WO2023125090A1 - Cell load adjustment method and related device thereof - Google Patents

Abstract

The present application provides a cell load adjustment method and a related device thereof, applied to the field of artificial intelligence, and capable of accurately adjusting the load degree of a target cell so as to keep each performance index of the target cell within a proper range and provide a higher-quality network service for users. The method of the present application comprises: acquiring feature information of the target cell and a load index of the target cell, the load index being used for indicating the load degree of the target cell; processing the feature information by means of a first target model to obtain model parameters of a second target model; acquiring the second target model on the basis of the model parameters; and processing the load index by means of the second target model to obtain an operation instruction, the operation instruction being used for adjusting the load index.

Description

A cell load adjustment method and related equipment

This application claims the priority of the Chinese patent application with the application number 202111630614.4 and the title of the invention "a cell load adjustment method and related equipment" submitted to the China Patent Office on December 28, 2021, the entire contents of which are incorporated by reference in this application.

technical field

The present application relates to the technical field of artificial intelligence (AI), in particular to a cell load adjustment method and related equipment.

Background technique

In a wireless cellular network, the uneven distribution of users leads to unbalanced load among cells (that is, areas covered by the wireless network). Due to the large number of users and high service demand in high-load cells, it is easy to cause insufficient network resources, and it is difficult to guarantee the quality of service for users. Correspondingly, due to the small number of users and small business demand in low-load cells, their network resources are not fully utilized.

At present, the means for adjusting the load level of the cell is mainly manual modification. However, manual modification relies on the manual experience of experts, and the factors considered are often single and not comprehensive enough to accurately adjust the load level of the cell, thus failing to keep various performance indicators of the cell within an appropriate range and provide users with Good enough web service.

Contents of the invention

The embodiment of the present application provides a cell load adjustment method and related equipment, which can accurately adjust the load level of the target cell, so as to keep various performance indicators of the target cell within an appropriate range, and provide users with a better network Serve.

The first aspect of the embodiments of the present application provides a method for adjusting a cell load, the method including:

When it is necessary to adjust the load level of the target cell, the characteristic information of the target cell and the load index of the target cell can be collected, wherein the characteristic information of the target cell can be used to indicate the current scene of the target cell, and the load index of the target cell can be used It is used to indicate the load level of the target cell (also called load balance level).

After obtaining the feature information of the target cell and the load index of the target cell, a first target model can be obtained, and the first target model is a trained neural network. Then, input the feature information of the target cell into the first target model, so as to perform a series of processing (for example, feature extraction, etc.) on the feature information of the target cell through the first target model to obtain model parameters of the second target model.

The second target model can be constructed based on the model parameters of the second target model, for example, the second target model can be a multi-layer perceptron. So far, the whole composed of the first target model and the second target model can be used to indicate the functional relationship between the load index of the target cell and the operation instruction for the target cell. Generally, the operation indication for the target cell is a monotonically increasing function of the load index of the target cell, and the increasing speed is determined by the characteristic information of the target cell.

After building the second target model, the load index of the target cell can be input into the second target model, so as to process the load index of the target cell through the second target model, and obtain the operation instruction for the target cell, and the operation instruction for the target cell It can be used to adjust the load index of the target cell.

It can be seen from the above method that after obtaining the characteristic information of the target cell and the load index of the target cell, the characteristic information can be processed through the first target model to obtain the model parameters of the second target model. Then, construct a second target model based on the model parameters of the second target model, and process the load index through the second target model to obtain an operation instruction for the target cell, which can be used to adjust the load index of the target cell. In the foregoing process, the model parameters of the second target model are obtained based on the characteristic information of the target cell, since the characteristic information of the target cell can be used to characterize various characteristics of the target cell (ie, the scene in which the target cell is located), then, in When using the second target model to process the load index of the target cell, various factors such as various characteristics of the target cell are considered, so the operation instructions for the target cell output by the second target model can be used to accurately adjust the target cell. The load level is used to keep various performance indicators of the target cell within an appropriate range and provide users with better network services.

Further, the whole formed by the first target model and the second target model can be used to indicate the functional relationship between the load index of the target cell and the operation instruction for the target cell, the functional relationship is usually a monotonically increasing or monotonically decreasing relationship, In line with the constraints of expert experience (equivalent to integrating expert experience into the model), the control strategy learned by the model is more in line with business logic.

In a possible implementation, the second target model includes the first sub-model and the second sub-model, and processing the load index through the second target model, and obtaining the operation instruction includes: processing the load index through the first sub-model , to obtain the first operation, if the load index is equal to the preset index threshold, then the first operation is used to not adjust the load index; process the load index through the second sub-model to obtain the second operation; for the first operation and the second The operation is weighted and summed to obtain the operation instruction. In the foregoing implementation manner, after obtaining the characteristic information of the target cell and the load index of the target cell, the first target model may be obtained. Then, the feature information of the target cell is input into the first target model, so as to perform a series of processing on the feature information of the target cell through the first target model to obtain model parameters of the first sub-model and model parameters of the second sub-model. Then, the first sub-model can be constructed based on the model parameters of the first sub-model, and the second sub-model can be constructed based on the model parameters of the second sub-model. Subsequently, the load index of the target cell can be input into the first sub-model and the second sub-model respectively, so as to process the load index of the target cell through the first sub-model, obtain the first operation for the target cell, and pass the second sub-model The second sub-model processes the load index of the target cell to obtain the second operation for the target cell. Finally, weighted summing is performed on the first operation on the target cell and the second operation on the target cell to obtain an operation instruction on the target cell. In the aforementioned process, the first sub-model and the second sub-model are constructed through the framework of the teacher-student network. The first sub-model adds expert constraints (that is, decides whether the target cell absorbs or releases users with preset index thresholds), The safety of the control strategy output by the model is guaranteed, while the second sub-model relaxes the expert constraints, learns and outputs the corresponding control strategy based on data, and weights the output of the two models as the final strategy, thus balancing the control strategy. safety and efficiency.

In a possible implementation, the first target model and the first sub-model are used to indicate the first functional relationship between the load indicator and the operation indicator, and the first target model and the second sub-model are used to indicate the load indicator and the operation A second functional relationship between the indicators, the second functional relationship is obtained by translating the first functional relationship, and both the first functional relationship and the second functional relationship are monotonically increasing or monotonically decreasing. In the foregoing implementation manner, since the first functional relationship and the second functional relationship are both monotonically increasing or monotonically decreasing, then the operation indication for the target cell obtained by performing weighted summation based on the first operation and the second operation , the functional relationship with the load index of the target cell is also a monotonically increasing relationship or a monotonically decreasing relationship, which is consistent with the expert experience strategy.

In a possible implementation manner, the magnitude of translation is determined based on feature information.

In a possible implementation manner, the operation instruction for the target cell is used to modify configuration parameters of the target cell, so as to adjust the load index. The characteristic information of the target cell includes at least one of the following: configuration parameters of the target cell; traffic statistics data of the target cell; configuration parameters of neighboring cells of the target cell; traffic statistics data of the neighboring cells. In this way, the feature information of the target cell includes the multi-dimensional features of the target cell, which can fully characterize the scene where the target cell is located.

In a possible implementation manner, the configuration parameters of the target cell may include at least one of the following information: antenna transmit power of the target cell, antenna downtilt angle of the target cell, antenna horizontal azimuth angle of the target cell, The reference signal receiving power (reference signal receiving power, RSRP) threshold for the user to start the inter-frequency handover measurement, the RSRP threshold for the user in the target cell to stop the inter-frequency handover measurement, and the specific frequency RSRP offset for the user in the target cell to trigger the inter-frequency handover procedure The RSRP offset of the specific neighboring cell for the user in the target cell to trigger the inter-frequency handover process, the RSRP threshold for the user in the target cell to start the intra-frequency handover measurement, the RSRP threshold for the user in the target cell to stop the intra-frequency handover measurement, and the target cell The specific frequency RSRP offset of the user in the target cell triggering the intra-frequency handover procedure and the specific neighbor cell RSRP offset of the user in the target cell triggering the intra-frequency handover procedure, etc. The configuration parameters of the neighbor cell may include at least one of the following information: the antenna transmit power of the neighbor cell, the antenna downtilt angle of the neighbor cell, the horizontal azimuth angle of the antenna of the neighbor cell, and the RSRP threshold for the user in the neighbor cell to initiate inter-frequency handover measurement , the RSRP threshold for the user in the neighboring cell to stop the inter-frequency handover measurement, the specific frequency RSRP offset for the user in the neighboring cell to trigger the inter-frequency handover process, the specific neighbor RSRP offset for the user in the neighboring cell to trigger the inter-frequency handover process, The RSRP threshold for the user in the neighbor cell to start the same-frequency handover measurement, the RSRP threshold for the user in the neighbor cell to stop the same-frequency handover measurement, the specific frequency RSRP offset for the user in the neighbor cell to trigger the same-frequency handover procedure, and the user in the neighbor cell Specific neighboring cell RSRP offset that triggers the intra-frequency handover process, etc.

In a possible implementation, the traffic statistics data of the target cell may include at least one of the following information: the average number of users per unit time period of the target cell, the average number of active users per unit time period of the target cell, and the average number of active users per unit time period of the target cell. The uplink traffic of the target cell, the proportion of low channel quality indicator (CQI) reports in the unit time period of the target cell, the proportion of data packets whose length in the unit time period of the target cell is less than the preset length threshold (also called the target cell The proportion of small packets per unit time period) and the average length of data packets per unit time period of the target cell, etc. The traffic statistics data of neighboring cells may include at least one of the following information: the average number of users per unit time period of neighboring cells, the average number of active users per unit time period of neighboring cells, the uplink traffic per unit time period of neighboring cells, the The proportion of the CQI report of the neighbor cell, the proportion of data packets whose length per unit time period of the neighbor cell is less than the preset length threshold (also called the proportion of small packets per unit time period of the neighbor cell) and the average number of data packets per unit time period of the neighbor cell length etc.

The second aspect of the embodiment of the present application provides a model training method, the method includes: obtaining the characteristic information of the first cell and the load index of the first cell, the load index of the first cell is used to indicate the load degree of the first cell ; Process the feature information through the first model to be trained to obtain the model parameters of the second model to be trained; obtain the second model to be trained based on the model parameters; process the load index of the first cell through the second model to be trained to obtain The operation instruction for the first cell is used to adjust the load index of the first cell; the operation instruction for the first cell and the operation instruction for the second cell are processed by the third target model, and the second cell is obtained A score and a second score, the first score is used to evaluate the impact of the operation indication of the first cell on the load index of the first cell, and the second score is used to evaluate the impact of the operation indication of the first cell on the performance index of the entire network, The operation instruction for the second cell is used to adjust the load index of the second cell, the first cell and the second cell are different cells; based on the first score and the second score, the model parameters of the first model to be trained are updated, Until the model training conditions are met, the first target model is obtained.

It can be seen from the above method that when using the critic model (that is, the third target model) to train the actor model (that is, the whole composed of the first target model and the second target model), the training goal of the actor model is to simultaneously optimize the local The output value of the network and the output value of the global network, based on the actor model trained in this way, has the performance of balancing local and global performance, that is, it can balance the impact of the load of a certain cell on the cell and the entire network. Influence, to achieve coordination between multiple cells.

In a possible implementation manner, the second model to be trained includes the first sub-model and the second sub-model, and the load index of the first cell is processed through the second target model, and the operation instruction for the first cell obtained includes: Process the load index of the first cell through the first sub-model to obtain the first operation, if the load index of the first cell is equal to the preset index threshold, the first operation is used to not adjust the load index of the first cell; by The second sub-model processes the load index of the first cell to obtain a second operation; performs weighted summation of the first operation and the second operation to obtain an operation instruction for the first cell.

In a possible implementation manner, the first model to be trained and the first sub-model are used to indicate a first functional relationship between the load index of the first cell and the operation indication for the first cell, and the first model to be trained and The second sub-model is used to indicate the second functional relationship between the load index of the first cell and the operation indication for the first cell, the second functional relationship is obtained by shifting the first functional relationship, the first functional relationship and The second functional relationship is a monotonically increasing relationship or a monotonically decreasing relationship.

In a possible implementation manner, the magnitude of translation is determined based on feature information.

In a possible implementation manner, the operation instruction for the first cell is used to modify the configuration parameters of the first cell to adjust the load index of the first cell; the feature information of the first cell includes at least one of the following: the first The configuration parameters of the cell; the traffic statistics data of the first cell; the configuration parameters of the neighbor cell of the first cell; the traffic statistics data of the neighbor cell.

In a possible implementation manner, the configuration parameters of the first cell may include at least one of the following information: antenna transmit power of the first cell, antenna downtilt angle of the first cell, antenna horizontal azimuth angle of the first cell, The RSRP threshold for the user in the first cell to start the inter-frequency handover measurement, the RSRP threshold for the user in the first cell to stop the inter-frequency handover measurement, the specific frequency RSRP offset for the user in the first cell to trigger the inter-frequency handover procedure, the first The RSRP offset of the specific neighboring cell for the user in the cell to trigger the inter-frequency handover process, the RSRP threshold for the user in the first cell to start the intra-frequency handover measurement, the RSRP threshold for the user in the first cell to stop the intra-frequency handover measurement, the first cell The specific frequency RSRP offset of the user in the first cell triggering the same-frequency handover procedure and the specific neighbor cell RSRP offset of the user in the first cell triggering the same-frequency handover procedure, etc. The configuration parameters of the neighbor cell may include at least one of the following information: antenna transmit power of the neighbor cell, antenna downtilt angle of the first cell, antenna horizontal azimuth angle of the first cell, user-initiated inter-frequency handover measurement in the neighbor cell RSRP threshold, RSRP threshold for users in neighboring cells to stop inter-frequency handover measurement, specific frequency RSRP offset for users in neighboring cells to trigger inter-frequency handover procedures, specific neighbor cell RSRP offset for users in neighboring cells to trigger inter-frequency handover procedures The RSRP threshold for the user in the neighbor cell to start the same-frequency handover measurement, the RSRP threshold for the user in the neighbor cell to stop the same-frequency handover measurement, the specific frequency RSRP offset for the user in the neighbor cell to trigger the same-frequency handover procedure, and the The user triggers the same-frequency handover process of the specific neighbor RSRP offset and so on.

In a possible implementation manner, the traffic statistics data of the first cell may include at least one of the following information: the average number of users in the unit time period of the first cell, the average number of active users in the unit time period of the first cell, the first The uplink traffic of the unit time period of the cell, the proportion of low channel quality indicator CQI reports in the unit time period of the first cell, and the proportion of data packets whose length in the unit time period of the first cell is less than the preset length threshold (also referred to as the first The proportion of small packets in the cell unit time period) and the average length of the data packets in the first cell unit time period, and so on. The traffic statistics data of the neighbor cells of the first cell may include at least one of the following information: the average number of users per unit time period of the neighbor cell, the average number of active users per unit time period of the neighbor cell, the uplink traffic per unit time period of the neighbor cell, the neighbor The proportion of CQI reports per unit time period of the cell, the proportion of data packets whose length per unit time period of the neighbor cell is less than the preset length threshold (also called the proportion of small packets per unit time period of the neighbor cell), and the proportion of packets per unit time period of the neighbor cell The average length of packets, etc.

The third aspect of the embodiment of the present application provides a device for adjusting cell load. The device for adjusting cell load includes: a first acquiring module, configured to acquire characteristic information of a target cell and a load index of the target cell, and the load index is used for Indicating the load level of the target cell; the first processing module is used to process the characteristic information through the first target model to obtain the model parameters of the second target model; the second acquisition module is used to obtain the second target model based on the model parameters; The second processing module is configured to process the load index through the second target model to obtain an operation instruction, and the operation instruction is used to adjust the load index.

It can be seen from the above device that after obtaining the characteristic information of the target cell and the load index of the target cell, the characteristic information can be processed through the first target model to obtain the model parameters of the second target model. Then, a second target model is constructed based on the model parameters of the second target model, and the load index is processed through the second target model to obtain an operation instruction for the target cell, and the operation instruction can be used to adjust the load index of the target cell. In the foregoing process, the model parameters of the second target model are obtained based on the characteristic information of the target cell, since the characteristic information of the target cell can be used to characterize various characteristics of the target cell (ie, the scene in which the target cell is located), then, in When using the second target model to process the load index of the target cell, various factors such as various characteristics of the target cell are considered, so the operation instructions for the target cell output by the second target model can be used to accurately adjust the target cell. The load level is used to keep various performance indicators of the target cell within an appropriate range and provide users with better network services.

In a possible implementation, the second target model includes the first sub-model and the second sub-model, and the second processing module is configured to: process the load index through the first sub-model to obtain the first operation, if the load If the indicator is equal to the preset indicator threshold, the first operation is used to not adjust the load indicator; the load indicator is processed through the second sub-model to obtain the second operation; the weighted sum of the first operation and the second operation is obtained to obtain the operation instruct.

In a possible implementation, the first target model and the first sub-model are used to indicate the first functional relationship between the load index and the first operation, and the first target model and the second sub-model are used to indicate the load index and The second functional relationship between the second operations, the second functional relationship is obtained by translating the first functional relationship, and both the first functional relationship and the second functional relationship are monotonically increasing or monotonically decreasing.

In a possible implementation manner, the magnitude of translation is determined based on feature information.

In a possible implementation, the operation instruction is used to modify the configuration parameters of the target cell to adjust the load indicator; the characteristic information of the target cell includes at least one of the following: configuration parameters of the target cell; traffic statistics data of the target cell; Configuration parameters of neighbor cells of the cell; traffic statistics data of neighbor cells.

In a possible implementation manner, the configuration parameters of the target cell may include at least one of the following information: antenna transmit power of the target cell, antenna downtilt angle of the target cell, antenna horizontal azimuth angle of the target cell, RSRP threshold for user to start inter-frequency handover measurement, RSRP threshold for user in target cell to stop inter-frequency handover measurement, specific frequency RSRP offset for user in target cell to trigger inter-frequency handover procedure, user in target cell to trigger inter-frequency handover Specific neighboring cell RSRP offset for the process, RSRP threshold for users in the target cell to start intra-frequency handover measurement, RSRP threshold for users in the target cell to stop intra-frequency handover measurement, specific frequency for users in the target cell to trigger the intra-frequency handover process The RSRP offset and the RSRP offset of the specific neighboring cell for the user in the target cell to trigger the intra-frequency handover procedure, etc. The configuration parameters of the neighbor cell may include at least one of the following information: the antenna transmit power of the neighbor cell, the antenna downtilt angle of the neighbor cell, the horizontal azimuth angle of the antenna of the neighbor cell, and the RSRP threshold for the user in the neighbor cell to initiate inter-frequency handover measurement , the RSRP threshold for the user in the neighboring cell to stop the inter-frequency handover measurement, the specific frequency RSRP offset for the user in the neighboring cell to trigger the inter-frequency handover process, the specific neighbor RSRP offset for the user in the neighboring cell to trigger the inter-frequency handover process, The RSRP threshold for the user in the neighbor cell to start the same-frequency handover measurement, the RSRP threshold for the user in the neighbor cell to stop the same-frequency handover measurement, the specific frequency RSRP offset for the user in the neighbor cell to trigger the same-frequency handover procedure, and the user in the neighbor cell Specific neighboring cell RSRP offset that triggers the intra-frequency handover process, etc.

In a possible implementation, the traffic statistics data of the target cell may include at least one of the following information: the average number of users per unit time period of the target cell, the average number of active users per unit time period of the target cell, and the average number of active users per unit time period of the target cell. The uplink traffic of the target cell, the proportion of low channel quality indicator CQI reports in the target cell unit time period, the proportion of data packets whose length in the target cell unit time period is less than the preset length threshold (also referred to as the number of small packets in the target cell unit time period ratio) and the average length of data packets in the unit time period of the target cell, etc. The traffic statistics data of the neighbor cell of the target cell may include at least one of the following information: the average number of users in the neighbor cell per unit time period, the average number of active users in the neighbor cell per unit time period, the uplink traffic in the neighbor cell unit time period, the neighbor cell The proportion of CQI reports per unit time period, the proportion of data packets whose length per unit time period of neighboring cells is less than the preset length threshold (also called the proportion of small packets per unit time period of neighboring cells) and the data of neighboring cells per unit time period The average length of packets, etc.

The fourth aspect of the embodiment of the present application provides a model training device, the model training device includes: a first acquisition module, used to acquire the characteristic information of the first cell and the load index of the first cell, the load index of the first cell Used to indicate the load level of the first cell; the first processing module is used to process the characteristic information of the first cell through the first model to be trained to obtain the model parameters of the second model to be trained; the second acquisition module is used to Obtain the second model to be trained based on the model parameters; the second processing module is used to process the load index of the first cell through the second model to be trained to obtain an operation instruction for the first cell, and use the operation instruction for the first cell For adjusting the load index of the first cell; the third processing module is used to process the operation instruction of the first cell and the operation instruction of the second cell through the third target model to obtain the first score and the second score, the first The score is used to evaluate the impact of the operation indication of the first cell on the load index of the first cell, the second score is used to evaluate the impact of the operation indication of the first cell on the performance index of the entire network, and the operation indication of the second cell is used for Adjusting the load index of the second cell, the first cell and the second cell are different cells; the update module is used to update the model parameters of the first model to be trained based on the first score and the second score until the model training is satisfied Conditions, get the first target model.

It can be seen from the above device that when using the critic model (ie, the third target model) to train the actor model (ie, the whole composed of the first target model and the second target model), the training goal of the actor model is to simultaneously optimize the local The output value of the network and the output value of the global network, based on the actor model trained in this way, has the performance of balancing local and global performance, that is, it can balance the impact of the load of a certain cell on the cell and the entire network. Influence, to achieve coordination between multiple cells.

In a possible implementation manner, the second model to be trained includes a first sub-model and a second sub-model, and the second processing module is configured to: process the load index of the first cell through the first sub-model to obtain the first sub-model One operation, if the load index of the first cell is equal to the preset index threshold, the first operation is used to not adjust the load index of the first cell; the load index of the first cell is processed through the second sub-model to obtain the second Operation: performing weighted summation on the first operation and the second operation to obtain an operation instruction for the first cell.

In a possible implementation manner, the first model to be trained and the first sub-model are used to indicate a first functional relationship between the load index of the first cell and the first operation for the first cell, and the first model to be trained and the second sub-model are used to indicate the second functional relationship between the load indicator of the first cell and the second operation for the first cell, the second functional relationship is obtained by shifting the first functional relationship, the first function Both the relationship and the second function relationship are monotonically increasing or monotonically decreasing.

In a possible implementation manner, the magnitude of translation is determined based on feature information.

In a possible implementation manner, the operation instruction for the first cell is used to modify the configuration parameters of the first cell to adjust the load index of the first cell; the feature information of the first cell includes at least one of the following: the first The configuration parameters of the cell; the traffic statistics data of the first cell; the configuration parameters of the neighbor cell of the first cell; the traffic statistics data of the neighbor cell.

In a possible implementation manner, the configuration parameters of the first cell may include at least one of the following information: antenna transmit power of the first cell, antenna downtilt angle of the first cell, antenna horizontal azimuth angle of the first cell, The RSRP threshold for the user in the first cell to start the inter-frequency handover measurement, the RSRP threshold for the user in the first cell to stop the inter-frequency handover measurement, the specific frequency RSRP offset for the user in the first cell to trigger the inter-frequency handover procedure, the first The RSRP offset of the specific neighboring cell for the user in the cell to trigger the inter-frequency handover process, the RSRP threshold for the user in the first cell to start the intra-frequency handover measurement, the RSRP threshold for the user in the first cell to stop the intra-frequency handover measurement, the first cell The specific frequency RSRP offset of the user in the first cell triggering the same-frequency handover procedure and the specific neighbor cell RSRP offset of the user in the first cell triggering the same-frequency handover procedure, etc. The configuration parameters of the neighbor cell may include at least one of the following information: the antenna transmit power of the neighbor cell, the antenna downtilt angle of the neighbor cell, the horizontal azimuth angle of the antenna of the neighbor cell, and the RSRP threshold for the user in the neighbor cell to initiate inter-frequency handover measurement , the RSRP threshold for the user in the neighboring cell to stop the inter-frequency handover measurement, the specific frequency RSRP offset for the user in the neighboring cell to trigger the inter-frequency handover process, the specific neighbor RSRP offset for the user in the neighboring cell to trigger the inter-frequency handover process, The RSRP threshold for the user in the neighbor cell to start the same-frequency handover measurement, the RSRP threshold for the user in the neighbor cell to stop the same-frequency handover measurement, the specific frequency RSRP offset for the user in the neighbor cell to trigger the same-frequency handover procedure, and the user in the neighbor cell Specific neighboring cell RSRP offset that triggers the intra-frequency handover process, etc.

In a possible implementation manner, the traffic statistics data of the first cell may include at least one of the following information: the average number of users in the unit time period of the first cell, the average number of active users in the unit time period of the first cell, the first The uplink traffic of the unit time period of the cell, the proportion of low channel quality indicator CQI reports in the unit time period of the first cell, and the proportion of data packets whose length in the unit time period of the first cell is less than the preset length threshold (also referred to as the first The proportion of small packets in the cell unit time period) and the average length of the data packets in the first cell unit time period, and so on. The traffic statistics data of the neighbor cells of the first cell may include at least one of the following information: the average number of users per unit time period of the neighbor cell, the average number of active users per unit time period of the neighbor cell, the uplink traffic per unit time period of the neighbor cell, the neighbor The proportion of CQI reports per unit time period of the cell, the proportion of data packets whose length per unit time period of the neighbor cell is less than the preset length threshold (also called the proportion of small packets per unit time period of the neighbor cell), and the proportion of packets per unit time period of the neighbor cell The average length of packets, etc.

The fifth aspect of the embodiment of the present application provides a cell load adjustment device, the device includes a memory and a processor; the memory stores codes, the processor is configured to execute the codes, and when the codes are executed, the cell load adjustment device Execute the method described in the first aspect or any possible implementation manner of the first aspect.

The sixth aspect of the embodiment of the present application provides a model training device, which includes a memory and a processor; the memory stores codes, the processor is configured to execute the codes, and when the codes are executed, the model training device executes as the second Aspect or the method described in any possible implementation manner of the second aspect.

A seventh aspect of the embodiments of the present application provides a circuit system, where the circuit system includes a processing circuit configured to execute the first aspect, any possible implementation manner of the first aspect, the second aspect, or the first aspect. The method described in any one possible implementation manner of the two aspects.

The eighth aspect of the embodiments of the present application provides a chip system, the chip system includes a processor, used to call the computer program or computer instruction stored in the memory, so that the processor executes the first aspect, the first aspect Any possible implementation manner, the second aspect, or the method described in any possible implementation manner in the second aspect.

In a possible implementation manner, the processor is coupled to the memory through an interface.

In a possible implementation manner, the chip system further includes a memory, where computer programs or computer instructions are stored.

The ninth aspect of the embodiments of the present application provides a computer storage medium, the computer storage medium stores a computer program, and when the program is executed by a computer, the computer implements any one of the possible methods of the first aspect and the first aspect. Implementation, the second aspect, or the method described in any possible implementation of the second aspect.

The tenth aspect of the embodiments of the present application provides a computer program product, the computer program product stores instructions, and when the instructions are executed by a computer, the computer implements any one of the possible implementations of the first aspect and the first aspect manner, the second aspect, or the method described in any one possible implementation manner of the second aspect.

In the embodiment of the present application, after acquiring the feature information of the target cell and the load index of the target cell, the feature information may be processed by the first target model to obtain model parameters of the second target model. Then, a second target model is constructed based on the model parameters of the second target model, and the load index is processed through the second target model to obtain an operation instruction for the target cell, and the operation instruction can be used to adjust the load index of the target cell. In the foregoing process, the model parameters of the second target model are obtained based on the characteristic information of the target cell, since the characteristic information of the target cell can be used to characterize various characteristics of the target cell (ie, the scene in which the target cell is located), then, in When using the second target model to process the load index of the target cell, various factors such as various characteristics of the target cell are considered, so the operation instructions for the target cell output by the second target model can be used to accurately adjust the target cell. The load level is used to keep various performance indicators of the target cell within an appropriate range and provide users with better network services.

Further, the whole formed by the first target model and the second target model can be used to indicate the functional relationship between the load index of the target cell and the operation instruction for the target cell, the functional relationship is usually a monotonically increasing or monotonically decreasing relationship, In line with the constraints of expert experience (equivalent to integrating expert experience into the model), the control strategy learned by the model is more in line with business logic.

Furthermore, the first sub-model and the second sub-model are constructed through the framework of the teacher-student network. The first sub-model adds expert constraints (that is, uses τ as the threshold to determine whether the target cell absorbs or releases users), which ensures The security of the control strategy output by the model, while the second sub-model relaxes the expert constraints, learns and outputs the corresponding control strategy based on data, and weights the output of the two models as the final strategy, thus balancing the security of the control strategy and efficiency.

Furthermore, when using the critic model (that is, the third target model) to train the actor model (that is, the whole composed of the first target model and the second target model), the training goal of the actor model is to simultaneously optimize the output of the local network value and the output value of the global network, based on the actor model trained in this way, it has the performance of balancing local and global performance, that is, it can balance the impact of a certain cell's load on the cell and the entire network after the load is adjusted, and realizes Coordination among multiple communities.

Description of drawings

Fig. 1 is a kind of structural schematic diagram of main frame of artificial intelligence;

FIG. 2a is a schematic structural diagram of a cell load adjustment system provided by an embodiment of the present application;

FIG. 2b is another schematic structural diagram of a cell load adjustment system provided by an embodiment of the present application;

FIG. 2c is a schematic diagram of related equipment for cell load adjustment provided by the embodiment of the present application;

FIG. 3 is a schematic diagram of the architecture of the system 100 provided by the embodiment of the present application;

FIG. 4 is a schematic flow diagram of a method for adjusting a cell load provided in an embodiment of the present application;

Fig. 5 is a schematic structural diagram of the actor model provided by the embodiment of the present application;

FIG. 6 is another schematic flowchart of a method for adjusting a cell load provided in an embodiment of the present application;

Fig. 7 is another schematic structural diagram of the actor model provided by the embodiment of the present application;

Fig. 8 is a schematic flow chart of the model training method provided by the embodiment of the present application;

FIG. 9 is another schematic flowchart of the model training method provided by the embodiment of the present application;

FIG. 10 is a schematic structural diagram of an apparatus for adjusting a cell load provided by an embodiment of the present application;

Fig. 11 is a schematic structural diagram of the model training device provided by the embodiment of the present application;

Fig. 12 is a schematic structural diagram of the execution device provided by the embodiment of the present application;

Fig. 13 is a schematic structural diagram of the training equipment provided by the embodiment of the present application;

FIG. 14 is a schematic structural diagram of a chip provided by an embodiment of the present application.

Detailed ways

The embodiment of the present application provides a cell load adjustment method and related equipment, which can accurately adjust the load level of the target cell, so as to keep various performance indicators of the target cell within an appropriate range, and provide users with a better network Serve.

The terms "first", "second" and the like in the specification and claims of the present application and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific sequence or sequence. It should be understood that the terms used in this way can be interchanged under appropriate circumstances, and this is merely a description of the manner in which objects with the same attribute are described in the embodiments of the present application. Furthermore, the terms "comprising" and "having", and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, product, or apparatus comprising a series of elements is not necessarily limited to those elements, but may include elements not expressly included. Other elements listed explicitly or inherent to the process, method, product, or apparatus.

In a wireless cellular network, the uneven distribution of users leads to unbalanced load among cells (that is, areas covered by the wireless network). Due to the large number of users and high service demand in high-load cells, it is easy to cause insufficient network resources, and it is difficult to guarantee the quality of service for users. Correspondingly, due to the small number of users and small business demand in low-load cells, their network resources are not fully utilized.

In order to adjust the load level of the cell, the configuration parameters of the cell can be modified so that the cell releases users or absorbs users (equivalent to reducing the load level of the cell or increasing the load level of the cell), thereby optimizing various performance indicators of the cell and providing users with Provide better network services.

When modifying the configuration parameters of the cell, the method adopted is usually manual modification. However, manual modification relies on the manual experience of experts, and the factors considered are often single and not comprehensive enough to accurately modify the configuration parameters of the cell so that the load level of the cell is within an appropriate range. All performance indicators are kept within an appropriate range to provide users with sufficient and high-quality network services.

In order to solve the above problems, an embodiment of the present application provides a cell load adjustment method, which can be implemented in combination with artificial intelligence (AI) technology. AI technology is a technical discipline that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence. AI technology obtains the best results by perceiving the environment, acquiring knowledge and using knowledge. In other words, artificial intelligence technology is a branch of computer science that attempts to understand the nature of intelligence and produce a new kind of intelligent machine that can respond in a similar way to human intelligence. Image processing using artificial intelligence is a common application of artificial intelligence.

First, describe the overall workflow of the artificial intelligence system, please refer to Figure 1, Figure 1 is a schematic structural diagram of the main framework of artificial intelligence, from the "intelligent information chain" (horizontal axis) and "IT value chain" (vertical axis) Two dimensions are used to elaborate the above artificial intelligence theme framework. Among them, the "intelligent information chain" reflects a series of processes from data acquisition to processing. For example, it can be the general process of intelligent information perception, intelligent information representation and formation, intelligent reasoning, intelligent decision-making, intelligent execution and output. In this process, the data has undergone a condensed process of "data-information-knowledge-wisdom". "IT value chain" reflects the value brought by artificial intelligence to the information technology industry from the underlying infrastructure of artificial intelligence, information (provided and processed by technology) to the systematic industrial ecological process.

(1) Infrastructure

The infrastructure provides computing power support for the artificial intelligence system, realizes communication with the outside world, and realizes support through the basic platform. Communicate with the outside through sensors; computing power is provided by smart chips (CPU, NPU, GPU, ASIC, FPGA and other hardware acceleration chips); the basic platform includes distributed computing framework and network and other related platform guarantees and supports, which can include cloud storage and Computing, interconnection network, etc. For example, sensors communicate with the outside to obtain data, and these data are provided to the smart chips in the distributed computing system provided by the basic platform for calculation.

(2) data

Data from the upper layer of the infrastructure is used to represent data sources in the field of artificial intelligence. The data involves graphics, images, voice, text, and IoT data of traditional equipment, including business data of existing systems and sensory data such as force, displacement, liquid level, temperature, and humidity.

(3) Data processing

Data processing usually includes data training, machine learning, deep learning, search, reasoning, decision-making, etc.

Among them, machine learning and deep learning can symbolize and formalize intelligent information modeling, extraction, preprocessing, training, etc. of data.

Reasoning refers to the process of simulating human intelligent reasoning in a computer or intelligent system, and using formalized information to carry out machine thinking and solve problems according to reasoning control strategies. The typical functions are search and matching.

Decision-making refers to the process of decision-making after intelligent information is reasoned, and usually provides functions such as classification, sorting, and prediction.

(4) General ability

After the above-mentioned data processing is performed on the data, some general capabilities can be formed based on the results of data processing, such as algorithms or a general system, such as translation, text analysis, computer vision processing, speech recognition, image processing identification, etc.

(5) Smart products and industry applications

Intelligent products and industry applications refer to the products and applications of artificial intelligence systems in various fields. It is the packaging of the overall solution of artificial intelligence, which commercializes intelligent information decision-making and realizes landing applications. Its application fields mainly include: intelligent terminals, intelligent transportation, Smart healthcare, autonomous driving, smart cities, etc.

Next, several application scenarios of this application are introduced.

Figure 2a is a schematic structural diagram of a cell load adjustment system provided by an embodiment of the present application, the system includes a network control center (also referred to as a data processing device), the network control center is connected to a plurality of wireless base stations, each wireless base station The area covered by the signal can be divided into multiple cells, and these cells constitute a wireless network that can provide communication services for users. The network control center can manage all the cells. For example, the network control center can collect the data of each cell, modify the configuration parameters of the cells, and monitor the status of each cell, etc.

The network control center receives the cell load adjustment request triggered by the network control center itself through the interactive interface, and then performs machine learning, deep learning, search, reasoning, decision-making and other methods of cell load through the memory for storing data and the processor for data processing. adjustment. The storage in the network control center can be a general term, including local storage and a database for storing historical data, and the database can be on the network control center or on other network servers.

In the cell load adjustment system shown in Figure 2a, the network control center can obtain a cell load adjustment request, and collect relevant information of a certain cell based on the request, and then, the network control center can process the relevant information of the cell, Thus, an operation for adjusting the load level of the cell is obtained. Exemplarily, the network control center can collect the characteristic information of a certain cell and the load index of the cell (used to indicate the load level of the cell), and then the network control center performs a series of processing on these information, so as to obtain the The operation of the load indicator of the cell.

In Fig. 2a, the network control center may execute the cell load adjustment method of the embodiment of the present application.

Fig. 2b is another schematic structural diagram of a cell load adjustment system provided by an embodiment of the present application. In Fig. 2b, the system includes user equipment and a network control center. Wherein, the user equipment includes smart terminals such as a mobile phone, a personal computer, or an information processing center. The user equipment is the initiator of the cell load adjustment, and as the initiator of the cell load adjustment request, usually a user (for example, an administrator of a wireless network, etc.) initiates the request through the user equipment.

In the image processing system shown in Figure 2b, the user equipment can receive the user's instruction and initiate a cell load adjustment request to the network control center, so that the network control center can realize the adjustment operation of the cell load, and the network control center can realize the adjustment process of the cell load Similar to FIG. 2a , reference may be made to the above description, and details are not repeated here.

In FIG. 2b, the network control center can also execute the cell load adjustment method of the embodiment of the present application.

FIG. 2c is a schematic diagram of related equipment for cell load adjustment provided by the embodiment of the present application.

The above-mentioned user equipment in FIG. 2a and FIG. 2b may specifically be the local device 301 or the local device 302 in FIG. 2c, and the data processing device in FIG. 2a may specifically be the execution device 210 in FIG. To store the data to be processed by the execution device 210, the data storage system 250 may be integrated on the execution device 210, or set on the cloud or other network servers.

The processors in Figure 2a and Figure 2b can perform data training/machine learning/deep learning through a neural network model or other models (for example, a model based on a support vector machine), and use the data to finally train or learn the model for the community. Relevant information is processed to obtain corresponding processing results.

FIG. 3 is a schematic diagram of the architecture of the system 100 provided by the embodiment of the present application. In FIG. 3, the execution device 110 is configured with an input/output (input/output, I/O) interface 112 for data interaction with external devices, and the user Data can be input to the I/O interface 112 through the client device 140, and the input data in this embodiment of the application may include: various tasks to be scheduled, callable resources, and other parameters.

When the execution device 110 preprocesses the input data, or when the calculation module 111 of the execution device 110 executes calculations and other related processing (such as implementing the function of the neural network in this application), the execution device 110 can call the data storage system 150 The data, codes, etc. in the system can be used for corresponding processing, and the data, instructions, etc. obtained by corresponding processing can also be stored in the data storage system 150 .

Finally, the I/O interface 112 returns the processing result to the client device 140, thereby providing it to the user.

It is worth noting that the training device 120 can generate corresponding target models/rules based on different training data for different goals or different tasks, and the corresponding target models/rules can be used to achieve the above-mentioned goals or complete the above-mentioned tasks , giving the user the desired result. Wherein, the training data may be stored in the database 130 and come from training samples collected by the data collection device 160 .

In the situation shown in FIG. 3 , the user can manually specify the input data, and the manual specification can be operated through the interface provided by the I/O interface 112 . In another case, the client device 140 can automatically send the input data to the I/O interface 112 . If the client device 140 is required to automatically send the input data to obtain the user's authorization, the user can set the corresponding authority in the client device 140 . The user can view the results output by the execution device 110 on the client device 140, and the specific presentation form may be specific ways such as display, sound, and action. The client device 140 can also be used as a data collection terminal, collecting the input data of the input I/O interface 112 and the output result of the output I/O interface 112 as new sample data, and storing them in the database 130 . Of course, it is also possible to collect without the client device 140, but the I/O interface 112 directly uses the input data input to the I/O interface 112 as shown in the figure and the output result of the output I/O interface 112 as a new sample The data is stored in database 130 .

It should be noted that FIG. 3 is only a schematic diagram of a system architecture provided by the embodiment of the present application, and the positional relationship between devices, devices, modules, etc. shown in the figure does not constitute any limitation. For example, in FIG. 3, the data The storage system 150 is an external memory relative to the execution device 110 , and in other cases, the data storage system 150 may also be placed in the execution device 110 . As shown in FIG. 3 , the neural network can be obtained by training according to the training device 120 .

An embodiment of the present application also provides a chip, the chip includes a neural network processor (NPU). The chip can be set in the execution device 110 shown in FIG. 3 to complete the computing work of the computing module 111 . The chip can also be set in the training device 120 shown in FIG. 3 to complete the training work of the training device 120 and output the target model/rule.

The neural network processor NPU is mounted on the main central processing unit (central processing unit, CPU) (host CPU) as a coprocessor, and the main CPU assigns tasks. The core part of the NPU is the operation circuit, and the controller controls the operation circuit to extract the data in the memory (weight memory or input memory) and perform operations.

In some implementations, the operation circuit includes multiple processing units (process engine, PE). In some implementations, the arithmetic circuit is a two-dimensional systolic array. The arithmetic circuit may also be a one-dimensional systolic array or other electronic circuitry capable of performing mathematical operations such as multiplication and addition. In some implementations, the arithmetic circuit is a general purpose matrix processor.

For example, suppose there is an input matrix A, a weight matrix B, and an output matrix C. The operation circuit fetches the data corresponding to the matrix B from the weight memory, and caches it on each PE in the operation circuit. The operation circuit takes the data of matrix A from the input memory and performs matrix operation with matrix B, and the obtained partial or final results of the matrix are stored in the accumulator.

The vector calculation unit can further process the output of the operation circuit, such as vector multiplication, vector addition, exponential operation, logarithmic operation, size comparison and so on. For example, the vector computing unit can be used for network calculations of non-convolution/non-FC layers in neural networks, such as pooling, batch normalization, local response normalization, etc.

In some implementations, the vector computation unit can store the processed output vectors to a unified register. For example, a vector computation unit may apply a non-linear function to the output of the arithmetic circuit, such as a vector of accumulated values, to generate activation values. In some implementations, the vector computation unit generates normalized values, merged values, or both. In some implementations, the vector of processed outputs can be used as an activation input to an operational circuit, eg, for use in a subsequent layer in a neural network.

Unified memory is used to store input data and output data.

The weight data directly transfers the input data in the external memory to the input memory and/or the unified memory through the storage unit access controller (direct memory access controller, DMAC), stores the weight data in the external memory into the weight memory, and stores the weight data in the unified memory Store the data in the external memory.

The bus interface unit (bus interface unit, BIU) is used to realize the interaction between the main CPU, DMAC and instruction fetch memory through the bus.

The instruction fetch buffer connected to the controller is used to store the instructions used by the controller;

The controller is used for invoking instructions cached in the memory to control the working process of the computing accelerator.

Generally, the unified memory, the input memory, the weight memory and the instruction fetch memory are all on-chip (On-Chip) memory, and the external memory is the memory outside the NPU, and the external memory can be a double data rate synchronous dynamic random access memory (double data rate synchronous dynamic random access memory, DDR SDRAM), high bandwidth memory (high bandwidth memory, HBM) or other readable and writable memory.

Since the embodiment of the present application involves the application of a large number of neural networks, in order to facilitate understanding, the following first introduces related terms and neural network related concepts involved in the embodiment of the present application.

(1) neural network

The neural network can be composed of neural units, and the neural unit can refer to an operation unit that takes xs and intercept 1 as input, and the output of the operation unit can be:

Wherein, s=1, 2, ... n, n is a natural number greater than 1, Ws is the weight of xs, and b is the bias of the neuron unit. f is the activation function of the neural unit, which is used to introduce nonlinear characteristics into the neural network to convert the input signal in the neural unit into an output signal. The output signal of this activation function can be used as the input of the next convolutional layer. The activation function may be a sigmoid function. A neural network is a network formed by connecting many of the above-mentioned single neural units, that is, the output of one neural unit can be the input of another neural unit. The input of each neural unit can be connected with the local receptive field of the previous layer to extract the features of the local receptive field. The local receptive field can be an area composed of several neural units.

The work of each layer in the neural network can be described by the mathematical expression y=a(Wx+b): From the physical level, the work of each layer in the neural network can be understood as through five kinds of input space (input vector set) to complete the transformation from the input space to the output space (that is, the row space of the matrix to the column space), these five operations include: 1. Dimension enhancement/dimension reduction; Translation; 5. "Bending". Among them, the operations of 1, 2, and 3 are completed by Wx, the operation of 4 is completed by +b, and the operation of 5 is realized by a(). The reason why the word "space" is used here is because the classified object is not a single thing, but a kind of thing, and space refers to the collection of all individuals of this kind of thing. Wherein, W is a weight vector, and each value in the vector represents the weight value of a neuron in this layer of neural network. The vector W determines the space transformation from the input space to the output space described above, that is, the weight W of each layer controls how to transform the space. The purpose of training the neural network is to finally obtain the weight matrix of all layers of the trained neural network (the weight matrix formed by the vector W of many layers). Therefore, the training process of the neural network is essentially to learn the way to control the spatial transformation, and more specifically, to learn the weight matrix.

Because we want the output of the neural network to be as close as possible to the value we really want to predict, we can compare the predicted value of the current network with the target value we really want, and then update each layer of neural network according to the difference between the two (Of course, there is usually an initialization process before the first update, that is, to pre-configure parameters for each layer in the neural network). For example, if the network's predicted value is high, adjust the weight vector to make it predict low Some, keep adjusting until the neural network can predict the desired target value. Therefore, it is necessary to pre-define "how to compare the difference between the predicted value and the target value", which is the loss function (loss function) or objective function (objective function), which is used to measure the difference between the predicted value and the target value important equation. Among them, taking the loss function as an example, the higher the output value (loss) of the loss function, the greater the difference, so the training of the neural network becomes a process of reducing the loss as much as possible.

(2) Back propagation algorithm

The neural network can use the error back propagation (back propagation, BP) algorithm to correct the size of the parameters in the initial neural network model during the training process, so that the reconstruction error loss of the neural network model becomes smaller and smaller. Specifically, passing the input signal forward until the output will generate an error loss, and updating the parameters in the initial neural network model by backpropagating the error loss information, so that the error loss converges. The backpropagation algorithm is a backpropagation movement dominated by error loss, aiming to obtain the optimal parameters of the neural network model, such as the weight matrix.

The method provided by this application is described below from the training side of the neural network and the application side of the neural network.

The model training method provided in the embodiment of the present application involves image processing, and can be specifically applied to data processing methods such as data training, machine learning, and deep learning. characteristic information, the load index of the target cell, etc.) to perform symbolic and formalized intelligent information modeling, extraction, preprocessing, training, etc., and finally obtain a trained neural network (such as the first target model and the first target model in the embodiment of the present application) second target model); and, the cell load adjustment method provided by the embodiment of the present application can use the above-mentioned trained neural network to input data (such as the characteristic information of the target cell in the cell load adjustment method of the embodiment of the present application , the load index of the target cell) into the trained neural network to obtain output data (such as the operation instructions in the cell load adjustment method of the embodiment of the present application, etc.). It should be noted that the model training method and the cell load adjustment method provided in the embodiment of this application are inventions based on the same idea, and can also be understood as two parts in a system, or two stages in an overall process: Such as model training phase and model application phase.

It is worth noting that the embodiment of the present application can be implemented based on the actor-critic model architecture in reinforcement learning, which includes the actor model and the critic model, wherein the actor model can be used to implement the cell load adjustment method provided by the embodiment of the present application (That is, the model application stage), the actor model contains two parts of the neural network, one part is the action network, and the other part is the weight network. The critic model can be used to implement the model training method provided by the embodiment of the present application (i.e. the model training stage), the purpose is to train the aforementioned actor model, and the critic model can also include two parts of the neural network, one part is a local network and the other part is a global network . For the convenience of explanation, the weight network is called the first target model, the action network is called the second target model, and the entire critic model is called the third target model.

In the following, the model application stage is first introduced. Figure 4 is a schematic flow chart of the method for adjusting the cell load provided by the embodiment of the present application. This method can be realized through the actor model, and its output is an operation (action). A schematic structural diagram of the actor model provided in the embodiment, as shown in FIG. 5, the actor model includes a first target model and a second target model, and the output end of the first target model acts on the second target model, which can also be understood as the first target model The output of one object model is used to construct a second object model. As shown in Figure 4, the method includes:

401. Acquire feature information of a target cell and a load index of the target cell, where the load index of the target cell is used to indicate a load degree of the target cell.

In this embodiment, when the load level of the target cell needs to be adjusted, the characteristic information c of the target cell and the load index x of the target cell can be collected, wherein the characteristic information c of the target cell can be used to indicate the current location of the target cell In the scenario, the load index x of the target cell may be used to indicate the load level of the target cell (also called the load balance level).

Specifically, the characteristic information c of the target cell may include at least one of the following information:

Configuration parameters of the target cell;

Traffic statistics data of the target cell;

configuration parameters of neighbor cells of the target cell;

Neighboring cell traffic statistics data, etc.

Further, the configuration parameters of the target cell may include at least one of the following information:

Antenna transmit power of the target cell;

Antenna downtilt of the target cell;

Antenna horizontal azimuth of the target cell;

The user in the target cell starts the reference signal receiving power (reference signal receiving power, RSRP) threshold value of inter-frequency handover measurement;

The RSRP threshold at which users in the target cell stop inter-frequency handover measurement;

The user in the target cell triggers the specific frequency RSRP offset of the inter-frequency handover procedure;

The user in the target cell triggers the RSRP offset of the specific neighboring cell for the inter-frequency handover process;

The user in the target cell starts the RSRP threshold of the same-frequency handover measurement;

The user in the target cell stops the RSRP threshold of the same-frequency handover measurement;

Specific frequency RSRP offset for users in the target cell to trigger intra-frequency handover procedures;

The user in the target cell triggers the specific neighbor cell RSRP offset of the intra-frequency handover process, etc.

Further, the statistics data of the target cell may include at least one of the following information:

The average number of users per unit time period in the target cell;

The average number of active users per unit time period in the target cell;

Uplink traffic per unit time period of the target cell;

The proportion of low channel quality indicator (channel quality indicator, CQI) reports in the unit time period of the target cell;

The ratio of packets whose length per unit time period of the target cell is less than the preset length threshold (also referred to as the ratio of small packets per unit time period of the target cell);

The average length of data packets in a unit time period of the target cell, etc.

Further, the configuration parameters of neighbor cells may include at least one of the following information:

Antenna transmit power of neighboring cells;

Antenna downtilt of neighboring cells;

Antenna horizontal azimuth of neighbor cell;

The user in the neighboring cell initiates the RSRP threshold of the inter-frequency handover measurement;

The RSRP threshold at which the user in the neighbor cell stops the inter-frequency handover measurement;

Specific frequency RSRP offset for users in neighboring cells to trigger inter-frequency handover procedures;

The RSRP offset of the specific neighbor cell triggering the inter-frequency handover process by the user in the neighbor cell;

The RSRP threshold for the user in the neighboring cell to start the same-frequency handover measurement;

The user in the neighbor cell stops the RSRP threshold of the same-frequency handover measurement;

Specific frequency RSRP offset for users in neighboring cells to trigger intra-frequency handover procedures;

The user in the neighbor cell triggers the specific neighbor cell RSRP offset of the intra-frequency handover procedure, etc.

Further, the traffic statistics data of neighbor cells may include at least one of the following information:

The average number of users per unit time period in neighboring cells;

The average number of active users per unit time period in neighboring cells;

Uplink traffic per unit time period of neighboring cells;

The proportion of CQI reports in the unit time period of neighboring cells;

The proportion of data packets whose length per unit time period of the neighbor cell is less than the preset length threshold (also referred to as the proportion of small packets per unit time period of the neighbor cell);

The average length of data packets in a unit time period of a neighboring cell, etc.

402. Process the feature information of the target cell by using the first target model to obtain model parameters of the second target model.

403. Acquire a second target model based on model parameters of the second target model.

After obtaining the characteristic information c of the target cell and the load index x of the target cell, the first target model can be obtained, and the first target model is a trained neural network. Then, input the feature information c of the target cell into the first target model, so as to perform a series of processing (for example, feature extraction, etc.) on the feature information c of the target cell through the first target model to obtain the model parameters of the second target model (It can also be called the weight of the second target model) W(c) and b(c). It should be noted that certain operations (eg, softplus activation function, etc.) can be performed in the first target model so that W(c) is a non-negative vector, that is, W(c)≥0.

Then, the second target model can be constructed based on the model parameters W(c) and b(c) of the second target model, and the second target model can be a multilayer perceptron (MLP). So far, the whole composed of the first target model and the second target model (that is, the whole actor model) can be used to indicate the functional relationship between the load index of the target cell and the operation instruction for the target cell (also can be understood as the first target model and the second target model can be expressed through the functional relationship), the functional relationship can be expressed as:

a=σ(f _M (x,c)) (2)

f _M (x, c) = MLP(W(c), b(c), x) (3)

In the above formula, a is the operation instruction for the target cell; f _M (x, c) is a monotonically increasing function of the load index x of the target cell, the model parameter W (c) is the slope of f _M (x, c), and the model The parameter b(c) is part of the remaining parameters of f _M (x, c); σ is a monotonically increasing function (for example, tanh activation function, etc.), which is used to constrain the value range of the operation a for the target cell to be [a _L , a _H ]. Then, since both f _M (x, c) and σ are monotonically increasing functions, the operation instruction a for the target cell is also a monotonically increasing function of the load index x of the target cell, and the increasing speed is determined by the characteristic information c of the target cell Decide.

It can be seen that when the load index x of the target cell is larger (that is, the load degree of the target cell is larger), the operation indication a for the target cell is larger, indicating that the target cell needs to release more users (or the target cell needs to release more users). absorb fewer users). When the load index x of the target cell is smaller (that is, the load degree of the target cell is smaller), the operation indication a for the target cell is smaller, indicating that the target cell needs to release fewer users (or the target cell needs to absorb more users ). It can be seen that the form of formula (2) is consistent with the expert experience strategy.

It should be understood that, in this embodiment, only the operation indication a of the target cell is a monotonically increasing function of the load index x of the target cell as an example for schematic illustration. In practical applications, the operation indication a of the target cell may also be The monotonically decreasing function of the load index x of , only needs to control the model parameter W(c)<0 or replace σ with a monotonically decreasing function. It can be understood that, in this case, when the load index x of the target cell is larger (that is, the load degree of the target cell is larger), the operation indication a for the target cell is smaller, indicating that the target cell needs to release more users (or the fewer users the target cell needs to absorb). When the load index x of the target cell is smaller (that is, the load degree of the target cell is smaller), the operation indication a for the target cell is larger, indicating that the target cell needs to release fewer users (or the target cell needs to absorb more users ).

It should also be understood that in this embodiment, only the softplus activation function and the tanh activation function are used for schematic illustration, which does not limit the types of activation functions in this application. In practical applications, other activation functions with similar properties can also be used.

404. Process the load index by using the second target model to obtain an operation instruction for the target cell, and the operation instruction for the target cell is used to adjust the load index of the target cell.

After constructing the second target model, the load index x of the target cell can be input into the second target model, so as to process the load index x of the target cell through the second target model to obtain the operation instruction a for the target cell, and for the target cell The operation indication a of can be used to adjust the load index x of the target cell.

Specifically, the load index x of the target cell is usually determined based on the traffic statistics data of the target cell and the neighbor cell. For example, the load index x of the target cell can be the average number of users per unit time period of the target cell and The ratio between the average number of users. For another example, the load index x of the target cell may be a ratio between the average number of active users per unit time period of the target cell and the average number of active users per unit time period of the neighbor cell. For another example, the load index x of the target cell may be a ratio between the uplink traffic of the target cell per unit time period and the uplink traffic of the neighbor cell per unit time period, and so on.

The traffic statistics data of the target cell used to determine the load index x of the target cell is often affected by one or some configuration parameters of the target cell, so the operation instruction a for the target cell can be used to modify the configuration parameters of the target cell to indirectly Adjust the load index x of the target cell accurately. For example, when the load index x of the target cell is the ratio between the average number of users in the unit time period of the target cell and the average number of users in the neighbor cell in the unit time period, since the average user in the unit time period of the target cell is affected by the The impact of the RSRP threshold for inter-frequency handover measurement (for example, if the RSRP threshold for users in the target cell to start inter-frequency handover measurement is set to be larger, the average number of users per unit time period in the target cell is smaller), then, for the target cell The operation instruction a can be used to modify the RSRP threshold for users in the target cell to start inter-frequency handover measurement, thereby affecting the value of the average user per unit time period of the target cell, so as to adjust the average number of users per unit time period of the target cell and the neighbor cell The ratio between the average number of users per unit time period.

Further, the operation instruction a for the target cell often represents the modification range of the configuration parameters of the target cell, that is, the adjustment range of the load index x of the target cell. Still as in the above example, when the operation instruction a for the target cell is larger, the RSRP threshold for users in the target cell to start inter-frequency handover measurement can be modified to be larger, so that the average number of users per unit time period in the target cell is less (that is, let the target cell release enough users), the greater the load reduction of the target cell is. It can be seen that after such adjustment, the adjusted load index x' of the target cell can be limited within an appropriate range.

To sum up, after adjusting the load index x of the target cell based on the operation instruction a for the target cell, the adjusted load index x` of the target cell can be kept within an appropriate range, so the performance indexes of the target cell ( For example, the average downlink perception rate of users in the unit time period of the target cell, the average uplink perception rate of users in the unit time period of the target cell, the average delay of sending data packets by users in the unit time period of the target cell, and the rate in the unit time period of the target cell The proportion of users below 5M, etc.) can also be optimized accordingly, so as to provide users in the target cell with better network services.

In the embodiment of the present application, after acquiring the feature information of the target cell and the load index of the target cell, the feature information may be processed by the first target model to obtain model parameters of the second target model. Then, a second target model is constructed based on the model parameters of the second target model, and the load index is processed through the second target model to obtain an operation instruction for the target cell, and the operation instruction can be used to adjust the load index of the target cell. In the foregoing process, the model parameters of the second target model are obtained based on the characteristic information of the target cell, since the characteristic information of the target cell can be used to characterize various characteristics of the target cell (ie, the scene in which the target cell is located), then, in When using the second target model to process the load index of the target cell, various factors such as various characteristics of the target cell are considered, so the operation instructions for the target cell output by the second target model can be used to accurately adjust the target cell. The load level is used to keep various performance indicators of the target cell within an appropriate range and provide users with better network services.

Further, the whole formed by the first target model and the second target model can be used to indicate the functional relationship between the load index of the target cell and the operation instruction for the target cell, the functional relationship is usually a monotonically increasing or monotonically decreasing relationship, In line with the constraints of expert experience (equivalent to integrating expert experience into the model), the control strategy learned by the model is more in line with business logic.

Fig. 6 is another schematic flowchart of the cell load adjustment method provided by the embodiment of the present application. This method can be realized through an actor model, and the output of the actor model is an operation. Fig. 7 is another example of the actor model provided by the embodiment of the present application. Schematic diagram of the structure, as shown in Figure 7, the actor model includes the first target model, the first sub-model (also called the teacher network) and the second sub-model (also called the student network), the output end of the first target model Acting on the first sub-model and the second sub-model can also be understood as the output of the first target model is used to construct the first sub-model and the second sub-model. As shown in Figure 6, the method includes:

601. Acquire feature information of a target cell and a load index of the target cell, where the load index is used to indicate a load degree of the target cell.

For the description of step 601, reference may be made to the related description of step 401 in the embodiment shown in FIG. 4 , and details are not repeated here.

602. Process the feature information by using the first target model to obtain model parameters of the first sub-model and model parameters of the second sub-model.

603. Acquire the first submodel based on the model parameters of the first submodel, and acquire the second submodel based on the model parameters of the second submodel.

After obtaining the characteristic information c of the target cell and the load index x of the target cell, the first target model can be obtained, and the first target model is a trained neural network. Then, input the feature information c of the target cell into the first target model, so as to perform a series of processing (for example, feature extraction, etc.) on the feature information c of the target cell through the first target model to obtain the model parameters of the first sub-model W(c), b(c), _hT (c) and _vT (c), and the model parameters of the second submodel W(c), b(c), _hS (c) and _vS (c ). It should be noted that certain operations (eg, softplus activation function, etc.) can be performed in the first target model to make W(c) a non-negative vector, that is, W(c)≥0.

Then, based on the model parameters W(c), b(c), h _T (c) and v _T (c) of the first sub-model, the first sub-model can be constructed, and based on the model parameters W(c) of the second sub-model , b(c), h _S (c) and v _S (c) can construct the second sub-model, and both the first sub-model and the second sub-model can be MLP. So far, the whole composed of the first target model and the first sub-model can be used to indicate the first functional relationship between the load index of the target cell and the first operation for the target cell (also can be understood as the first target model and the first The whole composed of sub-models can be expressed by this functional relationship), the first functional relationship can be expressed as:

a _T =σ(f _M (xh _T (c), c)+v _T (c)) (5)

f _M (xh _T (c), c) + v _T (c) = MLP (W (c), b (c), x h _T (c), v _T (c)) (6)

In the above formula, a _T is the first operation for the target cell; f _M (xh _T (c), c)+v _T (c) is a monotonically increasing function of the load index x of the target cell, and the model parameter W(c) is the slope of f _M (xh _T (c), c)+v _T (c), and the model parameter b(c) is part of the remaining parameters of f _M (x, c). It should be noted that f _M ( xh _T (c), c)+v _T (c) can be regarded as obtained by performing vertical translation and horizontal translation of f _M (x, c) shown in the aforementioned formula (3), where the model parameter h _T ( c) is the magnitude of horizontal translation, v _T (c) is the magnitude of vertical translation; σ is a monotonically increasing function (for example, tanh activation function, etc.), used to constrain the value of the first operation a _T for the target cell The range is [a _L , a _H ]. Then, since f _M (xh _T (c), c)+v _T (c) and σ are both monotonically increasing functions, the first operation a _T for the target cell is also a monotonically increasing function of the load index x of the target cell , and the increasing speed is determined by the characteristic information c of the target cell.

It is worth noting that the first functional relationship also needs to meet the following constraints:

a ₀ ＝σ(f _M (τ-h _T (c), c)+v _T (c)) (7)

In the above formula, τ is the preset index threshold. If the target cell is the target cell, it can be seen that when the load index x of the target cell = τ, the first operation a _T = a ₀ for the target cell, and the value of a ₀ is located in [a _L , a _H ], it means that the load index of the target cell does not need to be adjusted, that is, the target cell neither needs to absorb users nor release users.

Similarly, the whole composed of the first target model and the second sub-model can be used to indicate the second functional relationship between the load index of the target cell and the second operation for the target cell (also can be understood as the first target model and the second sub-model The whole composed of a sub-model can be expressed by this functional relationship), and the second functional relationship can be expressed as:

a _S ＝σ(f _M (xh _S (c), c)+v _S (c)) (8)

f _M (xh _S (c), c) + v _S (c) = MLP (W (c), b (c), xh _S (c), v _S (c)) (9)

In the above formula, a _S is the second operation for the target cell; f _M (xh _S (c), c)+v _S (c) is a monotonically increasing function of the load index x of the target cell, and the model parameter W(c) is the slope of f _M (xh _S (c), c)+v _S (c), and the model parameter b(c) is part of the remaining parameters of f _M (x, c). It should be noted that f _M ( xh _S (c), c)+v _S (c) can be regarded as obtained by performing vertical translation and horizontal translation of f _M (x, c) shown in the aforementioned formula (3), where the model parameter h _S ( c) is the magnitude of horizontal translation, v _S (c) is the magnitude of vertical translation; σ is a monotonically increasing function (for example, tanh activation function, etc.), used to constrain the value of the second operation a _S for the target cell The range is [a _L , a _H ]. Then, since f _M (xh _S (c), c)+v _S (c) and σ are both monotonically increasing functions, the second operation a _T for the target cell is also a monotonically increasing function of the load index x of the target cell , and the increasing speed is determined by the characteristic information c of the target cell.

It is worth noting that f _M (xh _T (c), c)+v _T (c) and f _M (xh _S (c), c)+v _S (c) are based on f _M (x, c) Then, f _M (xh _S (c), c)+v _S (c) can also be regarded as the result of translation based on f _M (xh _T (c), c)+v _T (c). obtained, and the magnitude of the translation is determined based on the characteristic information c of the target cell.

Then, the operation indication a for the target cell can be determined by the first operation a _T for the target cell and the second operation a _S for the target cell:

a＝w ₁ *a _T +w ₂ *a _S (10)

In the above formula, w ₁ and w ₂ are preset weight values, and their sizes can be set according to actual needs, and are not limited here.

Based on the formula (10), it can be seen that the operation instruction a for the target cell is also a monotonically increasing function of the load index x of the target cell, that is, when the load index x of the target cell is larger (that is, the load degree of the target cell is larger), for The greater the operation indication a of the target cell, it means that the target cell needs to release more users (or the target cell needs to absorb fewer users). When the load index x of the target cell is smaller (that is, the load degree of the target cell is smaller), the operation indication a for the target cell is smaller, indicating that the target cell needs to release fewer users (or the target cell needs to absorb more users ). It can be seen that the form of formula (2) is consistent with the expert experience strategy.

It should be understood that in this embodiment, only the relationship between the first functional relationship and the second functional relationship is monotonically increasing for schematic illustration. In practical applications, the first functional relationship and the second functional relationship may also be monotonically decreasing. , you only need to control the model parameter W(c)<0 or replace σ with a monotonically decreasing function. It can be understood that, in this case, the operation instruction a for the target cell is also a monotonically decreasing function of the load index x of the target cell, that is, when the load index x of the target cell is larger (that is, the load degree of the target cell The larger the value is, the smaller the operation indication a for the target cell is, indicating that the target cell needs to release more users (or the target cell needs to absorb fewer users). When the load index x of the target cell is smaller (that is, the load degree of the target cell is smaller), the operation indication a for the target cell is larger, indicating that the target cell needs to release fewer users (or the target cell needs to absorb more users ).

It should also be understood that in this embodiment, only the softplus activation function and the tanh activation function are used for schematic illustration, which does not limit the types of activation functions in this application. In practical applications, other activation functions with similar properties can also be used.

604. Process the load index of the target cell by using the first sub-model to obtain a first operation.

605. Process the load index of the target cell by using the second sub-model to obtain a second operation.

606. Perform weighted summation on the first operation and the second operation to obtain an operation instruction.

After constructing the first sub-module and the second sub-model, the load index x of the target cell can be input into the first sub-model and the second sub-model respectively, so as to process the load index x of the target cell through the first sub-model, The first operation a _T for the target cell is obtained, and the load index x of the target cell is processed through the second sub-model to obtain the second operation a _S for the target cell. Then, the first operation a _T for the target cell and the second operation a _S for the target cell are weighted and summed to obtain the operation instruction a for the target cell.

As for how to adjust the load index x of the target cell by using the operation instruction a for the target cell, refer to the relevant description of step 404 in the embodiment shown in FIG. 4 , which will not be repeated here.

In the embodiment of this application, the first sub-model and the second sub-model are constructed through the framework of the teacher-student network. The first sub-model adds expert constraints (that is, using τ as a threshold to determine whether the target cell absorbs or releases users), The safety of the control strategy output by the model is guaranteed, while the second sub-model relaxes the expert constraints, learns and outputs the corresponding control strategy based on data, and weights the output of the two models as the final strategy, thus balancing the control strategy. safety and efficiency.

The above is a detailed introduction to the model application phase, and the model training phase will be introduced below. Fig. 8 is a schematic flow chart of the model training method provided by the embodiment of the present application. As shown in Fig. 8, the method includes:

801. Acquire feature information of a first cell and a load index of the first cell, where the load index of the first cell is used to indicate a load degree of the first cell.

When it is necessary to train the first model to be trained (that is, the neural network to be trained), a batch of training samples can be obtained, that is, the feature information of the first cell used for training and the load index of the first cell.

In a possible implementation manner, the feature information of the first cell includes at least one of the following: configuration parameters of the first cell; traffic statistics data of the first cell; configuration parameters of neighbor cells of the first cell; data.

In a possible implementation, the configuration parameters include at least one of the following: antenna transmit power; antenna downtilt angle; antenna horizontal azimuth angle; the RSRP threshold for the user to start inter-frequency handover measurement; the RSRP threshold for the user to stop inter-frequency handover measurement ; Specific frequency RSRP offset for user triggering inter-frequency handover process; specific neighbor RSRP offset for user triggering inter-frequency handover process; RSRP threshold for user to start co-frequency handover measurement; RSRP threshold for user to stop co-frequency handover measurement; user triggered The frequency-specific RSRP offset of the same-frequency handover process; the specific neighbor cell RSRP offset of the user-triggered same-frequency handover process.

In a possible implementation, the traffic statistics data includes at least one of the following: the average number of users per unit time period; the average number of active users per unit time period; the uplink traffic per unit time period; the CQI report per unit time period Proportion; the proportion of data packets whose length per unit time period is less than the preset length threshold; the average length of data packets per unit time period.

For the description of the feature information of the first cell and the load index of the first cell, refer to the relevant description of step 401 in the embodiment shown in FIG. 4 , and details are not repeated here.

802. Process the feature information of the first cell by using the first model to be trained to obtain model parameters of the second model to be trained.

803. Acquire a second model to be trained based on model parameters of the second model to be trained.

804. Process the load index of the first cell by using the second model to be trained to obtain an operation instruction for the first cell, and the operation instruction for the first cell is used to adjust the load index of the first cell.

For descriptions of steps 802 to 804, reference may be made to relevant descriptions of steps 402 to 404 in the embodiment shown in FIG. 4 , and details are not repeated here.

805. Process the operation instruction for the first cell and the operation instruction for the second cell through the third target model to obtain a first score and a second score, and the first score is used to evaluate the operation instruction for the first cell on the first cell The second score is used to evaluate the impact of the operation indication of the first cell on the performance index of the entire network, and the operation indication of the second cell is used to adjust the load index of the second cell.

After the operation instruction for the first cell is obtained, the operation instruction for the second cell can also be obtained. It should be noted that the operation instruction for the second cell is used to adjust the load index of the second cell, and the operation instruction for the second cell The acquisition process of is also the same as the acquisition process of the operation indication for the first cell, which will not be repeated here. It should be noted that the first cell and the second cell are different cells in the entire network, and the number of the second cell may be one or more.

Then, a third target model (ie, the aforementioned critic model) can be obtained, and the third target model is a trained neural network. The third target model includes two neural networks, one part is a local network and the other part is a global network. Then, the characteristic information of the first cell, the load index of the first cell and the operation instruction for the first cell can be input into the value local network , to process the information through the local network to obtain a first score, and the first score is used to evaluate the impact of the operation instruction of the first cell on the load index of the first cell. At the same time, the feature information of the first cell, the operation instruction for the first cell, the feature information of the second cell and the operation instruction for the second cell can also be input into the global network to process these information through the global network , to obtain a second score, and the second score is used to evaluate the impact of the operation instruction of the first cell on the performance index (for example, rate, throughput, edge user proportion, etc.) of the entire network (network-wide).

806. Based on the first score and the second score, update the model parameters of the first model to be trained until the model training condition is met, and obtain the first target model.

Based on the first score and the second score, update the model parameters of the first model to be trained, and use the next batch of training samples to train the first model to be trained after the updated parameters (that is, re-execute steps 802 to 806) , until the model training conditions are met, and the first target model in the embodiment shown in FIG. 4 is obtained, which is equivalent to obtaining the second target model in the embodiment shown in FIG. 4 .

It is worth noting that the model training condition can be: the result of the weighted sum of the first score and the second score, after reducing to 80% of the peak value (of course, it can also be other percentages, there is no limit here) , stop training. The model training condition can also be: stop training after the first score drops to 80% of the peak value (of course, it can also be other percentages, which is not limited here). Of course, the model training conditions may also be other similar conditions.

In the embodiment of the present application, when using the critic model (that is, the third target model) to train the actor model (that is, the whole composed of the first target model and the second target model), the training goal of the actor model is to simultaneously optimize the local network The output value of the output value and the output value of the global network, based on the actor model trained in this way, has the performance of balancing local and global performance, that is, it can balance the impact of the load of a certain cell on the cell and the entire network. , to achieve coordination between multiple cells.

Fig. 9 is another schematic flowchart of the model training method provided by the embodiment of the present application. As shown in Fig. 9, the method includes:

901. Acquire feature information of a first cell and a load index of the first cell, where the load index of the first cell is used to indicate a load degree of the first cell.

For descriptions of step 901, reference may be made to relevant descriptions of step 801 in the embodiment shown in FIG. 8 , and details are not repeated here.

902. Process the feature information of the first cell by using the first model to be trained to obtain model parameters of the first sub-model and model parameters of the second sub-model.

903. Acquire the first submodel based on the model parameters of the first submodel, and acquire the second submodel based on the model parameters of the second submodel.

In a possible implementation manner, the first model to be trained and the first sub-model are used to indicate a first functional relationship between the load index of the first cell and the operation indication for the first cell, and the first model to be trained and The second sub-model is used to indicate the second functional relationship between the load index of the first cell and the operation indication for the first cell, the second functional relationship is obtained by shifting the first functional relationship, the first functional relationship and The second functional relationship is a monotonically increasing relationship or a monotonically decreasing relationship.

In a possible implementation manner, the magnitude of translation is determined based on feature information.

904. Process the load index of the first cell by using the first sub-model to obtain a first operation.

905. Process the load index of the first cell by using the second sub-model to obtain a second operation.

906. Perform weighted summation on the first operation and the second operation to obtain an operation instruction for the first cell, where the operation instruction for the first cell is used to adjust a load index of the first cell.

For descriptions of steps 902 to 906, reference may be made to relevant descriptions of steps 602 to 606 in the embodiment shown in FIG. 6 , and details are not repeated here.

907. Process the operation instruction for the first cell and the operation instruction for the second cell through the third target model to obtain a first score and a second score, and the first score is used to evaluate the operation instruction for the first cell on the first cell The second score is used to evaluate the impact of the operation indication of the first cell on the performance index of the entire network, and the operation indication of the second cell is used to adjust the load index of the second cell.

908. Based on the first score and the second score, update the model parameters of the first model to be trained until the model training condition is met, and obtain the first target model.

The first target model trained in this embodiment is the first target model in the embodiment shown in Figure 6. Similarly, after obtaining this model, it is equivalent to obtaining the first target model in the embodiment shown in Figure 6. submodel and a second submodel.

For descriptions of steps 907 to 908, reference may be made to relevant descriptions of steps 805 to 806 in the embodiment shown in FIG. 8 , and details are not repeated here.

In the embodiment of the present application, when using the critic model (that is, the third target model) to train the actor model (that is, the whole composed of the first target model and the second target model), the training goal of the actor model is to simultaneously optimize the local network The output value of the output value and the output value of the global network, based on the actor model trained in this way, has the performance of balancing local and global performance, that is, it can balance the impact of the load of a certain cell on the cell and the entire network. , to achieve coordination between multiple cells.

The above is a detailed description of the cell load adjustment method and model training method provided by the embodiment of the present application. The following will introduce the cell load adjustment device and model training device provided by the embodiment of the present application. FIG. 10 is a schematic structural diagram of a cell load adjusting device provided in the embodiment of the present application. As shown in FIG. 10 , the device includes:

The first acquiring module 1001 is configured to acquire characteristic information of the target cell and a load index of the target cell, where the load index is used to indicate the load degree of the target cell;

The first processing module 1002 is configured to process feature information through the first target model to obtain model parameters of the second target model;

The second acquisition module 1003 is configured to acquire a second target model based on model parameters;

The second processing module 1004 is configured to process the load index through the second target model to obtain an operation instruction, and the operation instruction is used to adjust the load index.

In the embodiment of the present application, after acquiring the feature information of the target cell and the load index of the target cell, the feature information may be processed by the first target model to obtain model parameters of the second target model. Then, a second target model is constructed based on the model parameters of the second target model, and the load index is processed through the second target model to obtain an operation instruction for the target cell, and the operation instruction can be used to adjust the load index of the target cell. In the foregoing process, the model parameters of the second target model are obtained based on the characteristic information of the target cell, since the characteristic information of the target cell can be used to characterize various characteristics of the target cell (ie, the scene in which the target cell is located), then, in When using the second target model to process the load index of the target cell, various factors such as various characteristics of the target cell are considered, so the operation instructions for the target cell output by the second target model can be used to accurately adjust the target cell. The load level is used to keep various performance indicators of the target cell within an appropriate range and provide users with better network services.

In a possible implementation manner, the second target model includes the first sub-model and the second sub-model, and the second processing module 1004 is configured to: process the load index through the first sub-model to obtain the first operation, if If the load index is equal to the preset index threshold, the first operation is used to not adjust the load index; the second operation is obtained by processing the load index through the second sub-model; the weighted sum of the first operation and the second operation is obtained Operating instructions.

In a possible implementation, the first target model and the first sub-model are used to indicate the first functional relationship between the load index and the first operation, and the first target model and the second sub-model are used to indicate the load index and The second functional relationship between the second operations, the second functional relationship is obtained by translating the first functional relationship, and both the first functional relationship and the second functional relationship are monotonically increasing or monotonically decreasing.

In a possible implementation manner, the magnitude of translation is determined based on feature information.

In a possible implementation, the operation instruction is used to modify the configuration parameters of the target cell to adjust the load indicator; the characteristic information of the target cell includes at least one of the following: configuration parameters of the target cell; traffic statistics data of the target cell; Configuration parameters of neighbor cells of the cell; traffic statistics data of neighbor cells.

In a possible implementation, the configuration parameters include at least one of the following: antenna transmit power; antenna downtilt angle; antenna horizontal azimuth angle; the RSRP threshold for the user to start inter-frequency handover measurement; the RSRP threshold for the user to stop inter-frequency handover measurement ; Specific frequency RSRP offset for user triggering inter-frequency handover process; specific neighbor RSRP offset for user triggering inter-frequency handover process; RSRP threshold for user to start co-frequency handover measurement; RSRP threshold for user to stop co-frequency handover measurement; user triggered The frequency-specific RSRP offset of the same-frequency handover process; the specific neighbor cell RSRP offset of the user-triggered same-frequency handover process.

In a possible implementation, the traffic statistics data includes at least one of the following: the average number of users per unit time period; the average number of active users per unit time period; the uplink traffic per unit time period; the CQI report per unit time period Proportion; the proportion of data packets whose length per unit time period is less than the preset length threshold; the average length of data packets per unit time period.

Fig. 11 is a schematic structural diagram of the model training device provided by the embodiment of the present application. As shown in Fig. 11, the device includes:

The first acquiring module 1101 is configured to acquire characteristic information of the first cell and a load index of the first cell, where the load index of the first cell is used to indicate the load degree of the first cell;

The first processing module 1102 is configured to process feature information through the first model to be trained to obtain model parameters of the second model to be trained;

The second acquiring module 1103 is configured to acquire a second model to be trained based on model parameters;

The second processing module 1104 is configured to process the load index of the first cell through the second model to be trained to obtain an operation instruction for the first cell, and the operation instruction for the first cell is used to adjust the load index of the first cell;

The third processing module 1105 is configured to process the operation instruction for the first cell and the operation instruction for the second cell through the third target model to obtain the first score and the second score, and the first score is used to evaluate the performance of the first cell The impact of the operation instruction on the load index of the first cell, the second score is used to evaluate the impact of the operation instruction of the first cell on the performance index of the entire network, and the operation instruction for the second cell is used to adjust the load index of the second cell, the first cell and the second cell are different cells;

The update module 1106 is configured to update the model parameters of the first model to be trained based on the first score and the second score until the model training conditions are met, and the first target model is obtained.

In the embodiment of the present application, when using the critic model (that is, the third target model) to train the actor model (that is, the whole composed of the first target model and the second target model), the training goal of the actor model is to simultaneously optimize the local network The output value of the output value and the output value of the global network, based on the actor model trained in this way, has the performance of balancing local and global performance, that is, it can balance the impact of the load of a certain cell on the cell and the entire network. , to achieve coordination between multiple cells.

In a possible implementation, the second model to be trained includes a first sub-model and a second sub-model, and the second processing module 1104 is configured to: process the load index of the first cell through the first sub-model to obtain The first operation, if the load index of the first cell is equal to the preset index threshold, the first operation is used to not adjust the load index of the first cell; the load index of the first cell is processed through the second sub-model to obtain the second Two operations: performing weighted summation on the first operation and the second operation to obtain an operation instruction for the first cell.

In a possible implementation manner, the first model to be trained and the first sub-model are used to indicate a first functional relationship between the load index of the first cell and the first operation for the first cell, and the first model to be trained and the second sub-model are used to indicate the second functional relationship between the load index of the first cell and the second operation for the first cell, the second functional relationship is obtained by shifting the first functional relationship, the first function Both the relationship and the second function relationship are monotonically increasing or monotonically decreasing.

In a possible implementation manner, the magnitude of translation is determined based on feature information.

In a possible implementation manner, the operation instruction for the first cell is used to modify the configuration parameters of the first cell to adjust the load index of the first cell; the feature information of the first cell includes at least one of the following: the first The configuration parameters of the cell; the traffic statistics data of the first cell; the configuration parameters of the neighbor cell of the first cell; the traffic statistics data of the neighbor cell.

In a possible implementation, the configuration parameters include at least one of the following: antenna transmit power; antenna downtilt angle; antenna horizontal azimuth angle; the RSRP threshold for the user to start inter-frequency handover measurement; the RSRP threshold for the user to stop inter-frequency handover measurement ; Specific frequency RSRP offset for user triggering inter-frequency handover process; specific neighbor RSRP offset for user triggering inter-frequency handover process; RSRP threshold for user to start co-frequency handover measurement; RSRP threshold for user to stop co-frequency handover measurement; user triggered The frequency-specific RSRP offset of the same-frequency handover process; the specific neighbor cell RSRP offset of the user-triggered same-frequency handover process.

In a possible implementation, the traffic statistics data includes at least one of the following: the average number of users per unit time period; the average number of active users per unit time period; the uplink traffic per unit time period; the CQI report per unit time period Proportion; the proportion of data packets whose length per unit time period is less than the preset length threshold; the average length of data packets per unit time period.

It should be noted that the information interaction and execution process between the modules/units of the above-mentioned device are based on the same concept as the method embodiment of the present application, and the technical effect it brings is the same as that of the method embodiment of the present application. The specific content can be Reference is made to the descriptions in the foregoing method embodiments shown in the embodiments of the present application, and details are not repeated here.

The embodiment of the present application also relates to an execution device, and FIG. 12 is a schematic structural diagram of the execution device provided in the embodiment of the present application. As shown in FIG. 12 , the execution device 1200 may specifically be a mobile phone, a tablet, a notebook computer, a smart wearable device, a server, etc., which is not limited here. Wherein, the apparatus for adjusting the cell load described in the embodiment corresponding to FIG. 10 may be deployed on the executing device 1200 to realize the function of adjusting the cell load in the embodiment corresponding to FIG. 4 or FIG. 6 . Specifically, the execution device 1200 includes: a receiver 1201, a transmitter 1202, a processor 1203, and a memory 1204 (the number of processors 1203 in the execution device 1200 may be one or more, and one processor is taken as an example in FIG. 12 ) , where the processor 1203 may include an application processor 12031 and a communication processor 12032 . In some embodiments of the present application, the receiver 1201, the transmitter 1202, the processor 1203, and the memory 1204 may be connected through a bus or in other ways.

The memory 1204 may include read-only memory and random-access memory, and provides instructions and data to the processor 1203 . A part of the memory 1204 may also include a non-volatile random access memory (non-volatile random access memory, NVRAM). The memory 1204 stores processors and operating instructions, executable modules or data structures, or their subsets, or their extended sets, wherein the operating instructions may include various operating instructions for implementing various operations.

The processor 1203 controls the operations of the execution device. In a specific application, various components of the execution device are coupled together through a bus system, where the bus system may include not only a data bus, but also a power bus, a control bus, and a status signal bus. However, for the sake of clarity, the various buses are referred to as bus systems in the figures.

The methods disclosed in the foregoing embodiments of the present application may be applied to the processor 1203 or implemented by the processor 1203 . The processor 1203 may be an integrated circuit chip, which has a signal processing capability. During implementation, each step of the above-mentioned method may be implemented by an integrated logic circuit of hardware in the processor 1203 or instructions in the form of software. The above-mentioned processor 1203 can be a general-purpose processor, a digital signal processor (digital signal processing, DSP), a microprocessor or a microcontroller, and can further include an application specific integrated circuit (application specific integrated circuit, ASIC), field programmable Field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The processor 1203 may implement or execute various methods, steps, and logic block diagrams disclosed in the embodiments of the present application. A general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like. The steps of the method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the field such as random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, register. The storage medium is located in the memory 1204, and the processor 1203 reads the information in the memory 1204, and completes the steps of the above method in combination with its hardware.

The receiver 1201 can be used to receive input digital or character information, and generate signal input related to performing device related settings and function control. The transmitter 1202 can be used to output digital or character information through the first interface; the transmitter 1202 can also be used to send instructions to the disk group through the first interface to modify the data in the disk group; the transmitter 1202 can also include a display device such as a display screen .

In this embodiment of the present application, in one case, the processor 1203 is configured to process the information of the target cell through the first target model in the embodiment corresponding to FIG. 4 or FIG. 8 , so as to adjust the load level of the target cell.

The embodiment of the present application also relates to a training device, and FIG. 13 is a schematic structural diagram of the training device provided in the embodiment of the present application. As shown in Figure 13, the training device 1300 is implemented by one or more servers, and the training device 1300 may have relatively large differences due to different configurations or performances, and may include one or more central processing units (central processing units, CPU) 1314 (eg, one or more processors) and memory 1332, and one or more storage media 1330 (eg, one or more mass storage devices) for storing application programs 1342 or data 1344. Wherein, the memory 1332 and the storage medium 1330 may be temporary storage or persistent storage. The program stored in the storage medium 1330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the training device. Furthermore, the central processing unit 1314 may be configured to communicate with the storage medium 1330 , and execute a series of instruction operations in the storage medium 1330 on the training device 1300 .

The training device 1300 can also include one or more power supplies 1326, one or more wired or wireless network interfaces 1350, one or more input and output interfaces 1358; or, one or more operating systems 1341, such as Windows Server™, Mac OS X™ , UnixTM, LinuxTM, FreeBSDTM and so on.

Specifically, the training device may execute the model training method in the embodiment corresponding to FIG. 8 or FIG. 9 .

The embodiment of the present application also relates to a computer storage medium, in which a program for signal processing is stored in the computer-readable storage medium, and when the program is run on the computer, the computer executes the steps performed by the aforementioned execution device, or, The computer is caused to perform the steps as performed by the aforementioned training device.

The embodiment of the present application also relates to a computer program product, where instructions are stored in the computer program product, and when executed by a computer, the instructions cause the computer to perform the steps performed by the aforementioned executing device, or cause the computer to perform the steps performed by the aforementioned training device. A step of.

The execution device, training device or terminal device provided in the embodiment of the present application may specifically be a chip. The chip includes: a processing unit and a communication unit. The processing unit may be, for example, a processor, and the communication unit may be, for example, an input/output interface, pins or circuits etc. The processing unit can execute the computer-executed instructions stored in the storage unit, so that the chips in the execution device execute the data processing methods described in the above embodiments, or make the chips in the training device execute the data processing methods described in the above embodiments. Optionally, the storage unit is a storage unit in the chip, such as a register, a cache, etc., and the storage unit may also be a storage unit located outside the chip in the wireless access device, such as only Read-only memory (ROM) or other types of static storage devices that can store static information and instructions, random access memory (random access memory, RAM), etc.

Specifically, please refer to FIG. 14. FIG. 14 is a schematic structural diagram of the chip provided by the embodiment of the present application. The chip can be represented as a neural network processor NPU 1400, and the NPU 1400 is mounted to the main CPU (Host CPU) as a coprocessor ), the tasks are assigned by the Host CPU. The core part of the NPU is the operation circuit 1403, and the operation circuit 1403 is controlled by the controller 1404 to extract matrix data in the memory and perform multiplication operations.

In some implementations, the operation circuit 1403 includes multiple processing units (Process Engine, PE). In some implementations, arithmetic circuit 1403 is a two-dimensional systolic array. The arithmetic circuit 1403 may also be a one-dimensional systolic array or other electronic circuits capable of performing mathematical operations such as multiplication and addition. In some implementations, arithmetic circuit 1403 is a general-purpose matrix processor.

For example, suppose there is an input matrix A, a weight matrix B, and an output matrix C. The operation circuit fetches the data corresponding to the matrix B from the weight storage 1402, and caches it in each PE in the operation circuit. The operation circuit takes the data of matrix A from the input memory 1401 and performs matrix operation with matrix B, and the obtained partial or final results of the matrix are stored in an accumulator 1408 .

The unified memory 1406 is used to store input data and output data. The weight data directly accesses the controller (Direct Memory Access Controller, DMAC) 1405 through the storage unit, and the DMAC is transferred to the weight storage 1402. The input data is also transferred to the unified memory 1406 through the DMAC.

The BIU is the Bus Interface Unit, that is, the bus interface unit 1413, which is used for the interaction between the AXI bus and the DMAC and the instruction fetch buffer (Instruction Fetch Buffer, IFB) 1409.

The bus interface unit 1413 (Bus Interface Unit, BIU for short), is used for the instruction fetch memory 1409 to obtain instructions from the external memory, and is also used for the storage unit access controller 1405 to obtain the original data of the input matrix A or the weight matrix B from the external memory.

The DMAC is mainly used to move the input data in the external memory DDR to the unified memory 1406 , to move the weight data to the weight memory 1402 , or to move the input data to the input memory 1401 .

The vector computing unit 1407 includes a plurality of computing processing units, and if necessary, further processes the output of the computing circuit 1403, such as vector multiplication, vector addition, exponent operation, logarithmic operation, size comparison and so on. It is mainly used for non-convolutional/fully connected layer network calculations in neural networks, such as Batch Normalization (batch normalization), pixel-level summation, upsampling of predicted label planes, etc.

In some implementations, the vector computation unit 1407 can store the vector of the processed output to the unified memory 1406 . For example, the vector calculation unit 1407 can apply a linear function; or, a non-linear function to the output of the operation circuit 1403, such as performing linear interpolation on the predicted label plane extracted by the convolution layer, and then for example, a vector of accumulated values to generate an activation value . In some implementations, the vector computation unit 1407 generates normalized values, pixel-level summed values, or both. In some implementations, the vector of processed outputs can be used as an activation input to operational circuitry 1403, eg, for use in subsequent layers in a neural network.

An instruction fetch buffer (instruction fetch buffer) 1409 connected to the controller 1404 is used to store instructions used by the controller 1404;

The unified memory 1406, the input memory 1401, the weight memory 1402 and the fetch memory 1409 are all On-Chip memories. External memory is private to the NPU hardware architecture.

Wherein, the processor mentioned above can be a general-purpose central processing unit, microprocessor, ASIC, or one or more integrated circuits for controlling the execution of the above-mentioned programs.

In addition, it should be noted that the device embodiments described above are only illustrative, and the units described as separate components may or may not be physically separated, and the components shown as units may or may not be A physical unit can be located in one place, or it can be distributed to multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in the drawings of the device embodiments provided in the present application, the connection relationship between the modules indicates that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines.

Through the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software plus necessary general-purpose hardware, and of course it can also be realized by special hardware including application-specific integrated circuits, dedicated CPUs, dedicated memories, Special components, etc. to achieve. In general, all functions completed by computer programs can be easily realized by corresponding hardware, and the specific hardware structure used to realize the same function can also be varied, such as analog circuits, digital circuits or special-purpose circuit etc. However, for this application, software program implementation is a better implementation mode in most cases. Based on this understanding, the essence of the technical solution of this application or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product is stored in a readable storage medium, such as a floppy disk of a computer , U disk, mobile hard disk, ROM, RAM, magnetic disk or optical disk, etc., including several instructions to make a computer device (which can be a personal computer, training device, or network device, etc.) execute the instructions described in various embodiments of the present application method.

In the above embodiments, all or part of them may be implemented by software, hardware, firmware or any combination thereof. When implemented using software, it may be implemented in whole or in part in the form of a computer program product.

The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the processes or functions according to the embodiments of the present application will be generated in whole or in part. The computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from a website, computer, training device, or data The center transmits to another website site, computer, training device or data center via wired (eg, coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (eg, infrared, wireless, microwave, etc.). The computer-readable storage medium may be any available medium that can be stored by a computer, or a data storage device such as a training device or a data center integrated with one or more available media. The available medium may be a magnetic medium (such as a floppy disk, a hard disk, or a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (Solid State Disk, SSD)), etc.

Claims (19)

A method for adjusting cell load, characterized in that the method comprises: Acquiring characteristic information of the target cell and a load index of the target cell, where the load index is used to indicate the load degree of the target cell; Processing the feature information through the first target model to obtain model parameters of the second target model; obtaining a second target model based on the model parameters; The load index is processed through the second target model to obtain an operation instruction, and the operation instruction is used to adjust the load index. The method according to claim 1, wherein the second target model includes a first sub-model and a second sub-model, and processing the load index through the second target model to obtain an operation instruction includes: Processing the load index through the first sub-model to obtain a first operation, if the load index is equal to a preset index threshold, the first operation is used to not adjust the load index; Processing the load index through a second sub-model to obtain a second operation; Perform weighted summation on the first operation and the second operation to obtain an operation instruction. The method according to claim 2, wherein the first target model and the first sub-model are used to indicate a first functional relationship between the load index and the first operation, and the first A target model and the second sub-model are used to indicate a second functional relationship between the load index and the second operation, and the second functional relationship is obtained by translating the first functional relationship , both the first functional relationship and the second functional relationship are monotonically increasing or monotonically decreasing. The method according to claim 3, wherein the magnitude of the translation is determined based on the feature information. The method according to any one of claims 1 to 4, wherein the operation instruction is used to modify the configuration parameters of the target cell to adjust the load indicator; the characteristic information of the target cell includes the following at least one of: configuration parameters of the target cell; Traffic statistics data of the target cell; configuration parameters of neighbor cells of the target cell; Traffic statistics data of the neighbor cell. The method according to any one of claims 1 to 5, wherein the configuration parameters include at least one of the following: Antenna transmit power; Antenna downtilt; Antenna horizontal azimuth; The reference signal received power RSRP threshold for the user to initiate inter-frequency handover measurement; The RSRP threshold at which the user stops inter-frequency handover measurement; The specific frequency RSRP offset of the user triggering the inter-frequency handover process; Specific neighboring cell RSRP offset for users to trigger the inter-frequency handover procedure; The RSRP threshold for the user to initiate the same-frequency handover measurement; The RSRP threshold at which the user stops the same-frequency handover measurement; Specific frequency RSRP offset for the user to trigger the intra-frequency handover process; The RSRP offset of a specific neighboring cell when the user triggers the intra-frequency handover procedure. The method according to any one of claims 4 to 6, wherein the traffic statistics data includes at least one of the following: Average number of users per unit time period; The average number of active users per unit time period; Upstream traffic per unit time period; The proportion of low channel quality indicator CQI reports per unit time period; The proportion of data packets whose length per unit time period is less than the preset length threshold; The average length of packets per unit time period. A method for model training, characterized in that the method comprises: Acquire feature information of the first cell and a load index of the first cell, where the load index of the first cell is used to indicate the load degree of the first cell; Processing the feature information through the first model to be trained to obtain model parameters of the second model to be trained; Obtaining a second model to be trained based on the model parameters; Process the load index of the first cell by the second model to be trained to obtain an operation instruction for the first cell, where the operation instruction for the first cell is used to adjust the load index of the first cell; Processing the operation instruction for the first cell and the operation instruction for the second cell through a third target model to obtain a first score and a second score, and the first score is used to evaluate the operation instruction for the first cell For the impact on the load index of the first cell, the second score is used to evaluate the impact of the operation indication of the first cell on the performance index of the entire network, and the operation indication for the second cell is used to adjust the The load indicator of the second cell, the first cell and the second cell are different cells; Based on the first score and the second score, the model parameters of the first model to be trained are updated until a model training condition is met to obtain a first target model. The method according to claim 8, wherein the second model to be trained includes a first sub-model and a second sub-model, and the load index of the first cell is processed through the second target model, Obtaining the operation instruction for the first cell includes: The load index of the first cell is processed through the first sub-model to obtain a first operation. If the load index of the first cell is equal to a preset index threshold, the first operation is used to not adjust the The load index of the first cell; processing the load index of the first cell through the second sub-model to obtain a second operation; Perform weighted summation on the first operation and the second operation to obtain an operation instruction for the first cell. The method according to claim 9, wherein the first model to be trained and the first sub-model are used to indicate the load index of the first cell and the first operation for the first cell The first functional relationship between the first model to be trained and the second sub-model is used to indicate the second functional relationship between the load index of the first cell and the second operation for the first cell , the second functional relationship is obtained by translating the first functional relationship, and both the first functional relationship and the second functional relationship are monotonically increasing or monotonically decreasing. The method according to claim 10, wherein the magnitude of the translation is determined based on the feature information. The method according to any one of claims 8 to 11, wherein the operation instruction for the first cell is used to modify the configuration parameters of the first cell to adjust the load index of the first cell; The feature information of the first cell includes at least one of the following: configuration parameters of the first cell; Traffic statistics data of the first cell; Configuration parameters of neighbor cells of the first cell; Traffic statistics data of the neighbor cell. The method according to any one of claims 8 to 12, wherein the configuration parameters include at least one of the following: Antenna transmit power; Antenna downtilt; Antenna horizontal azimuth; The RSRP threshold for the user to initiate inter-frequency handover measurement; The RSRP threshold at which the user stops inter-frequency handover measurement; The specific frequency RSRP offset of the user triggering the inter-frequency handover process; Specific neighboring cell RSRP offset for users to trigger the inter-frequency handover process; The RSRP threshold for the user to initiate the same-frequency handover measurement; The RSRP threshold at which the user stops the same-frequency handover measurement; Specific frequency RSRP offset for the user to trigger the intra-frequency handover process; The RSRP offset of a specific neighboring cell when the user triggers the intra-frequency handover process. The method according to any one of claims 11 to 13, wherein the traffic statistics data includes at least one of the following: Average number of users per unit time period; The average number of active users per unit time period; Upstream traffic per unit time period; The proportion of CQI reports per unit time period; The proportion of data packets whose length per unit time period is less than the preset length threshold; The average length of packets per unit time period. A cell load adjustment device, characterized in that the device comprises: A first obtaining module, configured to obtain characteristic information of a target cell and a load index of the target cell, where the load index is used to indicate a load degree of the target cell; The first processing module is configured to process the feature information through the first target model to obtain model parameters of the second target model; A second acquisition module, configured to acquire a second target model based on the model parameters; The second processing module is configured to process the load index through the second target model to obtain an operation instruction, and the operation instruction is used to adjust the load index. A model training device, characterized in that the device comprises: A first acquiring module, configured to acquire characteristic information of a first cell and a load index of the first cell, where the load index of the first cell is used to indicate a load degree of the first cell; The first processing module is used to process the feature information by the first model to be trained to obtain the model parameters of the second model to be trained; A second acquiring module, configured to acquire a second model to be trained based on the model parameters; The second processing module is configured to process the load index of the first cell through the second model to be trained to obtain an operation instruction for the first cell, and the operation instruction for the first cell is used to adjust the The load index of the first cell; The third processing module is configured to process the operation instruction for the first cell and the operation instruction for the second cell through a third target model to obtain a first score and a second score, and the first score is used to evaluate the The impact of the operation indication of the first cell on the load index of the first cell, the second score is used to evaluate the impact of the operation indication of the first cell on the performance index of the entire network, and the second score for the second cell The operation instruction is used to adjust the load index of the second cell, and the first cell and the second cell are different cells; An update module, configured to update the model parameters of the first model to be trained based on the first score and the second score, until the model training conditions are met, and a first target model is obtained. A cell load adjustment device, characterized in that the device includes a memory and a processor; the memory stores codes, the processor is configured to execute the codes, and when the codes are executed, the The device for adjusting cell load executes the method according to any one of claims 1-14. A computer storage medium, wherein the computer storage medium stores one or more instructions which, when executed by one or more computers, cause the one or more computers to implement any of claims 1 to 14. a method as described. A computer program product, characterized in that the computer program product stores instructions, and when the instructions are executed by a computer, the computer implements the method according to any one of claims 1 to 14.

Images