CN-116319385-B - Communication network load prediction method, device, equipment and readable storage medium

CN116319385BCN 116319385 BCN116319385 BCN 116319385BCN-116319385-B

Abstract

The application discloses a communication network load prediction method, a device, equipment and a readable storage medium, wherein the method comprises the steps of judging whether a load prediction model of a target communication network is suitable for the current working scene of the target communication network based on the historical actual load of the target communication network and the historical prediction load of the historical actual load; and if the load prediction model is not adapted, updating the load prediction model based on the near-segment historical load data of the target communication network, and predicting the target predicted load based on the updated load prediction model and the near-segment load data, wherein the near-segment historical load data is the historical load data of the target communication network in a second preset time period from the current moment. The model is updated in real time, so that the prediction model is more suitable for a real-time working scene, and a more accurate load prediction result is obtained.

Inventors

WANG LEIYU
XU XIAODONG
CHEN HAO
MA NAN
ZHANG PING

Assignees

鹏城实验室

Dates

Publication Date: 20260505
Application Date: 20230324

Claims (7)

1. A communication network load prediction method, characterized in that the communication network load prediction method comprises the following steps: Judging whether a load prediction model of a target communication network is adaptive to a current working scene of the target communication network or not based on a historical actual load of the target communication network and a historical prediction load of the historical actual load, wherein the load prediction model is a convolutional neural network; if the load prediction model is adapted, predicting a target predicted load of the target communication network at the next moment based on near-segment load data of the target communication and the load prediction model, wherein the near-segment load data is historical load data of the target communication network in a first preset time period before the current moment; If the load prediction model is not adapted, updating the load prediction model based on the near-segment historical load data of the target communication network, and predicting the target predicted load based on the updated load prediction model and the near-segment load data, wherein the near-segment historical load data is the historical load data of the target communication network within a second preset time period from the current moment, and the second preset time period is longer than the first preset time period; The target communication network comprises a plurality of communication base stations, and before the step of judging whether the load prediction model of the target communication network is suitable for the current working scene of the target communication network based on the historical actual load of the target communication network and the historical prediction load of the historical actual load, the communication network load prediction method further comprises the following steps: Acquiring historical load data of each communication base station as a training data set; extracting training samples from the training data set based on a preset time window and a time sequence of the training data set, wherein any one training sample comprises load data of different communication base stations at different moments in the preset time window under the same time period, and a label of any one training sample is the load data of the different communication base stations at the next moment of the corresponding time period of the training sample; Integrating load data adjacent in time sequence in the training sample into sample characteristics, and inputting the sample characteristics into the load prediction model to obtain the predicted load of the training sample; calculating a model prediction loss of the load prediction model based on the predicted load and the label of the training sample; iteratively updating the load prediction model based on the model prediction loss; the communication network load prediction method is applied to a digital twin network of the target communication network, any ground base station k in the target communication network, and a digital twin body in the digital twin network is as follows: , wherein, Representing the actual position coordinates of base station k, Indicating the communication service range of base station k, The communication load of base station k at time t is indicated, Indicating the decision action of base station k at time t, Indicating a distance from the current time interval of Is the communication load of base station k at the next moment, Indicating a distance from the current time interval of Before the step of determining whether the load prediction model of the target communication network is adapted to the current operating scenario of the target communication network based on the historical actual load of the target communication network and the historical predicted load of the historical actual load, the method further comprises: acquiring historical load data of the target communication network based on the digital twin network; And extracting the near-segment historical load data in real time by starting from the latest historical load data in the historical load data through a preset time sliding window.
2. The communication network load prediction method according to claim 1, wherein the step of judging whether the load prediction model of the target communication network is adapted to the current operation scenario of the target communication network based on the historical actual load of the target communication network and the historical predicted load of the historical actual load comprises: extracting vector features of the historical actual load to obtain a first load vector feature; extracting vector features of the historical predicted load to obtain a second load vector feature; calculating the similarity between the first load vector feature and the second load vector feature; if the similarity is larger than a preset similarity threshold, judging that the load prediction model is adapted to the current working scene of the target communication network; and if the similarity is smaller than or equal to the preset similarity threshold, judging that the load prediction model is not suitable for the current working scene of the target communication network.
3. The communication network load prediction method of claim 1, wherein the step of iteratively updating the load prediction model based on the model prediction loss comprises: judging whether the model prediction loss converges or not; if the model prediction loss is judged to be converged, the load prediction model training is completed; If the model prediction loss is judged not to be converged, updating parameters in the load prediction model based on the model prediction loss and a gradient descent method; And returning to execute the step of integrating the load data adjacent in time sequence in the training sample into sample characteristics and inputting the sample characteristics into the load prediction model based on the new training sample and the updated load prediction model to obtain the predicted load of the training sample and the subsequent steps.
4. The communication network load prediction method of claim 3, wherein the step of updating the load prediction model based on near-segment historical load data of the target communication network comprises: Acquiring near-segment historical load data of each target communication network as the training data set; and executing the step of extracting training samples from the training data set based on a preset time window and the time sequence of the training data set and the subsequent step.
5. A communication network load prediction apparatus, characterized in that the communication network load prediction apparatus comprises: The judging module is used for judging whether a load prediction model of a target communication network is suitable for the current working scene of the target communication network or not based on the historical actual load of the target communication network and the historical prediction load of the historical actual load, wherein the target communication network comprises a plurality of communication base stations, and the load prediction model is a convolutional neural network; the first prediction module is used for predicting a target predicted load of the next moment of the target communication network based on near-segment load data of the target communication and the load prediction model if the near-segment load data is adaptive, wherein the near-segment load data is historical load data of the target communication network in a first preset time period from the current moment; the second prediction module is configured to update the load prediction model based on near-segment historical load data of the target communication network if the load prediction model is not adapted, and predict the target predicted load based on the updated load prediction model and the near-segment load data, where the near-segment historical load data is historical load data of the target communication network within a second preset time period from the current time, and the second preset time period is longer than the first preset time period; The system comprises a first training module, a model prediction model, a model prediction loss calculation module, a model prediction model, a load prediction model updating module and a load prediction model updating module, wherein the first training module is used for acquiring historical load data of each communication base station as a training data set, extracting training samples from the training data set based on a preset time window and a time sequence of the training data set, wherein any training sample comprises load data of different communication base stations at different moments in the preset time window under the same time period, and the label of any training sample is load data of different communication base stations at the moment of the corresponding time period of the training sample; The acquisition module is used for acquiring historical load data of the target communication network based on a digital twin network, and extracting the near-segment historical load data in real time from the latest historical load data in the historical load data through a preset time sliding window, wherein any ground base station k in the target communication network is a digital twin body in the digital twin network: , wherein, Representing the actual position coordinates of base station k, Indicating the communication service range of base station k, The communication load of base station k at time t is indicated, Indicating the decision action of base station k at time t, Indicating a distance from the current time interval of Is the communication load of base station k at the next moment, Indicating a distance from the current time interval of The decision action of base station k at the next moment.
6. A communication network load predicting device comprising a memory, a processor and a communication network load predicting program stored on the memory and executable on the processor, the communication network load predicting program when executed by the processor implementing the steps of the communication network load predicting method according to any one of claims 1 to 4.
7. A readable storage medium, characterized in that the readable storage medium has stored thereon a communication network load prediction program, which when executed by a processor, implements the steps of the communication network load prediction method according to any of claims 1 to 4.

Description

Communication network load prediction method, device, equipment and readable storage medium Technical Field The present application relates to the field of communications technologies, and in particular, to a method, an apparatus, a device, and a readable storage medium for predicting a communications network load. Background The mobile communication network load level is a key performance indicator that affects communication network operation and maintenance decisions (such as network resource allocation, topology management, etc.). As the scale of the mobile network is continuously enlarged, the communication load levels of access nodes of the base station of the mobile communication network in different areas and in different time periods are obviously different, and the dynamic change characteristic is presented, so that the management difficulty and the management cost of the communication network are increased. And the prediction of the communication load condition of the future mobile communication network can provide prior information for network operation and maintenance decision, so that an effective coping decision scheme is formed in advance. There is a need for a method for accurately predicting future communication load of a mobile communication network for the current mobile network scale. The foregoing is provided merely for the purpose of facilitating understanding of the technical solutions of the present application and is not intended to represent an admission that the foregoing is prior art. Disclosure of Invention The application mainly aims to provide a communication network load prediction method, a device, equipment and a readable storage medium, aiming at improving the accuracy of future communication load prediction results of a mobile communication network. To achieve the above object, the present application provides a communication network load prediction method, including the steps of: Judging whether a load prediction model of a target communication network is suitable for the current working scene of the target communication network or not based on the historical actual load of the target communication network and the historical predicted load of the historical actual load; if the load prediction model is adapted, predicting a target predicted load of the target communication network at the next moment based on near-segment load data of the target communication and the load prediction model, wherein the near-segment load data is historical load data of the target communication network in a first preset time period before the current moment; If the load prediction model is not adapted, updating the load prediction model based on the near-segment historical load data of the target communication network, and predicting the target predicted load based on the updated load prediction model and the near-segment load data, wherein the near-segment historical load data is the historical load data of the target communication network within a second preset time period from the current moment, and the second preset time period is longer than the first preset time period. Further, the step of determining whether the load prediction model of the target communication network is adapted to the current working scenario of the target communication network based on the historical actual load of the target communication network and the historical predicted load of the historical actual load includes: extracting vector features of the historical actual load to obtain a first load vector feature; extracting vector features of the historical predicted load to obtain a second load vector feature; calculating the similarity between the first load vector feature and the second load vector feature; if the similarity is larger than a preset similarity threshold, judging that the load prediction model is adapted to the current working scene of the target communication network; and if the similarity is smaller than or equal to the preset similarity threshold, judging that the load prediction model is not suitable for the current working scene of the target communication network. Further, before the step of determining whether the load prediction model of the target communication network is adapted to the current operation scenario of the target communication network based on the historical actual load of the target communication network and the historical predicted load of the historical actual load, the method includes: Generating training samples based on historical load data of the target communication network; Integrating load data adjacent in time sequence in the training sample into sample characteristics, and inputting the sample characteristics into a load prediction model to obtain the predicted load of the training sample; calculating a model prediction loss of the load prediction model based on the predicted load and the label of the training sample; and iteratively updating the load predictio