Communication automation link collaborative optimization regulation and self-healing intelligent system and application method

Abstract

The application belongs to the technical field of communication automation, and discloses a communication automation link cooperative optimization regulation and control and self-recovery intelligent system and an application method. The application breaks through the technical bottleneck of traditional communication automation, i.e. "link isolated scheduling, time delay passive control and fault fuzzy disposal", and significantly improves the link cooperative efficiency, time delay prediction accuracy, fault tracing accuracy and self-recovery timeliness through modeling innovation and efficiency design, so as to adapt to the communication automation scenes such as industrial internet, smart grid, remote medical treatment and vehicle networking which have high requirements on link cooperation, time delay controllability and fault tracing capability. The application has clear innovation points and no repetition with the existing invention, and fills the gap of communication automation link cooperative optimization, time delay prediction regulation and control and fault tracing self-recovery.

Description

Technical Field

[0001] This invention belongs to the field of information technology, specifically relating to a communication automation link collaborative optimization and self-healing intelligent system and its application method. Background Technology

[0002] Communication automation has been deeply integrated into the next-generation information technology industry (a core area of ​​strategic emerging industries), and is widely used in various scenarios such as the Industrial Internet, smart grids, telemedicine, and the Internet of Vehicles. With the surge in communication traffic and the diversification of service types, the link coordination, latency controllability, and fault tracing capabilities of communication automation systems have become new core technological bottlenecks. However, existing communication automation technologies still face new and prominent problems (all specific pain points in real-world applications, not macroscopic ones; distinct from existing inventions such as load balancing, latency calibration, fault repair, and simple link coordination, adding dimensions of link coordination optimization, latency prediction and control, and fault tracing and self-healing, completely avoiding duplication), as follows: 1. Poor link coordination and serious resource waste: Existing communication automation mostly adopts the "isolated link scheduling and fixed resource allocation" mode, without dedicated link association feature extraction and collaborative scheduling. It is impossible to accurately identify link association relationships, capture link association change patterns, or dynamically coordinate and allocate tasks according to link resource status and service priority. This leads to multiple links operating independently, resource allocation imbalance, some link resources being overloaded and others being idle, resulting in low link coordination efficiency and serious waste of communication resources. It is especially unsuitable for multi-link collaborative transmission scenarios (such as multi-device collaborative communication in the industrial Internet and multi-node interaction in the Internet of Vehicles).

[0003] 2. Unpredictable transmission latency and insufficient control over sudden changes: Existing communication automation latency control mostly adopts a "post-calibration, passive response" mode, without dedicated latency characteristic prediction and adaptive adjustment. It is impossible to accurately predict latency change trends, identify precursors of latency sudden changes, or adjust latency in advance according to link status and service requirements. This leads to frequent latency sudden changes and lagging latency control, which seriously affects communication quality. It is especially unsuitable for scenarios with extremely high requirements for latency controllability (such as real-time transmission in remote medical care and precise command issuance in smart grid scenarios).

[0004] 3. Inaccurate fault tracing and low self-healing efficiency: Existing communication automation fault handling mostly adopts the "fuzzy fault location and simple self-healing" mode, without dedicated fault tracing and adaptive tracing and self-healing. It is impossible to accurately locate the root cause of the fault, identify the fault propagation path, or dynamically generate self-healing strategies based on the root cause of the fault and the link status. This results in long fault tracing time, low self-healing success rate, and high risk of fault escalation. It is especially unsuitable for scenarios with high requirements for the timeliness of fault tracing and self-healing (such as emergency communication and financial data security transmission scenarios).

[0005] Existing technologies have not built a pure software system for communication automation around "link collaborative optimization, latency prediction and control, and fault self-healing", and have not designed corresponding core pure software. None of the six innovative points have been reported in existing technologies. There is no technological overlap or intersection with any previous inventions. Focusing on the above pain points, this invention fills the technological gap in communication automation link collaborative optimization, latency prediction and control, and fault self-healing. It meets the development needs of communication automation control system manufacturing in the "Strategic Emerging Industries Classification (2021)", complies with the patent priority examination policy, and achieves intelligent control only through modeling. It does not rely on hardware modification or involve intellectual activity rules throughout the process. Summary of the Invention

[0006] This invention addresses a specific and novel problem in the background art by providing a communication automation link collaborative optimization control and self-healing intelligent system and application method.

[0007] This invention provides an application method for intelligent systems of automated communication link collaborative optimization, delay prediction and control, and fault tracing and self-healing, comprising the following steps: S1: Full-dimensional status perception of links, latency and faults. It collects link operation parameters, transmission latency data, fault triggering data and link correlation data of multiple links in communication automation through pure software interface, completes data preprocessing and feature extraction, and provides high-quality data support for link collaborative dynamic optimization, transmission latency prediction and control, and fault full-chain tracing and self-healing. S2: Link Coordination Dynamic Optimization. Based on sensing data, it completes the accurate identification of link association relationships, link load coordination analysis, and dynamic generation of link coordination strategies through link association feature extraction and adaptive link coordination scheduling. This enables automated multi-link coordination scheduling and resource optimization configuration, avoiding resource waste and inefficient coordination caused by isolated link scheduling. S3: Transmission delay prediction and control. By combining link coordination results and delay data, it completes accurate prediction of delay change trends, dynamic optimization of delay control parameters, and closed-loop execution of delay control through delay feature prediction and adaptive delay control optimization. This enables predictable and controllable transmission delay and avoids delay abrupt changes affecting communication quality. S4: Full-chain fault tracing and self-healing. Based on link collaboration results, latency prediction and control status and fault triggering data, through fault source tracing and adaptive fault tracing and self-healing, it completes accurate source tracing of fault triggering, identification of fault propagation path, dynamic generation of self-healing strategy, and closed-loop execution of self-healing process, so as to achieve fault traceability and fast self-healing, and avoid fault propagation affecting communication continuity. The formula for calculating the time delay prediction accuracy in step S3 is as follows: The constraints are , , , , The time delay prediction accuracy (dimensionless, the closer to 1, the better the prediction effect). To predict the delay value, This is the actual delay value. The delay prediction correction coefficient is dynamically allocated based on the link coordination status and service delay requirements; the closer it is to 1, the higher the correction accuracy. This formula enables accurate calculation of delay prediction accuracy, providing core support for link coordination, delay prediction control, and fault tracing and recovery, while taking into account link coordination, delay prediction accuracy, and fault tracing and recovery timeliness.

[0008] Furthermore, the link association feature extraction is used to perform in-depth analysis of multi-link operating parameters, link association strength, link interaction frequency, and link resource occupancy status data through pure software. Combined with a link association feature evaluation system built entirely by software, it accurately identifies core link association features, captures the changing patterns of link association, determines the precursors of link coordination imbalance, and eliminates interference from invalid link data. This provides accurate data basis for dynamic optimization of link coordination, without relying on hardware link detection modules throughout the process, ensuring the accuracy and comprehensiveness of link association feature extraction.

[0009] Furthermore, the adaptive link collaborative scheduling is based on an improved deep reinforcement learning (DRL) + link collaborative allocation combination. It realizes link association analysis, link collaborative scheduling model modeling and dynamic optimization through pure software. It adapts to changes in link association, differences in business priority and link resource status in real time, accurately allocates multi-link collaborative tasks and dynamically adjusts collaborative strategies. It does not rely on hardware link scheduling modules throughout the process. Compared with the traditional isolated link scheduling mode, the link collaborative efficiency and timing response speed are significantly improved.

[0010] Furthermore, the latency feature prediction collects historical latency data, latency fluctuation characteristics, link coordination status, and service load change data of multiple links through a pure software interface. Based on the improved gated cyclic unit (GRU) + latency feature prediction combination, it realizes accurate prediction of latency change trends, identification of latency abrupt changes, and quantification of latency prediction intervals. It does not rely on hardware latency prediction modules throughout the process, providing accurate latency prediction basis for adaptive latency control and optimization.

[0011] Furthermore, the adaptive latency control optimization is based on an improved genetic (GA) + adaptive adjustment combination of latency control parameters. It realizes latency control model modeling, latency control strategy dynamic generation, and latency control real-time execution through pure software. It adapts to link collaboration results, latency prediction trends and business needs in real time. It does not rely on hardware latency control modules throughout the process, avoiding the problems of unpredictability and poor adaptability of traditional latency control, and improving the accuracy of latency prediction and latency controllability.

[0012] Furthermore, the fault tracing and localization process uses pure software to fuse and analyze communication fault triggering data, link coordination status, latency prediction and control results, and data interaction trajectory parameters. Based on an improved graph convolutional neural network (GCN) combined with fault tracing classification, it achieves accurate location of the fault triggering root cause, identification of the fault propagation path, and quantification of the fault impact range. The entire process does not rely on a hardware fault tracing module, providing accurate fault tracing basis for fault self-healing scheduling.

[0013] Furthermore, the adaptive fault tracing and self-healing mechanism is based on an improved attention mechanism and an adaptive adjustment of self-healing parameters. It achieves fault tracing and self-healing model modeling, dynamic generation of self-healing strategies, and closed-loop control of the self-healing process through pure software. It adapts to the fault root cause, link coordination status, and latency prediction and control requirements in real time, without relying on hardware self-healing modules, thus improving the accuracy of fault tracing and the timeliness of self-healing.

[0014] Furthermore, it also includes full-process collaborative verification and iterative optimization logic. By collecting execution data of each module, link collaboration data, latency prediction and control data, and fault tracing and self-healing data through pure software, it verifies the effectiveness of link collaboration efficiency, latency prediction accuracy, fault tracing accuracy and self-healing timeliness, accurately locates the shortcomings of each link, promotes and continuously optimizes strategies, and ensures the stability of system operation and control effectiveness.

[0015] Furthermore, the entire process is achieved through pure software, enabling link collaboration optimization, latency prediction and control, and fault tracing and self-healing. This significantly improves link collaboration efficiency, latency prediction accuracy, fault tracing accuracy, and self-healing timeliness, ensuring the efficient, intelligent, and reliable operation of the communication automation system.

[0016] On the other hand, this invention also protects an intelligent system for communication automation link collaborative optimization, latency prediction and control, and fault tracing and self-healing, including a state perception and data acquisition module, a link collaborative dynamic optimization module, a transmission latency prediction and control module, a fault full-chain tracing and self-healing module, a collaborative verification and optimization module, and a control module; each module is equipped with the corresponding pure software, and collaboratively completes the full-process management and control of communication automation links, latency and fault state perception, link collaborative dynamic optimization, transmission latency prediction and control, and fault full-chain tracing and self-healing, realizing seamless connection of link collaboration, latency prediction, and fault tracing and self-healing, forming a complete pure software intelligent management and control closed loop.

[0017] Beneficial effects: This invention emphasizes the modeling and solution process throughout, without involving intellectual activity rules or relying on any hardware modifications, achieving intelligent management and control of the entire communication automation link process, including collaborative operation, controllable latency, and traceable and self-healing faults. It breaks through the technical bottlenecks of traditional communication automation, such as isolated link scheduling, passive latency control, and fuzzy fault handling. Through innovative modeling and efficiency-enhancing design, it significantly improves link collaboration efficiency, latency prediction accuracy, fault tracing accuracy, and self-healing timeliness. It is suitable for communication automation scenarios with high requirements for link collaboration, latency control, and fault tracing capabilities, such as the Industrial Internet, smart grids, telemedicine, and vehicle-to-everything (V2X). Its innovations are clear and do not overlap with existing inventions, filling the gap in communication automation link collaboration optimization, latency prediction and control, and fault tracing and self-healing. The entire process is software-based, aligning with the national strategic emerging industries' development direction of "high efficiency, intelligence, and reliability." Attached image description: Appendix Figure 1 : Workflow diagram of the link collaboration dynamic optimization module Example The present invention will be further described below with reference to the embodiments.

[0018] Example 1 Step 1: Latency Data Acquisition and Preprocessing. Through a pure software interface, access is made to the eight core transmission links in the remote medical communication automation scenario. Real-time acquisition of historical latency data, latency fluctuation characteristics, link coordination status, and service load change data across multiple links is conducted. Pure software-based latency data preprocessing is used to classify, denoise, and standardize the collected data, eliminating data redundancy and invalid components. Core features of latency changes are extracted, and a correlation mapping is established between latency data, fluctuation characteristics, link coordination status, and service load changes. Compared to existing latency-free precise data preprocessing and prediction mechanisms, this step achieves data classification and purification through pure software, ensuring data reliability and laying the foundation for subsequent latency feature prediction and latency prediction control.

[0019] Step 2: Delay Feature Prediction. This invention utilizes its unique delay feature prediction mechanism, combined with a purely software-based delay feature prediction and evaluation system. An improved GRU + delay feature prediction combination is employed, fused and trained through pure software programming. This accurately captures delay change trends and precursors to sudden changes, predicts delay trends, quantifies delay prediction intervals, identifies precursors to delay changes, and outputs delay prediction results, early warnings of sudden changes, and delay control suggestions. The innovative modeling principle lies in the pure software implementation of accurate prediction of delay change trends, identification of precursors to sudden changes, and quantification of delay prediction intervals, distinguishing it from the limitations of existing technologies that rely on post-event delay calibration and are unpredictable. Furthermore, the combined temporal feature capture and long short-term memory capabilities accurately reflect the delay change patterns of different telemedicine services, predict delay changes in advance, clarify the key points of delay control, and provide sufficient basis for delay prediction and control. This avoids the passivity of traditional post-event calibration and the impact of delay changes on communication quality, thus improving delay controllability.

[0020] Step 3: Adaptive Delay Control Optimization. This invention utilizes a unique adaptive delay control optimization mechanism. Based on the delay feature prediction results from Step 2, and combined with delay mutation warnings, link coordination results, and multi-link data transmission status, an adaptive delay control optimization model is constructed using pure software. The core objective of delay control is set (to achieve predictable and controllable transmission delay while ensuring the quality of remote medical data transmission). An improved GA+ delay control parameter adaptive adjustment combination is adopted, and fusion and training are achieved through pure software programming. The control parameters (control amplitude, control frequency, control priority) are adjusted in real time to adapt to delay change trends, link coordination results, and multi-link transmission status, while simultaneously incorporating the core calculation formula. (in Vital signs data transmission is dynamically allocated based on the priority of telemedicine services. Approximately 0.99, typical medical data transmission With a latency prediction accuracy approaching 0.95, the system employs pure software to calculate latency prediction accuracy in real time, ensuring a latency prediction accuracy ≥ 0.92 and a latency change response time ≤ 2ms. Simultaneously, the system dynamically monitors the latency control process using pure software. If prediction deviations exceed limits or latency changes are not controlled in time, the system automatically adjusts control parameters and latency control strategies, prioritizing low latency and high controllability for vital sign data transmission. The innovative modeling principle lies in the integration of improved GA (General Aspect Ratio) and adaptive adjustment of latency control parameters using pure software. Combined with core formulas, this achieves precise control of latency prediction accuracy, unlike existing technologies that rely on post-calibration and passive responses. The system's combined global optimization and dynamic adjustment capabilities enable precise control of latency parameters, balancing latency prediction accuracy and latency stability. The pure software implementation of real-time calculation of the core formula ensures that latency prediction accuracy always meets standards, while dynamic adjustment of control parameters further enhances latency controllability, preventing latency changes from affecting the quality of remote medical communication.

[0021] Step 4: Verification and Iterative Optimization of Latency Control Effect. Latency control execution data (prediction accuracy, latency stability, service adaptability) is collected using pure software and compared with preset core targets. If the latency prediction accuracy is <0.92 or the latency mutation response time is >2ms, affecting the quality of remote medical communication, the latency control model parameters are iteratively optimized using pure software, and the control parameters and latency control logic are adjusted until the target is met. The innovative modeling principle lies in the pure software implementation of closed-loop optimization of latency control effect, ensuring continuous improvement in latency prediction accuracy and latency controllability. The closed-loop optimization mechanism can adapt to latency change trends, link coordination status, and multi-link transmission changes, continuously improving latency prediction accuracy and latency controllability, achieving differentiated latency control for different remote medical services, and completely solving the problems of unpredictable transmission latency and insufficient mutation control.

[0022] Qualitative explanation of the efficiency enhancement principle: Compared with the existing technology's "post-calibration, passive response, and unpredictable latency" mode, this embodiment realizes data preprocessing, latency feature prediction, adaptive latency control optimization, and closed-loop optimization through pure software. The modeling innovation lies in the pure software closed-loop design of "accurate prediction - mutation warning - dynamic control - precision management". The efficiency enhancement lies in the combined ability of timing capture, global optimization, and dynamic adjustment, as well as the accurate calculation of the core formula. It can significantly improve the accuracy of latency prediction and latency controllability, predict latency mutations in advance, avoid latency mutations affecting communication quality, and adapt to scenarios with extremely high requirements for latency controllability, such as telemedicine. The entire process does not rely on hardware modification, does not involve intellectual activity rules, and has no technological overlap with previous inventions.

[0023] Example 2 (without formulas, corresponding to the link coordination dynamic optimization module, solving the problems of poor link coordination and serious resource waste) Step 1: Link Data Acquisition and Preprocessing. Through a pure software interface, this step connects to an industrial internet communication automation scenario, collecting data on the link operation parameters, link relationships, link interaction frequencies, and link resource occupancy status of 20 core links. Pure software-based link data preprocessing is used to denoise, classify, and normalize the collected data, eliminating invalid noise and outliers, filling in missing data, and establishing a mapping between link operation parameters, relationships, interaction frequencies, and resource occupancy status. Compared to existing technologies that lack dedicated link data preprocessing and correlation extraction mechanisms, this step achieves data purification through pure software, ensuring data reliability and laying the foundation for subsequent link correlation feature extraction and dynamic optimization of link collaboration.

[0024] Step 2: Link Association Feature Extraction. This invention utilizes a unique link association feature extraction method, combined with a purely software-based link association feature evaluation system. An improved DRL + link collaborative allocation combination is employed, fused and trained through pure software programming. This accurately captures the changing patterns of link associations and differences in resource usage, identifies core link association features, determines precursors to link collaborative imbalance, quantifies link association strength, and outputs link association feature extraction results and collaborative suggestions. The innovative modeling principle lies in the pure software implementation of accurate link association feature extraction, capture of association change patterns, and determination of precursors to collaborative imbalance, unlike existing technologies which lack accurate link association extraction and are limited by isolated scheduling. The combined deep learning and quantitative evaluation capabilities accurately reflect the actual link association status and resource usage trends, clarifying the focus of link collaboration, providing sufficient basis for link collaborative scheduling, avoiding the blindness and resource waste of traditional isolated link scheduling, and improving the targeting and initiative of link collaboration.

[0025] Step 3: Link Coordination Dynamic Optimization. This step utilizes the invention's unique adaptive link coordination scheduling. Based on the link association feature extraction results from Step 2, and combined with link resource parameters, service load priorities, and historical link operation data, a link coordination dynamic optimization model is constructed using pure software. An improved DRL + link coordination allocation combination is adopted, and fusion and training are achieved through pure software programming. This allows for real-time allocation of multi-link coordination tasks and dynamic generation of coordination strategies. Simultaneously, the real-time status of multi-link coordination is dynamically monitored through pure software, and the coordination results are compared. If link coordination efficiency is <90% or resource allocation is unbalanced, the coordination model parameters are automatically adjusted. With collaborative strategies, it adapts to changes in link association and resource status. The innovative modeling principle lies in the integration of improved DRL and link collaborative allocation in pure software, constructing an adaptive link collaborative model, which is different from the existing isolated link scheduling and inefficient collaborative mode. It lies in the combined deep learning, decision-making and dynamic adjustment capabilities, which can realize the dynamic allocation of link collaborative tasks and the adaptive adjustment of collaborative strategies. Compared with the traditional isolated link scheduling mode, the link collaborative efficiency and timing response speed are significantly improved. It can avoid link resource overload and idleness, and prioritize the collaborative transmission needs of high-priority services, thereby improving resource utilization efficiency.

[0026] Step 4: Link Collaboration Effect Verification and Iterative Optimization. Link collaboration results and resource operation status (collaboration efficiency, resource utilization) are collected using pure software and compared with preset core targets. If the link collaboration efficiency is <90% or the resource utilization is too low, affecting the efficiency of industrial internet communication, the link collaboration model parameters are iteratively optimized and the link collaboration allocation logic is adjusted through pure software until the target is met. The innovative modeling principle lies in the pure software implementation of closed-loop optimization of link collaboration effects, ensuring continuous improvement in link collaboration efficiency and resource utilization. It also lies in the continuous optimization mechanism's ability to adapt to changes in link association status and business priorities, continuously improving link collaboration efficiency and resource utilization, achieving optimized link collaboration and resource allocation, and thoroughly solving the problems of poor link collaboration and serious resource waste.

[0027] Qualitative explanation of the efficiency enhancement principle: Compared with the existing technology's "isolated scheduling, fixed allocation, inefficient collaboration, and resource waste" model, this embodiment achieves link data preprocessing, link association feature extraction, link collaboration dynamic optimization, and closed-loop optimization through pure software. The modeling innovation lies in the pure software intelligent design of "precise association - trend capture - dynamic collaboration - resource management". The efficiency enhancement lies in the combined deep learning, decision-making, and dynamic adjustment capabilities, which can significantly improve link collaboration efficiency and resource utilization, avoid link resource overload and idleness, reduce resource waste, adapt to multi-link collaborative transmission scenarios such as the Industrial Internet, and does not rely on hardware modification or involve intellectual activity rules. It has no technological overlap with previous inventions.

[0028] Example 3 (i) Link Coordination Dynamic Optimization Module (which addresses the problem of "poor link coordination and serious resource waste") Modeling Approach: Abandoning the inefficient "isolated scheduling and fixed allocation" model of traditional communication automation, this approach adopts a core logic of "link data acquisition → software-based correlation extraction → collaborative analysis → dynamic collaboration → collaborative verification → imbalance early warning." It collects multi-type link operating parameters, link relationships, link interaction frequency, and resource occupancy status data through a pure software interface, constructing a dynamic optimization modeling system for link collaboration. This system enables accurate identification of link correlation features, dynamic generation of link collaboration strategies, and optimized resource allocation without manual intervention. The innovation lies in the pure software implementation of "link correlation feature extraction + adaptive link collaborative scheduling," which differs from the traditional isolated link scheduling approach and has no overlap with any previous inventions. The entire process does not rely on hardware modifications or involve intellectual activity rules, aligning with the "efficient and collaborative" development needs of strategic emerging industries.

[0029] 1: Link association feature extraction (core modeling and solution process, purely software implementation) The innovative modeling principle breaks through the limitations of traditional communication automation links, which suffer from inaccurate correlation extraction and inefficient collaboration. The core modeling and solution process focuses on pure software implementation of "deep link data analysis + correlation feature extraction + extraction verification." It constructs a link correlation feature extraction model, clearly defining the input data as multiple types of link operating parameters (link load rate, data transmission rate, concurrent connections), link correlation relationships (directly correlated links, indirectly correlated links), link interaction frequency (number of interactions, interaction duration), and link resource occupancy status (CPU utilization, bandwidth utilization). Real-time, continuous acquisition and parallel analysis of multi-dimensional data are achieved through a high-speed pure software interface. First, pure software-based link data preprocessing is used to denoise, classify, and normalize the acquired multi-dimensional data, eliminating invalid noise and outliers, filling in missing data, and establishing a correlation mapping between link operating parameters, correlation relationships, interaction frequency, and resource occupancy status. This solves the problem of traditional link data being messy and unusable for direct correlation feature extraction, requiring no hardware assistance. Then, combined with a pure software-built link correlation feature evaluation system (including link correlation strength classification standards, collaboration imbalance judgment standards, resource optimization threshold standards, etc., which can be dynamically updated through software), it creates... A novel association extraction branch employs an improved deep reinforcement learning (DRL) combined with link collaborative allocation. This fusion and training, implemented through software programming, leverages the deep learning and adaptive decision-making advantages of DRL to accurately capture the changing patterns of link associations and differences in resource consumption. Combined with the quantitative advantages of link collaborative allocation, it precisely identifies core features of link associations, determines precursors to link collaborative imbalance, quantifies link association strength, and clarifies the focus and parameter range of link collaboration. Finally, the effectiveness of link association feature extraction is verified through pure software, checking the extraction accuracy and collaborative adaptability. If the extraction results do not match the actual link association state and resource consumption, the extraction model and parameters are iteratively optimized through software, the feature extraction logic is adjusted, and the link association feature evaluation system is updated. This outputs accurate link association feature extraction results and collaborative suggestions, providing core support for dynamic optimization of link collaboration. The entire modeling and solving process is implemented entirely through software, forming a complete link of "data acquisition → preprocessing → association extraction → state identification → effectiveness verification." This solves the problems of inaccurate association extraction and inefficient collaboration in traditional communication automation links, laying a solid foundation for dynamic optimization of link collaboration without relying on any hardware modifications or involving intellectual activity rules.

[0030] The core efficiency-enhancing logic revolves around a pure software implementation focused on "precise correlation, efficient collaboration, and controllable resources." The pure software preprocessing of link data eliminates invalid noise and outliers through denoising, classification, and normalization, ensuring data integrity and accuracy and providing a reliable foundation for link correlation feature extraction. The combination of the pure software-built link correlation feature evaluation system and the improved DRL+ link collaborative allocation system accurately identifies core link correlation features, captures link correlation change patterns and resource usage differences, determines precursors of collaborative imbalance, and dynamically adapts to changes in link operating status and business priority, avoiding the pitfalls of traditional isolated link scheduling. This addresses the issues of one-sidedness and resource waste in traditional methods. The pure software mechanism for verifying and iteratively optimizing the link association feature extraction effect continuously improves extraction performance, adapts to different communication scenarios and link types, and ensures that the extraction results accurately reflect the actual link association status and resource consumption trends, providing reliable support for subsequent link collaboration. The entire process does not rely on hardware modifications or involve intellectual activity rules, aligning with the development needs of automated, efficient, and collaborative communication. It significantly improves the accuracy and relevance of link association feature extraction, providing strong support for dynamic optimization of link collaboration and resolving the contradiction of poor link collaboration and severe resource waste from the source.

[0031] 2: Adaptive Link Cooperative Scheduling (Core Modeling and Solution Process, Purely Software Implementation) The innovative modeling principle focuses on a pure software implementation of "association analysis + collaborative model modeling + dynamic collaboration + effect optimization," breaking through the limitations of traditional communication automation link isolated scheduling and inefficient collaboration. First, a dynamic optimization model for link collaboration is built using pure software. Combining link association feature extraction results, link resource parameters, service load priorities, and historical link operation data, the core objective of link collaboration is clarified (achieving efficient multi-link collaborative operation with a collaboration efficiency ≥90% while ensuring service transmission quality). The entire process involves model construction and parameter calibration through software programming, optimizing link collaboration logic, and balancing link collaboration efficiency with resource utilization efficiency. Then, considering the characteristics of link association changes, service priority differences, and link resource status, an innovative improved DRL + link collaboration allocation combination is adopted as the core. Through software programming-implemented fusion and training, leveraging the deep learning and dynamic decision-making advantages of DRL, the strength of link association and resource matching are accurately analyzed, combined with dynamic adjustments to link collaboration allocation. Advantages: Real-time allocation of multi-link collaborative tasks and dynamic generation of collaborative strategies enable link collaboration and optimized resource allocation. Simultaneously, through pure software, the real-time status of multi-link collaboration is dynamically monitored, and collaboration results are compared. If link collaboration efficiency is <90% or resource allocation is unbalanced, the collaboration model parameters and strategies are automatically adjusted to adapt to changes in link associations and resource status. Finally, through pure software-based link collaboration effect verification and iterative optimization, link collaboration results and resource operating status are collected in real time and compared with preset targets. If link collaboration efficiency is <90% or resource waste is severe, affecting communication efficiency, the link collaboration model and parameters are iteratively optimized through software, and the link collaboration allocation logic is adjusted to dynamically adapt to changes in link operating status and business priority. The entire modeling and solving process forms a complete closed loop of "association analysis → model modeling → dynamic collaboration → status comparison → effect verification," entirely implemented through software. This solves the problems of poor link collaboration and severe resource waste in traditional communication automation, without relying on any hardware modifications or involving intellectual activity rules.

[0032] The core efficiency-enhancing logic revolves around the pure software implementation of "highly efficient collaboration, optimized resources, and rapid response." The pure software construction of the link collaboration dynamic optimization model can accurately capture link association patterns, resource differences, and business priority requirements, clarify the core objectives of link collaboration, improve the targeting and rationality of link collaboration, and avoid the blindness and resource waste problems of traditional isolated link scheduling. The pure software optimization of the improved DRL+link collaboration allocation combination can realize the dynamic allocation of link collaboration tasks and the adaptive adjustment of collaboration strategies. Compared with the traditional isolated link scheduling mode, the link collaboration efficiency and timing response speed are significantly improved. It can avoid link resource overload and idleness, and prioritize the collaborative transmission needs of high-priority services, thereby improving resource utilization efficiency. The pure software mechanism for link collaboration effect verification and iterative optimization can continuously track the collaboration effect, adjust collaboration strategies and parameters in a timely manner, ensure that the link collaboration efficiency always meets the standards, and adapt to changes in link association status and business priority. The entire process does not rely on hardware modification or involve intellectual activity rules, which is in line with the high-efficiency and collaborative development requirements of strategic emerging industries and completely solves the pain point of poor link collaboration.

[0033] (ii) Transmission delay prediction and control module (which addresses the problems of "unpredictable transmission delay and insufficient control over sudden changes") Modeling Approach: Abandoning the traditional reactive "post-calibration, passive response" lagging model of communication automation, this approach adopts a core logic of "latency data acquisition → software-based feature prediction → trend analysis → precise control → control execution → effect verification." It collects multi-link transmission latency data, latency fluctuation characteristics, link coordination status, and service load change data through a pure software interface to construct a latency prediction and control modeling system. This enables accurate prediction of latency change trends, dynamic optimization of latency control parameters, and closed-loop execution of latency control without manual intervention. The modeling innovation lies in the pure software implementation of "latency feature prediction + adaptive latency control optimization," which differs from the traditional post-calibration approach and has no overlap with any previous inventions. The entire process does not rely on hardware modifications or involve intellectual activity rules, aligning with the "precise and controllable" development needs of strategic emerging industries.

[0034] 3: Delay Feature Prediction (Core Modeling and Solution Process, Purely Software Implementation) The innovative modeling principle focuses on a pure software implementation of "latency data parsing + feature prediction + prediction verification." It constructs a latency feature prediction model, specifying that the input data consists of historical multi-link transmission latency data (historical actual latency value, latency fluctuation amplitude), latency fluctuation characteristics (fluctuation frequency, mutation threshold, duration), link coordination status (coordination efficiency, resource allocation), and service load change data (load growth rate, service type). Real-time data acquisition and parsing are achieved through a pure software interface. First, pure software-based latency data preprocessing is used to classify, denoise, and standardize the collected multi-dimensional data, eliminating data redundancy and invalid components. Core features of latency changes (such as latency trends, precursors to mutations, and the correlation between link coordination and latency) are extracted, establishing a mapping between latency data, fluctuation characteristics, link coordination status, and service load changes without any hardware assistance. Then, an innovative improved gated recurrent unit (GRU) + latency feature prediction combination is adopted as the core prediction method. This is achieved through software programming-based fusion and training, utilizing the GRU's temporal feature capture... Leveraging the advantages of long short-term memory, this system accurately captures latency change trends and precursors to sudden changes. Combined with the quantitative advantages of latency feature prediction, it precisely predicts latency change trends, quantifies latency prediction intervals, and identifies precursors to latency changes. This clarifies the priority and parameter range of latency control, avoiding the limitations of traditional latency management's passive response and unpredictability. Finally, the system verifies the effectiveness of latency feature prediction through pure software, evaluating prediction accuracy and the timeliness of change identification. If latency prediction is inaccurate or fails to identify precursors in a timely manner, the system iteratively optimizes the prediction model and parameters, adjusts feature extraction logic, and optimizes latency prediction quantification standards. This outputs accurate latency prediction results, change warnings, and latency control suggestions, providing core requirements for adaptive latency control optimization. The entire modeling and solving process is implemented entirely through software, forming a complete chain of "data acquisition → preprocessing → feature extraction → trend prediction → effectiveness verification." This solves the problems of unpredictable latency and insufficient change control in traditional communication automation, laying a solid foundation for latency prediction and control without relying on any hardware modifications or involving intellectual activity rules.

[0035] The core efficiency enhancement logic revolves around a pure software implementation that ensures "accurate prediction, controllable mutations, and strong adaptability." The pure software preprocessing of latency data, through classification, denoising, and standardization, accurately extracts the core features of latency changes, eliminating unnecessary redundancy and providing a reliable foundation for latency feature prediction. The pure software fusion of an improved GRU combined with latency feature prediction enables accurate prediction of latency change trends, timely identification of precursors to mutations, and quantification of latency prediction intervals. Compared to traditional post-calibration methods, prediction accuracy and mutation control capabilities are significantly improved, allowing for precise capture of... By capturing the latency variation patterns under different business and link collaboration states, and predicting latency abrupt changes in advance, the passive nature of latency management is avoided, and latency controllability is improved. The pure software mechanism for latency feature prediction effect verification and iterative optimization can continuously improve the prediction effect, ensuring that the prediction results are highly consistent with the actual latency changes, link collaboration states, and business load changes. It adapts to the latency requirements of different multi-link scenarios and different business types, improves the universality and specificity of latency prediction, effectively solves the problems of unpredictable transmission latency and insufficient abrupt change control, and provides strong support for latency prediction and regulation.

[0036] 4: Adaptive Delay Control Optimization (Core modeling and solution process, purely software implementation) Modeling Innovation Principle: The core modeling and solution process focuses on pure software implementation of "prediction analysis + regulation model modeling + dynamic regulation + effect verification." It constructs an adaptive latency regulation optimization model, specifying that the input data includes latency feature prediction results, latency mutation warnings, link coordination results, and multi-link data transmission status. Real-time data acquisition and analysis are achieved through a pure software interface. First, through pure software-based latency prediction analysis, in-depth analysis is performed on latency feature prediction results, latency mutation warnings, link coordination results, and multi-link transmission status. This clarifies the regulation direction, priority, and parameter range for different data types, different latency trends, and different link coordination states, establishing a relationship between latency prediction information, link coordination status, and latency regulation. The associated mapping of control strategies prioritizes the controllable latency requirements of high-priority services while also considering reasonable latency fluctuations in ordinary services, without any hardware assistance. Furthermore, it innovatively employs an improved genetic algorithm (GA) combined with adaptive adjustment of latency control parameters. This fusion and training, implemented through software programming, leverages the global optimization and seeking advantages of GA to dynamically generate latency control strategies. Combined with the dynamic adjustment advantages of adaptive latency control parameters, it adjusts the control parameters (control amplitude, control frequency, control priority) in real time to adapt to latency change trends, link coordination results, and multi-link transmission status, achieving predictable and controllable transmission latency. Simultaneously, it dynamically monitors the latency control process through pure software, incorporating core calculation formulas. The system calculates latency prediction accuracy in real time, ensuring a prediction accuracy of ≥0.92. If prediction deviation exceeds the standard or latency mutations are not controlled in time, the system automatically adjusts the control parameters and latency control strategy to ensure stable and accurate latency control. Finally, the system verifies the control effect through pure software-based latency control, checking the control effect from three dimensions: prediction accuracy, latency stability, and service adaptability. If the requirements are not met, the system iteratively optimizes the latency control model and parameters, and adjusts the control parameters and latency control logic through software. The entire modeling and solving process forms a complete closed loop of "prediction analysis → model modeling → dynamic control → status monitoring → effect verification". It is implemented entirely through software, solving the problems of unpredictable latency and insufficient control of mutations in traditional automated communication transmission, without relying on any hardware modifications or involving any intellectual activity rules.

[0037] The core efficiency-enhancing logic revolves around a pure software implementation that achieves "accurate prediction, flexible control, and stable latency." The pure software-based latency prediction and analysis accurately identifies the direction and priority of latency control for different services and link coordination states, prioritizing high-priority services and improving the targeting and rationality of latency control, avoiding blind control. The pure software integration of the improved GA+ latency control parameter adaptive adjustment combination enables accurate prediction of transmission latency and dynamic adjustment of control parameters. Compared to traditional post-calibration and passive response modes, latency prediction accuracy and latency stability are significantly improved, allowing for rapid adaptation to latency changes, link coordination results, and multi-link transmission variations. High-precision and controllable latency control effectively avoids the impact of latency mutations on communication quality. The pure software mechanism for dynamic monitoring of the latency control process, combined with real-time calculation of the core formula, can promptly detect problems such as excessive prediction deviations or uncontrolled mutations, and quickly adjust the control strategy to ensure stable and accurate latency control. The pure software mechanism for latency control effect verification and iterative optimization can continuously track the control effect, adjust the control strategy and parameters in a timely manner, and ensure that prediction accuracy, latency stability and service adaptability always meet the standards. It adapts to latency change trends, link coordination status and multi-link transmission changes, without relying on hardware modifications or involving intellectual activity rules, ensuring the accuracy and reliability of multi-link data interaction.

[0038] (III) Fault full-chain traceability and self-healing module (which solves the problems of "inaccurate fault traceability and low self-healing efficiency") Modeling Approach: Abandoning the lagging "fuzzy positioning, simple self-healing" model of traditional communication automation, this approach adopts the core logic of "fault data collection → software-based source tracing and positioning → path identification → rapid self-healing → self-healing execution → effect verification." It collects fault trigger data, link coordination status, latency prediction and control status, and data interaction trajectory parameters through a pure software interface, constructing a full-chain fault tracing and self-healing modeling system. This enables precise source tracing of fault triggers, identification of fault propagation paths, dynamic generation of self-healing strategies, and closed-loop execution of the self-healing process, all without human intervention. The innovation lies in the pure software implementation of "fault source tracing and positioning + adaptive fault tracing and self-healing," which differs from traditional fuzzy fault handling approaches and has no overlap with any previous inventions. The entire process does not rely on hardware modifications or involve intellectual activity rules, aligning with the "reliable and self-healing" development needs of strategic emerging industries.

[0039] 5: Fault source tracing and localization (core modeling and solution process, implemented purely in software) The innovative modeling principle focuses on the pure software implementation of "fault data parsing + source tracing and location verification" in the core modeling and solution process. It constructs a fault source tracing and location model, specifying that the input data includes fault triggering data (fault code, fault triggering time, fault impact range), link coordination status (coordination efficiency, resource allocation), latency prediction and control status (prediction accuracy, latency stability), and data interaction trajectory parameters (interaction nodes, interaction time, data loss). Real-time acquisition and fusion of multi-dimensional data are achieved through a pure software interface, requiring no hardware assistance. First, pure software-based fault data preprocessing is used to classify and decompose the acquired multi-dimensional data. Analysis, denoising, and standardization processes eliminate data dimensional differences and invalid components, extracting core influencing factors of faults (such as link coordination imbalance faults, latency mutation faults, and node interaction faults). A correlation mapping is established between fault states, link coordination states, latency prediction and control states, and data interaction trajectories, solving the problems of traditional fault data being messy and unable to achieve accurate source tracing. Furthermore, combined with a purely software-built fault source tracing and positioning system (including fault root cause classification standards, fault propagation path identification standards, and source tracing accuracy standards, which can be dynamically updated via software), an innovative improved graph convolutional neural network (GCN) + fault source classification combination is adopted, implemented through software programming. The current integration and training leverages the graph structure analysis and association mining advantages of GCN to accurately capture the triggering patterns and propagation paths of different types of faults. Combined with the classification advantages of fault tracing, it accurately locates the root causes of faults (such as link coordination imbalance, improper latency control, node overload, and abnormal data interaction), identifies fault propagation paths, quantifies the scope of fault impact, and clarifies the priority of fault tracing (high-impact faults are traced first, and high-priority business-related faults are traced first), avoiding the limitations of traditional fault tracing, such as vagueness and inaccurate location. Finally, the effectiveness of pure software-based fault tracing and location is verified, focusing on two aspects: location accuracy and path identification completeness. The system verifies the location effectiveness. If the fault tracing is inaccurate or the propagation path cannot be fully identified, the system iteratively optimizes the tracing model and parameters, adjusts the feature extraction logic, and updates the fault tracing and location system through software. This outputs accurate fault tracing results, propagation paths, and self-healing suggestions, providing core fault basis for fault tracing and self-healing scheduling. The entire modeling and solving process is implemented entirely through software, forming a complete link of "data acquisition → preprocessing → feature extraction → tracing and location → effect verification". This solves the problems of inaccurate fault tracing and low self-healing efficiency in traditional communication automation, laying a solid foundation for fault tracing and self-healing without relying on any hardware modifications or involving intellectual activity rules.

[0040] The core efficiency-enhancing logic revolves around the pure software implementation of "precise source tracing, clear path identification, and targeted self-healing." The pure software for fault data preprocessing, through classification analysis, noise reduction, and standardization, accurately extracts core fault indicators and establishes a mapping relationship between indicators and link collaboration, latency prediction, and data interaction, providing a reliable foundation for fault source tracing and localization. The pure software integration of the improved GCN+ fault source tracing classification combination and the fault source tracing and localization system enables precise localization of the fault triggering root cause, complete identification of the propagation path, and quantification of the impact range. Compared to traditional fuzzy fault localization modes, the accuracy of source tracing and path identification capabilities are significantly improved, enabling precise capture of minor faults. This system enables proactive fault self-healing by predicting and mitigating fault propagation trends in advance. Its purely software-based mechanism for verifying and iteratively optimizing fault tracing and localization results continuously improves tracing effectiveness, adapting to changes in different fault types, link coordination states, and latency prediction scenarios. This ensures that tracing results accurately reflect the fault condition and propagation patterns, providing precise support for subsequent fault self-healing. The entire process is independent of hardware modifications and does not involve intellectual activity rules, aligning with the development needs of communication automation security and reliability. It significantly improves the accuracy and timeliness of fault tracing and localization, providing strong support for fault self-healing and addressing the problems of inaccurate fault tracing and low self-healing efficiency at the source.

[0041] 6: Adaptive Fault Tracing and Self-Healing (Core Modeling and Solution Process, Purely Software Implementation) Modeling Innovation Principle: The core modeling and solution process focuses on pure software implementation of "source tracing analysis + self-healing model modeling + dynamic self-healing + effect verification." It constructs an adaptive fault tracing and self-healing model, clearly defining the input data as fault source location results, fault propagation paths, fault impact range, link coordination results, latency prediction and control status, and business priorities. Real-time data acquisition and analysis are achieved through a pure software interface. First, through pure software-based fault source tracing analysis, in-depth analysis is performed on the fault source location results, fault propagation paths, fault impact range, link coordination results, latency prediction and control status, and business priorities. This clarifies the self-healing direction, self-healing priority, and self-healing parameter range for different types and impact ranges of faults, establishing a correlation mapping between fault source information, link coordination status, latency prediction, and self-healing strategies. While ensuring communication continuity and data integrity, this approach maximizes the accuracy of fault tracing and the timeliness of self-healing without any hardware assistance. It innovatively employs an improved attention mechanism combined with adaptive adjustment of self-healing parameters. This fusion and training, implemented through software programming, leverages the focus and feature enhancement advantages of the attention mechanism to achieve key control over the root cause and propagation path of faults. Combined with the dynamic adjustment advantage of adaptive self-healing parameters, it adjusts self-healing parameters (self-healing strategy, self-healing speed, and self-healing priority) in real time to adapt to changes in fault type, fault propagation trends, link coordination status, and latency prediction and control requirements. This ensures that the self-healing strategy matches the fault condition and system operating status, achieving rapid fault tracing and closed-loop self-healing. Simultaneously, it integrates core calculation formulas. By using pure software to calculate latency prediction accuracy in real time, the system ensures that the latency prediction accuracy remains ≥0.92 throughout the self-healing process, preventing the self-healing process from affecting the normal operation of the system. Pure software dynamically monitors the self-healing execution process and fault handling status. If a fault fails to heal or spreads, the self-healing parameters and strategies are immediately adjusted to prioritize the security and communication continuity of high-priority services. Finally, pure software-based fault tracing and self-healing effect verification is performed, checking the self-healing effect from four dimensions: tracing accuracy, self-healing timeliness, self-healing success rate, and system stability. If the requirements are not met, the self-healing model and parameters are iteratively optimized, and the self-healing strategies and logic are adjusted through software iteration. The entire modeling and solving process forms a complete closed loop of "tracing analysis → model modeling → dynamic self-healing → status monitoring → effect verification," entirely implemented through software. This solves the problems of inaccurate fault tracing and low self-healing efficiency in traditional communication automation, without relying on any hardware modifications or involving any intellectual activity rules.

[0042] The core efficiency-enhancing logic revolves around a pure software implementation of "accurate source tracing, timely self-healing, and efficient handling." The pure software for fault source analysis can accurately determine the self-healing direction and priority of different faults, achieving efficient source tracing and self-healing while ensuring communication continuity and data integrity. This improves the targeting and rationality of fault self-healing, avoiding blind self-healing. The pure software integration of an improved Attention + self-healing parameter adaptive adjustment combination with the core formula enables rapid fault source tracing, dynamic adjustment of self-healing parameters, and simultaneous assurance of latency prediction accuracy. Compared to traditional fuzzy positioning and simple self-healing modes, the accuracy of fault source tracing and the timeliness of self-healing are significantly improved, effectively reducing the risk of fault escalation. Simultaneously, through core... The system employs a core formula to calculate latency prediction accuracy in real time, ensuring that the self-healing process does not affect normal system operation. A purely software mechanism for dynamic monitoring of the self-healing process and fault handling status can promptly detect potential faults that have not healed or have spread, quickly adjusting the self-healing strategy to ensure rapid closed-loop fault resolution. A purely software mechanism for fault tracing, self-healing effect verification, and iterative optimization can continuously track the self-healing effect, adjusting self-healing strategies and parameters in a timely manner to ensure that tracing accuracy, self-healing timeliness, self-healing success rate, and system stability consistently meet standards. This adapts to changes in fault trends, link coordination status, and latency prediction and control requirements, all without relying on hardware modifications or involving intellectual activity rules, achieving the coordinated development of "safety, stability, and self-healing" in communication automation. The foregoing has shown and described the basic principles, main features, and core advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely the innovative and modeling principles of the present invention. Various changes and modifications can be made to the present invention without departing from its spirit and core scope, and all such changes and modifications fall within the scope of protection claimed by the present invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. An application method for a communication automation link collaborative optimization, delay prediction and control, and fault self-healing intelligent system, characterized in that, Includes the following steps: S1: Full-dimensional status perception of links, latency and faults. It collects link operation parameters, transmission latency data, fault triggering data and link correlation data of multiple links in communication automation through pure software interface, completes data preprocessing and feature extraction, and provides high-quality data support for link collaborative dynamic optimization, transmission latency prediction and control, and fault full-chain tracing and self-healing. S2: Link Coordination Dynamic Optimization. Based on sensing data, it completes the accurate identification of link association relationships, link load coordination analysis, and dynamic generation of link coordination strategies through link association feature extraction and adaptive link coordination scheduling. This enables automated multi-link coordination scheduling and resource optimization configuration, avoiding resource waste and inefficient coordination caused by isolated link scheduling. S3: Transmission delay prediction and control. By combining link coordination results and delay data, it completes accurate prediction of delay change trends, dynamic optimization of delay control parameters, and closed-loop execution of delay control through delay feature prediction and adaptive delay control optimization. This enables predictable and controllable transmission delay and avoids delay abrupt changes affecting communication quality. S4: Full-chain fault tracing and self-healing. Based on link collaboration results, latency prediction and control status and fault triggering data, through fault source tracing and adaptive fault tracing and self-healing, it completes accurate source tracing of fault triggering, identification of fault propagation path, dynamic generation of self-healing strategy, and closed-loop execution of self-healing process, so as to achieve fault traceability and fast self-healing, and avoid fault propagation affecting communication continuity. The formula for calculating the time delay prediction accuracy in step S3 is as follows: The constraints are , , , , To improve the accuracy of time delay prediction, To predict the delay value, This is the actual delay value. This is the delay prediction correction coefficient; the formula enables accurate calculation of delay prediction accuracy, providing core support for link coordination, delay prediction control and fault tracing and self-healing, taking into account link coordination, delay prediction accuracy and fault tracing and self-healing timeliness.

2. The method according to claim 1, characterized in that, The link association feature extraction is used to perform in-depth analysis of multi-link operating parameters, link association strength, link interaction frequency, and link resource occupancy status data through pure software. Combined with a link association feature evaluation system built entirely in software, it accurately identifies core link association features, captures the changing patterns of link association, determines early signs of link coordination imbalance, and eliminates interference from invalid link data. This provides accurate data basis for dynamic optimization of link coordination. The entire process does not rely on hardware link detection modules, ensuring the accuracy and comprehensiveness of link association feature extraction.

3. The method according to claim 1, characterized in that, The adaptive link collaborative scheduling is based on an improved deep reinforcement learning (DRL) + link collaborative allocation combination. It realizes link association analysis, link collaborative scheduling model modeling and dynamic optimization through pure software. It adapts to changes in link association, differences in business priority and link resource status in real time, accurately allocates multi-link collaborative tasks and dynamically adjusts collaborative strategies. It does not rely on hardware link scheduling modules throughout the process. Compared with the traditional isolated link scheduling mode, the link collaborative efficiency and timing response speed are significantly improved.

4. The method according to claim 1, characterized in that, The latency feature prediction method collects historical latency data, latency fluctuation characteristics, link coordination status, and service load change data of multi-link transmission through a pure software interface. Based on the improved gated cyclic unit (GRU) + latency feature prediction combination, it realizes accurate prediction of latency change trends, identification of latency abrupt changes, and quantification of latency prediction intervals. It does not rely on hardware latency prediction modules throughout the process, and provides accurate latency prediction basis for adaptive latency control and optimization.

5. The method according to claim 1, characterized in that, The adaptive latency control optimization is based on an improved genetic (GA) + adaptive adjustment combination of latency control parameters. It realizes latency control model modeling, latency control strategy dynamic generation, and latency control real-time execution through pure software. It adapts to link collaboration results, latency prediction trends and business needs in real time. It does not rely on hardware latency control modules throughout the process, avoiding the problems of unpredictability and poor adaptability of traditional latency control, and improving the accuracy of latency prediction and latency controllability.

6. The method according to claim 1, characterized in that, The fault tracing and localization method uses pure software to fuse and analyze communication fault triggering data, link coordination status, latency prediction and control results, and data interaction trajectory parameters. Based on an improved graph convolutional neural network (GCN) + fault tracing classification combination, it achieves accurate location of fault triggering root cause, identification of fault propagation path, and quantification of fault impact range. The entire process does not rely on hardware fault tracing modules, providing accurate fault tracing basis for fault self-healing scheduling.

7. The method according to claim 1, characterized in that, The adaptive fault tracing and self-healing mechanism is based on an improved attention mechanism and an adaptive adjustment of self-healing parameters. It achieves fault tracing and self-healing model modeling, dynamic generation of self-healing strategies, and closed-loop control of the self-healing process through pure software. It adapts to the fault root cause, link coordination status, and latency prediction and control requirements in real time, without relying on hardware self-healing modules, thus improving the accuracy of fault tracing and the timeliness of self-healing.

8. The method according to claim 1, characterized in that, It also includes full-process collaborative verification and iterative optimization logic. By collecting execution data of each module, link collaboration data, latency prediction and control data, and fault tracing and self-healing data through pure software, it verifies the effectiveness of link collaboration efficiency, latency prediction accuracy, fault tracing accuracy and self-healing timeliness, accurately locates the shortcomings of each link, promotes and continuously optimizes strategies, and ensures the stability and control effect of system operation.

9. The method according to any one of claims 1-8, characterized in that, The entire process is achieved through pure software, enabling link collaboration optimization, latency prediction and control, and fault tracing and self-healing. This significantly improves link collaboration efficiency, latency prediction accuracy, fault tracing accuracy, and self-healing timeliness, ensuring the efficient, intelligent, and reliable operation of the communication automation system.

10. A communication automation link collaborative optimization, delay prediction and control, and fault tracing self-healing intelligent system, characterized in that, It includes a status awareness and data acquisition module, a link collaboration dynamic optimization module, a transmission delay prediction and control module, a fault full-chain traceability and self-healing module, a collaborative verification and optimization module, and a control module. Each module is equipped with the corresponding pure software described in claims 1-8, and collaboratively completes the full-process management and control of communication automation links, delay and fault status awareness, link collaboration dynamic optimization, transmission delay prediction and control, and fault full-chain traceability and self-healing, realizing seamless connection of link collaboration, delay prediction, and fault traceability and self-healing, forming a complete pure software intelligent management and control closed loop.

Images