A parameter determination method, apparatus and communication device

Abstract

The application provides a parameter determination method, device and communication equipment. The method comprises the following steps: acquiring network data and a pre-scheduling parameter candidate set, wherein the pre-scheduling parameter candidate set comprises multiple groups of pre-scheduling parameters; taking the multiple groups of pre-scheduling parameters and the network data as inputs of a pre-trained prediction model, performing pre-scheduling performance prediction, and obtaining a predicted performance parameter of each group of pre-scheduling parameters; and determining a target pre-scheduling parameter based on the predicted performance parameter of each group of pre-scheduling parameters. The application can improve the scheduling effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication, and in particular to a parameter determination method and device and communication equipment. BACKGROUND

[0002] When a terminal (User Equipment, UE) has uplink data to send, it first needs to send a scheduling request to the base station, and the terminal can only start sending data after receiving uplink grant information sent by the base station. Thus, the terminal will cause a delay in uplink transmission from sending a scheduling request to obtaining uplink grant. In order to reduce the delay in uplink transmission, the base station will actively and periodically send uplink grant information to the terminal, and no longer need to wait for the terminal to send a scheduling request, thereby reducing the delay from sending a scheduling request to obtaining uplink grant by the terminal. However, the current uplink pre-scheduling parameters (such as pre-scheduling duration and pre-scheduling scheduling period) are pre-configured, resulting in poor scheduling effect. SUMMARY

[0003] The present application provides a parameter determination method and device and communication equipment to solve the problem of poor scheduling effect.

[0004] In a first aspect, an embodiment of the present application provides a parameter determination method, characterized in that comprising:

[0005] obtaining network data and a pre-scheduling parameter candidate set, wherein the pre-scheduling parameter candidate set comprises a plurality of groups of pre-scheduling parameters;

[0006] using the plurality of groups of pre-scheduling parameters and the network data as inputs of a pre-trained prediction model to perform pre-scheduling performance prediction to obtain a predicted performance parameter of each group of pre-scheduling parameters;

[0007] determining a target pre-scheduling parameter based on the predicted performance parameter of each group of pre-scheduling parameters.

[0008] In a second aspect, an embodiment of the present application further provides a parameter determination device, characterized in that comprising:

[0009] a first obtaining module configured to obtain network data and a pre-scheduling parameter candidate set, wherein the pre-scheduling parameter candidate set comprises a plurality of groups of pre-scheduling parameters;

[0010] a prediction module configured to use the plurality of groups of pre-scheduling parameters and the network data as inputs of a pre-trained prediction model to perform pre-scheduling performance prediction to obtain a predicted performance parameter of each group of pre-scheduling parameters;

[0011] a determination module configured to determine a target pre-scheduling parameter based on the predicted performance parameter of each group of pre-scheduling parameters.

[0012] In a third aspect, an embodiment of the present application further provides a communication device, comprising: a transceiver, a memory, a processor, and a program stored in the memory and capable of running on the processor; the processor is configured to read the program in the memory to implement the steps in the method according to the first aspect of the present application.

[0013] In a fourth aspect, an embodiment of the present application further provides a readable storage medium, the readable storage medium stores a program, and the program is executed by a processor to implement the steps in the method according to the first aspect of the present application.

[0014] In the embodiment of the present application, the present network data and the candidate set of pre-scheduling parameters are obtained, the candidate set of pre-scheduling parameters comprises multiple groups of pre-scheduling parameters; the multiple groups of pre-scheduling parameters and the present network data are taken as inputs of a pre-trained prediction model to perform pre-scheduling performance prediction, so as to obtain a predicted performance parameter of each group of pre-scheduling parameters; and a target pre-scheduling parameter is determined based on the predicted performance parameter of each group of pre-scheduling parameters. That is, the predicted performance parameter of each group of pre-scheduling parameters can be predicted in real time according to the present network data, so as to determine the target pre-scheduling parameter and improve the scheduling effect. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without any creative labor.

[0016] Figure 1 is a flowchart of a parameter determination method provided by an embodiment of the present application;

[0017] Figure 2 is a schematic diagram of an uplink pre-scheduling optimization method provided by an embodiment of the present application;

[0018] Figure 3 is a schematic diagram of another uplink pre-scheduling optimization method provided by an embodiment of the present application;

[0019] Figure 4 is a schematic diagram of a prediction model provided by an embodiment of the present application;

[0020] Figure 5 is a schematic diagram of terminal auxiliary information reporting provided by an embodiment of the present application;

[0021] Figure 6 is a structural schematic diagram of a parameter determination device provided by an embodiment of the present application;

[0022] Figure 7is a structural schematic diagram of a communication device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0024] The terms "first", "second", and the like in the embodiments of the present application are used to distinguish similar objects, and do not necessarily indicate a specific order or a chronological sequence. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to include only those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to the process, method, product, or device. In addition, "and / or" is used in the present application to represent at least one of the connected objects, for example, A and / or B and / or C represents 7 cases including A alone, B alone, C alone, A and B both exist, B and C both exist, A and C both exist, and A, B and C all exist.

[0025] The parameter determination method provided in the present application can be applied to a base station, a terminal device or other electronic devices, that is, when the execution subject of the parameter determination method provided in the present application is a base station, the base station can directly perform pre-scheduling based on the target pre-scheduling parameter; when the execution subject of the parameter determination method provided in the present application is a terminal device, the base station can perform pre-scheduling based on the target pre-scheduling parameter after the terminal device sends the target pre-scheduling parameter to the base station; when the execution subject of the parameter determination method provided in the present application is other electronic devices, the base station can perform pre-scheduling based on the target pre-scheduling parameter after the electronic device sends the target pre-scheduling parameter to the base station.

[0026] The parameter determination method provided in the present application will be described below.

[0027] Please refer to Figure 1 , Figure 1 is a flowchart of a parameter determination method provided in an embodiment of the present application, as shown in Figure 1 , comprising the following steps:

[0028] Step 101, obtaining network data and a pre-scheduling parameter candidate set, the pre-scheduling parameter candidate set including multiple groups of pre-scheduling parameters.

[0029] The current network data can be understood as real-time data of the current network. For different execution subjects, the specific data of the current network data can be different. For example, when executed by a base station, the current network data is corresponding data on the base station side; when executed by a terminal device, the current network data is corresponding data on the terminal side; and when executed by other electronic devices, the base station side data or the terminal side data can be selected. For example, when the execution subject is a base station, the current network data can include downlink service conditions or downlink and received uplink service conditions. Specifically, the current network data can include a Quality of Service (QoS) attribute of a service, an arrival time of the service, an arrival interval, and the like. If the service can be associated and identified, the current network data can include a service type. The current network data can also include service prediction conditions of other models.

[0030] The pre-scheduling parameter candidate set can be determined in advance. Each group of pre-scheduling parameters can include a pre-scheduling duration, a pre-scheduling period (which can be represented by a scheduling interval or a scheduling number), and a pre-scheduling data volume, and the like. To obtain multiple groups of pre-scheduling parameters, each parameter can be traversed in a small step. For example, the scheduling interval can be traversed in 1 slot, the pre-scheduling duration can be traversed in 5 ms, and the pre-scheduling data volume can be traversed in 1 byte, and the like. Multiple groups of pre-scheduling parameters can be formed by permutation and combination of different parameter values, thereby improving the accuracy of pre-scheduling parameter candidate set configuration. Alternatively, a certain number of groups of pre-scheduling parameters can be formed according to experience values, thereby reducing the complexity and calculation amount of performance prediction of multiple groups of pre-scheduling parameters. Regardless of whether the execution subject is a base station, a terminal device, or other electronic devices, the pre-scheduling parameter candidate set can be agreed upon by the terminal device and the base station in the standard, or the base station can send the pre-scheduling parameter candidate set to the terminal device through signaling.

[0031] In step 102, the multiple groups of pre-scheduling parameters and the current network data are used as inputs of a pre-trained prediction model to perform pre-scheduling performance prediction, so as to obtain a predicted performance parameter of each group of pre-scheduling parameters.

[0032] The prediction model can be obtained based on a supervised learning algorithm such as a neural network, a support vector machine, or a random forest. For example, when the prediction model is a neural network model, the neural network model includes an input layer, a hidden layer, and an output layer. During the training process, the neural network model can be iteratively trained using training samples to determine parameters of the neural network model.

[0033] The prediction performance parameter can be used to represent the scheduling effect after pre-scheduling using the pre-scheduling parameter, for example, can include scheduling delay, terminal power consumption and data padding rate, etc. The scheduling delay is the delay from sending a scheduling request by the terminal device to obtaining an uplink grant. The terminal power consumption can be obtained by testing and counting using a current meter. The data padding rate can be obtained by dividing the amount of padding data in the transmitted data by the total amount of data, or by using the uplink and downlink padding rate, wherein the uplink and downlink padding rate = 1 - PDCP (Packet Data Convergence Protocol) throughput / PHY (Physical layer) throughput.

[0034] Optionally, the prediction model is obtained by training in the following manner:

[0035] The prediction model is iteratively trained using training sample data.

[0036] The training sample data includes input data and output data. The input data includes the pre-scheduling parameter candidate set, service condition data, channel quality data and terminal energy consumption parameter. The output data includes at least one of the following:

[0037] Scheduling delay, terminal energy consumption and data padding rate.

[0038] It can be understood that, in the case of the execution subject being a base station, the channel quality data can include channel quality indication (CQI) reported by the terminal, rank indication (RI), precoding matrix indication (PMI), interference information, etc. In the case of the execution subject being a terminal device, the channel quality data can include measured reference signal received power (RSRP), signal to interference plus noise ratio (SINR), reference signal received quality (RSRQ), Doppler frequency offset estimation information, etc.

[0039] The terminal energy consumption parameter can specifically include a processing timing (K0 / K1 / K2) of the terminal, a connected discontinuous reception (C-DRX) parameter configuration in a connected state, whether a subset bandwidth (BWP) adaptive switching function is enabled, a multiple input multiple output (MIMO) layer configuration, a carrier aggregation (CA) condition, an orthogonal frequency division multiplexing (OFDM) coding mode, a resource allocation condition, and the like.

[0040] Optionally, the output data can further include a data padding rate, to measure the resource utilization rate of scheduling.

[0041] In this embodiment, the prediction model is iteratively trained using training sample data; the training sample data includes input data and output data, the input data includes the candidate set of pre-scheduling parameters, service condition data, channel quality data, and terminal energy consumption parameters, and the output data includes at least one of the following: scheduling delay, terminal energy consumption, and data padding rate. By iteratively training the prediction model using training sample data, the prediction model is fully generalized, which can improve the prediction ability of the prediction model. Furthermore, by predicting the output data after scheduling using the input data and determining the target pre-scheduling parameter based on the output data, the scheduling delay and / or terminal energy consumption and / or data padding rate can be considered in the scheduling process, thereby reducing the scheduling delay and / or terminal energy consumption and / or data padding rate.

[0042] Optionally, the iteratively training the prediction model using training sample data includes:

[0043] obtaining a prediction performance parameter output by the prediction model in the iterative training process;

[0044] using a loss function to obtain a loss value of the output prediction performance parameter, and updating parameters of the prediction model using the loss value until the loss function converges.

[0045] It can be understood that the prediction model is constructed based on a supervised learning algorithm, and the parameters of the prediction model can be adjusted in a training process by using the training sample data to improve the prediction performance of the prediction model. For example, when the prediction model is a classifier constructed based on a support vector machine, the parameters of the classifier can be adjusted through the iterative training; when the prediction model is a neural network model constructed based on a neural network, the parameters of the neural network model can be adjusted through the iterative training.

[0046] In this embodiment, the prediction accuracy of the prediction model can be further improved by obtaining the prediction performance parameters output by the prediction model in the iterative training process, using a loss function to obtain the loss value of the output prediction performance parameters, and updating the parameters of the prediction model using the loss value until the loss function converges.

[0047] In step 103, the target pre-scheduling parameter is determined based on the prediction performance parameters of each group of pre-scheduling parameters.

[0048] The target pre-scheduling parameter can be a group of pre-scheduling parameters in the pre-scheduling candidate set. The target pre-scheduling parameter with the best performance in the pre-scheduling candidate set can be determined based on the obtained prediction performance parameters of each group of pre-scheduling parameters. It can be understood that the group of pre-scheduling parameters with the best performance in the pre-scheduling candidate set can be determined based on the prediction performance parameters. For example, when the prediction performance parameters include latency and terminal power consumption, the prediction latency and terminal power consumption of each group of pre-scheduling parameters can be directly weighted and summed, and the maximum value is compared to obtain a group of pre-scheduling parameters corresponding to the maximum value as the target pre-scheduling parameter. The weight of each parameter in the weighted sum can be determined according to an empirical value or multiple tests, which is not limited in the present application.

[0049] In the case that the base station receives the target pre-scheduling parameter sent by the terminal or other electronic devices, the base station can locally store the pre-scheduling parameter candidate set. The base station can only obtain the identifier of the target pre-scheduling parameter, and then perform pre-scheduling according to the target pre-scheduling parameter. The identifier can be obtained by numbering each group of pre-scheduling parameters in the pre-scheduling parameter candidate set. The base station can also directly obtain the target pre-scheduling parameter, and then perform pre-scheduling using the target pre-scheduling parameter.

[0050] Optionally, the target pre-scheduling parameter includes any one of the following:

[0051] The target pre-scheduling parameter at the terminal level;

[0052] The target pre-scheduling parameter at the cell level;

[0053] a target pre-scheduling parameter of a quality of service class identifier (QCI) level;

[0054] a target pre-scheduling parameter of a 5G quality of service identifier (5QI) level.

[0055] It can be understood that the target pre-scheduling parameter of the terminal level is that the base station corresponding to each terminal device has a corresponding target pre-scheduling parameter for pre-scheduling configuration and execution; the target pre-scheduling parameter of the cell level is that all terminal devices in the target cell use the same target pre-scheduling parameter for pre-scheduling configuration and execution, that is, the target pre-scheduling parameter is the pre-scheduling parameter used by the base station for the target cell; for the target pre-scheduling parameter of the quality of service class identifier (QCI) level, the base station uses the same target pre-scheduling parameter for pre-scheduling configuration and execution for the terminal devices corresponding to the same QCI; for the target pre-scheduling parameter of the 5G quality of service identifier (5QI) level, the base station uses the same target pre-scheduling parameter for pre-scheduling configuration and execution for the terminal devices corresponding to the same 5QI.

[0056] In addition, in the case where the target pre-scheduling parameter includes the target pre-scheduling parameter of the cell level, the target pre-scheduling parameter corresponding to each terminal device in the target cell can also be obtained before the base station performs pre-scheduling, and the target pre-scheduling parameter suitable for the target cell is determined based on the obtained target pre-scheduling parameter corresponding to each terminal device in the target cell, for example: the pre-scheduling parameter with the highest proportion of the target pre-scheduling parameter corresponding to the terminal device in the target cell can be directly selected as the target pre-scheduling parameter of the cell level. Alternatively, the predicted performance parameter when each pre-scheduling parameter corresponding to each terminal device in the target cell is obtained can also be obtained before the base station performs pre-scheduling, and the group of pre-scheduling parameters with the best execution effect in the target cell is selected as the target pre-scheduling parameter of the cell level.

[0057] In this embodiment, the target pre-scheduling parameter corresponding to the pre-scheduling of the terminal level, the cell level, the QCI level, or the 5QI level can be used.

[0058] Optionally, after the target pre-scheduling parameter is determined based on the predicted performance parameter of each group of pre-scheduling parameters in step 103, the method can further include the following steps:

[0059] indicating the base station to perform pre-scheduling according to the target pre-scheduling parameter;

[0060] In the case where the pre-scheduling is of the cell level, the target pre-scheduling parameter is used by the base station to perform pre-scheduling on all terminals in the target cell.

[0061] In the case that the pre-scheduling is terminal level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on the first terminal;

[0062] In the case that the pre-scheduling is QCI level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on the second terminal;

[0063] In the case that the pre-scheduling is 5QI level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on the third terminal.

[0064] In the case that the execution subject is the base station, the corresponding module can be directly internally generated code to indicate the execution of pre-scheduling according to the target pre-scheduling parameter; in the case that the execution subject is the terminal device or other electronic device, the base station can be instructed to perform pre-scheduling according to the target pre-scheduling parameter by sending instructions. In addition, in the case that the execution subject is the terminal device, the target pre-scheduling parameter can also be sent by directly reusing the process of reporting terminal assistance information.

[0065] In the case that the pre-scheduling is cell level, the base station performs pre-scheduling on all terminal devices in the target cell according to the target pre-scheduling parameter, simplifying the operation of pre-scheduling. It can be understood that in the case that the pre-scheduling is cell level, the target pre-scheduling parameter can make as many terminal devices in the target cell as possible to obtain optimal scheduling effect, and the determination of the target pre-scheduling parameter can be understood as an optimization process. Specifically, the predicted performance parameters of all terminal devices can be summed to obtain the optimal target pre-scheduling parameter; or the predicted performance parameter of each terminal device can be obtained first to obtain a set of pre-scheduling parameters, and then the set of pre-scheduling parameters with the largest proportion is counted as the target pre-scheduling parameter; or other optimization algorithms can be used to obtain the target pre-scheduling parameter.

[0066] In the case that the pre-scheduling is terminal level, the base station can perform pre-scheduling on the pre-scheduling parameter configured for each first terminal device; in the case that the pre-scheduling is QCI level, if a plurality of second terminal devices correspond to the same QCI, the base station can use the same pre-scheduling parameter to perform pre-scheduling on the plurality of second terminal devices corresponding to the same QCI; in the case that the pre-scheduling is 5QI level, if a plurality of third terminal devices correspond to the same 5QI, the base station can use the same pre-scheduling parameter to perform pre-scheduling on the plurality of third terminal devices corresponding to the same 5QI. That is, in the case that the pre-scheduling is QCI level or 5QI level, the target pre-scheduling parameter can also be used to pre-schedule a plurality of terminal devices corresponding to the same QCI or 5QI.

[0067] In the embodiment, when the pre-scheduling is at the cell level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on all terminals in the target cell; when the pre-scheduling is at the terminal level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on the first terminal; when the pre-scheduling is at the QCI level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on the second terminal; and when the pre-scheduling is at the 5QI level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on the third terminal. That is, the base station can perform pre-scheduling according to the level of pre-scheduling, thereby expanding the applicable range of pre-scheduling.

[0068] In addition, when the pre-scheduling is at the cell level, the target pre-scheduling parameter of each terminal in the target cell can also be obtained in advance, and when the base station performs pre-scheduling, the base station determines the target pre-scheduling parameter at the cell level according to the target pre-scheduling parameter of each terminal in the target cell, and then performs pre-scheduling on all terminals in the target cell according to the target pre-scheduling parameter at the cell level.

[0069] In the embodiment, the present network data and a pre-scheduling parameter candidate set are obtained, the pre-scheduling parameter candidate set includes multiple groups of pre-scheduling parameters; the multiple groups of pre-scheduling parameters and the present network data are taken as inputs of a pre-trained prediction model to perform pre-scheduling performance prediction, so as to obtain a predicted performance parameter of each group of pre-scheduling parameters; and a target pre-scheduling parameter is determined based on the predicted performance parameter of each group of pre-scheduling parameters. That is, the predicted performance parameter of each group of pre-scheduling parameters can be predicted in real time according to the present network data, so as to determine the target pre-scheduling parameter, thereby improving the scheduling effect.

[0070] In addition, the target pre-scheduling parameter is determined in the pre-scheduling candidate set, so that the number of times of changing the scheduling parameter caused by too large real-time change range based on the present network data can be reduced.

[0071] Optionally, when the target pre-scheduling parameter includes a terminal-level pre-scheduling parameter, before the target pre-scheduling parameter is determined based on the predicted performance parameter of each group of pre-scheduling parameters in step 103, the method can further include the following steps:

[0072] The predicted performance parameter of each group of pre-scheduling parameters corresponding to multiple terminals is obtained.

[0073] The target pre-scheduling parameter is determined based on the predicted performance parameter of each group of pre-scheduling parameters, including:

[0074] The target pre-scheduling parameter corresponding to each terminal is determined based on the predicted performance parameter of each group of pre-scheduling parameters corresponding to the multiple terminals.

[0075] In a case that the target pre-scheduling parameter comprises a terminal-level pre-scheduling parameter, the base station can perform pre-scheduling according to the corresponding target pre-scheduling parameter for each terminal, so as to improve the scheduling effect of each terminal and reduce network resource waste and terminal power consumption.

[0076] In the embodiment, in a case that the target pre-scheduling parameter comprises a terminal-level pre-scheduling parameter, the target pre-scheduling parameter corresponding to each terminal can be determined based on the predicted performance parameter of each group of pre-scheduling parameters corresponding to the plurality of terminals, so as to improve the scheduling effect of each terminal.

[0077] Optionally, the determining of the target pre-scheduling parameter corresponding to each terminal based on the predicted performance parameter of each group of pre-scheduling parameters corresponding to the plurality of terminals comprises:

[0078] obtaining an initial pre-scheduling parameter corresponding to a target terminal, the initial pre-scheduling parameter belonging to the pre-scheduling parameter candidate set;

[0079] iteratively updating the initial pre-scheduling parameter corresponding to the target terminal based on a genetic algorithm to obtain an updated pre-scheduling parameter corresponding to the target terminal in an updating process;

[0080] obtaining the fitness of the updated pre-scheduling parameter corresponding to the target terminal based on the predicted performance parameter of each group of pre-scheduling parameters corresponding to the plurality of terminals;

[0081] determining the target pre-scheduling parameter corresponding to the target terminal based on the fitness, the target pre-scheduling parameter corresponding to the target terminal being a group of pre-scheduling parameters whose fitness satisfies a first preset condition in the updated pre-scheduling parameter obtained in the updating process;

[0082] wherein, the target terminal is any terminal of the plurality of terminals.

[0083] The first preset condition can be set in cooperation with the genetic algorithm, for example, the first preset condition can be set as an adaptability less than that of other pre-scheduling parameters, that is, the target pre-scheduling parameter is a group of pre-scheduling parameters with the minimum adaptability. In the genetic algorithm, each terminal and the corresponding initial pre-scheduling parameter can be used as a first generation chromosome. In the case of terminal-level pre-scheduling, a plurality of terminals and a plurality of initial pre-scheduling parameters corresponding to the plurality of terminals can constitute a plurality of first generation chromosomes. Taking the existence of n first generation chromosomes as an example, the adaptability of the n first generation chromosomes can be calculated using an adaptability function. If the adaptability of m first generation chromosomes is less than a preset threshold, n-m first generation chromosomes with adaptability greater than the preset threshold can be selected to enter the next evolution. Through the crossover of the chromosomes in the evolution process, n-m second generation chromosomes can be obtained. The m first generation chromosomes are copied to obtain m second generation chromosomes, thereby forming n second generation chromosomes. The above steps can be repeated to realize multiple rounds of evolution. The number of evolutions can be set in advance, thereby obtaining n chromosomes with the minimum adaptability, that is, the target pre-scheduling parameter corresponding to each terminal can be determined. Correspondingly, the adaptability function can be a function positively correlated with the predicted performance parameter. The number of iterations of the iterative update can be set in advance. Each update obtains a group of updated pre-scheduling parameters of the target terminal. The updated pre-scheduling parameter can be the initial pre-scheduling parameter of the target terminal (that is, the adaptability of the initial pre-scheduling parameter of the target terminal is small, and the correspondence between the target terminal and the initial pre-scheduling parameter remains unchanged when the initial pre-scheduling parameter is updated). The updated pre-scheduling parameter can also be the initial pre-scheduling parameter of another terminal (that is, the adaptability of the initial pre-scheduling parameter of the target terminal is large, and the initial pre-scheduling parameter needs to be updated to another pre-scheduling parameter with smaller adaptability based on the genetic algorithm). After the iterative update, a group of pre-scheduling parameters with the minimum adaptability can be selected as the target pre-scheduling parameter corresponding to the target terminal.

[0084] It can be understood that the first preset condition can also be an adaptability greater than that of other pre-scheduling parameters, that is, the target pre-scheduling parameter is a group of pre-scheduling parameters with the maximum adaptability. Then, in the genetic algorithm, after the adaptability of n first generation chromosomes is calculated using the adaptability function, n-m first generation chromosomes with adaptability less than a preset threshold can be selected to enter the next evolution. The above steps can be repeated to realize multiple rounds of evolution, and n chromosomes with the maximum adaptability can be selected as the target pre-scheduling parameter corresponding to each terminal. Correspondingly, the adaptability function can be a function negatively correlated with the predicted performance parameter.

[0085] In this embodiment, the genetic algorithm can be used to quickly determine the target pre-scheduling parameters corresponding to the plurality of terminals.

[0086] Optionally, the step of determining the target pre-scheduling parameter based on the predicted performance parameter of each group of pre-scheduling parameters in step 103 can specifically include:

[0087] determining a performance measurement parameter of the multiple groups of pre-scheduling parameters based on the predicted performance parameter of each group of pre-scheduling parameters;

[0088] determining the target pre-scheduling parameter based on the performance measurement parameter, the target pre-scheduling parameter being a group of pre-scheduling parameters in the multiple groups of pre-scheduling parameters whose performance measurement parameter satisfies a second preset condition.

[0089] The performance measurement parameter can be used to measure the scheduling effect of each group of pre-scheduling parameters, so as to determine the target pre-scheduling parameter with the optimal scheduling effect. For example, when the second preset condition is that the performance measurement parameter is the smallest, the performance measurement parameter of each group of pre-scheduling parameters can be directly calculated by weighted summation of each parameter in the group, so that the group of pre-scheduling parameters with the smallest performance measurement parameter is determined as the target pre-scheduling parameter. For example, when the scheduling effect decision function P=a* scheduling delay+b* terminal power consumption+c* data padding rate is used to determine the performance measurement parameter, a, b, and c are all greater than 0, and the performance measurement parameter of each group of pre-scheduling parameters can be calculated. Alternatively, when a, b, and c are all less than 0, the second preset condition can be that the performance measurement parameter is the largest.

[0090] In addition, when the pre-scheduling is at the cell level, the target pre-scheduling parameters of multiple terminals can be obtained in advance, and the target pre-scheduling parameter with the largest proportion can be selected as the target pre-scheduling parameter at the cell level. Alternatively, the performance measurement parameter sum of each group of pre-scheduling parameters for the multiple terminals can be obtained by traversing the candidate set of pre-scheduling parameters, and the target pre-scheduling parameter with the largest performance measurement parameter sum can be selected as the target pre-scheduling parameter at the cell level.

[0091] In this embodiment, the performance measurement parameter of the multiple groups of pre-scheduling parameters is determined based on the predicted performance parameter of each group of pre-scheduling parameters, and the target pre-scheduling parameter is determined based on the performance measurement parameter, the target pre-scheduling parameter being a group of parameters in the multiple groups of pre-scheduling parameters whose performance measurement parameter satisfies the second preset condition. Thus, when the target pre-scheduling parameter is determined according to the predicted performance parameter of each group of pre-scheduling parameters based on real-time network data, the scheduling effect can be measured by the performance measurement parameter, and the scheduling effect can be further improved.

[0092] The various optional embodiments introduced in the embodiments of the present application can be combined with each other to implement, or can be implemented alone, and the embodiments of the present application do not limit the same.

[0093] For the convenience of understanding, the specific embodiments are as follows:

[0094] The application provides an uplink pre-scheduling parameter optimization method, which models a pre-scheduling process by using an artificial intelligence method based on collected and processed network data, and optimizes parameters of uplink pre-scheduling based on the model and outputs the parameters to a base station to perform a pre-scheduling process.

[0095] The uplink pre-scheduling parameter optimization method can be performed at a base station side, as shown in FIG. 1. Figure 2 The uplink pre-scheduling parameter optimization method can be performed at a terminal side, as shown in FIG. 2. Figure 3 When the uplink pre-scheduling parameter optimization method is performed at the terminal side, step 3a, i.e., the terminal feeds back the optimal pre-scheduling parameters to the base station, needs to be added compared with the base station side. Specifically, the uplink pre-scheduling parameter optimization method can be embedded into an existing module of the base station or the terminal, for example, a Building Base band Unit (BBU) of the base station or a master control, or a baseband module of the terminal. In addition, the method can also be modularly implemented, which can be an independent component externally connected to the base station or the terminal, or an embedded independent component.

[0096] The uplink pre-scheduling parameter optimization method can specifically include the following steps:

[0097] Step 1, data collection and processing:

[0098] Data input: can include service conditions, uplink pre-scheduling candidate set, channel quality related data, terminal energy consumption related data, etc.

[0099] Service conditions: for the base station side, mainly the downlink service conditions or the downlink and received uplink service conditions; for example, including the QoS attributes of the service, the arrival time of the service, the arrival interval, etc., if the service can be associated and identified, the service type can be input, and other model service prediction conditions can also be input.

[0100] Uplink pre-scheduling candidate set: the key parameters of uplink pre-scheduling include pre-scheduling duration, pre-scheduling interval, pre-scheduling data volume. (It can also be similar parameters with the same effect, for example, the scheduling interval can be calculated or replaced by the scheduling times) When input, as many different parameter values as possible can be input. For example, each parameter is traversed with a small step, such as the scheduling interval is traversed with 1 slot, the pre-scheduling duration is traversed with 5 ms granularity, the pre-scheduling data volume is traversed with 1 byte, etc., in order to obtain a more accurate optimized candidate set. In practice, due to the limitation of frame structure and considering some existing test experience, a limited number of uplink pre-scheduling candidates can be set to reduce complexity and computational load. Hereinbelow, an embodiment is given, but it is not limited to this setting. Among them, the limited candidate set can be agreed by the terminal and the base station in the standard, or the base station can know the terminal through signaling.

[0101] For example: the pre-scheduling duration can be selected from: 20 ms, 30 ms, 40 ms, 50 ms, 80 ms, 100 ms, 120 ms, 140 ms, 160 ms, etc.

[0102] The pre-scheduling interval can be selected from 1 slot (or 2 times within 5 ms), 5 slots (or 1 time or 2 times within 5 ms), 10 slots (or 1 time within 10 ms), etc.

[0103] The pre-scheduling data volume can be selected from: 20B, 40B, 80B, 100B, 150B, 500B, 1500B, etc.

[0104] As shown in Table 1, according to the test results, a desired pre-scheduling candidate set is given, and the options of pre-scheduling parameters can be increased or decreased according to the actual situation.

[0105] Table 1. Uplink pre-scheduling parameter candidate set

[0106]

[0107]

[0108] Channel quality related data:

[0109] Base station side: it can include CQI, RI, PMI information, interference information, etc. reported by the terminal;

[0110] Terminal side: it can include measured RSRP, SINR, RSRQ, Doppler frequency offset estimation information, etc.

[0111] Terminal energy consumption related data: including: terminal processing timing (K0 / K1 / K2), C-DRX parameter configuration, whether the BWP adaptive switching function is turned on, MIMO layer configuration, CA case, OFDM encoding mode, resource allocation, etc.

[0112] Step 2, pre-scheduling modeling based on artificial intelligence:

[0113] Input data: can be determined based on the data input of step 1.

[0114] Output data: including latency, terminal power consumption, data padding rate (the amount of padding data in the transmitted data divided by the total amount of transmitted data, or uplink and downlink padding rate, i.e. padding rate = 1-PDCP throughput / PHY throughput)

[0115] Among them, in terms of data for modeling training, latency, power consumption and data padding rate can be obtained by testing, for example, latency and data padding rate can be counted, and power consumption can be tested and counted by ammeter. In addition, it can also be: collect in the existing network to enrich the training data set, including latency and data padding rate; part of the terminal power consumption can be measured in real time by ammeter, and part of the terminal power consumption can be modeled according to the existing power consumption model, for example, based on the number of times the terminal is scheduled, terminal service condition, terminal transmit power, etc.

[0116] Among them, the pre-scheduling modeling can be offline modeling, real-time modeling, or can be specified to update the model once in a certain period, for example: it can be updated once in a second, minute, hour, day or week; Or train once and don't need to update again, apply to all terminals, base stations or independent components.

[0117] Among them, the construction of the above uplink pre-scheduling model can adopt the commonly used neural network, support vector machine or random forest and other supervised learning algorithms in machine learning algorithm, taking neural network modeling as an example, a neural network model needs to be constructed and trained. As shown in Figure 4 The present application provides a 2-layer neural network modeling scheme, the input layer vector a is related to the input data; the dimension (neuron data) of the hidden layer vector y can be set to any non-zero natural number; the output layer can include 2 dimensions (latency and power consumption), or 3 dimensions (latency, power consumption and data padding rate). The activation function of the hidden layer of the neural network can also be preset, such as sigmoid function (sigmoid growth curve) or others.

[0118] In the model training stage, the input data can be obtained through network side collection, the output data can be obtained through measurement or calculation, and the power consumption can be obtained through auxiliary measurement by ammeter. In the training process, the loss function related to the weight matrix W / V can also be defined, so that the training target is to minimize the loss function; after training with high-density data, the weight matrix and the fitting function of input and output are obtained, so that the model is fully generalized and has prediction ability.

[0119] Through the above process, modeling can be obtained: F (delay, terminal power consumption, padding rate) = f (service situation, uplink pre-scheduling candidate set, channel quality, terminal energy consumption related data).

[0120] Step 3, pre-scheduling parameter optimization:

[0121] After modeling, the model will be input with the actual data collected in real time from the actual network to optimize the pre-scheduling parameters. The optimization methods include user level optimization and cell level optimization.

[0122] The user level optimization can include the following methods:

[0123] Method 1: First, based on the modeling model obtained in step 2, input the network data and the pre-scheduling candidate set to obtain the F value corresponding to each pre-scheduling parameter. Second, define the optimization decision function to determine the optimal solution. For example, P = a*delay + b*terminal power consumption + c*padding rate, and the scheme with the smallest P is determined as the optimal scheme.

[0124] Method 2: Based on the modeling model in step 2, a genetic algorithm is used for optimization. This problem can be summarized as a mathematical problem of how to select m pre-scheduling parameters for n users to obtain the optimal performance.

[0125] In the initial stage of the algorithm, a set of feasible solutions can be randomly generated, which can be considered as selecting a set of pre-scheduling parameters, i.e. the first generation of chromosomes. Then, the fitness of each chromosome is calculated using the fitness function, and the fitness can be evaluated using the function P (i.e. the smaller the P, the better). And calculate the probability of each chromosome being selected in the next evolution according to the fitness. The probability of chromosome i being selected = the fitness of chromosome i / the sum of the fitness of all chromosomes (the greater the probability, the better). Then is the "evolution" process, through "crossing", n-m chromosomes are generated, through "copying", m chromosomes are generated, and the fitness of the n chromosomes and the probability of being selected next time are calculated respectively. The above steps are the process of one evolution, and the number of evolutions can be limited to do multiple evolutions. Finally, the optimal pre-scheduling parameters in m pre-scheduling parameters are selected for n users through the algorithm.

[0126] The cell-level optimization can specifically include the following methods:

[0127] In some scenarios, for the sake of simplicity, the parameters are still set at the cell level, and all users finally use the same parameters. Therefore, it is required to make more users obtain the optimal effect.

[0128] The method 1 of the user-level optimization can be used to traverse the pre-scheduling candidate set, obtain the performance of each user under the corresponding pre-scheduling parameter, and then sum the performance of all users, and the optimal is the cell-level pre-scheduling parameter. The n pre-scheduling parameters with the largest proportion in the n selections of n users can be obtained, and the pre-scheduling parameter i with the largest proportion is taken as the optimal cell-level pre-scheduling parameter.

[0129] Step 3a, feedback of the optimal pre-scheduling parameter:

[0130] If the above modeling is performed at the base station side, the next pre-scheduling process can be directly performed.

[0131] If the uplink pre-scheduling parameter optimization is performed at the terminal side, a step needs to be added, and the terminal needs to feed back the optimal pre-scheduling parameter to the base station. As shown in FIG. 8, the terminal can reuse the reporting process of the user terminal assistance information (UAI) as shown in FIG. 9, and the content to be fed back is added in the UAI: the content to be fed back can include the optimal uplink pre-scheduling parameter number, or directly include the recommended uplink pre-scheduling parameter, including the key parameters of the continuous pre-scheduling duration, the scheduling interval and the scheduling data volume. Figure 5 Step 4, pre-scheduling process decision and execution:

[0132] If the optimization is performed at the base station side, the base station performs the uplink pre-scheduling configuration and execution according to the optimal uplink pre-scheduling parameter output by the above model. The configuration and execution can be performed according to the optimization results of the user or can be uniformly configured and executed according to the cell-level optimization results. (Special, the parameter set number 0 represents that the uplink pre-scheduling operation is not performed.)

[0133] If the optimization is performed at the terminal side, and is user-level, the base station performs the uplink pre-scheduling configuration and execution according to the optimal uplink pre-scheduling parameter recommended by the terminal; if it is cell-level, the base station needs to further process and decide the parameter recommended by the terminal. For example, the distribution of the parameters recommended by the terminal can be counted, that is, the distribution of the parameter recommendation and the terminal number, and the uplink pre-scheduling parameter with the most parameter recommendations is selected for execution.

[0134] Step 5, model updating.

[0135]

[0136] ​The process can also include a model updating process, that is, after the base station performs the pre-scheduling process, the obtained data can be fed back to the data collection module for further training of the model.

[0137] In the embodiments of the present application, based on the collected and processed live network data, artificial intelligence is used to model the pre-scheduling process, and the constructed pre-scheduling model is used to predict the pre-scheduling candidate set to perform pre-scheduling parameter optimization, so that the base station can use the obtained optimal pre-scheduling parameters for pre-scheduling configuration and execution, thereby improving the scheduling effect of pre-scheduling.

[0138] Referring to Figure 6 , Figure 6 is a structural schematic diagram of a parameter determination apparatus provided by the embodiments of the present application. As shown in Figure 6 , the parameter determination apparatus 600 comprises:

[0139] A first acquisition module 601 is configured to acquire live network data and a pre-scheduling parameter candidate set, wherein the pre-scheduling parameter candidate set comprises a plurality of groups of pre-scheduling parameters.

[0140] A prediction module 602 is configured to use the plurality of groups of pre-scheduling parameters and the live network data as inputs of a pre-trained prediction model to perform pre-scheduling performance prediction to obtain a predicted performance parameter of each group of pre-scheduling parameters.

[0141] A determination module 603 is configured to determine a target pre-scheduling parameter based on the predicted performance parameter of each group of pre-scheduling parameters.

[0142] Optionally, the target pre-scheduling parameter comprises any one of the following:

[0143] a terminal-level target pre-scheduling parameter;

[0144] a cell-level target pre-scheduling parameter;

[0145] a quality of service classification identification code (QCI)-level target pre-scheduling parameter;

[0146] a 5G quality of service classification identifier (5QI)-level target pre-scheduling parameter.

[0147] Optionally, in the case where the target pre-scheduling parameter comprises a terminal-level pre-scheduling parameter, the parameter determination apparatus 600 can further comprise:

[0148] A second acquisition module is configured to acquire a predicted performance parameter of each group of pre-scheduling parameters corresponding to a plurality of terminals;

[0149] The prediction module 602 can specifically comprise:

[0150] The first determining unit is configured to determine a target pre-scheduling parameter corresponding to each terminal based on the predicted performance parameter of each group of pre-scheduling parameters corresponding to the plurality of terminals.

[0151] Optionally, the determining module 603 can specifically include:

[0152] The first obtaining unit is configured to obtain an initial pre-scheduling parameter corresponding to the target terminal, the initial pre-scheduling parameter belonging to the pre-scheduling parameter candidate set;

[0153] The updating unit is configured to iteratively update the initial pre-scheduling parameter corresponding to the target terminal based on a genetic algorithm to obtain an updated pre-scheduling parameter corresponding to the target terminal in an updating process;

[0154] The second obtaining unit is configured to obtain a fitness of the updated pre-scheduling parameter corresponding to the target terminal based on the predicted performance parameter of each group of pre-scheduling parameters corresponding to the plurality of terminals;

[0155] The second determining unit is configured to determine the target pre-scheduling parameter corresponding to the target terminal based on the fitness, the target pre-scheduling parameter corresponding to the target terminal being a pre-scheduling parameter whose fitness satisfies a first preset condition among the updated pre-scheduling parameters obtained in the updating process;

[0156] The target terminal is any terminal of the plurality of terminals.

[0157] Optionally, the predicting module 602 can specifically include:

[0158] The third determining unit is configured to determine a performance measurement parameter of the plurality of groups of pre-scheduling parameters based on the predicted performance parameter of each group of pre-scheduling parameters;

[0159] The fourth determining unit is configured to determine the target pre-scheduling parameter based on the performance measurement parameter, the target pre-scheduling parameter being a group of pre-scheduling parameters whose performance measurement parameter satisfies a second preset condition among the plurality of groups of pre-scheduling parameters.

[0160] Optionally, the parameter determining apparatus 600 can further include:

[0161] The indicating module is configured to instruct a base station to perform pre-scheduling according to the target pre-scheduling parameter;

[0162] When the pre-scheduling is at a cell level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on all terminals in a target cell.

[0163] When the pre-scheduling is at a terminal level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on a first terminal.

[0164] When the pre-scheduling is at the QCI level, the target pre-scheduling parameters are used by the base station to perform pre-scheduling on the second terminal;

[0165] When the pre-scheduling is at the 5QI level, the target pre-scheduling parameters are used by the base station to perform pre-scheduling on the third terminal.

[0166] Optionally, the prediction model can be trained in the following manner:

[0167] The prediction model is iteratively trained using training sample data;

[0168] The training sample data includes input data and output data. The input data includes the pre-scheduling parameter candidate set, service status data, channel quality data, and terminal power consumption parameters. The output data includes at least one of the following:

[0169] Scheduling latency, terminal power consumption, and data fill rate.

[0170] Optionally, the iterative training of the prediction model using training sample data may specifically include:

[0171] Obtain the prediction performance parameters output by the prediction model during the iterative training process;

[0172] The loss function is used to obtain the loss value of the prediction performance parameters of the output, and the loss value is used to update the parameters of the prediction model until the loss function converges.

[0173] The parameter determination device 600 can realize the embodiments of this application. Figure 1 The various processes in the method embodiments, and the ways to achieve the same beneficial effects, will not be repeated here to avoid repetition.

[0174] This application also provides a communication device. Because the principle by which the communication device solves the problem is similar to that in the embodiments of this application... Figure 1 The image recognition method shown is similar; therefore, the implementation of this communication device can be found in the implementation of the method, and repeated details will not be elaborated further. For example... Figure 7 As shown, the communication device in this embodiment includes a processor 700, configured to read a program from a memory 720 and execute the following processes:

[0175] Acquire live network data and a candidate set of pre-scheduling parameters, wherein the candidate set of pre-scheduling parameters includes multiple sets of pre-scheduling parameters;

[0176] The multiple sets of pre-scheduling parameters and the live network data are used as inputs to a pre-trained prediction model to perform pre-scheduling performance prediction, so as to obtain the prediction performance parameters of each set of pre-scheduling parameters.

[0177] determine a target pre-scheduling parameter based on the predicted performance parameter of each set of pre-scheduling parameters.

[0178] a transceiver 710 for receiving and transmitting data under the control of the processor 700.

[0179] wherein, in Figure 7 The bus architecture can include any number of interconnected buses and bridges, specifically, various circuits linking the one or more processors represented by the processor 700 and the memory represented by the memory 720. The bus architecture can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art, and thus, are not further described herein. The bus interface provides an interface. The transceiver 710 can be a plurality of elements, i.e., including a transmitter and a receiver, providing a means for communicating with various other apparatuses over a transmission medium. The processor 700 is responsible for managing the bus architecture and general processing, and the memory 720 can store data used by the processor 700 in performing operations.

[0180] Optionally, the target pre-scheduling parameter comprises:

[0181] a terminal-level target pre-scheduling parameter;

[0182] a cell-level target pre-scheduling parameter;

[0183] a quality of service class identifier (QCI)-level target pre-scheduling parameter;

[0184] a 5G quality of service class identifier (5QI)-level target pre-scheduling parameter.

[0185] Optionally, in the case that the target pre-scheduling parameter comprises a terminal-level pre-scheduling parameter, the processor 700 is further configured to read a program in the memory 720 and perform the following steps:

[0186] obtain a predicted performance parameter of each set of pre-scheduling parameters corresponding to a plurality of terminals;

[0187] The determination of the target pre-scheduling parameter based on the predicted performance parameter of each set of pre-scheduling parameters comprises:

[0188] determination of a target pre-scheduling parameter corresponding to each terminal based on the predicted performance parameter of each set of pre-scheduling parameters corresponding to the plurality of terminals.

[0189] Optionally, the determination of a target pre-scheduling parameter corresponding to each terminal based on the predicted performance parameter of each set of pre-scheduling parameters corresponding to the plurality of terminals comprises:

[0190] obtaining an initial pre-scheduling parameter corresponding to a target terminal, the initial pre-scheduling parameter belonging to the candidate set of pre-scheduling parameters;

[0191] updating the initial pre-scheduling parameter corresponding to the target terminal based on a genetic algorithm to obtain an updated pre-scheduling parameter corresponding to the target terminal in an updating process;

[0192] obtaining the fitness of the updated pre-scheduling parameter corresponding to the target terminal based on the predicted performance parameter of each group of pre-scheduling parameters corresponding to the plurality of terminals;

[0193] determining the target pre-scheduling parameter corresponding to the target terminal based on the fitness, the target pre-scheduling parameter corresponding to the target terminal being a pre-scheduling parameter whose fitness satisfies a first preset condition among the updated pre-scheduling parameters obtained in the updating process;

[0194] The target terminal is any terminal of the plurality of terminals.

[0195] Optionally, the determining the target pre-scheduling parameter based on the predicted performance parameter of each group of pre-scheduling parameters comprises:

[0196] determining a performance measurement parameter of each group of pre-scheduling parameters based on the predicted performance parameter of each group of pre-scheduling parameters;

[0197] determining the target pre-scheduling parameter based on the performance measurement parameter, the target pre-scheduling parameter being a group of pre-scheduling parameters whose performance measurement parameter satisfies a second preset condition among the plurality of groups of pre-scheduling parameters.

[0198] Optionally, the processor 700 is further configured to read a program in the memory 720 and perform the following steps:

[0199] indicating the base station to perform pre-scheduling according to the target pre-scheduling parameter;

[0200] In the case of cell-level pre-scheduling, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on all terminals in a target cell.

[0201] In the case of terminal-level pre-scheduling, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on a first terminal.

[0202] In the case of QCI-level pre-scheduling, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on a second terminal.

[0203] In the case of 5QI-level pre-scheduling, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on a third terminal.

[0204] Optionally, the prediction model is obtained by training in the following manner:

[0205] iteratively training the prediction model using training sample data.

[0206] The training sample data includes input data and output data, the input data includes the candidate set of pre-scheduling parameters, service condition data, channel quality data and terminal energy consumption parameters, and the output data includes at least one of the following:

[0207] a scheduling delay, terminal energy consumption and data filling rate.

[0208] Optionally, the iterative training of the prediction model using the training sample data comprises:

[0209] obtaining a prediction performance parameter output by the prediction model in the iterative training process;

[0210] using a loss function to obtain a loss value of the output prediction performance parameter, and updating parameters of the prediction model using the loss value until the loss function converges.

[0211] The communication device provided by the embodiment of the application can execute the method shown in the above Figure 1 The implementation principle and technical effects of the method embodiment are similar, and details are not described herein.

[0212] The embodiment of the application further provides a readable storage medium, the readable storage medium stores a program, and the program is executed by a processor to implement each process of the method embodiment of the application and achieve the same technical effects. Figure 1 To avoid repetition, details are not described herein.

[0213] In several embodiments provided in the application, it should be understood that the disclosed method and device can be implemented in other ways. For example, the above-described device embodiments are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0214] In addition, each functional unit in each embodiment of the application can be integrated in one processing unit, or each unit can be physically included separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of hardware plus software functional unit.

[0215] The integrated unit in the form of software function unit can be stored in a computer readable storage medium. The software function unit is stored in a storage medium, and includes a plurality of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform part of the steps of the transceiving method described in various embodiments of the present application. The foregoing storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various storage medium capable of storing program codes.

[0216] The above describes the preferred embodiments of the present application. It should be noted that, for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which should also be considered as the protection scope of the present application.

Claims

1. A parameter determination method characterized by, The method comprises: obtaining online data and a candidate set of pre-scheduling parameters, the candidate set of pre-scheduling parameters comprising a plurality of groups of pre-scheduling parameters; inputting the plurality of groups of pre-scheduling parameters and the online data into a pre-trained prediction model to perform pre-scheduling performance prediction to obtain a predicted performance parameter of each group of pre-scheduling parameters; determining a target pre-scheduling parameter based on the predicted performance parameter of each group of pre-scheduling parameters; wherein the online data comprises real-time data of a current network, and each group of pre-scheduling parameters comprises at least one of a pre-scheduling duration, a pre-scheduling period, and a pre-scheduling data volume; the determining of the target pre-scheduling parameter based on the predicted performance parameter of each group of pre-scheduling parameters comprises: determining a performance measurement parameter of the plurality of groups of pre-scheduling parameters based on the predicted performance parameter of each group of pre-scheduling parameters; determining the target pre-scheduling parameter based on the performance measurement parameter, the target pre-scheduling parameter being a group of pre-scheduling parameters in the plurality of groups of pre-scheduling parameters that satisfies a second preset condition.

2. The method of claim 1, wherein, The target pre-scheduling parameter comprises any one of: a terminal-level target pre-scheduling parameter; a cell-level target pre-scheduling parameter; a quality of service classification identifier (QCI)-level target pre-scheduling parameter; a 5G quality of service classification identifier (5QI)-level target pre-scheduling parameter.

3. The method of claim 2, wherein, In a case where the target pre-scheduling parameter comprises a terminal-level pre-scheduling parameter, the method further comprises, before the determining of the target pre-scheduling parameter based on the predicted performance parameter of each group of pre-scheduling parameters: obtaining a predicted performance parameter of each group of pre-scheduling parameters corresponding to a plurality of terminals; the determining of the target pre-scheduling parameter based on the predicted performance parameter of each group of pre-scheduling parameters comprises: determining a target pre-scheduling parameter corresponding to each terminal based on the predicted performance parameter of each group of pre-scheduling parameters corresponding to the plurality of terminals.

4. The method of claim 3, wherein, The determining of the target pre-scheduling parameter corresponding to each terminal based on the predicted performance parameter of each group of pre-scheduling parameters corresponding to the plurality of terminals comprises: obtaining an initial pre-scheduling parameter corresponding to a target terminal, the initial pre-scheduling parameter belonging to the candidate set of pre-scheduling parameters; iteratively updating the initial pre-scheduling parameter corresponding to the target terminal based on a genetic algorithm to obtain an updated pre-scheduling parameter corresponding to the target terminal in an updating process; obtaining a fitness of the updated pre-scheduling parameter corresponding to the target terminal based on the predicted performance parameter of each group of pre-scheduling parameters corresponding to the plurality of terminals; determining the target pre-scheduling parameter corresponding to the target terminal based on the fitness, the target pre-scheduling parameter corresponding to the target terminal being a pre-scheduling parameter in the updated pre-scheduling parameter obtained in the updating process that satisfies a first preset condition; wherein the target terminal is any one of the plurality of terminals.

5. The method of any one of claims 1 to 4, wherein, After the determining of the target pre-scheduling parameter based on the predicted performance parameter of each group of pre-scheduling parameters, the method further comprises: instructing a base station to perform pre-scheduling according to the target pre-scheduling parameter; in a case where the pre-scheduling is cell-level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on all terminals in a target cell. In the case that the pre-scheduling is terminal level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on the first terminal; In the case that the pre-scheduling is QCI level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on the second terminal; In the case that the pre-scheduling is 5QI level, the target pre-scheduling parameter is used for the base station to perform pre-scheduling on the third terminal.

6. The method of claim 1, wherein, The prediction model is trained in the following manner: The prediction model is iteratively trained using training sample data; The training sample data includes input data and output data, the input data includes the pre-scheduling parameter candidate set, service condition data, channel quality data and terminal energy consumption parameter, and the output data includes at least one of the following: scheduling delay, terminal energy consumption and data filling rate.

7. The method of claim 6, wherein, The iterative training of the prediction model using training sample data includes: obtaining the prediction performance parameter output by the prediction model in the iterative training process; using a loss function to obtain the loss value of the output prediction performance parameter, and updating the parameters of the prediction model using the loss value until the loss function converges.

8. A parameter determination apparatus characterized by comprising: including: The first acquisition module is configured to acquire network data and a pre-scheduling parameter candidate set, and the pre-scheduling parameter candidate set includes multiple groups of pre-scheduling parameters; The prediction module is configured to use the multiple groups of pre-scheduling parameters and the network data as inputs of a pre-trained prediction model to perform pre-scheduling performance prediction to obtain a prediction performance parameter of each group of pre-scheduling parameters; The determination module is configured to determine a target pre-scheduling parameter based on the prediction performance parameter of each group of pre-scheduling parameters; The network data includes real-time data of the current network, and each group of pre-scheduling parameters includes at least one of the following: pre-scheduling duration, pre-scheduling period and pre-scheduling data volume; The determination of the target pre-scheduling parameter based on the prediction performance parameter of each group of pre-scheduling parameters includes: determining a performance measurement parameter of the multiple groups of pre-scheduling parameters based on the prediction performance parameter of each group of pre-scheduling parameters; determining the target pre-scheduling parameter based on the performance measurement parameter, the target pre-scheduling parameter being a group of pre-scheduling parameters in the multiple groups of pre-scheduling parameters whose performance measurement parameter satisfies a second preset condition.

9. A communication device comprising: The transceiver, the memory, the processor and the program stored on the memory and executable on the processor; characterized in that, The processor is configured to read the program in the memory to implement the steps in the method of any one of claims 1 to 7.

10. A readable storage medium, characterized by, The readable storage medium stores a program, and the program is executed by the processor to implement the steps in the method of any one of claims 1 to 7.

Images