Multi-channel video stream cooperative transmission method based on dynamic priority

Abstract

The invention relates to the technical field of resource allocation, in particular to a multi-channel video stream cooperative transmission method based on dynamic priority, which comprises the following steps of: acquiring task context characteristic parameters, system resource states, user behavior responses and computing node load data in real time, and constructing a priority allocation model to carry out pattern recognition to generate real-time task priority. And establishing a task scheduling strategy generation model, and performing resource allocation by adopting a priority weighting efficiency evaluation function, a memory-computing unit occupancy rate prediction matrix and a gradient optimization target under an adjustment resource constraint condition to form an initial scheduling strategy. And performing multi-dimensional parameter fusion analysis on the data through an adaptive optimization decision model, generating a resource redistribution correction vector, dynamically adjusting an initial scheduling strategy by adopting an online iterative optimization mechanism, and outputting a final real-time optimization task scheduling strategy. The task priority dynamic evaluation and the closed-loop optimization of the resource allocation are realized, and the efficient cooperative execution of the multi-channel video processing task is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of resource allocation, and in particular to a method for cooperative transmission of multiple video streams based on dynamic priorities. Background Art

[0002] With the rapid development of general computing and digital information processing technology, complex application systems that need to concurrently process multiple real-time, interactive data streams are becoming increasingly popular, such as remote collaboration platforms that require low-latency responses, real-time monitoring and analysis systems, multi-user interactive simulation environments, etc. They usually run on general-purpose digital data processing systems, and their performance is highly dependent on the effective management of internal computing resources.

[0003] Currently, in order to achieve efficient and orderly processing of these heterogeneous data streams, existing digital data processing systems usually adopt a series of computing resource scheduling strategies (e.g., fixed priority or round-robin scheduling).

[0004] However, when dealing with application scenarios with strong dynamics and drastic changes in task load, existing technologies still have significant bottlenecks. First, in the case of dynamic changes or high competition in system computing resources, task scheduling and resource allocation algorithms are often application-independent, and it is difficult to distinguish and adjust the priorities of different processing tasks (and their associated data flows) according to their real-time importance under specific application logic. Secondly, the existing system-level priority management mechanism is usually decoupled from high-level application status information, and lacks a mechanism for real-time perception and integration of this information. This results in the system scheduler being unable to use this valuable contextual information to dynamically and finely adjust the relative priorities of concurrently executed tasks. Furthermore, the efficiency of existing methods in the collaborative execution of concurrent tasks needs to be improved. When the processing system needs to manage and execute a large number of data processing jobs at the same time, the failure to intelligently allocate resources and control execution flow based on the dynamic priority differences between tasks may cause delays in critical tasks and affect the overall processing throughput and response speed of the system.

[0005] Therefore, how to design a dynamic task priority evaluation model that can deeply integrate application layer context information, and based on this model, implement intelligent and collaborative computing resource management and task scheduling control methods at the operating system or system software level to ensure the timely completion of key processing tasks, optimize system resource utilization, and improve overall system performance (such as throughput and responsiveness) has become a technical problem that needs to be urgently solved in the field of digital data processing system design and operating system optimization. Summary of the invention

[0006] The object of the present invention is to provide a multi-channel video stream collaborative transmission method based on dynamic priority. By collecting real-time task context feature parameters, system resource status, user behavior responses, and computing node load data, a priority allocation model is constructed for pattern recognition to generate real-time task priorities. A task scheduling strategy generation model is established, and a priority weighted efficiency evaluation function, a memory-computation unit occupancy prediction matrix, and a resource constraint gradient optimization target are used for resource allocation to form an initial scheduling strategy. Through an adaptive optimization decision model, multi-dimensional parameter fusion analysis is performed on system resources, user behavior, and node load to generate a resource reallocation correction vector, and an online iterative optimization mechanism is used to dynamically adjust the initial strategy, and finally a real-time optimized task scheduling strategy is output. The closed-loop optimization of task priority dynamic evaluation and resource allocation is realized, ensuring the efficient collaborative execution of multi-channel video processing tasks.

[0007] To achieve the above object, the present invention provides the following technical solutions: A multi-channel video stream collaborative transmission method based on dynamic priority, comprising: Real-time collection of task context feature parameters, system resource status parameters, user behavior response data, and computing node load parameters associated with multi-channel video data processing tasks, where the task context feature parameters include user operation focus coordinates, application event trigger sequences, and predefined task rule sets; Construct a priority allocation model to perform pattern recognition on the task context feature parameters to generate real-time priorities for each multi-channel video data processing task; Establish a task scheduling strategy generation model, and perform processing resource allocation on the multi-channel video data processing tasks based on the real-time priorities to generate an initial scheduling strategy; the construction method of the task scheduling strategy generation model includes: generating a priority weighted efficiency evaluation function, predicting a memory-computation unit occupancy matrix, and adjusting the gradient optimization target under resource constraints; Construct an adaptive optimization decision model to perform multi-dimensional parameter fusion analysis on the system resource status parameters, user behavior response data, and computing node load parameters to generate a resource reallocation correction vector; perform online iterative optimization on the initial scheduling strategy according to the resource reallocation correction vector to generate a final real-time optimized task scheduling strategy.

[0008] Preferably, constructing the priority allocation model includes a feature extraction and fusion layer, a pattern recognition layer, and a priority mapping layer; The feature extraction and fusion layer preprocesses the task context feature parameters, extracts spatio-temporal features from the user operation focus coordinates, extracts temporal pattern features from the application event trigger sequence, and extracts semantic features from the predefined task rule set. By fusing the spatio-temporal features, temporal pattern features, and semantic features, a comprehensive task context state feature is generated; the pattern recognition layer classifies and clusters the fused comprehensive task context state features to identify task patterns; the priority mapping layer maps the task patterns to priority scores according to the mapping function to generate the real-time priorities of various video data processing tasks.

[0009] Preferably, the process by which the priority mapping layer generates the real-time priorities of various video data processing tasks includes: Obtaining the task patterns recognized and output by the pattern recognition layer; invoking the mapping function, where the mapping function is the correspondence between task patterns and priority scores; the mapping function is obtained by using a machine learning model to fit the relationship between task patterns and priority scores; inputting the obtained task patterns into the mapping function for calculation, assigning a corresponding priority score to each task pattern, and generating the real-time priorities of various video data processing tasks.

[0010] Preferably, the task scheduling policy generation model includes a resource requirement assessment layer, a resource requirement prediction layer, and an optimization objective adjustment layer; The resource requirement assessment layer obtains the weighted efficacy expected value of each video data processing task according to the real-time priority and task pattern, in combination with the priority weighted efficacy evaluation function; the resource requirement prediction layer uses a scheduling algorithm, based on historical task context feature parameters, and uses the predicted memory-computation unit occupancy rate matrix to predict the predicted resource requirements of each video data processing task in the next period; the predicted resource requirements include memory and computation unit resource amounts; the optimization objective adjustment layer aims to maximize the overall weighted efficacy expected value, while satisfying the adjusted resource constraint conditions, and based on the predicted resource requirements, uses gradient optimization to solve the processing resource allocation scheme for each task, generating the initial scheduling policy, where the initial scheduling policy includes the proportion of computation units assigned to each task, the memory size, and the processing time slice.

[0011] Preferably, the generation process of the priority weighted efficacy evaluation function includes: a basic efficacy metric standard for quantifying the processing efficiency and contribution degree to user experience of a single video data processing task under the same resources; integrating the basic efficacy metric standard and the real-time priority to generate an association model of the real-time priority to the basic efficacy amount; solidifying the association model into a calculation rule, outputting the weighted efficacy expected value, and setting the parameters in the function according to the optimization objective and experimental data; The process of generating the predicted memory-computation unit occupancy rate matrix includes: obtaining historical data of various video data processing tasks during operation, including the amount of memory and computation unit resources actually allocated within a specific time period and the actual peak, average memory occupancy, and computation unit utilization rate generated under the corresponding resource allocation; based on the historical data, applying statistical analysis to identify and establish a quantitative prediction relationship model between the combination of task context feature parameters and future memory and computation unit occupancy rates; analyzing the results of the prediction relationship model to form the predicted memory-computation unit occupancy rate matrix.

[0012] Preferably, the adaptive optimization decision model includes a state evaluation layer and an adjustment strategy generation layer; The state evaluation layer receives the system resource status parameters, user behavior response data, and computing node load parameters in real time; adopts multi-dimensional parameter fusion technology to comprehensively analyze the input real-time data, generate a multi-channel video stream collaborative transmission vector, and identify potential risk points; the adjustment strategy generation layer generates the resource reallocation correction vector based on the potential risk points, and the resource reallocation correction vector includes clear indications of the amount of computing resources, memory quotas, network bandwidth priorities that need to be increased and / or decreased for specific high-priority and / or low-priority video data processing tasks, as well as adjusting global parameters.

[0013] Preferably, the process of generating the resource reallocation correction vector specifically includes: applying the resource reallocation correction vector to the initial scheduling strategy; adjusting the corresponding resource allocation parameters in the initial scheduling strategy according to the adjustment direction and amplitude indicated by each component in the resource reallocation correction vector; during the adjustment process, checking in real time whether the adjusted resource allocation exceeds the total constraint of the existing resources; if it exceeds the constraint, perform rebalancing according to the preset rules; the online iterative optimization is to repeat all steps in each scheduling cycle or a predetermined time interval, and adjust the scheduling strategy of the previous cycle based on the latest resource reallocation correction vector, and output the optimal real-time task scheduling strategy for the current cycle.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. Based on the dynamic priority allocation and collaborative scheduling method, the present invention realizes multi-dimensional information fusion and precise resource allocation by collecting task context features, resource status, and user behavior data in real time. This method constructs a feature extraction and fusion layer, a pattern recognition layer, and a priority mapping layer, and uses machine learning to automatically fit the relationship between task patterns and priority scores, effectively overcoming the defects of rigid resource configuration and long response delay in traditional static scheduling schemes. It greatly improves the video data processing efficiency, reduces the scheduling complexity, realizes precise grading and resource optimization, ensures the high efficiency and reliability of the transmission process, and optimizes the user experience and overall performance.

[0015] 2. The present invention introduces an adaptive optimization decision model to achieve iterative optimization of online resource reallocation and scheduling strategies. This model comprehensively analyzes resource status, user behavior feedback, and computing node load, uses multi-dimensional parameter fusion technology to identify potential risks, and generates a resource reallocation correction vector, thereby realizing dynamic resource adjustment for each video task. Automatically regulating the scheduling strategy according to real-time collected data effectively alleviates the bottleneck problem caused by resource tension, ensures the smoothness and anti-interference ability of the video transmission process, and significantly improves the overall network collaboration performance and operation reliability.

[0016] 3. The present invention comprehensively applies statistical analysis, multi-dimensional parameter fusion, and gradient optimization algorithms to construct a matrix for predicting memory and computing unit occupancy rates, realizing quantitative prediction of task resource requirements. By real-time monitoring historical operation data and based on the optimization objective of the priority-weighted efficiency evaluation function, an initial scheduling strategy is automatically generated and iteratively corrected online to ensure that each task obtains optimal allocation under the specified resource constraints. This not only improves resource utilization efficiency but also can adaptively adjust according to dynamic changes to ensure network load balance and overall efficiency improvement. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic flowchart of a multi-channel video stream collaborative transmission method based on dynamic priority provided by the present invention; Figure 2 It is a schematic structural diagram of a priority allocation model provided by the present invention; Figure 3 It is a schematic structural diagram of a task scheduling strategy generation model provided by the present invention; Figure 4 It is a schematic flowchart of dynamic priority collaborative transmission provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0019] With the rapid development of network communication technology and multimedia processing technology, network-based real-time video interaction systems have been widely applied in many fields such as distance education, video conferencing, telemedicine, and collaborative design. Especially in scenarios such as remote interactive teaching and online classrooms, in order to simulate a real teaching environment and enhance interactivity, it is often necessary to simultaneously collect, transmit, process, and present video streams from multiple sources.

[0020] However, the existing multi-channel video stream transmission methods still have significant technical bottlenecks when dealing with complex interactive scenarios. First, when network bandwidth resources are limited or fluctuating violently, traditional bandwidth allocation and quality adjustment strategies (such as average allocation, priority based on simple rules) are often difficult to effectively distinguish the actual importance of different video streams in a specific application context. Secondly, the priority management mechanism of the existing technology is usually relatively static or overly simplified, lacking in-depth perception of the specific business logic and real-time interactive status of the upper-layer application. Furthermore, the efficiency of the existing methods in multi-stream collaborative transmission needs to be improved. When a large number of video streams need to be transmitted and received simultaneously, the system tends to adjust each stream independently or based on the overall network conditions, and fails to make full use of the priority differences between the streams for collaborative optimization. Therefore, how to design a method that can perceive the application scenario and user interaction intention, dynamically determine the priority of multi-channel video stream transmission, and perform intelligent and collaborative resource allocation and transmission control based on this priority, so as to improve the transmission quality and overall user experience of key video streams under the condition of limited network resources, has become a technical problem that needs to be solved in the current field of multi-channel video stream transmission technology.

[0021] Embodiment 1: See also Figure 1 The present invention provides a method for cooperative transmission of multiple video streams based on dynamic priority, and the technical solution is as follows: Real-time acquisition of task context feature parameters, system resource status parameters, user behavior response data, and computing node load parameters associated with multi-channel video data processing tasks, wherein the task context feature parameters include user operation focus coordinates, application event triggering sequences, and predefined task rule sets; Constructing a priority allocation model to perform pattern recognition on the task context feature parameters and generate real-time priorities of various video data processing tasks; Further, the priority allocation model is constructed including a feature extraction fusion layer, a pattern recognition layer and a priority mapping layer, see Figure 2 ; The feature extraction and fusion layer preprocesses the task context feature parameters, extracts spatiotemporal features from the user operation focus coordinates, extracts timing pattern features from the application event trigger sequence, and extracts semantic features from the predefined task rule set, and generates comprehensive task context state features by fusing the spatiotemporal features, timing pattern features, and semantic features; the pattern recognition layer classifies and clusters the fused comprehensive task context state features to identify task patterns; the priority mapping layer maps the task patterns into priority scores based on a mapping function to generate real-time priorities for each video data processing task.

[0022] In this embodiment, by collecting in real time the task context feature parameters, system resource status parameters, user behavior response data, and computing node load parameters associated with the multi-channel video data processing tasks, and constructing a priority assignment model to perform pattern recognition on the task context feature parameters, the real-time priorities of each channel of video data processing tasks are generated. This method can comprehensively perceive the system operating environment and user requirements, providing a rich data basis for subsequent priority assignment and resource scheduling. This multi-source data collection mechanism can dynamically adapt to changing application scenarios and user behaviors, improving the intelligent level and response ability.

[0023] Further, the process by which the priority mapping layer generates the real-time priorities of each channel of video data processing tasks includes: Obtain the task patterns recognized and output by the pattern recognition layer; call the mapping function, where the mapping function is the correspondence between the task patterns and the priority scores; the mapping function is obtained by using a machine learning model to fit the relationship between the task patterns and the priority scores; input the obtained task patterns into the mapping function for calculation, assign a corresponding priority score to each task pattern, and generate the real-time priorities of each channel of video data processing tasks.

[0024] In this embodiment, the priority mapping layer uses a machine learning model to automatically fit the mapping relationship between the task patterns and the priority scores, realizing the dynamic generation of task priorities. Overcoming the limitations of traditional static priority assignment, it can quickly adjust the priorities according to the changes in real-time task patterns, ensuring that critical tasks are preferentially supported by resources in complex and changing application scenarios. Combining the multi-dimensional data collected in real time and the priority assignment model, this method can accurately identify the importance and urgency of tasks, providing a scientific basis for subsequent resource allocation and scheduling, and significantly improving the response speed and processing efficiency for critical tasks.

[0025] Establish a task scheduling strategy generation model, perform processing resource allocation on the multi-channel video data processing tasks based on the real-time priorities, and generate an initial scheduling strategy; the construction method of the task scheduling strategy generation model includes: generating a priority weighted efficiency evaluation function, predicting the memory-computation unit occupancy rate matrix, and adjusting the gradient optimization objective under resource constraints; Further, the task scheduling strategy generation model includes a resource demand evaluation layer, a resource demand prediction layer, and an optimization objective adjustment layer, refer to Figure 3 ; The resource requirement assessment layer obtains the weighted efficiency expectation value of each video data processing task according to the real-time priority and task mode, in combination with the priority weighted efficiency evaluation function; the resource requirement prediction layer uses a scheduling algorithm, based on the historical task context feature parameters, and uses the predicted memory-computation unit occupancy rate matrix to predict the predicted resource requirements of each video data processing task in the next period; the predicted resource requirements include the memory and computation unit resource amounts; the optimization target adjustment layer aims to maximize the overall weighted efficiency expectation value, while satisfying the adjusted resource constraint conditions, and based on the predicted resource requirements, uses gradient optimization to solve the processing resource allocation scheme for each task, and generates the initial scheduling strategy, where the initial scheduling strategy includes the proportion of computation units allocated to each task, the memory size, and the processing time slice.

[0026] In this embodiment, the task scheduling strategy generation model generates an initial scheduling strategy based on the real-time priority and task mode through the collaborative work of the resource requirement assessment layer, the resource requirement prediction layer, and the optimization target adjustment layer. It comprehensively considers the priority and resource requirements of tasks, uses the gradient optimization algorithm to allocate resources under resource constraint conditions, and realizes the refinement and scientificization of resource allocation. On the basis of real-time data collection and dynamic generation of priorities, the task scheduling strategy generation model further converts the priorities into specific resource allocation schemes, ensuring that high-priority tasks obtain more resource support, thereby improving the collaborative execution efficiency of multi-channel video processing tasks and the overall performance of the system.

[0027] Further, the generation process of the priority weighted efficiency evaluation function includes: a basic efficiency measurement standard for quantifying the processing efficiency and the contribution degree to the user experience of a single-channel video data processing task when obtaining the same resources; integrating the basic efficiency measurement standard and the real-time priority to generate an association model of the real-time priority to the basic efficiency amount; solidifying the association model into a calculation rule, outputting the weighted efficiency expectation value, and setting the parameters in the function according to the optimization target and experimental data; The process of generating the predicted memory-computation unit occupancy rate matrix includes: obtaining the historical data of each video data processing task during operation, including the memory and computation unit resource amounts actually allocated during a specific time period and the actual memory occupancy peak values, mean values, and computation unit usage rates generated under the corresponding resource allocations; based on the historical data, applying statistical analysis to identify and establish a quantitative prediction relationship model between the combination of task context feature parameters and the future memory and computation unit resource occupancy rates; analyzing the results of the prediction relationship model to form the predicted memory-computation unit occupancy rate matrix.

[0028] In this embodiment, by constructing a priority-weighted efficiency evaluation function and a predicted memory-computation unit occupancy rate matrix, a quantitative evaluation of the task processing effect and resource requirements is achieved. The priority-weighted efficiency evaluation function combines the basic efficiency metric and real-time priority to ensure a high degree of matching between resource allocation and task importance; the prediction matrix accurately predicts the future resource occupancy rate through historical data analysis, providing a scientific basis for resource allocation. On the basis of priority assignment and task scheduling strategy generation, an efficiency evaluation and resource requirement prediction mechanism is further introduced to achieve dynamic optimization and forward-looking adjustment of resource allocation, significantly improving resource utilization efficiency and task execution stability.

[0029] Construct an adaptive optimization decision model to perform multi-dimensional parameter fusion analysis on the system resource state parameters, user behavior response data, and computing node load parameters, and generate a resource reallocation correction vector; perform online iterative optimization on the initial scheduling strategy according to the resource reallocation correction vector to generate a final real-time optimized task scheduling strategy.

[0030] Furthermore, the adaptive optimization decision model includes a state evaluation layer and an adjustment strategy generation layer; The state evaluation layer receives the system resource state parameters, user behavior response data, and computing node load parameters in real time; uses multi-dimensional parameter fusion technology to comprehensively analyze the input real-time data, generates a multi-channel video stream collaborative transmission vector, and identifies potential risk points; the adjustment strategy generation layer generates the resource reallocation correction vector based on the potential risk points, and the resource reallocation correction vector includes a clear indication of the amount of computing resources, memory quota, network bandwidth priority, and adjustment of global parameters that need to be increased and / or decreased for specific high-priority and / or low-priority video data processing tasks.

[0031] In this embodiment, through the multi-dimensional parameter fusion analysis of the state evaluation layer and the adjustment strategy generation layer, the adaptive optimization decision model can perceive the changes in the system operation state and user behavior in real time and generate a resource reallocation correction vector. This vector guides the dynamic adjustment of resources for high-priority and low-priority tasks to ensure that critical tasks are supported first when resources are scarce. On the basis of priority assignment, task scheduling, and resource requirement prediction, an adaptive optimization mechanism is further introduced to achieve real-time adjustment and optimization of resource allocation, significantly enhancing the adaptability and stability in complex network environments.

[0032] Further, the process of generating the resource reallocation correction vector specifically includes: applying the resource reallocation correction vector to the initial scheduling policy; adjusting the corresponding resource allocation parameters in the initial scheduling policy according to the adjustment direction and amplitude indicated by each component in the resource reallocation correction vector; during the adjustment process, continuously check whether the adjusted resource allocation exceeds the total amount constraint of the existing resources; if it exceeds the constraint, perform rebalancing according to the preset rules to ensure the feasibility of resource allocation; the online iterative optimization is to repeat all steps in each scheduling cycle or at a predetermined time interval, and adjust the scheduling policy of the previous cycle based on the latest resource reallocation correction vector to achieve continuous optimization of the scheduling policy with the dynamic changes of the system state and user behavior, and output the optimal real-time task scheduling policy for the current cycle.

[0033] In this embodiment, through the online iterative optimization mechanism, the system state and user behavior can be continuously monitored, and the scheduling policy can be dynamically adjusted according to the latest resource reallocation correction vector. It ensures that the resource allocation can quickly adapt to real-time changes and realizes the continuous optimization of the scheduling policy. Based on real-time data collection, dynamic generation of priorities, task scheduling, performance evaluation, and adaptive optimization, closed-loop control is achieved through online iterative optimization, ensuring the stability and efficiency of the collaborative transmission of multiple video streams, and significantly improving the overall performance and user experience.

[0034] The present invention realizes forward-looking resource allocation through the prediction matrix and gradient optimization. Compared with the average allocation mode of round-robin scheduling, it significantly improves the resource utilization efficiency and the ability to guarantee critical tasks. Specifically, refer to Table 1.

[0035] Table 1 Comparison table of resource allocation efficiency Comparison metrics Round-robin scheduling algorithm The solution of the present invention Improvement range / Explanation Resource allocation response time Fixed-cycle round-robin, slow response Online iterative optimization, real-time adjustment of resource allocation strategy Response time shortened by 70% Resource waste rate High (average allocation leads to inefficiency) Fine-grained allocation based on performance evaluation and gradient optimization Waste rate reduced by 55% Guarantee of high-priority tasks Unable to distinguish priorities Weighted performance evaluation function ensures that high-priority tasks obtain more resources Satisfaction rate of critical task resources increased by 85% Embodiment 1 provides a method for collaborative transmission of multiple video streams based on dynamic priorities. Through real-time collection of multi-dimensional data, construction of a priority allocation model, generation of a task scheduling policy, and adaptive optimization decision-making, it realizes the deep integration of task priority evaluation and resource allocation. It can dynamically sense changes in the application scenario and quickly adjust the resource allocation strategy to ensure the efficient collaborative execution of multiple video processing tasks in a complex network environment. The present invention constructs a highly intelligent collaborative transmission framework for multiple video streams by integrating technologies such as task context awareness, dynamic priority evaluation, resource demand prediction, and online iterative optimization. Compared with traditional static scheduling methods, this solution has significantly improved in terms of resource utilization efficiency, task execution efficiency, and system stability. Especially in scenarios where the network bandwidth is limited or the load fluctuates violently, this method can give priority to ensuring the quality of critical video streams, optimize the overall system performance, provide reliable technical support for multi-user interaction scenarios such as remote collaboration and real-time monitoring, and greatly improve the user experience and system operation efficiency.

[0036] Embodiment 2: The method for collaborative transmission of multiple video streams based on dynamic priority provided by the present invention, for the specific flowchart, please refer to Figure 4 , and the specific solution is as follows: Real-time collect task context feature parameters, system resource status parameters, user behavior response data, and computing node load parameters associated with the multi-channel video data processing task. The task context feature parameters include user operation focus coordinates, application event trigger sequences, and predefined task rule sets; The user operation focus coordinates record the operation focus position and timestamp t of the user in real time through an input device (such as a mouse or touch screen), forming a time series data sequence, which reflects the current focus of the user; The application event trigger sequence records events within the application, such as button clicks and video switches, forming an event sequence. Each event includes an event type (such as "play" or "pause") and the occurrence time; The predefined task rule set defines priority rules according to the application scenario, such as "in a video conference, the video stream of the speaker has a higher priority than that of the audience", and is stored in the form of rules, such as {Rule 1: "Speaker first", Rule 2: "High-definition mode first"}.

[0037] The system resource status parameters include CPU usage rate, memory occupancy, and network bandwidth; The CPU usage rate obtains the overall CPU usage rate and the usage rate of each core through the system API; the memory occupancy records the total memory, the used memory, and the available memory; the network bandwidth monitors the upload and download speeds and the bandwidth utilization rate.

[0038] The user behavior response data includes quality feedback and interaction frequency; Quality feedback: The feedback submitted by the user through the interface, such as "video stuttering" or "audio-video out of sync"; the interaction frequency counts the number of user operations within a unit time, such as clicking the mouse 5 times and typing on the keyboard 3 times within 1 minute.

[0039] The computing node load parameters include node CPU load, node memory load, and node network load.

[0040] Construct a priority allocation model to perform pattern recognition on the task context feature parameters and generate the real-time priorities of each channel of video data processing tasks; Furthermore, constructing the priority allocation model includes a feature extraction and fusion layer, a pattern recognition layer, and a priority mapping layer; The feature extraction and fusion layer preprocesses the task context feature parameters, extracts spatio-temporal features from the user operation focus coordinates, extracts temporal pattern features from the application event trigger sequence, and extracts semantic features from the predefined task rule set. By fusing the spatio-temporal features, temporal pattern features, and semantic features, a comprehensive task context state feature is generated; Spatio-temporal features of user operation focus coordinates: Calculate the focus movement speed: , in pixels per second, where is the position of the user's operation focus at the th feature at time , is the horizontal and vertical coordinates under the th feature, with the lower left corner of the screen as the origin, is the focus movement speed under the th feature.

[0041] Calculate the dwell time : If the coordinate change is less than the threshold (e.g., 5 pixels) for k consecutive time points, then the dwell time .

[0042] Temporal pattern features of the application event trigger sequence: Use a sliding window (window size w) to calculate the event frequency : .

[0043] Recognition pattern: If the event sequence conforms to a predefined pattern (e.g., "click 3 times continuously"), it is marked as a specific pattern.

[0044] Semantic features of the predefined task rule set: Convert the rules into feature vectors, for example, using one-hot encoding: {"speaker first": [1, 0], "high-definition mode first": [0, 1]}.

[0045] Feature fusion: Weighted sum the above features to form a comprehensive task context state feature vector F: ; where is the weight of the corresponding vector, and R is the rule feature vector.

[0046] The pattern recognition layer classifies and clusters the fused comprehensive task context state feature F to identify the task pattern; Use the K-means algorithm to divide F into k clusters, and each cluster represents a task pattern M (such as "high-priority interaction mode", "low-priority background mode").​​

[0047] Output the current task mode , indicates "high-priority interaction mode"; The priority mapping layer maps the task mode into a priority score according to the mapping function, and generates the real-time priority of each video data processing task.

[0048] Furthermore, the process of the priority mapping layer generating the real-time priority of each video data processing task includes: Obtain the task mode recognized and output by the pattern recognition layer; call the mapping function, where the mapping function is the correspondence between the task mode and the priority score; the mapping function uses a machine learning model to fit the relationship between the task mode and the priority score (such as a multi-layer perceptron ); input the obtained task mode into the mapping function for calculation, and assign a corresponding priority score to each task mode , generating the real-time priority of each video data processing task.

[0049] Establish a task scheduling policy generation model, allocate processing resources to the multi-channel video data processing tasks based on the real-time priority, and generate an initial scheduling policy; the construction method of the task scheduling policy generation model includes: generating a priority weighted efficiency evaluation function, predicting the memory-computation unit occupancy matrix, and adjusting the gradient optimization objective under resource constraints; Furthermore, the task scheduling policy generation model includes a resource demand evaluation layer, a resource demand prediction layer, and an optimization objective adjustment layer; The resource demand evaluation layer obtains the weighted efficiency expected value of each video data processing task according to the real-time priority and the task mode, in combination with the priority weighted efficiency evaluation function; the resource demand prediction layer uses a scheduling algorithm, based on historical task context feature parameters, and uses the predicted memory-computation unit occupancy matrix to predict the predicted resource demand of each video data processing task in the next period; the predicted resource demand includes memory and computation unit resource amounts; the optimization objective adjustment layer aims to maximize the overall weighted efficiency expected value, while meeting the adjusted resource constraints, and based on the predicted resource demand, uses gradient optimization to solve the processing resource allocation scheme for each task, and generates the initial scheduling policy, where the initial scheduling policy includes the proportion of computation units allocated to each task, the memory size, and the processing time slice.

[0050] Furthermore, the generation process of the priority weighted efficiency evaluation function includes: a basic efficiency metric standard for quantifying the processing efficiency of a single-channel video data processing task under the same resources and the degree of contribution to the user experience : ; where is the throughput of task i, is the latency.

[0051] Based on the above-mentioned basic performance metrics and real-time priorities, an association model between real-time priorities and basic performance is generated; the association model is solidified into calculation rules, and a weighted performance expectation value is output, where is the real-time priority of task i, and the parameters in the function are set according to the optimization goal and experimental data; The process of generating the predicted memory-computation unit occupancy rate matrix includes: obtaining the historical data of various video data processing tasks during operation, including the amount of memory and computation unit resources actually allocated during a specific time period and the actual memory occupancy peaks, means, and computation unit utilization rates under the corresponding resource allocations; based on the historical data, applying statistical analysis (such as linear regression) to identify and establish a quantitative prediction relationship model between the combination of task context feature parameters and future memory and computation unit resource occupancy rates; analyzing the results of the prediction relationship model to form the predicted memory-computation unit occupancy rate matrix; where the memory occupancy rate and the computation unit occupancy rate can be predicted by linear regression.

[0052] The optimization target adjustment layer aims to maximize the overall weighted performance expectation value, while satisfying the adjusted resource constraint conditions, and based on the predicted resource requirements, uses gradient optimization to solve the processing resource allocation scheme for each task, generating the initial scheduling strategy, where the initial scheduling strategy includes the proportion of computation units, memory size, and processing time slices allocated to each task; Maximize the overall weighted performance expectation value , while satisfying the resource constraint .

[0053] The resources allocated to task i, is the total resource.

[0054] Solve using gradient descent: , constraint .

[0055] Table 2 Comparison table of system resource utilization Resource type Static scheduling utilization rate Utilization rate of the solution of the present invention Utilization rate improvement CPU 79% 88% 9% Memory 70% 82% 12% Network bandwidth 50% 68% 18% The present invention maximizes resource utilization through multi-dimensional parameter fusion and gradient optimization goals, and is particularly outstanding in scenarios with limited network bandwidth. The data is specifically referred to Table 2.

[0056] Construct an adaptive optimization decision model to perform multi-dimensional parameter fusion analysis on the system resource status parameters, user behavior response data, and computing node load parameters, and generate a resource reallocation correction vector; according to the resource reallocation correction vector, perform online iterative optimization on the initial scheduling strategy to generate a final real-time optimized task scheduling strategy.

[0057] Further, the adaptive optimization decision model includes a state evaluation layer and an adjustment strategy generation layer; The state evaluation layer receives the system resource status parameters, user behavior response data, and computing node load parameters in real time; adopts multi-dimensional parameter fusion technology to comprehensively analyze the input real-time data, generates a multi-channel video stream collaborative transmission vector, and identifies potential risk points; the adjustment strategy generation layer generates the resource reallocation correction vector based on the potential risk points The resource reallocation correction vector includes clearly indicating the amount of computing resources, memory quota, network bandwidth priority, and adjusting global parameters that need to be increased and / or decreased for specific high-priority and / or low-priority video data processing tasks.

[0058] Resource reallocation correction vector The calculation formula is: ; Where is the current memory occupancy rate, is the current computing unit occupancy rate, and are adjustment coefficients.

[0059] Further, the process of generating the resource reallocation correction vector specifically includes: applying the resource reallocation correction vector to the initial scheduling strategy; adjusting the corresponding resource allocation parameters in the initial scheduling strategy according to the adjustment direction and amplitude indicated by each component in the resource reallocation correction vector; during the adjustment process, check in real time whether the adjusted resource allocation exceeds the total constraint of the existing resources; if it exceeds the constraint, perform rebalancing according to the preset rules to ensure the feasibility of the resource allocation; the online iterative optimization is to repeat all steps in each scheduling cycle or a predetermined time interval, and adjust the scheduling strategy of the previous cycle based on the latest resource reallocation correction vector to achieve continuous optimization of the scheduling strategy with the dynamic changes of the system state and user behavior, and output the optimal real-time task scheduling strategy of the current cycle.

[0060] The present invention significantly improves users' satisfaction with video smoothness, response speed, and multi-task collaboration through a dynamic priority and adaptive optimization mechanism.

[0061] Although embodiments of the present invention have been shown and described, those of ordinary skill in the art will appreciate that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for collaborative transmission of multiple video streams based on dynamic priorities, characterized in that, Including: Real-time collect task context feature parameters, system resource status parameters, user behavior response data, and computing node load parameters associated with multi-channel video data processing tasks. The task context feature parameters include user operation focus coordinates, application event trigger sequences, and predefined task rule sets; Construct a priority allocation model to perform pattern recognition on the task context feature parameters and generate real-time priorities for each multi-channel video data processing task; Establish a task scheduling policy generation model, and based on the real-time priorities, allocate processing resources to the multi-channel video data processing tasks to generate an initial scheduling policy; The construction method of the task scheduling policy generation model includes: generating a priority weighted efficiency evaluation function, predicting a memory-computation unit occupancy rate matrix, and adjusting the gradient optimization objective under resource constraints; Construct an adaptive optimization decision model to perform multi-dimensional parameter fusion analysis on the system resource status parameters, user behavior response data, and computing node load parameters, and generate a resource reallocation correction vector; according to the resource reallocation correction vector, perform online iterative optimization on the initial scheduling policy to generate a final real-time optimized task scheduling policy.

2. The multi-channel video stream cooperative transmission method based on dynamic priority according to claim 1, wherein: Constructing the priority allocation model includes a feature extraction and fusion layer, a pattern recognition layer, and a priority mapping layer; The feature extraction and fusion layer preprocesses the task context feature parameters, extracts spatio-temporal features from the user operation focus coordinates, extracts temporal pattern features from the application event trigger sequences, and extracts semantic features from the predefined task rule sets. By fusing the spatio-temporal features, temporal pattern features, and semantic features, a comprehensive task context state feature is generated; the pattern recognition layer classifies and clusters the fused comprehensive task context state features to identify task patterns; the priority mapping layer maps the task patterns to priority scores according to a mapping function to generate real-time priorities for each multi-channel video data processing task.

3. The multi-channel video stream cooperative transmission method based on dynamic priority according to claim 2, wherein: The process of the priority mapping layer generating real-time priorities for each multi-channel video data processing task includes: Obtain the task patterns recognized and output by the pattern recognition layer; call the mapping function, where the mapping function is the correspondence between task patterns and priority scores; the mapping function is obtained by using a machine learning model to fit the relationship between task patterns and priority scores; input the obtained task patterns into the mapping function for calculation, and assign a corresponding priority score to each task pattern to generate real-time priorities for each multi-channel video data processing task.

4. The multi-channel video stream cooperative transmission method based on dynamic priority according to claim 1, wherein: The task scheduling policy generation model includes a resource requirement evaluation layer, a resource requirement prediction layer, and an optimization objective adjustment layer; The resource requirement assessment layer obtains the weighted efficiency expectation value of each video data processing task according to the real-time priority and task mode, in combination with the priority weighted efficiency evaluation function; the resource requirement prediction layer uses a scheduling algorithm, based on the historical task context feature parameters, and uses the predicted memory-computation unit occupancy rate matrix to predict the predicted resource requirements of each video data processing task in the next period; the predicted resource requirements include memory and computation unit resource amounts; the optimization target adjustment layer aims to maximize the overall weighted efficiency expectation value, while satisfying the adjusted resource constraint conditions, and based on the predicted resource requirements, uses gradient optimization to solve the processing resource allocation scheme for each task, and generates the initial scheduling strategy, where the initial scheduling strategy includes the proportion of computation units, memory size, and processing time slices allocated to each task.

5. The method for collaborative transmission of multiple video streams based on dynamic priority according to claim 4, characterized in that: The generation process of the priority weighted efficiency evaluation function includes: a basic efficiency measurement standard for quantifying the processing efficiency and contribution degree to user experience of a single video data processing task under the same resources; integrating the basic efficiency measurement standard and the real-time priority to generate an association model of the real-time priority to the basic efficiency amount; solidifying the association model into a calculation rule, outputting the weighted efficiency expectation value, and setting the parameters in the function according to the optimization target and experimental data; The process of generating the predicted memory-computation unit occupancy rate matrix includes: obtaining the historical data of each video data processing task during operation, including the memory and computation unit resource amounts actually allocated within a specific time period and the actual memory occupancy peak values, mean values, and computation unit usage rates under the corresponding resource allocations; based on the historical data, applying statistical analysis to identify and establish a quantitative prediction relationship model between the combination of task context feature parameters and future memory and computation unit resource occupancy rates; analyzing the results of the prediction relationship model to form the predicted memory-computation unit occupancy rate matrix.

6. The method for collaborative transmission of multiple video streams based on dynamic priority according to claim 1, characterized in that: The adaptive optimization decision model includes a state evaluation layer and an adjustment strategy generation layer; The state evaluation layer receives the system resource state parameters, user behavior response data, and computing node load parameters in real time; Adopting a multi-dimensional parameter fusion technology, comprehensively analyzing the input real-time data to generate a collaborative transmission vector of multiple video streams and identifying potential risk points; The adjustment strategy generation layer generates the resource reallocation correction vector according to the potential risk points, where the resource reallocation correction vector includes clearly indicating the amount of computing resources, memory quotas, network bandwidth priorities that need to be increased and / or decreased for specific high-priority and / or low-priority video data processing tasks, and adjusting global parameters.

7. The method for collaborative transmission of multiple video streams based on dynamic priority according to claim 1, characterized in that: The process of generating the resource reallocation correction vector specifically includes: applying the resource reallocation correction vector to the initial scheduling policy; adjusting the corresponding resource allocation parameters in the initial scheduling policy according to the adjustment direction and amplitude indicated by each component in the resource reallocation correction vector; during the adjustment process, checking in real time whether the adjusted resource allocation exceeds the total amount constraint of the existing resources; if it exceeds the constraint, perform rebalancing according to the preset rules; the online iterative optimization is to repeat all steps in each scheduling cycle or at a predetermined time interval, and adjust the scheduling policy of the previous cycle based on the latest resource reallocation correction vector, and output the optimal real-time task scheduling policy for the current cycle.

Images