SA network load prediction method, device and electronic equipment
By acquiring and analyzing the target sample data of the SA network and using the prediction model to predict and warn load indicators, the challenges of SA network load management are solved and the Internet access experience and network quality of 5G customers are ensured.
Patent Information
- Application Number
- CN202110688324.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-06-21
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2041-06-21
AI Technical Summary
How to effectively predict and manage loads in SA networks to ensure the Internet access experience of 5G customers, especially when the SA network load situation is closely related to the Internet access rate of 5G customers.
By obtaining the target sample data of the SA network, the corresponding load indicators are determined, and the first prediction model and the second prediction model are used to predict the load indicators in combination with the target sample data, and early warning is made based on the predicted value to avoid load abnormalities.
Through prediction and early warning mechanisms, SA network load abnormalities are handled in a timely manner, ensuring the Internet experience of 5G customers, and improving users' experience of 5G networks.
Smart Images

Figure CN115515171B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of information and communication technology, and in particular to a load prediction method, device and electronic equipment for an SA network. Background Art
[0002] The fifth generation mobile communication technology (5th Generation Mobile Communication Technology, 5G) is a new generation of broadband mobile communication technology with high speed, low latency and large connection. 5G is divided into two types of networking, namely non-standalone networking (NSA) and standalone networking (SA). Among them, SA refers to the construction of a new network, including new base stations, backhaul links and core networks, so as to realize all the characteristics and functions of the 5G network and provide users with a high-speed Internet experience. For SA networks, the load situation in the SA network is closely related to the Internet speed of 5G customers. How to predict the load in the SA network to ensure the Internet experience of 5G customers is a technical problem that technicians in this field need to solve. Summary of the invention
[0003] The purpose of the embodiments of the present application is to provide a load prediction method, device and electronic device for an SA network, so as to predict the load in the SA network and ensure the Internet experience of 5G customers.
[0004] In order to solve the above technical problems, the embodiments of the present application are implemented as follows:
[0005] In a first aspect, an embodiment of the present application provides a load prediction method for an SA network, including:
[0006] Obtain target sample data in the SA network;
[0007] Based on the target sample data, determining a corresponding load indicator in the SA network;
[0008] Using at least a portion of the target sample data as an input variable of a first prediction model, predicting a predicted value of the indicator-related quantity at a specified future time;
[0009] Using the predicted value of the index-related quantity as an input variable of the second prediction model, predicting the first index value of the load index at the specified future time;
[0010] An early warning is issued to the SA network according to the first indicator value.
[0011] In a second aspect, an embodiment of the present application provides a load prediction device for an SA network, including:
[0012] An acquisition module, used to acquire target sample data in the SA network;
[0013] A determination module, configured to determine a corresponding load index for the SA network based on the target sample data;
[0014] A first prediction module, configured to use at least a portion of the target sample data as an input variable of a first prediction model to predict a predicted value of an indicator-related quantity at a specified time in the future;
[0015] A second prediction module, configured to use the predicted value of the indicator-related quantity as an input variable of a second prediction model to predict a first indicator value of the load indicator at the specified future time;
[0016] An early warning module is used to issue an early warning to the SA network according to the first indicator value.
[0017] In a third aspect, an embodiment of the present application provides an electronic device, comprising a processor, a communication interface, a memory and a communication bus; wherein the processor, the communication interface and the memory communicate with each other through a bus; the memory is used to store computer programs; the processor is used to execute the programs stored in the memory to implement the steps of the load prediction method for the SA network as described in the first aspect.
[0018] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the load prediction method for the SA network as described in the first aspect are implemented.
[0019] It can be seen from the technical solution provided by the above embodiments of the present application that the target sample data in the SA network is obtained, and based on the target sample data, the corresponding load index is determined for the SA network, and at least a part of the target sample data is used as the input variable of the first prediction model to predict the predicted value of the index-related quantity at a specified time in the future. The predicted value of the index-related quantity is used as the input variable of the second prediction model to predict the first index value of the load index at a specified time in the future, and the SA network is warned according to the first index value. The first prediction model and the second prediction model are combined with the target sample data to predict the first index value of the load index at a specified time in the future, and then the SA network is warned according to the first index value. Therefore, when the predicted load index in the SA network is abnormal, an early warning can be issued, so that the maintenance personnel can deal with it in time, avoiding the abnormal load of the SA network, ensuring the Internet experience of 5G customers, and improving the user experience of the 5G network. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0021] Figure 1 A schematic diagram of a first flow chart of a load forecasting method for an SA network provided in an embodiment of the present application;
[0022] Figure 2 An architectural diagram of a first prediction model provided in an embodiment of the present application;
[0023] Figure 3 A second flow chart of the load forecasting method for the SA network provided in an embodiment of the present application;
[0024] Figure 4 A third flow chart of the load forecasting method for the SA network provided in an embodiment of the present application;
[0025] Figure 5 A schematic diagram of the module composition of the load prediction device of the SA network provided in an embodiment of the present application;
[0026] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0027] The embodiments of the present application provide a load prediction method, device and electronic device for an SA network to predict the load in the SA network and ensure the Internet experience of 5G customers.
[0028] In order to enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.
[0029] In some scenarios, the SA network architecture includes 5G virtualization functions (VNFs). VNFs can be virtualized network elements to implement functions in the SA network. For example, the access and mobility management function (AMF) can correspond to the mobility management network element, which is responsible for user mobility and access management, the session management function (SMF) can correspond to the session management network element, which is responsible for user session management, the policy control function (PCF) can correspond to the policy control network element, which is responsible for policy control, the user plane function (UPF) can correspond to the user plane network element, which is responsible for user plane routing and forwarding functions, the unified data management function (UDM) can correspond to the unified data management network element, which is responsible for unified data processing, including user identification, user contract data and authentication data, the billing function (CHF) can correspond to the billing network element, which is responsible for online billing, offline billing and converged billing, the network slice selection function (NSSF) can correspond to the network slice selection network element, which is responsible for managing network slice-related information, and the network open function (NSF) can correspond to the network slice selection network element. ExposureFunction, NEF), which can correspond to the network open network element, responsible for managing the open network data, network repository function (NRF), which can correspond to the network storage network element, responsible for registering and managing various network functions, application function (AF), which can correspond to the application network element, responsible for various services of the application layer, user equipment (UE), data network (DN), which can correspond to the data network element, responsible for operator services, Internet access services or third-party services, radio access network (RAN) such as gNodeB base station, etc.
[0030] Among them, UE is connected to AMF through N1 interface, RAN is connected to AMF through N2 interface, RAN is connected to UPF through N3 interface, UPF is connected to SMF through N4 interface, UPF is connected to DN through N6 interface, UPF and UPF can be connected through N9 interface (such as N9a, N9c interface, etc.). PCF and AF are connected through N5 interface.
[0031] For SA networks, the SA network load is closely related to the Internet speed of 5G customers. How to predict the load in the SA network to ensure the Internet experience of 5G customers is a technical problem that technicians in this field need to solve.
[0032] To this end, the embodiments of the present application provide a technical solution that can solve the above-mentioned problems, and the details can be found in the following content.
[0033] For example, Figure 1 As shown, an embodiment of the present application provides a load prediction method for an SA network, and the executor of the method may be a server, wherein the server may be an independent server or a server cluster composed of multiple servers, and the server may be a server capable of performing load prediction for the SA network.
[0034] The load forecasting method of the SA network may specifically include the following steps:
[0035] In S101, target sample data in the SA network is acquired.
[0036] Specifically, the target sample data in the SA network include but are not limited to: 5G terminal data (such as the total number of 5G terminals, 5G terminal promotion plan and user replacement rhythm, the proportion of 5G soft switches on 5G terminals, the SA network standard priority version upgrade progress of 5G terminals, etc.), gNodeB data (such as the number of gNodeB base stations and coverage areas, gNodeB base station construction plan, etc.), 5G VNF network element business volume (such as AMF registered users, AMF attachment, 5G base station number, SMF peak session number, PCF contracted users, SMF / UPF peak service quality (Quality of Service, QoS) flow number, N3 / N6 interface traffic, maximum number of allocated addresses, N3 / N9a / N9c / N6 interface receiving a set of General Packet Radio Service (GPRS) tunneling protocol (GPRS Tunnelling The data includes the number of 5G packets in the GTP (GTP) protocol, the number of calls per unit time in the CHF, and the number of 5G users on the UDM), signaling resources (such as the number of signaling times and signaling load), capacity-related statistical data (such as the registered capacity license of core network elements such as AMF, SMF, UDM, PCF, the N3 / N6 interface bandwidth and maximum data throughput of the UPF, the maximum number of QoS flows of the SMF / UPF, etc.), and the usage of telecom cloud resources (that is, the usage of computing resources, storage resources, and network resources, such as disk usage, server CPU utilization, and memory utilization).
[0037] In a possible implementation, after acquiring the target sample data in the SA network and before determining the corresponding load index for the SA network based on the target sample data, the following steps are further included:
[0038] The target sample data is preprocessed, including data cleaning and null value interpolation.
[0039] Specifically, the target sample data is cleaned to remove data that does not conform to the actual situation of the existing network to improve the accuracy of the target sample data and further improve the prediction accuracy of the prediction model. For example, data with a weekday user number fluctuation exceeding the first threshold (such as 10%) or a traffic fluctuation exceeding the second threshold (such as 15%), and data with a holiday user number fluctuation exceeding the third threshold (such as 20%) or a traffic fluctuation exceeding the fourth threshold (such as 30%) are removed. Null value interpolation is performed on the target data sample, that is, interpolation is performed through the median, mean or mode.
[0040] In S102 , a corresponding load index is determined for the SA network based on the target sample data.
[0041] Specifically, the core network elements in the SA network play an important role in the user's Internet access speed. In a possible implementation, when calculating the load index in the SA network, the load index of the core network elements in the SA network can be used as the load index in the SA network. Specifically, based on the types of target sample data above, capacity indicators such as AMF registration capacity utilization, AMF base station management capacity utilization, SMF session capacity peak utilization, SMF peak QoS flow capacity utilization, UPF peak QoS flow capacity utilization, UDM user capacity utilization and PCF user capacity utilization can be calculated, as well as load indicators such as address pool peak utilization, UPF data throughput capacity utilization, traffic bandwidth utilization (such as N3 / N6 interface traffic bandwidth utilization), signaling load, and telecom cloud resource utilization.
[0042] AMF registration capacity utilization, AMF base station management capacity utilization, SMF session capacity peak utilization, SMF peak QoS flow capacity utilization, UPF peak QoS flow capacity utilization, UDM user capacity utilization, CHF call record processing capacity utilization and PCF user capacity utilization can be collectively referred to as capacity utilization.
[0043] Among them, AMF registration capacity utilization = number of AMF registered users / AMF registration capacity License. AMF registration capacity utilization is a key capacity indicator of AMF. If the AMF registration capacity utilization is too high, users may not be able to successfully register to the 5G SA network.
[0044] AMF base station management capacity utilization = number of 5G base stations connected to AMF / maximum number of supported base stations. The number of base stations supported by AMF for connection is usually limited, so the capacity of AMF base station management needs to be considered when building a new gNodeB base station.
[0045] SMF session capacity peak utilization = peak PDU session number / maximum PDU session number. SMF session capacity peak utilization is the core capacity load indicator of SMF and is a content that must be paid attention to in network planning to understand whether there will be business overflow. Among them, in the 5G network, the information interaction between the user terminal and the network is carried out through the PDU session. The PDU session provides data connectivity between the 5G UE and DN. All data traffic of the UE must be carried through the PDU session. Before the UE sends data traffic, the PDU session should be established first.
[0046] SMF peak QoS flow capacity utilization = peak QoS flow number / maximum QoS flow number.
[0047] UPF peak QoS flow capacity utilization = peak QoS flow number / maximum QoS flow number. A PDU session can include up to 64 QoS flows, and each QoS flow has different TFT parameters. The SMF peak QoS flow capacity utilization and UPF peak QoS flow capacity utilization are mainly used to determine whether there is a service overflow situation.
[0048] UDM user capacity utilization = number of 5G users / maximum number of storable mobile subscriber numbers (MobileSubscriber International ISDN Number, MSISDN), which is used to understand the trend of UDM storage capacity utilization and prepare for network expansion in advance.
[0049] PCF user capacity utilization = number of PCF subscribed users / maximum number of PCF subscribed users, which is the ratio of the number of subscribed users actually stored in the PCF UDR to its capacity. This indicator is used to understand whether there are any business restrictions.
[0050] CHF call bill processing capacity utilization = CHF call bill volume per unit time / maximum call bill processing capacity. Call bills are the source of revenue for telecom operators. Correct handling of call bills is important. Call bill backlogs may result in restricted user services and revenue losses. Therefore, it is necessary to pay attention to CHF call bill processing capacity utilization to avoid overload operation.
[0051] Address pool peak utilization = maximum number of allocated addresses / total number of IP addresses in the address pool, which is the ratio of the maximum number of addresses actually used (allocated) in the address pool to the total number of addresses in the address pool. This indicator is used to understand whether there is any service restriction.
[0052] UPF data throughput capacity utilization = (number of GTP packet bytes received by N3 interface + number of GTP packet bytes received by N9a interface + number of GTP packet bytes received by N9c interface + number of bytes received by N6 interface) * 8 / (maximum data throughput rate * 1000). This indicator is used to understand whether there is a bottleneck in the data throughput capacity of the entire UPF machine.
[0053] In S103, at least a portion of the target sample data is used as an input variable of the first prediction model to predict the index-related quantity predicted value at a specified time in the future.
[0054] Specifically, the first prediction model is a trained prediction model, and the first prediction model can be a BP neural network model. The BP neural network model adopts a three-layer structure, and its structure is as follows: Figure 2 As shown, where x i They represent input variables, i = 1, 2, 3, ..., n, where n is the number of input variables. For example, when n is 7, the input variables of the BP neural network model can be the total number of 5G terminals at a certain point in time, the 5G terminal promotion plan and the user replacement rhythm, the 5G soft switch opening ratio of 5G terminals, the version upgrade progress of the SA network format priority of 5G terminals, the coverage of gNodeB, AMF
[0055] ^Number of registered users and N6 interface traffic. That is, the input layer of the BP neural network model has 7 neurons, and Y is the output of the BP neural network model, which is the indicator-related quantity at a specified time in the future (such as the predicted value of the number of 5G users and the predicted value of 5G traffic). The number of neurons in the middle layer of the BP neural network model can be 14, the hidden layer uses the sigmoid function, and the output layer function uses y=x.
[0056] The input and output relationship of the BP neural network model can be expressed as follows:
[0057]
[0058] is the weight from the mth neuron in the middle layer to the neuron in the output layer, tansig represents the sigmoid function, is the weight from the neuron corresponding to the i-th value of the input to the m-th neuron in the middle layer, is the threshold of the mth neuron in the middle layer, b o is the threshold of the output layer neurons. Indicates the predicted value of the indicator-related quantity being predicted.
[0059] In a possible implementation, at least a portion of the target sample data is used as an input variable of the first prediction model, and the predicted value of the indicator-related quantity at a specified time in the future is predicted to include:
[0060] 5G terminal data (total number of 5G terminals, 5G terminal promotion plan and user replacement rhythm, 5G soft switch opening ratio of 5G terminals, version upgrade progress of 5G terminal SA network standard priority), gNodeB base station coverage data, number of AMF registered users and N6 interface traffic are used as input variables of the BP neural network model.
[0061] Predict the number of 5G users and 5G traffic. Predict the number of 5G users and 5G traffic at a specified time in the future.
[0062] Taking the predicted value of the index-related quantity as the input variable of the second prediction model, the first index value of the predicted load index at a specified time in the future includes:
[0063] The predicted value of the number of 5G users and the predicted value of 5G traffic are used as input variables of the binary linear regression model to predict the first indicator value of the load index at a specified time in the future.
[0064] In a possible implementation, in order to reduce the prediction error of the BP neural network model and improve the prediction accuracy, the BP neural network model will back-propagate data from the output layer to the input layer, readjust the value of the weight w, and stop training until the prediction error of the BP neural network model reaches the minimum, thus completing the training of the BP neural network model. After predicting the indicator-related value of the load indicator at a specified time in the future, the method also includes:
[0065] Based on the predicted value of the indicator-related quantity and the actual value of the indicator-related quantity in the target sample data, the accuracy score of the first prediction model is calculated.
[0066] Specifically, the accuracy score of the first prediction model can be calculated based on the following formula:
[0067]
[0068] Among them, R 2 represents the accuracy score of the first prediction model, represents the predicted value of the indicator-related quantity output by the first prediction model, Indicates the actual value of the indicator-related quantity, N S Represents the sample size input into the first prediction model.
[0069] If the accuracy score does not meet the preset requirements, re-predict the indicator-related quantities at the specified future time.
[0070] In S104, the predicted value of the index-related quantity is used as an input variable of the second prediction model to predict the first index value of the load index at a specified time in the future.
[0071] Specifically, as indicator-related quantities, there is a significant linear correlation between them and the load indicators. For example, the number of 5G users and 5G traffic have a significant linear correlation with various load indicators.
[0072] In one possible implementation, the classic binary linear regression model is used as the second prediction model to reflect the linear relationship, that is, after the BP neural network model is used to obtain the predicted values of the number of 5G users and 5G traffic at a specified time in the future, the predicted values of each load indicator are further calculated based on the binary linear regression model. The details are as follows:
[0073] Assume that the dependent variable Y (each load index) and the independent variable X (the predicted value of the index-related quantity, such as X1 and X2 representing the predicted value of the number of 5G users and the predicted value of 5G traffic, respectively) satisfy the following formula:
[0074] Y=β0+β1X1+β2X2+ε
[0075] Among them, β0, β1 and β2 are binary regression parameters, ε is the statistical error, which obeys the normal distribution with mean zero, that is, ε:N(0,σ 2 ).
[0076] By conducting n experiments on (X1, X2, Y), we can obtain a sample with a capacity of n and a finite sample model, as shown below:
[0077]
[0078] According to the least squares method, the least squares estimate of β can be obtained As shown below:
[0079]
[0080] The above formula is the binary regression prediction model for load indicators such as capacity utilization, address pool utilization, traffic bandwidth utilization, signaling load, data throughput, and telecom cloud resource utilization. Through this binary regression prediction model, the least squares estimate X is a known quantity, and the first index value of each load index can be predicted through the binary regression model.
[0081] In S105, an early warning is issued to the SA network according to the first indicator value.
[0082] Specifically, whether the load indicators corresponding to the first indicator values are abnormal is determined according to the magnitude of the first indicator values, and an early warning is issued when the load indicators are abnormal.
[0083] In one possible implementation, Figure 3 As shown, S105 specifically includes the following steps:
[0084] In S1051, for the SA network, a corresponding load level is determined according to the first indicator value.
[0085] Specifically, in one possible implementation, the load level can be divided into five levels. The first load level can be expressed as the overall resource shortage of the SA network. In the first load level, indicator parameters such as capacity utilization, address pool peak utilization, traffic bandwidth utilization, signaling load, UPF data throughput capacity utilization, and telecom cloud resource utilization are all at a high level (i.e., they exceed their respective corresponding warning thresholds). The second load level can be expressed as the network capabilities of the SA network (such as core network elements), which can meet the user experience. The capacity utilization, address pool peak utilization, traffic bandwidth utilization, signaling load, and UPF data throughput capacity utilization are at a low level (none exceeds their respective corresponding warning thresholds), and the telecom cloud resource utilization is at a high level (exceeds the corresponding warning thresholds).
[0086] The third load level can be expressed as insufficient software capacity license or address pool resources (such as core network elements) of the SA network, which may lead to limited 5G services. In the third load level, there is a situation where at least one load indicator in the capacity utilization and / or address pool peak utilization load indicators exceeds the warning threshold, but the traffic bandwidth utilization, signaling load, UPF data throughput capacity and telecom cloud resource utilization are all at a low level (none exceeds their corresponding warning thresholds).
[0087] The fourth load level may indicate that the traffic bandwidth or data throughput or forwarding processing capacity in the SA network is insufficient (such as core network elements). In this fourth load level, the telecom cloud resource utilization, software capacity license and address pool resources are sufficient, but at least one load indicator of the N3 / N6 interface traffic bandwidth utilization, signaling load or UPF data throughput capacity utilization exceeds the warning threshold.
[0088] The fifth load level may be represented as a surplus of resources in the SA network, in which all load indicators are lower than the warning threshold.
[0089] In S1052, an early warning work order corresponding to the load level is issued.
[0090] After determining the corresponding load level, an early warning work order corresponding to each load level is issued. Specifically, the early warning work order carries the abnormal load index in the corresponding load level. According to the level of the load level, the load level and the early warning work order can be marked, and the importance of each load level is marked in turn according to the level of the load level. The lower the load level, the higher the importance of the early warning work order, so that maintenance personnel can handle it first. For example, the early warning work order corresponding to the first load level is marked with the number 1, indicating the response level, such as the first level response, indicating that it is the most important early warning work order and needs to be processed urgently. Correspondingly, the second load level to the fifth load level are marked as 2-5, indicating the response levels of the second level response, the third level response, the fourth level response and the fifth level response, respectively. Among them, when the early warning work order is marked as 5, the resources of the SA network are abundant and no early warning work order can be issued.
[0091] It can be seen from the technical solution provided by the above embodiments of the present application that the first indicator value of the load indicator at a specified time in the future is predicted by combining the first prediction model and the second prediction model with the target sample data, and then the SA network is warned according to the first indicator value. Therefore, when the predicted load indicator of the SA network is abnormal, an early warning can be issued, so that the maintenance personnel can deal with it in time, avoiding the abnormal load of the SA network, ensuring the Internet experience of 5G customers, and improving the user experience of the 5G network.
[0092] For example, Figure 4 As shown, an embodiment of the present application provides a load prediction method for an SA network, and the executor of the method may be a server, wherein the server may be an independent server or a server cluster composed of multiple servers, and the server may be a server capable of performing load prediction for the SA network.
[0093] The load forecasting method of the SA network may specifically include the following steps:
[0094] In S401, target sample data in the SA network is obtained.
[0095] In S402, based on the target sample data, a corresponding load index is determined for the SA network.
[0096] In S403, for the target sample data, a second index value of the load index at the time of collecting the target sample data is calculated.
[0097] Specifically, the target sample data is collected at the collection time, and then the second indicator value of each load indicator at the collection time is calculated according to the calculation method of each load indicator in S102.
[0098] In a possible implementation, after S403, the following steps are further included:
[0099] The load index corresponding to the second index value exceeding the threshold is marked.
[0100] The marked load index and the load index corresponding to the second index value that does not exceed the threshold are stored.
[0101] Specifically, the load index whose second index value exceeds the threshold value may be marked with the mark word "Abnormal", so that it is convenient for maintenance personnel to directly find the abnormal load index. In addition, the second index that does not exceed the threshold value may be marked with the mark word "normal", so that maintenance personnel can distinguish between abnormal load index and normal load index.
[0102] In S404, for the SA network, a corresponding load level is determined according to the second indicator value.
[0103] In S405, an early warning corresponding to the load level is issued.
[0104] In S406, at least a portion of the target sample data is used as an input variable of the first prediction model to predict the indicator-related quantity predicted value at a specified future time.
[0105] In S407, the predicted value of the index-related quantity is used as an input variable of the second prediction model to predict the first index value of the load index at a specified time in the future.
[0106] In S408, an early warning is issued to the SA network according to the first indicator value.
[0107] It is worth noting that S401 to S402 and S406 to S408 have the same or similar implementation methods as the above-mentioned embodiments S101 to S105, which can refer to each other and the embodiments of the present application will not be repeated here.
[0108] It can be seen from the technical solution provided by the above embodiments of the present application that the first indicator value of the load indicator at a specified time in the future is predicted by combining the first prediction model and the second prediction model with the target sample data, and then the SA network is warned according to the first indicator value. Therefore, when the predicted load indicator of the SA network is abnormal, it can be processed in time to avoid abnormal load in the SA network, ensuring the Internet experience of 5G customers and improving the user experience of the 5G network.
[0109] In addition, the index value of the load index at the time of collection of the target sample data can also be calculated. When the load index of the SA network calculated at the time of collection is abnormal, it can be processed in time and the problem of abnormal load index can be quickly solved, ensuring the Internet experience of 5G customers and improving the user experience of the 5G network.
[0110] Corresponding to the load prediction method of the SA network provided in the above embodiment, based on the same technical concept, the embodiment of the present application also provides a load prediction device of the SA network, Figure 5 A schematic diagram of the module composition of a load prediction device for an SA network provided in an embodiment of the present application, wherein the load prediction device for the SA network is used to perform Figures 1 to 4 The load forecasting method of SA network described in Figure 5 As shown, the load prediction device of the SA network includes: an acquisition module 501, a determination module 502, a first prediction module 503, a second prediction module 504, and an early warning module 505.
[0111] The acquisition module 501 is used to acquire target sample data in the SA network.
[0112] The determination module 502 is used to determine the corresponding load index in the SA network based on the target sample data.
[0113] The first prediction module 503 is used to use at least a part of the target sample data as an input variable of the first prediction model to predict the indicator-related quantity predicted value at a specified time in the future.
[0114] The second prediction module 504 is used to use the predicted value of the index-related quantity as an input variable of the second prediction model to predict the first index value of the load index at a specified time in the future.
[0115] The warning module 505 is used to issue a warning to the SA network according to the first indicator value.
[0116] It can be seen from the technical solution provided by the above embodiments of the present application that the target sample data in the SA network is obtained, and based on the target sample data, the corresponding load index is determined for the SA network, and at least a part of the target sample data is used as the input variable of the first prediction model to predict the predicted value of the index-related quantity at a specified time in the future. The predicted value of the index-related quantity is used as the input variable of the second prediction model to predict the first index value of the load index at a specified time in the future, and the SA network is warned according to the first index value. The first prediction model and the second prediction model are combined with the target sample data to predict the first index value of the load index at a specified time in the future, and then the SA network is warned according to the first index value. Therefore, when the predicted load index in the SA network is abnormal, an early warning can be issued, so that the maintenance personnel can deal with it in time, avoiding the abnormal load of the SA network, ensuring the Internet experience of 5G customers, and improving the user experience of the 5G network.
[0117] Optionally, the core network element load prediction device of the SA network further includes: a calculation module (not shown in the figure), a determination module 2 (not shown in the figure), and an early warning module 2 (not shown in the figure).
[0118] The calculation module is used to calculate the second indicator value of the load indicator at the time of collecting the target sample data.
[0119] Determination module 2 is used to determine the corresponding load level for the SA network according to the second indicator value.
[0120] Early warning module 2 is used to issue early warnings corresponding to the load level.
[0121] Optionally, the load prediction device of the SA network further includes: a marking module (not shown in the figure) and a storage module (not shown in the figure).
[0122] The marking module is used to mark the load index corresponding to the second index value exceeding the threshold.
[0123] The storage module is used to store the marked load indicators.
[0124] Optionally, the load prediction device of the SA network further includes: a preprocessing module (not shown in the figure).
[0125] The preprocessing module is used to preprocess the target sample data, and the preprocessing includes data cleaning and null value interpolation.
[0126] Optionally, the warning module 505 includes: a determining unit and an issuing unit.
[0127] The determination unit is used to determine a corresponding load level for the SA network according to the first indicator value.
[0128] The issuing unit is used to issue an early warning work order corresponding to the load level.
[0129] Optionally, the first prediction model is a BP neural network model, the second prediction model is a binary linear regression model, and the first prediction module 503 includes:
[0130] The prediction unit is used to use the 5G terminal data, gNodeB base station coverage data, the number of AMF registered users and the N6 interface traffic at a predetermined time as input variables of the BP neural network model to predict the number of 5G users and the predicted value of 5G traffic at a specified time in the future.
[0131] The second prediction module 504 includes:
[0132] The prediction unit uses the predicted value of the number of 5G users and the predicted value of 5G traffic as input variables of the binary linear regression model to predict the first indicator value of the load indicator at a specified time in the future.
[0133] Optionally, the load prediction device of the SA network further includes: a calculation module (not shown in the figure) and a third prediction module (not shown in the figure).
[0134] The calculation module is used to calculate the accuracy score of the first prediction model based on the predicted value of the indicator-related quantity and the actual value of the indicator-related quantity in the target sample data.
[0135] The third prediction module is used to re-predict the indicator-related quantity at a specified time in the future when the accuracy score does not meet the preset requirements.
[0136] The SA network-based load prediction device provided in the embodiment of the present application can implement each process in the embodiment corresponding to the above-mentioned SA network load prediction method. To avoid repetition, it will not be repeated here.
[0137] It should be noted that the load prediction device of the SA network provided in the embodiment of the present application and the load prediction method of the SA network provided in the embodiment of the present application are based on the same application concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned SA network load prediction method, and the repeated parts will not be repeated.
[0138] Corresponding to the load prediction method of the SA network provided in the above embodiment, based on the same technical concept, the embodiment of the present application further provides an electronic device, which is used to execute the load prediction method of the SA network. Figure 6 A schematic diagram of the structure of an electronic device for implementing each embodiment of the present application is shown in FIG. Figure 6As shown. The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors 601 and memory 602, and the memory 602 may store one or more storage applications or data. Among them, the memory 602 may be a temporary storage or a permanent storage. The application stored in the memory 602 may include one or more modules (not shown in the figure), and each module may include a series of computer executable instructions in the electronic device.
[0139] Furthermore, the processor 601 may be configured to communicate with the memory 602 to execute a series of computer executable instructions in the memory 602 on the electronic device. The electronic device may also include one or more power supplies 603, one or more wired or wireless network interfaces 604, one or more input and output interfaces 605, and one or more keyboards 606.
[0140] Specifically in this embodiment, the electronic device includes a processor, a communication interface, a memory and a communication bus. The processor, the communication interface and the memory communicate with each other through the bus. The memory is used to store computer programs. The processor is used to execute the program stored in the memory to implement the following method steps:
[0141] Get the target sample data in the SA network.
[0142] Based on the target sample data, the corresponding load index is determined for the SA network.
[0143] At least a portion of the target sample data is used as an input variable of the first prediction model to predict the indicator-related quantity predicted value at a specified future time.
[0144] The predicted value of the index-related quantity is used as the input variable of the second prediction model to predict the first index value of the load index at a specified time in the future.
[0145] An early warning is issued to the SA network according to the first indicator value.
[0146] It can be seen from the technical solution provided by the above embodiments of the present application that the target sample data in the SA network is obtained, and based on the target sample data, the corresponding load index is determined for the SA network, and at least a part of the target sample data is used as the input variable of the first prediction model to predict the predicted value of the index-related quantity at a specified time in the future. The predicted value of the index-related quantity is used as the input variable of the second prediction model to predict the first index value of the load index at a specified time in the future, and the SA network is warned according to the first index value. The first prediction model and the second prediction model are combined with the target sample data to predict the first index value of the load index at a specified time in the future, and then the SA network is warned according to the first index value. Therefore, when the predicted load index in the SA network is abnormal, an early warning can be issued, so that the maintenance personnel can deal with it in time, avoiding the abnormal load of the SA network, ensuring the Internet experience of 5G customers, and improving the user experience of the 5G network.
[0147] This embodiment also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:
[0148] Get the target sample data in the SA network.
[0149] Based on the target sample data, the corresponding load index is determined for the SA network.
[0150] At least a portion of the target sample data is used as an input variable of the first prediction model to predict the indicator-related quantity predicted value at a specified future time.
[0151] The predicted value of the index-related quantity is used as the input variable of the second prediction model to predict the first index value of the load index at a specified time in the future.
[0152] An early warning is issued to the SA network according to the first indicator value.
[0153] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, devices, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.
[0154] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0155] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0156] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0157] In a typical configuration, an electronic device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0158] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.
[0159] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0160] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0161] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, devices or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.
[0162] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.
Claims
1. A load forecasting method for SA network, characterized in that: The method comprises: Obtain target sample data in the SA network; Based on the target sample data, determining a corresponding load indicator for the SA network, the load indicator including a capacity utilization rate; Using at least a portion of the target sample data as an input variable of a first prediction model, predicting a predicted value of the indicator-related quantity at a specified future time; Using the predicted value of the index-related quantity as an input variable of the second prediction model, predicting the first index value of the load index at the specified future time; Providing an early warning to the SA network according to the first indicator value; The first prediction model is a BP neural network model, the second prediction model is a binary linear regression model, and the predicted value of the index-related quantity at a specified time in the future includes: Using 5G terminal data, gNodeB base station coverage data, the number of AMF registered users and N6 interface traffic as input variables of the BP neural network model; Predict the number of 5G users and 5G traffic at a specified time in the future. The method of using the predicted value of the index-related quantity as an input variable of the second prediction model to predict the first index value of the load index at the specified future time includes: Using the predicted value of the number of 5G users and the predicted value of 5G traffic as input variables of the binary linear regression model, predict the first indicator value of the load index at the specified future time.
2. The method according to claim 1, characterized in that: After determining the corresponding load index for the SA network based on the target sample data, the method further includes: For the target sample data, calculating a second index value of the load index at the time of collecting the target sample data; For the SA network, determining a corresponding load level according to the second indicator value; An early warning corresponding to the load level is issued.
3. The method according to claim 2, characterized in that After calculating the second index value of the load index at the time of collecting the target sample data, the method further includes: marking the load index corresponding to the second index value exceeding the threshold; The marked load index is stored.
4. The method according to claim 1, characterized in that: After acquiring the target sample data in the SA network, the method further includes: The target sample data is preprocessed, and the preprocessing includes data cleaning processing and null value interpolation processing.
5. The method according to any one of claims 1 to 4, characterized in that: The providing an early warning to the SA network according to the first indicator value includes: For the SA network, determining a corresponding load level according to the first indicator value; Issue an early warning work order corresponding to the load level.
6. The method according to claim 1, characterized in that After predicting the first index value of the load index at the future specified time, the method further includes: Calculating the accuracy score of the first prediction model based on the predicted value of the indicator-related quantity and the actual value of the indicator-related quantity in the target sample data; When the accuracy score does not meet the preset requirement, the indicator-related quantity at a specified future time is re-predicted.
7. A load forecasting device for SA network, characterized in that: The device comprises: An acquisition module, used to acquire target sample data in the SA network; A determination module, configured to determine a corresponding load index for the SA network based on the target sample data, wherein the load index includes a capacity utilization rate; A first prediction module, configured to use at least a portion of the target sample data as an input variable of a first prediction model to predict a predicted value of an indicator-related quantity at a specified time in the future; A second prediction module, configured to use the predicted value of the indicator-related quantity as an input variable of a second prediction model to predict a first indicator value of the load indicator at the specified future time; An early warning module, configured to issue an early warning to the SA network according to the first indicator value; The first prediction model is a BP neural network model, and the second prediction model is a binary linear regression model; In the first prediction module, the predicted value of the indicator-related quantity at a specified future time includes: Using 5G terminal data, gNodeB base station coverage data, the number of AMF registered users and N6 interface traffic as input variables of the BP neural network model; Predict the number of 5G users and 5G traffic at a specified time in the future. In the second prediction module, using the predicted value of the indicator-related quantity as an input variable of the second prediction model, predicting the first indicator value of the load indicator at the specified future time includes: Using the predicted value of the number of 5G users and the predicted value of 5G traffic as input variables of the binary linear regression model, predict the first indicator value of the load index at the specified future time.
8. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus; wherein the processor, the communication interface and the memory communicate with each other through the bus; the memory is used to store computer programs; the processor is used to execute the programs stored in the memory to implement the steps of the load forecasting method for the SA network as described in any one of claims 1-6.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the load prediction method for the SA network according to any one of claims 1 to 6 are implemented.
Citation Information
Patent Citations
Base station load early warning method and device based on deep learning
CN110662245A
Method and device for base station load early warning
CN111132190A
Anti-disaster network control system and device and method and program
JP2015215835A
Workload estimation for mobile device feature integration
US20150274180A1
Burst load prediction method and device, storage medium and electronic device
WO2020135510A1
