable parameter, takes other feature vectors as fixed parameters into the function, calculates the maximum value of the function, and ensures that the time delay offset value obtained by (1) the service success rate meeting the service threshold (such as 95%) requirement (2) is minimum, and the configuration parameter vector is the optimal configuration parameter vector obtained by the algorithm.
Step five: the parameter decision module feeds back the optimal configuration parameter vector as a response value to the service platform;
step six: the service platform returns the optimal configuration parameter vector value and response values required by other services to the client;
step seven: after receiving the return message, the client adjusts the business receiving and transmitting model according to the optimal configuration parameter vector pushed by the platform side, and then starts the next business flow. Through optimizing the parameters, the effects of avoiding the network tide effect and improving the service success rate are achieved.
In the scheme, the service platform also periodically transmits the service data (feature vector) reported by the received client to the machine learning service platform, continuously updates the data source of machine learning, realizes the self-adaptive update of the service success rate prediction model, and ensures the high availability of the model function.
The embodiment of the invention determines the configuration parameters for finally adjusting the service model by the Internet of things terminal based on the service success rate prediction model obtained by training the machine learning method and the feature vector reported by the Internet of things terminal, and can update the model in time based on the latest reported feature vector, thereby being a complete closed-loop service self-adaptive optimization process, being capable of better adapting to continuous changes of service scenes, being capable of self-adaptively correcting when the conditions of expanding service scale, updating service process, newly adding other Internet of things products and the like occur, and being rapidly applied to the existing service process by a closed-loop adjustment mechanism, thereby ensuring that the service can operate under the optimal configuration parameters. Compared with a custom business optimization means aiming at specific products, the method has the advantages of stronger adaptability, lower maintenance cost and better optimization effect.
In addition, the embodiment of the invention provides a method for establishing a prediction model for the service success rate by taking the service success rate as a main decision basis, obtaining a simplified function of the configuration parameter and the service success rate based on the model, and obtaining an optimal value of the configuration parameter by solving a function maximum under a specified condition. The configuration parameters obtained based on the decision mechanism are the schemes with highest efficiency and lowest terminal energy consumption on the basis of ensuring the success rate of the service, can maximally utilize the network bearing capacity, and simultaneously save the terminal energy consumption.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the present invention.
Claims (9)
1. A tidal effect avoidance method, the method comprising:
when the service runs, the feature vector data reported by the terminal of the Internet of things is classified according to the base station identifier, and a corresponding service success rate prediction model is inquired based on the classification result; each type of feature vector data corresponds to one base station and corresponds to one business success rate prediction model;
determining configuration parameters for enabling functions of the business success rate prediction model to obtain maximum values based on the feature vector data of each type and the corresponding business success rate prediction model;
and transmitting the configuration parameters to the Internet of things terminal, and adjusting a service model and sending service data in a peak shifting mode by the Internet of things terminal.
2. The method of claim 1, wherein prior to the operation of the service, the method further comprises:
based on the feature vector reported by the Internet of things terminal during test operation, training by combining with a logistic regression LR method to obtain a service success rate prediction model; different service success rate prediction models correspond to different base stations.
3. The method of claim 2, wherein the feature vector reported by the terminal of the internet of things based on the test run time is trained by combining an LR method to obtain a service success rate prediction model, and the method comprises the following steps:
dividing a set of feature vectors reported by an Internet of things terminal in at least one test run period to obtain a data set S for training and a data set T for testing;
training a preset business success rate prediction model by utilizing the data set S for training to obtain a group of fitted parameters;
and generalizing the service success rate prediction model obtained through training by utilizing the data set T for testing to obtain the service success rate prediction model.
4. The method according to claim 1 or 2, wherein the feature vector reported by the internet of things terminal includes the following parameters:
x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, …, xn, and y; wherein, the liquid crystal display device comprises a liquid crystal display device,
the method comprises the steps that x1 is data packet length, x2 is QoS grade, x3 is transmission time, x4 is transmission delay offset, x5 is concurrent terminal number, x6 is end-to-end service delay, x7 is retransmission times, x8 is signal-to-interference-plus-noise ratio (SINR), x9 is Reference Signal Received Power (RSRP), and x10 to xn are extensible fields; and y is the success rate of the service.
5. The method of claim 1, wherein the configuration parameters include the following parameters:
e 0 ’，e 1 ’，e 2 ’，e 3 ’，…e n 'A'; wherein, the liquid crystal display device comprises a liquid crystal display device,
said e 0 ' to maximize the transmission delay offset of the function of the traffic success rate prediction model,
said e 1 ' to maximize the number of traffic retransmissions as a function of the traffic success rate prediction model,
said e 2 ' to maximize the function of the traffic success rate prediction model,
said e 3 ' to e n ' an extensible influence factor that maximizes the function of the business success rate prediction model.
6. The method according to claim 1, characterized in that the method further comprises:
and when the service runs, updating the trained service success rate prediction model according to a preset period based on the feature vector reported by the Internet of things terminal in real time.
7. A tidal effect avoidance apparatus, the apparatus comprising:
the determining module is used for classifying the feature vector data reported by the terminal of the Internet of things according to the base station identifier and inquiring the corresponding service success rate prediction model according to the classification result when the service runs; each type of feature vector data corresponds to one base station and corresponds to one business success rate prediction model;
determining configuration parameters for enabling functions of the business success rate prediction model to obtain maximum values based on the feature vector data of each type and the corresponding business success rate prediction model;
and the sending module is used for sending the configuration parameters to the Internet of things terminal, and is used for the Internet of things terminal to adjust the service model and send service data in a peak shifting mode.
8. A tidal effect avoidance apparatus, the apparatus comprising: a processor and a memory for storing a computer program capable of running on the processor,
wherein the processor is adapted to perform the steps of the method of any of claims 1-6 when the computer program is run.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, carries out the steps of the method according to any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910000461.1A CN111405583B (en) | 2019-01-02 | 2019-01-02 | Tidal effect avoidance method, tidal effect avoidance device and computer-readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910000461.1A CN111405583B (en) | 2019-01-02 | 2019-01-02 | Tidal effect avoidance method, tidal effect avoidance device and computer-readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111405583A CN111405583A (en) | 2020-07-10 |
| CN111405583B true CN111405583B (en) | 2023-06-30 |
Family
ID=71435896
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910000461.1A Active CN111405583B (en) | 2019-01-02 | 2019-01-02 | Tidal effect avoidance method, tidal effect avoidance device and computer-readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111405583B (en) |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105516325A (en) * | 2015-12-18 | 2016-04-20 | 内蒙古农业大学 | Cloud load balancing method for carrying out elastic expansion and traffic distribution expansion according to application load |
| CN107277927B (en) * | 2016-04-08 | 2023-04-25 | 中兴通讯股份有限公司 | Processing method of user text perception service and network equipment |
| CN106304290A (en) * | 2016-08-12 | 2017-01-04 | 辛建芳 | Internet of Things cooperative node Poewr control method based on N strategy |
| CN107220217A (en) * | 2017-05-31 | 2017-09-29 | 北京京东尚科信息技术有限公司 | Characteristic coefficient training method and device that logic-based is returned |
| CN107172623A (en) * | 2017-06-01 | 2017-09-15 | 国网浙江省电力公司嘉兴供电公司 | A kind of TD LTE230 wireless private networks frequency resource planing method |
| CN108769164B (en) * | 2018-05-17 | 2019-06-21 | 重庆电子工程职业学院 | A kind of dispositions method of marine monitoring Internet of Things three-dimensional space probe node |
| CN108471625A (en) * | 2018-07-03 | 2018-08-31 | 徐州立讯信息科技有限公司 | Internet of Things radio communication quality automatic monitoring method and the equipment for realizing this method |
-
2019
- 2019-01-02 CN CN201910000461.1A patent/CN111405583B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN111405583A (en) | 2020-07-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110809306B (en) | Terminal access selection method based on deep reinforcement learning | |
| US8073381B2 (en) | Radio communication system and base station | |
| CN110139325B (en) | Network parameter tuning method and device | |
| CN112492698A (en) | Multilink low-delay communication method, device, storage medium and electronic device | |
| EP2150081A2 (en) | Base station and resources securing method | |
| CN105992252A (en) | Processing method and apparatus for context of UE | |
| CN113904948B (en) | 5G network bandwidth prediction system and method based on cross-layer multidimensional parameters | |
| CN105636062A (en) | Cognitive radio network transmission learning method for moderate business services | |
| KR20190042708A (en) | Adaptive Modulation and Coding Method and Base Station | |
| CN111405583B (en) | Tidal effect avoidance method, tidal effect avoidance device and computer-readable storage medium | |
| US7512411B2 (en) | Radio communication system, radio network controller, mobile station and down link transmission power control method | |
| CN110875756B (en) | Method and equipment for automatically adjusting transmitting power in frequency hopping communication | |
| US20140314013A1 (en) | Method and apparatus for mobile terminal to switch base station | |
| CN111130698B (en) | Wireless communication receiving window prediction method and device and wireless communication equipment | |
| CN114826427B (en) | Optical network receiving terminal | |
| CN108702584A (en) | Service optimization processing method, equipment and system | |
| Khuntia et al. | An actor-critic reinforcement learning for device-to-device communication underlaying cellular network | |
| US20120021798A1 (en) | Communication device and its transmission power control method in radio communications system | |
| KR20110009195A (en) | Radio communication device and radio communication method | |
| CN108476545A (en) | Control method, device and the base station of radio resource control RRC state | |
| CN113765703A (en) | Data transmission switching control method and system of intelligent gateway | |
| CN106533629A (en) | Switching method of delay mode, switching device, terminal and base station | |
| CN103812627B (en) | A kind of for sending the method for schedule information, device and base station | |
| CN103596216B (en) | Method and apparatus for determining cell capacity enhancing capability | |
| CN104429144A (en) | A resource scheduling method, an information interaction method and a device |
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 |