OTA flow dynamic capacity expansion method and system

Abstract

The invention discloses an OTA flow dynamic capacity expansion method and system, relates to the field of communication, and realizes active monitoring of states such as package surplus, terminal activity, network switching and the like by deploying an Applet module in a user SIM card, and generates quantitative state data. And when a preset triggering condition is satisfied, initiating a capacity expansion request to an operator cloud platform. The platform combines user historical behaviors, position information and network load, and carries out traffic peak prediction and capacity expansion grade determination based on a multi-task graph reasoning model. And after the capacity expansion benefit is evaluated through the causal reasoning module, an OTA instruction is generated and issued. And the SIM card temporarily authorizes, records and feeds back according to the data, so that accurate, control and intelligent capacity expansion control is realized.

Description

Technical Field

[0001] This specification relates to the field of communications, and more specifically, this application relates to an OTA (Over-The-Air) traffic dynamic expansion method and system. Background Technology

[0002] With the rapid development of mobile internet services, users' demand for real-time and flexible data resources is constantly increasing. Traditional data plans typically use a fixed-limit preset mechanism. Once users approach their data limit, they need to manually purchase additional data through the client, which leads to problems such as interrupted experience, delayed response, and wasted resources. Although some operators support adjusting plans by issuing configuration commands through OTA (Over-The-Air) channels, these mechanisms mostly rely on users actively initiating requests or manual intervention from customer service, lacking the ability to automatically perceive and predict the user's actual status.

[0003] On the other hand, existing traffic expansion mechanisms are usually based on static policies or rules, such as triggering expansion when the remaining traffic falls below a fixed threshold. However, these mechanisms do not consider user behavior characteristics, network load status, or contextual information, which can easily lead to problems such as inaccurate expansion and unreasonable authorization. In addition, for new users or users with sparse data, the system has difficulty accurately modeling their expansion needs, resulting in technical bottlenecks such as cold start incompatibility and large prediction bias.

[0004] In recent years, although some machine learning models have been introduced on the cloud side for traffic behavior prediction, the following shortcomings still exist: (1) Most models do not integrate the joint modeling capabilities of spatial movement patterns and time trends, making it difficult to capture the behavioral evolution patterns of users in different regions and time periods; (2) The expansion strategy formulation process does not make full use of user experience feedback after expansion, and cannot achieve experience-driven strategy optimization; (3) There is a lack of quantitative evaluation mechanism for whether expansion is "worthwhile", which may lead to redundant use of resources or ineffective expansion without actual improvement in user experience.

[0005] Therefore, there is an urgent need for a dynamic traffic expansion method that can achieve edge-cloud collaborative perception, support scenario understanding and adaptive control, and take into account the adaptability to new users and the evaluation of experience benefits, so as to improve operational efficiency and user satisfaction. Summary of the Invention

[0006] The summary section introduces a series of simplified concepts, which will be further explained in detail in the detailed description section. This summary section is not intended to limit the key and essential technical features of the claimed technical solution, nor is it intended to determine the scope of protection of the claimed technical solution.

[0007] Firstly, this application proposes a method for dynamically expanding OTA traffic, including: An applet module for monitoring data usage status is deployed in the user's SIM card to obtain local data plan usage information and generate quantitative status data. The applet module uses a lightweight statistical processing method to count events and accumulate windows for highly dynamic data, including network switching and session frequency, and forms low-dimensional quantitative indicators locally to reduce the SIM card's computing overhead. When the above-mentioned quantified status data meets the preset trigger conditions, the above-mentioned Applet module uploads the expansion request to the operator's cloud platform through the terminal; Based on the aforementioned capacity expansion request, the aforementioned operator cloud platform combines historical traffic usage data, user behavior characteristics, geographic location information, and current network load to perform traffic demand prediction and credit policy evaluation, and generate capacity expansion decisions. Based on the above expansion decision, generate OTA instructions with temporary expansion parameters; The above OTA command is sent to the SIM card through the OTA channel so that the SIM card can update its internal data plan configuration and activate temporary data authorization according to the command. During the temporary authorization period, the SIM card records the actual usage and periodically reports it to the aforementioned operator's cloud platform to support dynamic adjustment or renewal control.

[0008] In one feasible implementation, the above-mentioned acquisition of local package usage information and generation of quantified status data includes: Read the user's SIM card's internal data plan file to determine the remaining data allowance; Obtain terminal environment switching information, including network standard switching records and cell ID change frequency; Capture user data session startup frequency; Record the startup cycle change information of the terminal, including changes in the activation cycle and sleep cycle; Based on the changes in the remaining traffic quota, terminal environment switching information, user data session startup frequency, and startup cycle information, quantitative status data is generated by merging these factors according to predefined weights.

[0009] In one feasible implementation, it further includes: In high mobility scenarios, increase the weight of the aforementioned terminal environment switching information; In fixed hotspot area scenarios, increase the weight of the above data session startup frequency and the above startup cycle change information; In low-activity scenarios at night, increase the weight of the remaining traffic quota mentioned above.

[0010] In one feasible implementation, the aforementioned expansion request includes the aforementioned quantized status data. Based on the aforementioned capacity expansion request, the aforementioned operator cloud platform, combined with historical traffic usage data, user behavior characteristics, geographic location information, and current network load, performs traffic demand prediction and credit policy evaluation to generate a capacity expansion decision, including: Based on the aforementioned quantitative status data, historical traffic usage data, user behavior characteristics, geographic location information, and current network load, the aforementioned operator cloud platform uses a dynamic attention-guided multi-task graph inference model to predict and generate expansion decisions, including future peak traffic for users, expansion urgency levels, and recommended authorization validity periods. The aforementioned dynamic attention-guided multi-task graph inference models include: Behavior graph construction module: Constructs a behavior event graph based on the above user behavior characteristics, and injects the above geographical location information and the above network load as node attributes into the graph structure; The graph attention processing module is used to perform graph neural network processing on the above-mentioned behavioral event graph to extract the spatial behavior evolution tensor; The temporal feature extraction module is used to perform multi-channel temporal convolution on the above-mentioned quantized state data and the above-mentioned historical traffic usage data to extract temporal feature tensors; The multi-task decoding module is used to fuse the aforementioned spatial behavior evolution tensor and the aforementioned temporal feature tensor and input them into the multi-task decoding structure to generate expansion decisions including future user traffic peaks, expansion urgency levels, and recommended authorization validity periods.

[0011] In one feasible implementation, the graph constructed by the behavior graph construction module is a heterogeneous graph structure. The node types of the heterogeneous graph structure include user behavior events, geographic location nodes, time nodes, network load level nodes, and package level nodes. The edges of the heterogeneous graph structure represent behavior transfer and scenario causal relationships. The aforementioned behavior graph construction module includes a dynamic node generation unit, which is used to generate corresponding temporary nodes in real time and inject them into the aforementioned heterogeneous graph structure when a sudden change in user behavior is detected.

[0012] In one feasible approach, the graph attention processing module adopts a spatiotemporal graph convolutional network structure to simultaneously capture the temporal evolution and spatial dependencies of the behavioral graph. The behavioral offset extracted by the graph attention processing module is also used to drive the adaptive authorization strategy. When a high offset scenario is detected, the expansion parameters are automatically adjusted. The adjustment of the expansion parameters includes extending the authorization validity period or increasing the expansion amount.

[0013] In one feasible implementation, the graph attention processing module includes a main processing branch for extracting the current user behavior graph embedding representation and an auxiliary branch for extracting the behavior graph embedding representation of similar user groups. The feature difference between the main processing branch and the auxiliary branch represents a behavior offset, which is used to improve the determination result of the urgency level of capacity expansion. In one feasible implementation, the aforementioned multi-task decoding module integrates a reinforcement learning mechanism based on the Actor-Critic architecture; The Actor network is used to fuse the aforementioned spatial behavior evolution tensor and the aforementioned temporal feature tensor and input them into the multi-task decoding structure to generate expansion decisions including future user traffic peaks, expansion urgency levels, and recommended authorization validity periods. The Critic network is used to receive the current state characteristics and the aforementioned scaling decision, and to estimate its state value function; The cloud platform constructs reward signals based on user behavior feedback after scaling up. These reward signals include whether the scaled-up data has been exhausted, the frequency of business usage, user subjective interaction behavior, and abnormal usage indicators, in order to simultaneously optimize the parameters of the aforementioned Actor network and Critic network.

[0014] In one feasible implementation, the aforementioned Actor-Critic architecture integrates a meta-learning mechanism. In scenarios involving new users or cold-start users, the cloud platform performs graph comparison learning to calculate similarity scores based on the structural features of the user behavior graph and the historical behavior graph set. It also selects multiple optimal matching historical users as a source policy candidate set and uses a policy distillation and local parameter freezing mechanism to fuse and transfer the candidate source policies to the current Actor network. During the integration and migration process, a dynamic behavior offset detection module is introduced to assess the migration adaptability based on the rate of change in the new user's policy behavior. When the adaptability is lower than a preset threshold, the policy fine-tuning and freeze control process is automatically triggered.

[0015] In one feasible implementation, the above-mentioned multi-task graph reasoning model further includes a causal reasoning decision suppression module, the specific steps of which include: A cause-effect graph structure is established based on users' historical expansion behavior, traffic usage after expansion, and experience indicators. The cause-effect graph structure includes expansion behavior processing variables, experience indicator result variables, covariates, and mediator variable nodes. Extract expanded and unexpanded samples and corresponding multi-dimensional experience feedback data from the user group for use in training the causal inference model; Based on the current user status, the expected user experience score is predicted by the causal inference model under the conditions of performing capacity expansion and not performing capacity expansion, and the experience improvement value between the two is calculated. The experience improvement value includes the results of a two-layer causal evaluation system. The results of the two-layer causal evaluation system include the immediate experience improvement result after the user capacity expansion determined based on the immediate experience index, and the long-term value improvement result of the capacity expansion determined based on the user retention rate and package upgrade tendency index. The experience enhancement value is compared with a preset intervention benefit threshold; When the experience improvement value is lower than the preset intervention benefit threshold, the generation of the current expansion command is automatically suppressed, or the expansion quota and authorization validity period are reduced.

[0016] Secondly, this application proposes an OTA (Over-The-Air) traffic dynamic expansion system, comprising: The acquisition unit is used to deploy an applet module in the user's SIM card to monitor data usage status, in order to obtain local data plan usage information and generate quantitative status data. The upload unit is used to upload the expansion request to the operator's cloud platform via the terminal when the above-mentioned quantized status data meets the preset trigger conditions. The first generation unit is used by the aforementioned operator cloud platform to perform traffic demand prediction and credit policy evaluation based on the aforementioned expansion request, combined with historical traffic usage data, user behavior characteristics, geographical location information and current network load, and generate expansion decisions. The second generation unit is used to generate OTA instructions with temporary expansion parameters based on the above expansion decision. The startup unit is used to send the above-mentioned OTA instructions to the SIM card through the OTA channel, so that the SIM card updates its internal package configuration according to the above-mentioned instructions and starts temporary data authorization; The monitoring unit is used to record the actual usage of the SIM card during the temporary authorization period and periodically report it to the aforementioned operator's cloud platform to support dynamic adjustment or renewal control.

[0017] In summary, this invention, by deploying an Applet module in the SIM card, proactively monitors multi-dimensional states including remaining data allowance, data session frequency, terminal activity cycle, and network environment switching, and generates quantified state vectors. This enables proactive, end-user-initiated expansion needs, avoiding complete reliance on manual user triggering and significantly improving the automation and timeliness of data services. The operator's cloud platform integrates information such as historical user data usage, current status, geographic location, and network load to construct a dynamic attention-guided multi-task graph inference model. By combining graph neural networks and temporal convolutional network structures, it predicts future user data usage needs and urgency levels from both spatial evolution and temporal trend perspectives, enabling more refined and scenario-aware expansion strategy generation. Before generating expansion commands, a causal inference module performs counterfactual simulations of the user experience under the "expansion or not" scenario. If expansion does not significantly improve the experience, it automatically suppresses expansion or lowers the authorization level, fundamentally addressing the disconnect between expansion behavior and user perception in traditional methods and improving resource utilization efficiency. The multi-task decoding module introduces a reinforcement learning mechanism based on the Actor-Critic architecture. It uses actual user behavior after capacity expansion (such as whether the authorized quota is exhausted, subjective satisfaction, and abnormal traffic patterns) as reward signals to update the model's policy, forming a closed-loop process of "perception-prediction-feedback-optimization," significantly improving the policy's adaptability and long-term service quality. By constructing a user behavior graph comparison learning mechanism, it can quickly match the policy network of similar historical users in new user scenarios. Combined with policy distillation and local freeze migration, it solves the cold start problem of traditional models being unable to effectively predict when data is insufficient, expanding the model's applicability. During the capacity expansion authorization period, the SIM card continuously records traffic usage and periodically transmits it back to the cloud platform for decision-making on whether to terminate, renew, or adjust the quota early, achieving dynamic resource control and avoiding over-authorization. In summary, this invention significantly improves the system's intelligence, expansion efficiency, and user experience assurance capabilities by combining edge-side proactive perception, cloud-side deep prediction, multi-modal information fusion, and causal decision suppression.

[0018] The OTA traffic dynamic expansion method proposed in this application, along with other advantages, objectives, and features of this application, will be partly apparent from the following description and partly understood by those skilled in the art through research and practice of this application. Attached Figure Description

[0019] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit this specification. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1This is a flowchart illustrating an OTA traffic dynamic expansion method provided in an embodiment of this application. Figure 2 This application provides a flowchart illustrating a method for obtaining local package usage information and generating quantified status data in an embodiment of the present application. Figure 3 This is a flowchart illustrating the specific steps of a causal reasoning decision suppression module provided in an embodiment of this application. Figure 4 This is a structural schematic diagram of an OTA traffic dynamic expansion system provided in an embodiment of this application. Detailed Implementation

[0020] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus. The technical solutions of the embodiments of this application will now be clearly and completely described in conjunction with the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them.

[0021] Please see Figure 1 This is a flowchart illustrating an OTA traffic dynamic expansion method provided in an embodiment of this application, which may specifically include: S110. Deploy an Applet module in the user's SIM card to monitor data usage status, obtain local data package usage information and generate quantitative status data. The Applet module performs event counting and window accumulation on high dynamic data including network switching and session frequency through lightweight statistical processing, and forms low-dimensional quantitative indicators locally to reduce SIM card computing overhead. S120. When the above-mentioned quantized state data meets the preset triggering conditions, the above-mentioned Applet module uploads the expansion request to the operator's cloud platform through the terminal. S130. Based on the aforementioned capacity expansion request, the above-mentioned operator cloud platform combines historical traffic usage data, user behavior characteristics, geographic location information, and current network load to perform traffic demand prediction and credit policy evaluation, and generate a capacity expansion decision. S140. Generate an OTA instruction with temporary expansion parameters based on the above expansion decision; S150: Send the above OTA command to the SIM card through the OTA channel so that the SIM card can update the internal package configuration according to the above command and start temporary data authorization; S160. During the temporary authorization period, the SIM card records the actual usage and periodically reports it to the aforementioned operator's cloud platform to support dynamic adjustment or renewal control.

[0022] For example, the following is a specific implementation of the present invention, which aims to provide an OTA-based dynamic traffic expansion method. This method combines user SIM card status awareness with the intelligent prediction mechanism of the operator's cloud platform to achieve on-demand, real-time, and reliable authorization control of user data packages. Specifically, it includes the following process: First, an applet module with proactive state awareness is pre-installed in the user's SIM card. This module has the ability to periodically or event-drivenly read local data plan status information. Specifically, the applet can extract the remaining data allowance for the current plan from the file structure within the SIM card, identify the remaining time of the current plan period, and simultaneously monitor network mode switching records, cell ID switching frequency, user data session initiation frequency, and changes in the terminal's active and dormant cycles. These multi-dimensional state data are fused according to preset weights to generate "quantified state data," which comprehensively reflects the user's current data usage pressure and contextual network environment characteristics in numerical form. The applet module uses lightweight statistical processing to perform event counting and window accumulation on highly dynamic data, including network switching and session frequency, and forms low-dimensional quantitative indicators locally, thereby effectively reducing the SIM card's computational and storage overhead while ensuring monitoring accuracy.

[0023] When the aforementioned quantified status data meets one or more triggering conditions (e.g., remaining traffic is below a threshold, data usage frequency increases abnormally, network environment switches frequently, etc.), the Applet will proactively construct a capacity expansion request message and report it to the operator's cloud platform via the terminal communication link. This capacity expansion request carries the generated quantified status data and user identification information to support the cloud platform's subsequent decision-making process.

[0024] Upon receiving a capacity expansion request, the operator's cloud platform initiates an intelligent prediction and strategy evaluation process. The platform extracts behavioral characteristics such as traffic consumption patterns, typical active periods, and usage preferences from the user's historical traffic usage records. These characteristics are then combined with the quantified status data carried in the capacity expansion request, as well as the user's current geographical location information and access network load level, and fed into a dynamically attention-guided multi-task graph inference model.

[0025] The model first constructs a heterogeneous graph structure based on the user's behavioral event flow using a behavior graph construction module. This graph includes user behavior nodes, geographic location nodes, network load level nodes, and package level nodes. Nodes are connected by edges based on behavior transition paths and contextual causal relationships, and real-time environmental attributes are injected. A graph attention processing module performs graph neural network inference on this graph to generate a spatial behavior evolution tensor. Simultaneously, a temporal feature extraction module uses a multi-channel temporal convolution structure to extract temporal trend feature tensors from current state data and historical traffic. A multi-task decoding module fuses these two types of tensors and outputs three types of prediction results: predicted peak traffic for the user in the future, predicted urgency level of capacity expansion, and recommended temporary authorization validity period.

[0026] If a user's current behavior deviates from their historical behavior, the platform can further activate the causal reasoning suppression module. This module constructs a causal graph structure based on the user's expansion history and experience feedback metrics (such as download success rate, webpage response latency, video stuttering rate, etc.), and performs counterfactual estimations of the experience improvement value for both the "expansion" and "non-expansion" paths in the current state. If this value is lower than the set intervention benefit threshold, the platform will suppress the current expansion request or automatically downgrade the authorization level.

[0027] After the platform finally confirms permission for the capacity expansion, it will generate an OTA command containing temporary expansion parameters (such as additional data allowance, limited time window, and usage area) based on the above prediction results. This OTA command is sent to the user terminal via an OTA channel (such as SMS, BIP, or HTTP channel), and the terminal guides the SIM card to receive the command. After receiving the OTA command, the Applet module first performs digital signature verification and authorization ID verification to ensure the legality of the command; if the verification passes, the SIM card automatically updates its internal plan configuration according to the command content, thereby activating the temporary capacity expansion status.

[0028] During the temporary capacity expansion period, the Applet module continuously monitors the actual usage behavior of authorized traffic, constructing a usage feedback vector. This feedback includes the percentage of capacity expansion quota used, the distribution of usage time periods, whether there is resource waste or sudden surges, network stability, and whether the user triggers requests again during the authorization period. The Applet module reports the feedback vector to the cloud platform at policy-defined intervals (e.g., every 30 minutes or each time 10% usage is reached). Based on this feedback, the platform executes dynamic adjustment logic, including: determining whether to extend the capacity expansion validity period; whether to terminate the capacity expansion status early; whether to include the current behavior in the reward signal of the reinforcement learning model for optimizing subsequent strategies; and whether to update the "capacity expansion credit rating" and behavior tags in the user profile.

[0029] To improve the accuracy of the prediction model in scenarios involving new users or cold-start users when initial data is insufficient, the platform further introduces a reinforcement learning mechanism based on the Actor-Critic architecture and integrates a meta-learning module. Specifically, the platform uses a graph contrastive learning mechanism to identify a set of historical users whose behavior graph structure is similar to that of the new user, forming a candidate set of migration strategies. These strategy networks are then transferred to the initial parameters of the current user's Actor network through strategy distillation and local freezing. The platform also evaluates the fitness of the strategies based on the rate of change in user behavior. When the migration strategy is ineffective, it can automatically trigger a fine-tuning or freezing control mechanism to ensure that the new user strategy converges quickly and stably.

[0030] In summary, this invention, by deploying an Applet module in the SIM card, proactively monitors multi-dimensional states including remaining data allowance, data session frequency, terminal activity cycle, and network environment switching, and generates quantified state vectors. This enables proactive, end-user-initiated expansion needs, avoiding complete reliance on manual user triggering and significantly improving the automation and timeliness of data services. The operator's cloud platform integrates information such as historical user data usage, current status, geographic location, and network load to construct a dynamic attention-guided multi-task graph inference model. By combining graph neural networks and temporal convolutional network structures, it predicts future user data usage needs and urgency levels from both spatial evolution and temporal trend perspectives, enabling more refined and scenario-aware expansion strategy generation. Before generating expansion commands, a causal inference module performs counterfactual simulations of the user experience under the "expansion or not" scenario. If expansion does not significantly improve the experience, it automatically suppresses expansion or lowers the authorization level, fundamentally addressing the disconnect between expansion behavior and user perception in traditional methods and improving resource utilization efficiency. The multi-task decoding module introduces a reinforcement learning mechanism based on the Actor-Critic architecture. It uses actual user behavior after capacity expansion (such as whether the authorized quota is exhausted, subjective satisfaction, and abnormal traffic patterns) as reward signals to update the model's policy, forming a closed-loop process of "perception-prediction-feedback-optimization," significantly improving the policy's adaptability and long-term service quality. By constructing a user behavior graph comparison learning mechanism, it can quickly match the policy network of similar historical users in new user scenarios. Combined with policy distillation and local freeze migration, it solves the cold start problem of traditional models being unable to effectively predict when data is insufficient, expanding the model's applicability. During the capacity expansion authorization period, the SIM card continuously records traffic usage and periodically transmits it back to the cloud platform for decision-making on whether to terminate, renew, or adjust the quota early, achieving dynamic resource control and avoiding over-authorization. In summary, this invention significantly improves the system's intelligence, expansion efficiency, and user experience assurance capabilities by combining edge-side proactive perception, cloud-side deep prediction, multi-modal information fusion, and causal decision suppression.

[0031] In one feasible implementation, such as Figure 2 As shown, step S110 above, which obtains local package usage information and generates quantified status data, includes: S1101. Read the user's SIM card internal package file to determine the remaining data allowance; S1102. Obtain terminal environment switching information, wherein the terminal environment switching information includes network standard switching records and cell ID change frequency; S1103, Capture user data session startup frequency; S1104. Record the startup cycle change information of the terminal, wherein the startup cycle change information includes changes in the activation cycle and the sleep cycle. S1105. Based on the changes in the remaining traffic quota, terminal environment switching information, user data session startup frequency, and startup cycle change information, quantitative status data is generated by fusing them according to predefined weights.

[0032] In one feasible implementation, it further includes: In high mobility scenarios, increase the weight of the aforementioned terminal environment switching information; In fixed hotspot area scenarios, increase the weight of the above data session startup frequency and the above startup cycle change information; In low-activity scenarios at night, increase the weight of the remaining traffic quota mentioned above.

[0033] For example, to implement an intelligent sensing mechanism for user data usage status and a pre-judgment mechanism for triggering capacity expansion requests, an app module (applet) is deployed in the user's SIM card. This app acquires multi-dimensional status indicators from the terminal side and generates quantitative status data reflecting the current usage status. The process includes the following specific steps: First, after the Applet module is started on the SIM card side, it executes the following sub-process: S1101. Read remaining data allowance information: The Applet module accesses the EF-DATA-CAPACITY file inside the SIM card or the operator-defined exclusive service file to extract information such as the remaining data allowance, used data, and the current validity period of the current data plan, which is used to assess the user's data plan usage progress.

[0034] S1102. Obtaining Terminal Environment Switching Information: The Applet module listens to network environment parameters reported by the terminal, including network standard switching records (such as the frequency reduction during switching from 5G to 4G) and the frequency of cell ID changes, thereby determining whether the user is in a high-speed mobile or weak coverage environment. The above information can be transmitted from the terminal to the SIM card through standard interfaces (such as CAT, BIP).

[0035] S1103. Capture user data session initiation frequency: The Applet module records the frequency of data channel establishment on the terminal side, that is, the triggering frequency of events such as PDP context activation and IP layer data stream opening, which is used to characterize the user's data request density and behavior intensity.

[0036] S1104. Record terminal startup cycle change information: The module continuously observes the change pattern of the terminal's activation cycle and sleep cycle (e.g., shortening of activation duration or increase of wake-up frequency) to identify the current usage status of high-frequency wake-up and active data pull.

[0037] S1105. Fusion and Generation of Quantitative State Data: The Applet module standardizes the multiple indicators acquired in S1101 to S1104 and, based on the predefined weight configuration issued by the operator platform, fuses them to generate a set of multi-dimensional quantitative state data. For example, the following feature vector form can be constructed: Q=[R_data,F_switch,F_session,R_awake]×W Where: R_data represents the remaining data allowance percentage of the package, F_switch represents the frequency of network type or cell ID switching per unit time, F_session represents the frequency of data session initiation per unit time, R_awake represents the intensity of changes in terminal activity cycles, and W is a dynamic weight vector that is adaptively adjusted according to the current scenario.

[0038] Furthermore, to improve the environmental adaptability and discriminative ability of this state data, this embodiment introduces a scene-adaptive weight adjustment strategy, specifically including: In high mobility scenarios, if rapid changes in cell ID and frequent switching of network standards are detected, and the user is determined to be in a transportation scenario (such as high-speed rail, subway, etc.), the system will automatically increase the weight of "terminal environment switching information" (F_switch) to enhance the ability to expand response to potential outage risks. In fixed hotspot areas, if a user's GPS location remains stable for a long time and the data session frequency remains high, the system will appropriately increase the weights of "data session initiation frequency" (F_session) and "activation cycle change" (R_awake) to identify user behavior patterns that are highly active but stably persisting. In low-activity scenarios at night, if it is currently nighttime and the frequency of terminal activity and sessions has decreased significantly, the system will focus on the package balance information (R_data), that is, increase its weight, to ensure that users will not experience service interruption due to the exhaustion of their package before the next possible activation peak.

[0039] The final generated quantized state data Q will be uploaded to the cloud platform via the terminal as part of the expansion request, supporting subsequent predictive analysis, expansion strategy formulation, and OTA command generation. This method not only achieves intelligent state aggregation on the terminal side, but also improves the discriminativeness and generalization ability of the state vector by combining scenario understanding strategies, effectively ensuring the initiative, adaptability, and resource accuracy of the expansion behavior.

[0040] In one feasible implementation, the aforementioned expansion request includes the aforementioned quantized status data. Based on the aforementioned capacity expansion request, the aforementioned operator cloud platform, combined with historical traffic usage data, user behavior characteristics, geographic location information, and current network load, performs traffic demand prediction and credit policy evaluation to generate a capacity expansion decision, including: Based on the aforementioned quantitative status data, historical traffic usage data, user behavior characteristics, geographic location information, and current network load, the aforementioned operator cloud platform uses a dynamic attention-guided multi-task graph inference model to predict and generate expansion decisions, including future peak traffic for users, expansion urgency levels, and recommended authorization validity periods. The aforementioned dynamic attention-guided multi-task graph inference models include: Behavior graph construction module: Constructs a behavior event graph based on the above user behavior characteristics, and injects the above geographical location information and the above network load as node attributes into the graph structure; The graph attention processing module is used to perform graph neural network processing on the above-mentioned behavioral event graph to extract the spatial behavior evolution tensor; The temporal feature extraction module is used to perform multi-channel temporal convolution on the above-mentioned quantized state data and the above-mentioned historical traffic usage data to extract temporal feature tensors; The multi-task decoding module is used to fuse the aforementioned spatial behavior evolution tensor and the aforementioned temporal feature tensor and input them into the multi-task decoding structure to generate expansion decisions including future user traffic peaks, expansion urgency levels, and recommended authorization validity periods.

[0041] In one feasible implementation, the graph constructed by the behavior graph construction module is a heterogeneous graph structure. The node types of the heterogeneous graph structure include user behavior events, geographic location nodes, time nodes, network load level nodes, and package level nodes. The edges of the heterogeneous graph structure represent behavior transfer and scenario causal relationships. The aforementioned behavior graph construction module includes a dynamic node generation unit, which is used to generate corresponding temporary nodes in real time and inject them into the aforementioned heterogeneous graph structure when a sudden change in user behavior is detected.

[0042] In one feasible approach, the graph attention processing module adopts a spatiotemporal graph convolutional network structure to simultaneously capture the temporal evolution and spatial dependencies of the behavioral graph. The behavioral offset extracted by the graph attention processing module is also used to drive the adaptive authorization strategy. When a high offset scenario is detected, the expansion parameters are automatically adjusted. The adjustment of the expansion parameters includes extending the authorization validity period or increasing the expansion amount.

[0043] In one feasible implementation, the aforementioned multi-task decoding module integrates a reinforcement learning mechanism based on the Actor-Critic architecture; The Actor network is used to fuse the aforementioned spatial behavior evolution tensor and the aforementioned temporal feature tensor and input them into the multi-task decoding structure to generate expansion decisions including future user traffic peaks, expansion urgency levels, and recommended authorization validity periods. The Critic network is used to receive the current state characteristics and the aforementioned scaling decision, and to estimate its state value function; The cloud platform constructs reward signals based on user behavior feedback after scaling up. These reward signals include whether the scaled-up data has been exhausted, the frequency of business usage, user subjective interaction behavior, and abnormal usage indicators, in order to simultaneously optimize the parameters of the aforementioned Actor network and Critic network.

[0044] In one feasible implementation, the aforementioned Actor-Critic architecture integrates a meta-learning mechanism. In scenarios involving new users or cold-start users, the cloud platform performs graph comparison learning to calculate similarity scores based on the structural features of the user behavior graph and the historical behavior graph set. It also selects multiple optimal matching historical users as a source policy candidate set and uses a policy distillation and local parameter freezing mechanism to fuse and transfer the candidate source policies to the current Actor network. During the integration and migration process, a dynamic behavior offset detection module is introduced to assess the migration adaptability based on the rate of change in the new user's policy behavior. When the adaptability is lower than a preset threshold, the policy fine-tuning and freeze control process is automatically triggered.

[0045] For example, the behavior graph construction module organizes users' multi-source behaviors and scene elements into a heterogeneous graph. .in, It also includes multiple types of nodes such as user behavior events, geographic location nodes, time nodes, network load level nodes, and package level nodes; This indicates the causal relationship between behavior transfer and the scene; For relation type The adjacency matrix below. For each node Assigning attribute vectors (Including event time, location code, load / package level, etc.), constituting the node feature tensor To enhance the ability to characterize sudden behaviors, the module incorporates a dynamic node generation unit: when abnormal drift occurs in behavior statistics within a continuous window (e.g., session initiation frequency, location migration rate, or cell ID handover rate exceeding a threshold relative to the moving average z-score),... Temporary nodes are generated instantly. And add inbound / outbound edges to it according to the detected triggering relationship, and assign attributes accordingly. Record the type of incident, the time of occurrence, and a contextual summary to enable the heterogeneous graph to be updated online.

[0046] The graph attention processing module employs a spatiotemporal graph convolutional network (ST-GCN) to jointly model spatial dependencies and temporal evolution. Let the time-unfolded input be... In the first layer, spatial convolution first processes each relation... Adjacency using symmetric normalization To spread, we obtain: , Then, a one-dimensional convolution is performed in the time dimension (kernel length). To capture sequence dependencies:

[0047] Stacking After the layer, the embedding is obtained from the current target user subgraph through the readout operator. Meanwhile, to improve robustness, auxiliary representations from similar user groups are introduced. (Obtained from the matched similar user subgraph via a homogeneous network). Through Calculate the difference vector between the two.

[0048] Characterize the "behavioral offset" and use it for adaptive authorization in downstream policies. When a high offset scenario is detected (e.g.) At that time, the system pressed

[0049] Adaptive adjustment of license validity period, capacity expansion level, and peak budget, among which This is the adjustment coefficient.

[0050] The multi-task decoding module integrates an Actor-Critic-based reinforcement learning mechanism, fusing the spatial behavior evolution tensor and the temporal feature tensor into a state. ,in It can be obtained by extracting the quantized state time series using TCN / 1D-CNN. For context (such as network load, geographical partition, package constraints). The Actor output action is: .

[0051] Critic estimation or After actually issuing authorization and collecting feedback, the reward is constructed as follows: .

[0052] Among them, exhausted This indicates whether the expansion is consumed in a timely and reasonable manner; freq is the increase in business usage frequency; feedback is the mapping value of the user's subjective interaction rating / complaint reverse signal; anomaly_drop is the measure of abnormal behavior rate or sudden drop in quality; and overgrant is a penalty for over-provisioning waste. Weights obtained through experience or offline parameter tuning. An advantage function is used. Perform policy gradient update Critic is trained using TD error.

[0053] To address new users / cold starts, the Actor-Critic architecture integrates a meta-learning mechanism. On the cloud side, the new user graph is first processed... With historical user graph collection The similarity score for image comparison learning is:

[0054] and with The weighted fusion candidate source strategy parameters are initialized. , Distillation phase student strategies The applied temperature is The KL constraint is: .

[0055] Parameters of some layers are frozen to stabilize the migration. A dynamic behavior offset detection module is introduced during the migration to... Assess fitness ( (for smoothing terms), when Automatically trigger the fine-tuning of the small learning rate and the unlocking / relocking process of the freeze strategy until the fitness is restored.

[0056] For the number of nodes, For time steps, The initial feature dimension; For relationship The adjacency matrix, Its degree matrix, For unit array; For the first level of relation Learnable weights Nonlinear activation; and These are graph embeddings for the current user and similar groups, respectively. This is the behavior offset; Control the threshold and intensity of adaptive authorization; For Actor parameters, For Critic parameters; As a reward weight; The similarity temperature coefficient, For source strategy fusion weights; For migration fitness index, The threshold for triggering further fine-tuning.

[0057] Through the above mechanism, the system can take the robust spatiotemporal representation of behavior provided by ST-GCN as the core, combine offset-driven adaptive authorization with Actor-Critic closed-loop optimization, realize dynamic, personalized and interpretable decision-making on expansion parameters (authorization validity period, expansion level and peak budget), and maintain good convergence and business performance under cold start and sudden scene changes.

[0058] In one feasible implementation, the aforementioned multi-task graph reasoning model further includes a causal reasoning decision inhibition module, such as... Figure 3 As shown, the specific steps of the above causal reasoning decision suppression module include: S210. Establish a cause-effect graph structure based on users' historical expansion behavior, traffic usage after expansion, and experience indicators. The cause-effect graph structure includes expansion behavior processing variables, experience indicator result variables, covariates, and mediator variable nodes. S220. Extract expanded and non-expanded samples and corresponding multi-dimensional experience feedback data from the user group for use in training the causal inference model. S230. Based on the current user status, predict the expected user experience score under the conditions of performing capacity expansion and not performing capacity expansion through the above causal reasoning model, and calculate the experience improvement value between the two. The experience improvement value includes the result of the two-layer causal evaluation system. The result of the two-layer causal evaluation system includes the result of the immediate experience improvement after the user capacity expansion based on the immediate experience index, and also includes the result of the long-term value improvement of the capacity expansion based on the user retention rate and package upgrade tendency index. S240. Compare the above-mentioned experience enhancement value with the preset intervention benefit threshold; S250. When the experience improvement value is lower than the above-mentioned preset intervention benefit threshold, the generation of the current expansion command is automatically suppressed, or the expansion quota and authorization validity period are reduced.

[0059] For example, the multi-task graph inference model also includes a causal inference decision suppression module. This module is used to further determine whether to issue an expansion command after generating an expansion decision by modeling the causal relationship between expansion behavior and user experience, thereby avoiding resource redundancy, invalid expansion, or expansion with negative effects.

[0060] First, the cloud platform constructs a cause-effect graph structure based on users' historical expansion records. This cause-effect graph uses whether a user received an expansion operation as the processing variable. The user experience metrics after the expansion (such as average download speed, service completion rate, user satisfaction score, etc.) are used as the outcome variables. The causal graph is constructed by introducing variables related to user behavior, environment, and package level as covariates (C), and some mediator variables (M) that may be located between expansion and experience. .

[0061] Secondly, representative control samples (expansion / non-expansion) are extracted from the historical user base, and their multi-dimensional experience feedback data is labeled (such as whether users quickly exhaust their quotas after expansion, network congestion response, changes in complaint rate, login frequency, etc.) to construct a causal inference model for training. Models such as Dragonnet (two-branch causal inference network), TARNet (counterfactual contrast network), and GNN-Causal (graph causal inference structure) can be used to model the potential outcomes.

[0062] Then, for the user currently initiating the expansion request, the model inputs its current state feature vector. Calculate the following two results respectively: Predicted experience score under expanded capacity:

[0063] Predicted experience score without scaling:

[0064] Under the two-tiered causal evaluation system, short-term and long-term experience improvements are calculated: 1. Short-term benefit improvement (instant experience)

[0065] in, It is mainly determined by real-time experience metrics, such as download speed, latency, and packet loss rate.

[0066] 2. Long-term value enhancement

[0067] in, It is determined by long-term value metrics, such as user retention rate, subscription upgrade tendency, and long-term satisfaction.

[0068] 3. Overall Experience Enhancement Value The system weights and merges short-term and long-term improvement results:

[0069] in, The adjustable weight represents the proportion of attention given to immediate experience versus long-term value.

[0070] Subsequently, the increase value Compared with the preset intervention benefit threshold Comparison: like This indicates that the expansion action will significantly improve the user experience, allowing the current expansion strategy to continue. like If this occurs, the suppression mechanism will be triggered, and one of the following two control methods can be executed: directly suppress the generation of expansion instructions (i.e., do not issue OTA expansion instructions); or adjust the authorization parameters of the expansion instructions, such as: reducing the expansion quota, shortening the validity period of the temporary authorization, or downgrading the expansion level.

[0071] Through the above methods, this embodiment can not only predict expansion needs based on user behavior, but also determine the actual necessity of expansion through causal analysis, so that the system can avoid problems such as excessive expansion or inefficient authorization, and improve the intelligence and economy of the expansion decision system.

[0072] like Figure 4 As shown, this application proposes an OTA traffic dynamic expansion system 10, comprising: The acquisition unit 11 is used to deploy an Applet module in the user's SIM card to monitor data usage status, so as to obtain local data plan usage information and generate quantitative status data. Upload unit 12 is used to upload the expansion request to the operator cloud platform through the terminal when the above-mentioned quantized state data meets the preset trigger conditions. The first generation unit 13 is used by the above-mentioned operator cloud platform to perform traffic demand prediction and credit policy evaluation based on the above-mentioned expansion request, combined with historical traffic usage data, user behavior characteristics, geographical location information and current network load, and generate expansion decision. The second generation unit 14 is used to generate OTA instructions with temporary expansion parameters based on the above expansion decision. The startup unit 15 is used to send the above-mentioned OTA command to the SIM card through the OTA channel, so that the SIM card updates the internal package configuration according to the above-mentioned command and starts temporary data authorization; The monitoring unit 16 is used to record the actual usage of the SIM card during the temporary authorization period and periodically report it to the aforementioned operator cloud platform to support dynamic adjustment or renewal control.

[0073] Understandably, the system proposed in the second aspect is capable of performing any of the methods described in the first aspect.

[0074] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for dynamic capacity expansion based on OTA traffic, characterized in that, include: An applet module for monitoring data usage status is deployed in the user's SIM card to obtain local data plan usage information and generate quantitative status data. The applet module performs event counting and window accumulation on highly dynamic data, including network switching and session frequency, through a lightweight statistical processing method, and forms low-dimensional quantitative indicators locally to reduce the SIM card's computing overhead. When the quantized state data meets the preset triggering conditions, the Applet module uploads a capacity expansion request to the operator's cloud platform through the terminal; Based on the expansion request, the operator cloud platform combines historical traffic usage data, user behavior characteristics, geographic location information, and current network load to perform traffic demand prediction and credit policy evaluation, and generate expansion decisions. OTA instructions with temporary expansion parameters are generated based on the expansion decision; The OTA command is sent to the SIM card through the OTA channel, so that the SIM card updates its internal data plan configuration and initiates temporary data authorization according to the command; During the temporary authorization period, the SIM card records the actual usage and periodically reports it to the operator's cloud platform to support dynamic adjustment or renewal control.

2. The method for dynamic capacity expansion based on OTA traffic according to claim 1, characterized in that, The process of obtaining local package usage information and generating quantified status data includes: Read the user's SIM card's internal data plan file to determine the remaining data allowance; Obtain terminal environment switching information, wherein the terminal environment switching information includes network standard switching records and cell ID change frequency; Capture user data session startup frequency; Record the startup cycle change information of the terminal, wherein the startup cycle change information includes changes in the activation cycle and the sleep cycle; Quantitative status data is generated by fusing the remaining traffic quota, terminal environment switching information, user data session startup frequency, and startup cycle change information according to predefined weights.

3. The method for dynamic capacity expansion based on OTA traffic according to claim 2, characterized in that, Also includes: In high mobility scenarios, the weight of the terminal environment switching information is increased; In fixed hotspot area scenarios, the weights of the data session initiation frequency and the initiation cycle change information are increased; In low-activity scenarios at night, increase the weight of the remaining traffic quota.

4. The method for dynamic capacity expansion based on OTA traffic according to claim 2, characterized in that, The expansion request includes the quantized status data. Based on the capacity expansion request, the operator's cloud platform, combined with historical traffic usage data, user behavior characteristics, geographical location information, and current network load, performs traffic demand prediction and credit policy evaluation to generate a capacity expansion decision, including: The operator cloud platform, based on the quantified status data, historical traffic usage data, user behavior characteristics, geographic location information, and current network load, uses a dynamic attention-guided multi-task graph inference model to predict and generate expansion decisions, including future peak traffic for users, expansion urgency level, and recommended authorization validity period. The dynamic attention-guided multi-task graph inference model includes: Behavior graph construction module: Constructs a behavior event graph based on the user behavior characteristics, and injects the geographic location information and network load as node attributes into the graph structure; The graph attention processing module is used to perform graph neural network processing on the behavior event graph to extract the spatial behavior evolution tensor. The temporal feature extraction module is used to perform multi-channel temporal convolution on the quantized state data and the historical traffic usage data to extract temporal feature tensors; The multi-task decoding module is used to fuse the spatial behavior evolution tensor and the temporal feature tensor and input them into the multi-task decoding structure to generate expansion decisions including future user traffic peaks, expansion urgency levels, and recommended authorization validity periods.

5. The method for dynamic capacity expansion based on OTA traffic according to claim 4, characterized in that, The graph constructed by the behavior graph construction module is a heterogeneous graph structure. The node types of the heterogeneous graph structure include user behavior events, geographical location nodes, time nodes, network load level nodes, and package level nodes. The edges of the heterogeneous graph structure represent behavior transitions and scene causal relationships. The behavior graph construction module includes a dynamic node generation unit, which is used to generate corresponding temporary nodes in real time and inject them into the heterogeneous graph structure when a sudden change in user behavior is detected.

6. The method for dynamic capacity expansion based on OTA traffic according to claim 4, characterized in that, The graph attention processing module adopts a spatiotemporal graph convolutional network structure to simultaneously capture the temporal evolution and spatial dependencies of the behavioral graph. The behavioral offset extracted by the graph attention processing module is also used to drive the adaptive authorization strategy. When a high offset scenario is detected, the expansion parameters are automatically adjusted, including extending the authorization validity period or increasing the expansion amount.

7. The method for dynamic capacity expansion based on OTA traffic according to claim 4, characterized in that, The multi-task decoding module integrates a reinforcement learning mechanism based on the Actor-Critic architecture; The Actor network is used to fuse the spatial behavior evolution tensor and the temporal feature tensor and input them into the multi-task decoding structure to generate expansion decisions including future user traffic peaks, expansion urgency levels, and recommended authorization validity periods. The Critic network is used to receive the current state features and the scaling decision, and to estimate its state value function; The cloud platform constructs a reward signal based on user behavior feedback information after expansion. The reward signal includes whether the expanded data has been exhausted, the frequency of business usage, user subjective interaction behavior, and abnormal usage indicators, so as to simultaneously optimize the parameters of the Actor network and the Critic network.

8. The method for dynamic capacity expansion based on OTA traffic according to claim 7, characterized in that, The Actor-Critic architecture integrates a meta-learning mechanism. In scenarios involving new users or cold-start users, the cloud platform performs graph comparison learning based on the structural features of the user behavior graph and the historical behavior graph set to calculate the similarity score. It selects multiple optimal matching historical users as the source policy candidate set and uses a policy distillation and local parameter freezing mechanism to fuse and transfer the candidate source policies to the current Actor network. During the integration and migration process, a dynamic behavior offset detection module is introduced to assess the migration adaptability based on the rate of change in the new user's policy behavior. When the adaptability is lower than a preset threshold, the policy fine-tuning and freeze control process is automatically triggered.

9. The method for dynamic capacity expansion based on OTA traffic according to claim 4, characterized in that, The multi-task graph inference model further includes a causal inference decision suppression module, the specific steps of which include: A cause-effect graph structure is established based on users' historical expansion behavior, traffic usage after expansion, and experience indicators. The cause-effect graph structure includes expansion behavior processing variables, experience indicator result variables, covariates, and mediator variable nodes. Extract expanded and unexpanded samples and corresponding multi-dimensional experience feedback data from the user group for use in training the causal inference model; Based on the current user status, the expected user experience score is predicted by the causal inference model under the conditions of performing capacity expansion and not performing capacity expansion, and the experience improvement value between the two is calculated. The experience improvement value includes the results of a two-layer causal evaluation system. The results of the two-layer causal evaluation system include the immediate experience improvement result after the user capacity expansion determined based on the immediate experience index, and the long-term value improvement result of the capacity expansion determined based on the user retention rate and package upgrade tendency index. The experience enhancement value is compared with a preset intervention benefit threshold; When the experience improvement value is lower than the preset intervention benefit threshold, the generation of the current expansion command is automatically suppressed, or the expansion quota and authorization validity period are reduced.

10. A dynamic capacity expansion system based on OTA traffic, characterized in that, include: The acquisition unit is used to deploy an applet module in the user's SIM card to monitor data usage status, in order to obtain local data plan usage information and generate quantitative status data. The upload unit is used so that when the quantized state data meets the preset trigger conditions, the Applet module uploads the expansion request to the operator's cloud platform through the terminal; The first generation unit is used by the operator cloud platform to perform traffic demand prediction and credit policy evaluation based on the expansion request, combined with historical traffic usage data, user behavior characteristics, geographical location information and current network load, and generate expansion decisions. The second generation unit is used to generate OTA instructions with temporary expansion parameters based on the expansion decision. The startup unit is used to send the OTA command to the SIM card through the OTA channel, so that the SIM card updates the internal package configuration according to the command and starts temporary data authorization; The monitoring unit is used to record the actual usage of the SIM card during the temporary authorization period and periodically report it to the operator's cloud platform to support dynamic adjustment or renewal control.

Images