Scheduling method and device, storage medium, program product and computer equipment

Abstract

The invention discloses a scheduling method and device, a storage medium, a program product and computer equipment, and the method comprises the steps: carrying out the processing of a collection end, generating merged data matched with user equipment, and generating network strategy information corresponding to the user equipment through a strategy management end according to the merged data through a strategy generation model, according to the method, the network strategy can be generated timely and intelligently, the adaptability between the network strategy and the user equipment is improved, then the target platform can respond to the network strategy information to generate the corresponding scheduling strategy information and send the strategy instruction to the second network element, and the second network element executes the scheduling operation. Therefore, the network policy scheduling of the user equipment can be completed, so that the flexibility, intelligence and timeliness of the network policy scheduling can be improved, and the real-time network service requirements of the terminal can be responded more timely, accurately and reliably; and the model can be generated according to the feedback data optimization strategy corresponding to the scheduling operation so as to improve the prediction accuracy and reliability of the model.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a scheduling method, apparatus, storage medium, program product, and computer equipment. Background Technology

[0002] In related technologies, the network policy corresponding to the terminal is usually a policy that is pre-configured for the terminal. When configuring the policy, it is generally done by following fixed / static network policy configuration rules and combining them with the relevant historical data of the terminal.

[0003] However, such pre-configured network policies have low real-time adaptability to terminals, making it difficult to respond to the real-time network service needs of terminals in a timely, accurate, and reliable manner. Summary of the Invention

[0004] To address the aforementioned technical problems, embodiments of this application propose a scheduling method, apparatus, storage medium, program product, and computer equipment, which can improve the flexibility, intelligence, and timeliness of network policy scheduling, thereby enabling more timely, accurate, and reliable responses to the real-time network service needs of terminals.

[0005] In a first aspect, embodiments of this application provide a scheduling method applicable to a policy management terminal. The method includes: receiving merged data from a collection terminal, wherein the merged data is generated by the collection terminal through preset processing of raw data collected from a first network element, the preset processing including merging processing, and the raw data being generated by the first network element based on user behavior data sent by a user equipment; generating network policy information corresponding to the user equipment using a policy generation model based on the merged data; sending the network policy information to a target platform, wherein the network policy information is used to instruct the target platform to generate scheduling policy information corresponding to the network policy information, and sending a policy instruction to a second network element according to the scheduling policy information, wherein the policy instruction is used to instruct the second network element to perform a scheduling operation; receiving feedback data returned by the second network element, wherein the feedback data is generated based on the execution result corresponding to the scheduling operation; and optimizing the policy generation model based on the feedback data.

[0006] Secondly, embodiments of this application provide a scheduling method applicable to a target platform. The method includes: receiving network policy information generated and sent by a policy management terminal, wherein the network policy information is information corresponding to the user equipment generated by a policy generation model based on merged data from a collection terminal and matched with user equipment; generating scheduling policy information corresponding to the network policy information; sending a policy instruction to a second network element according to the scheduling policy information, wherein the policy instruction is used to instruct the second network element to perform a scheduling operation, and generating feedback data according to the execution result corresponding to the scheduling operation, and returning the feedback data to the policy management terminal, wherein the feedback data is used to optimize the policy generation model.

[0007] Thirdly, embodiments of this application provide a scheduling device suitable for a policy management terminal, the device comprising: The first receiving module is used to receive merged data from the acquisition end, wherein the merged data is generated by the acquisition end through preset processing of the original data acquired from the first network element, the preset processing including merging processing, and the original data is generated by the first network element based on user behavior data sent by the user equipment. The first processing module is used to generate network policy information corresponding to the user equipment based on the merged data using a policy generation model; The first sending module is used to send the network policy information to the target platform, wherein the network policy information is used to instruct the target platform to generate scheduling policy information corresponding to the network policy information, and to send a policy instruction to the second network element according to the scheduling policy information, wherein the policy instruction is used to instruct the second network element to perform a scheduling operation; The first receiving module is further configured to receive feedback data returned by the second network element, wherein the feedback data is generated based on the execution result corresponding to the scheduling operation; The first processing module is further configured to optimize the strategy generation model based on the feedback data.

[0008] Fourthly, embodiments of this application provide a scheduling device suitable for a target platform, the device comprising: The second receiving module is used to receive network policy information generated and sent by the policy management terminal. The network policy information is information corresponding to the user equipment generated by the policy generation model based on merged data from the acquisition terminal and matched with the user equipment. The second processing module is used to generate scheduling policy information corresponding to the network policy information; The second sending module is used to send a policy instruction to the second network element according to the scheduling policy information. The policy instruction is used to instruct the second network element to perform a scheduling operation, generate feedback data according to the execution result of the scheduling operation, and return the feedback data to the policy management terminal. The feedback data is used to optimize the policy generation model.

[0009] Fifthly, embodiments of this application provide a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method described in any of the preceding claims.

[0010] Sixthly, embodiments of this application provide a computer program product, including computer instructions that, when executed by a processor, implement the steps of the method described in any of the preceding claims.

[0011] In a seventh aspect, embodiments of this application provide a computer device including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the steps of the method described in any of the preceding claims.

[0012] In summary, the embodiments of this application have at least the following beneficial effects: In this embodiment, the acquisition end first processes and generates merged data matching the user equipment. Then, the policy management end uses the policy generation model based on the merged data to generate network policy information corresponding to the user equipment. This allows for timely and intelligent generation of network policies and improves the adaptability between the generated network policies and the user equipment. Subsequently, the target platform can generate corresponding scheduling policy information in response to the network policy information and send policy instructions to the corresponding second network element to enable the second network element to perform scheduling operations. This completes the network policy scheduling for the user equipment, improving the flexibility, intelligence, and timeliness of network policy scheduling. This allows for more timely, accurate, and reliable responses to the real-time network service needs of the terminal. Finally, the policy generation model can be optimized based on the feedback data corresponding to the scheduling operations to further improve the predictive accuracy and reliability of the model. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of terminal network policy adjustment in the related technologies provided in the embodiments of this application; Figure 2 This is a flowchart illustrating the scheduling method provided in an embodiment of this application; Figure 3 This is a schematic diagram of the data collection thread on the NFVI server performing data collection as provided in the embodiments of this application; Figure 4This is a schematic diagram of the file data provided in the embodiments of this application; Figure 5 This is a schematic diagram of terminal network policy adjustment provided in an embodiment of this application; Figure 6 This is another flowchart illustrating the scheduling method provided in the embodiments of this application; Figure 7 This is a schematic diagram of the scheduling device provided in an embodiment of this application; Figure 8 This is another structural schematic diagram of the scheduling device provided in the embodiments of this application; Figure 9 This is a schematic diagram of the computer device provided in the embodiments of this application. Detailed Implementation

[0014] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments / examples are only a part of the embodiments / examples of this application, and not all of the embodiments / examples. Based on the embodiments / examples in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0015] In the description of this application, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "multiple" means two or more. In the description of this application, the term "comprising" and its variations are open-ended, meaning "including but not limited to." The term "based on" means "at least partially based on." The term "according to" means "at least partially according to." The term "one embodiment / example" means "at least one embodiment / example"; the term "another embodiment / example" means "at least one additional embodiment / example"; the term "some embodiments / examples" means "at least some embodiments / examples."

[0016] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0017] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this application is for the purpose of describing specific embodiments only and is not intended to limit the application. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0018] In some cases, the adjustment of terminal network policies in related technologies requires multi-stage collaboration throughout the entire process from requirement proposal to network element policy configuration. The distribution of manual intervention steps and automated breakpoints (i.e., steps difficult to complete automatically by computer equipment) directly affects overall efficiency and accuracy. See also... Figure 1 The specific implementation method of this process is as follows.

[0019] 1) Strategy Definition Phase: Business departments confirm requirements and define strategy operation scenarios (such as "high-speed rail, hospitals, exhibition halls and other regional scenarios"), clarify business categories (marketing, management control, security, reminders, etc.) and corresponding strategies.

[0020] 2) Strategy Preprocessing Stage: For the target operational scenario, target customer groups are identified based on user profiles and business preferences. These customer groups are created by combining customer tags accumulated through the Intelligent Operation Platform (IOP) according to fixed rules, and the update frequency can be once a day.

[0021] 3) Policy configuration phase: Based on the identified customer groups, configure the deployment area, priority, execution method, and other information of the business policies; at the same time, set the trigger conditions (time, location, user status, terminal type, etc.) and execution actions of the network element policies, and generate policy codes through the network service management system (POP-R).

[0022] 4) Policy Execution Phase: Business departments collaborate with the network and billing sides via email to associate policy codes, billing rules, and network element policies in the billing rule configuration platform and network element policy configuration platform (POP-D) to complete the policy implementation preparation.

[0023] 5) Strategy Implementation Phase: The Customer Big Data Operation Platform (IOP) matches marketing strategies based on strategy codes and implements network strategy adjustments for the target customer group.

[0024] As can be seen, the various systems in the above process (IOP, POP-R, POP-D, etc.) achieve collaboration through offline data interaction (such as daily updated customer tags and static policy coding), lacking the ability to make real-time dynamic adjustments. Specifically, this process has one or more of the following problems.

[0025] 1) Insufficient customer segmentation and strategy differentiation: The target customer group is created based on fixed rules and cannot be further segmented based on real-time data; if a differentiated strategy is to be implemented for the customer group, the corresponding number of customer groups must be pre-made, which is extremely inflexible.

[0026] 2) Static policies and delayed optimization: Network element policies are configured as static rules through the network service management system. The execution results need to be transmitted to the big data platform for offline analysis on a daily basis via SFTP (Secure File Transfer Protocol). Policy optimization requires manual re-initiation of the process and cannot be adjusted in real time.

[0027] 3) Data updates and policy implementation are not timely: Customer tags and policy rules are refreshed daily, and customer segmentation and policy matching are based on offline data, which cannot respond to real-time business needs (such as sudden network load or changes in user behavior).

[0028] In summary, this technology results in low real-time adaptability between pre-configured network policies and terminals, making it difficult to respond to the real-time network service needs of terminals in a timely, accurate, and reliable manner.

[0029] Firstly, see [the following] Figure 2 The diagram shows a flowchart of a scheduling method provided in an embodiment of this application. The method is applicable to the policy management end and includes steps S201-S205, as detailed below.

[0030] S201, Receive merged data from the acquisition end, wherein the merged data is generated by the acquisition end through preset processing of the original data collected from the first network element, the preset processing includes merging processing, and the original data is generated by the first network element based on user behavior data sent by the user equipment.

[0031] In some examples, the first network element can be an NWDAF (Network Data Analytics Function) network element, or it can be any other core network element that can interact with user equipment; no single limitation is made here.

[0032] See in some examples Figure 3The user equipment (user) can send user behavior data to the first network element (the user behavior data can be used to indicate the network behavior generated by the user equipment during network use, and the user behavior data can be generated and sent in real time based on the network behavior); the first network element can generate raw data (or raw network data) based on the user behavior data, or it can generate raw data based on the user behavior data and the network data of other network elements. The other network elements can include at least one of the following: UPF (User Plane Function) and PCF (Policy Control Function). The network data of the other network elements can include the O-domain data of the other network elements; the acquisition end can collect the above-mentioned raw data from the first network element, perform preset processing on the raw data to obtain merged data, and then send the merged data to the policy management end. The preset processing can include merging processing. In addition, the preset processing can also include data cleaning processing performed before merging processing.

[0033] In some examples, this raw data can be used to characterize the user experience of a user's device.

[0034] In some examples, the data fields contained in the original data may include any one or more of the fields in Table 1 below.

[0035] Table 1

[0036] See in some examples Figure 3The data acquisition endpoint can refer to an NFVI (Network Functions Virtualization Infrastructure) server. This NFVI server can deploy a policy-based data acquisition module (a multi-threaded task within a container / virtual machine, also known as a acquisition thread running on the NFVI server). This endpoint can monitor the raw data (or user-aware data, which may include user identifier Gpsi, application type SubAppId, network metrics such as latency / bandwidth / packet loss, etc., specific fields can be found in Table 1 above) of the first network element (NWDAF network element) in real time via a Kafka interface. The acquisition thread of this endpoint can automatically identify the IP address and type of the first network element by specifying a Kafka Topic, in order to dynamically maintain interface protocol and port information. For example, this policy-based data acquisition module can also have a built-in data cleaning microservice. Therefore, after obtaining the raw data, this data cleaning microservice can be used to clean the raw data (e.g., remove outliers, remove duplicate data (which can be identified by timestamp hash comparison)). In addition, a sliding window algorithm can be used to mark and statistically analyze abnormal data.

[0037] S202, Based on the merged data, network policy information corresponding to the user equipment is generated using a policy generation model.

[0038] In some examples, the policy generation model can be an artificial intelligence model deployed in the policy management terminal, or it can be an artificial intelligence model deployed on other servers that can be called by the policy management terminal; there is no single limitation here.

[0039] In some examples, a pre-trained policy generation model can be used to determine network policy information based on merged data. This model can be a trained model capable of predicting network policy information as its output and using merged data as its input. During training, sample merged data (which also carries the expected policy label representing the corresponding expected network policy information) can be used as sample data to obtain the predicted network policy information generated by the model based on this sample data. The difference between this predicted network policy information and the expected network policy information represented by the label is used to calculate the loss value using a general loss function. A general training algorithm (e.g., gradient descent) is then used to train the model based on this loss value, enabling the trained model to possess the aforementioned capabilities. It is easy to understand that the sample data in this embodiment can be experimental data obtained in advance through multiple corresponding experiments.

[0040] S203, the network policy information is sent to the target platform, wherein the network policy information is used to instruct the target platform to generate scheduling policy information corresponding to the network policy information, and to send a policy instruction to the second network element according to the scheduling policy information, wherein the policy instruction is used to instruct the second network element to perform a scheduling operation.

[0041] In some examples, the target platform could be the aforementioned customer big data operations platform, IOP.

[0042] In some examples, the scheduling operation may refer to the second network element scheduling at least some network elements related to the user equipment in order to achieve scheduling optimization for the network policy used by the user equipment.

[0043] S204, receive feedback data returned by the second network element, wherein the feedback data is generated based on the execution result corresponding to the scheduling operation.

[0044] In some examples, after performing scheduling operations, the second network element can collect network information and / or user information related to user equipment, and then generate the feedback data.

[0045] S205, Optimize the strategy generation model based on the feedback data.

[0046] In some examples, the strategy management end can use this feedback data to optimize the strategy generation model. For instance, when the strategy generation model is a large model, the feedback data can be used to fine-tune the large model, thereby achieving real-time iterative optimization of the strategy across the entire chain, forming a closed loop across the entire chain, and ensuring the real-time performance and adaptability of the network and marketing strategies.

[0047] In one optional implementation, the original data includes N file data, where N is a positive integer, and the method for generating the merged data includes: The N files are grouped to obtain at least one group of files, wherein the files classified into the same group have the same specified information, which includes user identifier, first application identifier and monitoring start time. Determine the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantees for each set of file data. Based on the at least one set of file data and their respective monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts, the merged data is generated through the merging process.

[0048] In some examples, file data can refer to experience forms generated by the acquisition device at preset time intervals (typically 0.5s-5min) according to hardware parameters. These forms can be TXT files. See here. Figure 4 Since the StartTime, the monitoring start time when a user enters the monitoring process, is an instantaneous value, and the duration of the user's guaranteed service (communication service guarantee duration) may exceed a preset time interval, the StartTime value for the same user during the same monitoring process is generally unique. The Timestamp will be represented as different values ​​in different files according to the time period. In this case, this embodiment can merge multi-period data from the same user (users with the same identifier in the same group of file data) to facilitate subsequent processing by the policy management terminal.

[0049] In some examples, during the merging process, the merging process can be used to merge specific field information in the same set of file data, as well as the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantees corresponding to the set of file data. For example, the specific field information can refer to the fields with the "retained" description shown in Table 2 below.

[0050] Table 2

[0051] In some examples, the specific grouping method for file data can be as follows: take the file data with the same user identifier Gpsi, the same first application identifier SubAppId, and the same monitoring start time StartTime from N file data (i.e., N files) and treat them as the same group of file data.

[0052] In one optional implementation, determining the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts corresponding to each set of file data includes: Based on the N sets of file data, determine the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts for each set of file data.

[0053] In some examples, the addition of new calculated data for each set of file data, such as monitoring end time EndTime, monitoring duration DurationTime, communication service guarantee duration ExpOptStartTime, and communication service guarantee count ExpOptcount, can help improve the efficiency of merged processing.

[0054] In some examples, the EndTime field can be calculated as follows: take the last TimeStamp field from the TimeStamp fields of each file in the same set of file data, and denote it as EndTime. This EndTime can be UTC (Coordinated Universal Time), such as: 2024-01-25 12:24:00.

[0055] In some examples, the DurationTime field can be calculated as follows: Take the last TimeStamp field from the TimeStamp field of each file in the same set of file data (i.e., file data with the same user ID Gpsi, the same first application ID SubAppId, and the same monitoring start time StartTime), and the first StartTime field of that set of file data. Then, calculate the DurationTime based on the last TimeStamp field and the first StartTime field, i.e., DurationTime = (the last TimeStamp field) - (the first StartTime field). The unit of DurationTime can be seconds, such as 63s.

[0056] In some examples, the specific calculation method for the communication service guarantee duration field expOptDurationTime can be as follows: Take the last TimeStamp field from the TimeStamp field of each file in the same set of file data, and the first ExpOptStartTime from the set of file data. Then, calculate expOptDurationTime based on the last TimeStamp field and the first ExpOptStartTime, that is, expOptDurationTime = (the last TimeStamp field) - (the first ExpOptStartTime). The unit of expOptDurationTime can be seconds, such as 63s.

[0057] In some examples, the specific calculation method for the ExpOptcount field of communication service guarantee count can be as follows: take the ExpOptFlag of each file data in the same group of file data, and take the number of file data with ExpOptFlag of 1 (i.e. the number of files) as ExpOptcount.

[0058] In some examples, the merged data generated during the merge process may also include any one or more fields as shown in Table 3 below.

[0059] Table 3

[0060] In some examples, the data acquisition terminal can asynchronously store the merged data into the database, so that the merged data can be sent to the policy management terminal.

[0061] In one alternative implementation, the policy generation model is a large model, and at least a portion of the merged data is used to map scene information representing the network state. The scene information is used to construct prompt words, which are used to prompt the large model to output the network policy information.

[0062] In some examples, the prompt word can be used to guide the large model to select and output the network policy information from a pre-defined policy library. In this way, the scenario information evaluation model can be combined with prompt word engineering to transform network data into an understandable scenario description (e.g., poor network quality - guaranteed - improved quality - game), driving the large model to output the optimal adjustment plan from the pre-defined policy library. This achieves intelligent transformation from data to decision, improving the efficiency and accuracy of policy generation.

[0063] In some examples, the policy management end can map at least a portion of the merged data to scenario information, construct prompt words based on the scenario information, and then call the large model to generate and output network policy information based on the prompt words.

[0064] In one optional implementation, the scene information includes multi-level classification information, or the scene information includes multi-level classification information and a second application identifier; The classification information at each level corresponds to a set of preset network state classification indicators.

[0065] In this embodiment, a hierarchical classification model with multi-level classification, such as "network state change (level 1) - guarantee state (level 2) - quality change (level 3)," is used. Combined with quantitative indicators (AppQuality, latency, packet loss rate, bandwidth, etc.) to define the judgment rules for each category, a precise characterization of the network scenario can be achieved, providing a structured basis for policy generation.

[0066] In some examples, the multi-level classification information may include three-level classification information, and the preset network state classification metric may include a KQI (Key Quality Indicator). Thus, the multi-level classification information may include the KQI code corresponding to that KQI.

[0067] For example, the first-level category in this three-level classification information can be represented as shown in Table 4 below.

[0068] Table 4

[0069] For example, the second-level classification in this three-level classification information can be represented as shown in Table 5 below.

[0070] Table 5

[0071] For example, the three-level classification in this three-level classification information can be represented as shown in Table 6 below.

[0072] Table 6

[0073] Thus, using the schemes shown in Tables 1, 2, and 3 above, a unique KOI value can be calculated to generate scenario information. For example, this KQI value can be a 10-digit fixed field, where the first two digits represent the primary category, indicating three scenarios: poor network quality, network mitigation, and end of protection. Digits 3-5 represent the secondary category, indicating whether the user's network is in a pre-set protection state. For a specific network state, such as poor network quality - protected - improved quality - game, the KQI value is 0100101001. The specific structure of this scenario information can be seen in Table 7 below. Furthermore, the multi-level classification information can be combined with the second application identifier (APP value) to obtain the scenario information. Generally, the second application identifier can indicate the APP category, while the first application identifier can indicate the APP subcategory. An APP category can include at least one APP subcategory.

[0074] Table 7

[0075] As shown above, prompts can be constructed based on the meaning represented by the scenario information. For example, the prompt could be "If you were a network optimization engineer, please optimize the network for the above scenario" + "Current network status is" + "Scenario information (KQI scenario)". Further, the input could be "If you were a network optimization engineer, please optimize the network for the above scenario. We have preset several methods for adjusting network element strategies. Please provide the adjustment methods and their numbers.", "Current network status is", "Poor network quality - Guaranteed - Improved quality - Game". The large model output is "04 - Reduce 5QI value, adjust guarantee level", as shown in Table 8 below.

[0076] Table 8

[0077] Thus, the network policy information output by the large model can be composed of 10 KQI bits + 2 adjustment bits, such as 0100101001 04.

[0078] In one optional implementation, the scheduling policy information is generated based on the network policy information and user characteristic information matching the user equipment; the user characteristic information is stored in the target platform.

[0079] In some examples, this user characteristic information may include a user profile. This user characteristic information / user profile can be determined based on ARPU (Average Revenue Per User), DOU (Data Usage), and / or demand preferences, etc.

[0080] In one alternative implementation, the scheduling operation is used to adjust network policies, which include one or more of the following: Policy Control Function (PCF) policy; User Plane Function (UPF) policy.

[0081] In some examples, the adjusted network policy may be a network policy associated with the user device.

[0082] In some examples, the policy control function PCF policy can be the policy used by the PCF associated with the user equipment.

[0083] In some examples, the UPF policy may be the policy used by the UPF associated with the user device.

[0084] In one optional implementation, the scheduling strategy information includes user group identifiers and service product identifiers; The user group identifier and the service product identifier are used together to generate a recommendation instruction, which is used to instruct the sending of corresponding service recommendation information to the user device.

[0085] In some examples, the user group identifier can be used to represent the customer group ID, and the business product identifier can be used to represent the product ID of the business product used by the user device.

[0086] In some examples, a user demand mining engine can be deployed in the target platform. The scheduling strategy information can be generated using this user demand mining engine based on network policy information and user characteristic information. For example, the user demand mining engine can be implemented in various ways, such as through multi-dimensional model frameworks like clustering, collaborative filtering, and user information graphs. By inputting network policy information and user characteristic information (such as product ordering relationships, user ARPU, location, trajectory, etc.), it can mine the demand relationships between users and one or more types of products, including package traffic packages, benefit products, and new services, to generate a person-product relationship result table. This forms a user group identifier (e.g., a 5-digit customer group ID) and a business product identifier (e.g., a 5-digit product ID) related to the one or more types of products, thereby generating scheduling strategy information.

[0087] In some examples, the scheduling policy information can be a unique policy code, with the format "KQI code - adjustment number (e.g., as shown in Table 8 above) - user group identifier (e.g., customer group ID) - business product identifier (product ID)" (e.g., 0100102001 04 01031 61001). After obtaining the scheduling policy information, the target platform can parse it, that is, parse the unique policy code, to extract key data such as the adjustment number, the currently used app, and network status, facilitating user-level adjustments to network and marketing policies for user devices.

[0088] In some examples, the target platform can generate a policy instruction carrying scheduling policy information and send it to the second network element, so that the second network element can perform scheduling operations according to the scheduling policy information. See, for example, [example missing]. Figure 5 The second network element can be an intelligent network element. This scheduling operation can include scheduling PCF network elements and / or UPF network elements and / or NWDAF network elements to perform their respective corresponding policy adjustments (for example, the aforementioned adjustment number 02 indicates an increase of 5QI value and adjustment of the protection level) to achieve network optimization. For example, taking the PCF network element as an example, the aforementioned adjustment numbers 01-05 can be preset in the PCF network element. When the target platform decodes the unique policy code and obtains the adjustment number, it can insert the adjustment number into the PCF's user subscription table according to fields such as user ID, policy number, effective time, and expiration time. The PCF will sequentially read and delete the data, insert the result into the history table, and execute the corresponding policy, thereby realizing single-user-level policy adjustment. Other UPF network elements and / or NWDAF network elements can also be scheduled in the same way.

[0089] See in some examples Figure 5After receiving the scheduling strategy information, the target platform can also send marketing outreach instructions to other touchpoints. Once the target platform receives the marketing opportunities contained in the scheduling strategy information, it can generate corresponding scripts using its marketing capabilities and reach users through other touchpoints such as real-time SMS (the target platform can send marketing outreach instructions to other touchpoints related to real-time SMS, which instruct these touchpoints to send real-time SMS messages carrying the corresponding scripts to users). For example, it can push acceleration package discounts to users with poor gaming experiences.

[0090] Specifically, after obtaining the scheduling strategy information, the target platform can acquire information such as the type of app the user is using, network quality, customer group, and product ID, and then fine-tune the existing user strategy. For example, if it needs to send marketing SMS messages about acceleration packages to users, it can be adjusted to send marketing SMS messages about game acceleration packages.

[0091] Secondly, see Figure 6 The diagram shows a flowchart of a scheduling method provided in an embodiment of this application. The method is applicable to the target platform and includes steps S601-S603, as detailed below.

[0092] S601, Receive network policy information generated and sent by the policy management terminal, wherein the network policy information is information corresponding to the user equipment generated by the policy generation model based on merged data from the acquisition terminal and matched with the user equipment.

[0093] S602, Generate scheduling policy information corresponding to the network policy information.

[0094] S603, a policy instruction is sent to the second network element according to the scheduling policy information, wherein the policy instruction is used to instruct the second network element to perform a scheduling operation, and feedback data is generated according to the execution result corresponding to the scheduling operation, and the feedback data is returned to the policy management terminal, wherein the feedback data is used to optimize the policy generation model.

[0095] In one optional implementation, the original data includes N file data sets, where N is a positive integer. The method for generating the merged data includes: grouping the N file data sets to obtain at least one group of file data sets, wherein the file data sets classified into the same group have the same specified information, including a user identifier, a first application identifier, and a monitoring start time; determining the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts for each group of file data sets; and generating the merged data sets based on the at least one group of file data sets and their respective monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts through the merging process.

[0096] In one optional implementation, determining the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantees corresponding to each set of file data includes: determining the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantees corresponding to each set of file data based on the N sets of file data.

[0097] In one alternative implementation, the policy generation model is a large model, and at least a portion of the merged data is used to map scene information representing the network state. The scene information is used to construct prompt words, which are used to prompt the large model to output the network policy information.

[0098] In one optional implementation, the scene information includes multi-level classification information, or the scene information includes multi-level classification information and a second application identifier; the multi-level classification information corresponds one-to-one with multiple preset network status classification indicators.

[0099] In one optional implementation, the scheduling policy information is generated based on the network policy information and user characteristic information matching the user equipment; the user characteristic information is stored in the target platform.

[0100] In one alternative implementation, the scheduling operation is used to adjust network policies, which include one or more of the following: Policy Control Function (PCF) policy; User Plane Function (UPF) policy.

[0101] In one optional implementation, the scheduling strategy information includes a user group identifier and a service product identifier; the user group identifier and the service product identifier are used together to generate a recommendation instruction, which is used to instruct the sending of corresponding service recommendation information to the user equipment.

[0102] Thirdly, correspondingly, the embodiments of this application also provide a scheduling device capable of implementing all processes of the scheduling method provided in any of the embodiments of the first aspect above.

[0103] See Figure 7 The diagram illustrates the structure of a scheduling device provided in an embodiment of this application. This scheduling device is suitable for a policy management terminal and includes: The first receiving module 701 is used to receive merged data from the acquisition end, wherein the merged data is generated by the acquisition end through preset processing of the original data acquired from the first network element, the preset processing including merging processing, and the original data is generated by the first network element based on user behavior data sent by the user equipment. The first processing module 702 is used to generate network policy information corresponding to the user equipment based on the merged data and using a policy generation model. The first sending module 703 is used to send the network policy information to the target platform, wherein the network policy information is used to instruct the target platform to generate scheduling policy information corresponding to the network policy information, and to send a policy instruction to the second network element according to the scheduling policy information, wherein the policy instruction is used to instruct the second network element to perform a scheduling operation. The first receiving module 701 is further configured to receive feedback data returned by the second network element, wherein the feedback data is generated based on the execution result corresponding to the scheduling operation; The first processing module 702 is further configured to optimize the strategy generation model based on the feedback data.

[0104] In one optional implementation, the original data includes N file data sets, where N is a positive integer. The method for generating the merged data includes: grouping the N file data sets to obtain at least one group of file data sets, wherein the file data sets classified into the same group have the same specified information, including a user identifier, a first application identifier, and a monitoring start time; determining the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts for each group of file data sets; and generating the merged data sets based on the at least one group of file data sets and their respective monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts through the merging process.

[0105] In one optional implementation, determining the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantees corresponding to each set of file data includes: determining the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantees corresponding to each set of file data based on the N sets of file data.

[0106] In one alternative implementation, the policy generation model is a large model, and at least a portion of the merged data is used to map scene information representing the network state. The scene information is used to construct prompt words, which are used to prompt the large model to output the network policy information.

[0107] In one optional implementation, the scene information includes multi-level classification information, or the scene information includes multi-level classification information and a second application identifier; the multi-level classification information corresponds one-to-one with multiple preset network status classification indicators.

[0108] In one optional implementation, the scheduling policy information is generated based on the network policy information and user characteristic information matching the user equipment; the user characteristic information is stored in the target platform.

[0109] In one alternative implementation, the scheduling operation is used to adjust network policies, which include one or more of the following: Policy Control Function (PCF) policy; User Plane Function (UPF) policy.

[0110] In one optional implementation, the scheduling strategy information includes a user group identifier and a service product identifier; the user group identifier and the service product identifier are used together to generate a recommendation instruction, which is used to instruct the sending of corresponding service recommendation information to the user equipment.

[0111] Fourthly, correspondingly, embodiments of this application also provide a scheduling device capable of implementing all processes of the scheduling method provided in any of the embodiments of the second aspect above.

[0112] See Figure 8 The diagram illustrates the structure of a scheduling device provided in an embodiment of this application. This scheduling device is applicable to a target platform and includes: The second receiving module 801 is used to receive network policy information generated and sent by the policy management terminal, wherein the network policy information is information corresponding to the user equipment generated by the policy generation model based on merged data from the acquisition terminal and matched with the user equipment. The second processing module 802 is used to generate scheduling policy information corresponding to the network policy information; The second sending module 803 is used to send a policy instruction to the second network element according to the scheduling policy information. The policy instruction is used to instruct the second network element to perform a scheduling operation, generate feedback data according to the execution result of the scheduling operation, and return the feedback data to the policy management terminal. The feedback data is used to optimize the policy generation model.

[0113] In one optional implementation, the original data includes N file data sets, where N is a positive integer. The method for generating the merged data includes: grouping the N file data sets to obtain at least one group of file data sets, wherein the file data sets classified into the same group have the same specified information, including a user identifier, a first application identifier, and a monitoring start time; determining the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts for each group of file data sets; and generating the merged data sets based on the at least one group of file data sets and their respective monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts through the merging process.

[0114] In one optional implementation, determining the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantees corresponding to each set of file data includes: determining the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantees corresponding to each set of file data based on the N sets of file data.

[0115] In one alternative implementation, the policy generation model is a large model, and at least a portion of the merged data is used to map scene information representing the network state. The scene information is used to construct prompt words, which are used to prompt the large model to output the network policy information.

[0116] In one optional implementation, the scene information includes multi-level classification information, or the scene information includes multi-level classification information and a second application identifier; the multi-level classification information corresponds one-to-one with multiple preset network status classification indicators.

[0117] In one optional implementation, the scheduling policy information is generated based on the network policy information and user characteristic information matching the user equipment; the user characteristic information is stored in the target platform.

[0118] In one alternative implementation, the scheduling operation is used to adjust network policies, which include one or more of the following: Policy Control Function (PCF) policy; User Plane Function (UPF) policy.

[0119] In one optional implementation, the scheduling strategy information includes a user group identifier and a service product identifier; the user group identifier and the service product identifier are used together to generate a recommendation instruction, which is used to instruct the sending of corresponding service recommendation information to the user equipment.

[0120] Fifthly, embodiments of this application provide a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method described in any of the preceding claims.

[0121] Sixthly, embodiments of this application provide a computer program product, including computer instructions that, when executed by a processor, implement the steps of the method described in any of the preceding claims.

[0122] In a seventh aspect, embodiments of this application provide a computer device including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the steps of the method described in any of the preceding claims.

[0123] See Figure 9 The computer device of this embodiment includes a processor 901, a memory 902, and a computer program, such as a scheduler, stored in the memory 902 and executable on the processor 901. When the processor 901 executes the computer program, it implements the steps described in the various scheduling method embodiments above. Exemplarily, the computer program can be divided into one or more modules / units, which are stored in the memory 902 and executed by the processor 901 to complete this application. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the computer device. The computer device can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. The computer device may include, but is not limited to, the processor 901 and the memory 902. Those skilled in the art will understand that the schematic diagram is merely an example of a computer device and does not constitute a limitation on the computer device. It may include more or fewer components than illustrated, or combine certain components, or different components. For example, the computer device may also include input / output devices, network access devices, buses, etc.

[0124] In summary, the embodiments of this application have at least the following beneficial effects: In this embodiment, the acquisition end first processes and generates merged data matching the user equipment. Then, the policy management end uses the policy generation model based on the merged data to generate network policy information corresponding to the user equipment. This allows for timely and intelligent generation of network policies and improves the adaptability between the generated network policies and the user equipment. Subsequently, the target platform can generate corresponding scheduling policy information in response to the network policy information and send policy instructions to the corresponding second network element to enable the second network element to perform scheduling operations. This completes the network policy scheduling for the user equipment, improving the flexibility, intelligence, and timeliness of network policy scheduling. This allows for more timely, accurate, and reliable responses to the real-time network service needs of the terminal. Finally, the policy generation model can be optimized based on the feedback data corresponding to the scheduling operations to further improve the predictive accuracy and reliability of the model.

[0125] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications are also considered to be within the scope of protection of this application.

Claims

1. A scheduling method, characterized in that, Applicable to policy management terminals, the method includes: Receive merged data from the acquisition end, wherein the merged data is generated by the acquisition end through preset processing of the original data collected from the first network element, the preset processing including merging processing, and the original data is generated by the first network element based on user behavior data sent by the user equipment; Based on the merged data, network policy information corresponding to the user equipment is generated using a policy generation model; The network policy information is sent to the target platform, wherein the network policy information is used to instruct the target platform to generate scheduling policy information corresponding to the network policy information, and to send a policy instruction to the second network element according to the scheduling policy information, wherein the policy instruction is used to instruct the second network element to perform a scheduling operation; Receive feedback data returned by the second network element, wherein the feedback data is generated based on the execution result corresponding to the scheduling operation; The strategy generation model is optimized based on the feedback data.

2. The method according to claim 1, characterized in that, The original data includes N files, where N is a positive integer. The method for generating the merged data includes: The N files are grouped to obtain at least one group of files, wherein the files classified into the same group have the same specified information, which includes user identifier, first application identifier and monitoring start time. Determine the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantees for each set of file data. Based on the at least one set of file data and their respective monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts, the merged data is generated through the merging process.

3. The method according to claim 2, characterized in that, The determination of the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts corresponding to each group of file data includes: Based on the N sets of file data, determine the monitoring end time, monitoring duration, communication service guarantee duration, and number of communication service guarantee attempts for each set of file data.

4. The method according to claim 1, characterized in that, The strategy generation model is a large model, and at least a portion of the merged data is used to map scene information representing the network state. The scene information is used to construct prompt words, which are used to prompt the large model to output the network strategy information.

5. The method according to claim 4, characterized in that, The scene information includes multi-level classification information, or the scene information includes multi-level classification information and a second application identifier; The classification information at each level corresponds to a set of preset network state classification indicators.

6. The method according to claim 1, characterized in that, The scheduling strategy information is generated based on the network strategy information and user characteristic information matching the user equipment. The user characteristic information is stored in the target platform.

7. The method according to claim 1, characterized in that, The scheduling operation is used to adjust network policies, and the adjusted network policies include one or more of the following: Policy control function PCF policy; User-facing functionality UPF strategy.

8. The method according to claim 1, characterized in that, The scheduling strategy information includes user group identifiers and service product identifiers; The user group identifier and the service product identifier are used together to generate a recommendation instruction, which is used to instruct the sending of corresponding service recommendation information to the user device.

9. A scheduling method, characterized in that, Applicable to the target platform, the method includes: Receive network policy information generated and sent by the policy management terminal, wherein the network policy information is information corresponding to the user device generated by the policy generation model based on merged data from the acquisition terminal and matched with the user device; Generate scheduling policy information corresponding to the network policy information; According to the scheduling policy information, a policy instruction is sent to the second network element, wherein the policy instruction is used to instruct the second network element to perform a scheduling operation, and to generate feedback data according to the execution result corresponding to the scheduling operation, and to return the feedback data to the policy management terminal, wherein the feedback data is used to optimize the policy generation model.

10. A scheduling device, characterized in that, The device, applicable to policy management, includes: The first receiving module is used to receive merged data from the acquisition end, wherein the merged data is generated by the acquisition end through preset processing of the original data acquired from the first network element, the preset processing including merging processing, and the original data is generated by the first network element based on user behavior data sent by the user equipment. The first processing module is used to generate network policy information corresponding to the user equipment based on the merged data using a policy generation model; The first sending module is used to send the network policy information to the target platform, wherein the network policy information is used to instruct the target platform to generate scheduling policy information corresponding to the network policy information, and to send a policy instruction to the second network element according to the scheduling policy information, wherein the policy instruction is used to instruct the second network element to perform a scheduling operation; The first receiving module is further configured to receive feedback data returned by the second network element, wherein the feedback data is generated based on the execution result corresponding to the scheduling operation; The first processing module is further configured to optimize the strategy generation model based on the feedback data.

11. A scheduling device, characterized in that, Suitable for a target platform, the device includes: The second receiving module is used to receive network policy information generated and sent by the policy management terminal. The network policy information is information corresponding to the user equipment generated by the policy generation model based on merged data from the acquisition terminal and matched with the user equipment. The second processing module is used to generate scheduling policy information corresponding to the network policy information; The second sending module is used to send a policy instruction to the second network element according to the scheduling policy information. The policy instruction is used to instruct the second network element to perform a scheduling operation, generate feedback data according to the execution result of the scheduling operation, and return the feedback data to the policy management terminal. The feedback data is used to optimize the policy generation model.

12. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1-9.

13. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the method described in any one of claims 1-9.

14. A computer device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the method of any one of claims 1-9.

Images