Dynamic networking load balancing method and system for multi-node intelligent converged terminals

Abstract

This application provides a dynamic network load balancing method and system for multi-node intelligent fusion terminals, relating to the field of load balancing technology. The method includes: extracting a first load balancing decision link, a second load balancing decision link, and a feedback optimization link from the dynamic network load balancing decision network; obtaining a multi-source decision operation dataset based on the multi-node intelligent fusion terminal and calculating a first time complexity; obtaining a feedback optimization decision operation dataset corresponding to the multi-source decision operation dataset and calculating a second time complexity; inputting the first and second time complexities into a network parameter optimizer, and optimizing the parameters of the dynamic network load balancing decision network based on the output first or second network optimization parameters. This application enables dynamic evaluation and optimization of multi-node load balancing complexity, improving scheduling response, resource utilization, and network operational stability.

Description

Technical Field

[0001] This application relates to the field of load balancing technology, and in particular to a dynamic networking load balancing method and system for multi-node intelligent fusion terminals. Background Technology

[0002] With the rapid development of multi-node intelligent terminals and dynamic networking technologies, the demand for task scheduling and resource allocation in communication networks is increasing. The increase in the number of nodes and the complexity of service types have led to highly dynamic changes in network operation. Factors such as node resource utilization, task load distribution, and network link status have a significant impact on overall network performance, requiring the system to complete complex scheduling calculations and load balancing analyses in a short period of time.

[0003] Currently, existing load balancing methods mainly rely on fixed scheduling strategies or static optimization algorithms, lacking the ability to perform refined and dynamic analysis and processing of the real-time operating status and load distribution of each node in the network. When there are a large number of nodes or complex task distribution, existing methods are prone to computational latency and uneven resource allocation. At the same time, they cannot effectively predict the impact of load changes on the overall network performance, resulting in some nodes being overloaded or idle, thus affecting the overall service quality of the network.

[0004] In summary, existing technologies suffer from a lack of refined analysis and dynamic evaluation of the computational complexity and real-time scheduling characteristics of load balancing in multi-node, dynamically changing network environments. This leads to uneven resource allocation and task scheduling delays under high load conditions, further impacting network operating efficiency, service continuity, and overall system performance. Summary of the Invention

[0005] The purpose of this application is to provide a dynamic networking load balancing method and system for multi-node intelligent fusion terminals, in order to solve the technical problems in the prior art where the lack of refined analysis and dynamic evaluation of the computational complexity and real-time scheduling characteristics of load balancing in multi-node, dynamically changing network environments leads to uneven resource allocation and task scheduling delays under high load conditions, which further affects the network's operating efficiency, service continuity and overall system performance.

[0006] In view of the above problems, this application provides a dynamic networking load balancing method and system for multi-node intelligent fusion terminals.

[0007] In a first aspect, this application provides a dynamic network load balancing method for multi-node intelligent fusion terminals, implemented through a dynamic network load balancing system for multi-node intelligent fusion terminals. The method includes: extracting a first load balancing decision link and a second load balancing decision link from the dynamic network load balancing decision network, wherein the second load balancing decision link is connected to a feedback optimization link of the first load balancing decision link; obtaining a multi-source decision operation dataset based on the first load balancing decision link and calculating a first time complexity corresponding to the multi-source decision operation dataset; obtaining a feedback optimization decision operation dataset based on the second load balancing decision link and calculating a second time complexity corresponding to the feedback optimization decision operation dataset; inputting the first time complexity and the second time complexity into a network parameter optimizer, and optimizing the parameters of the dynamic network load balancing decision network according to the output first or second network optimization parameters.

[0008] Preferably, the dynamic networking load balancing method of the multi-node intelligent fusion terminal further includes: obtaining the previous multi-source decision operation dataset based on the first load balancing decision link, calculating the third time complexity corresponding to the previous multi-source decision operation dataset; and inputting the first time complexity, the second time complexity, and the third time complexity into the network parameter optimizer to re-output the first network optimization parameter or the second network optimization parameter.

[0009] Preferably, the dynamic network load balancing method of the multi-node intelligent fusion terminal further includes: obtaining the dataset size parameter of the multi-source decision-making operation dataset, wherein the multi-source decision-making operation dataset includes the resource operation status, task allocation information, network topology status information, and real-time load distribution of each node; processing the decision algorithm type of the multi-source decision-making operation dataset according to the first load balancing decision link; and calculating the time complexity of the dataset size parameter and the decision algorithm type using a time complexity calculation function to obtain a first time complexity.

[0010] Preferably, the dynamic network load balancing method for the multi-node intelligent fusion terminal further includes: obtaining the number of k key execution steps of the decision algorithm type, wherein the key execution steps include at least node comparison steps, task and node matching calculations, queue operation processing, and iteration cost calculation steps; extracting the number of participating nodes, the number of tasks to be assigned, the feature vector dimension, and the number of constraints of the dataset size parameter; and calculating the functional relationship between the number of the k key execution steps and the corresponding dataset size parameter based on the time complexity calculation function to obtain the first time complexity.

[0011] Preferably, the dynamic networking load balancing method of the multi-node intelligent fusion terminal further includes: mapping the number of times of the k key execution steps to function expressions with the number of participating nodes, the number of tasks to be assigned, the dimension of feature vectors, and the number of constraints as independent variables; summing the k function expressions corresponding to the number of times of the k key execution steps to obtain a comprehensive function expression of the total number of times; and performing a time complexity function O(f(n)) transformation on the comprehensive function expression to obtain a first time complexity.

[0012] Preferably, the dynamic networking load balancing method of the multi-node intelligent fusion terminal further includes: calculating the second time complexity corresponding to the feedback optimization decision running dataset using the time complexity calculation function, and calculating the third time complexity corresponding to the previous multi-source decision running dataset using the time complexity calculation function; wherein, the feedback optimization decision running dataset includes a set of feedback correction nodes participating in the feedback optimization decision running, a set of feedback correction tasks, feedback path status information, and feedback optimization constraint information.

[0013] Preferably, the dynamic networking load balancing method for the multi-node intelligent fusion terminal further includes: extracting the dataset size parameter of the feedback optimization decision running dataset; obtaining the feedback optimization algorithm type for the second load balancing decision link processing the feedback optimization decision running dataset; obtaining the number of m key feedback execution steps corresponding to the feedback optimization algorithm type, wherein the key feedback execution steps include at least a feedback path backtracking step, a local redistribution calculation step, a feedback constraint verification step, and a feedback iterative update step; and calculating the functional relationship between the number of the m key feedback execution steps and the corresponding dataset size parameter based on the time complexity calculation function to obtain the first time complexity.

[0014] Preferably, the dynamic network load balancing method of the multi-node intelligent fusion terminal further includes: if the first time complexity is greater than the second time complexity, performing network parameter optimization with the first load balancing decision link as the input parameter variable and the time complexity difference between the first time complexity and the second time complexity as the optimization target, and outputting a first network optimization parameter; if the first time complexity is less than the second time complexity, performing network parameter optimization with the second load balancing decision link as the input parameter variable and the time complexity difference between the first time complexity and the second time complexity as the optimization target, and outputting a second network optimization parameter; if the first time complexity is equal to the second time complexity, not activating the network parameter optimizer.

[0015] Preferably, the dynamic network load balancing method of the multi-node intelligent fusion terminal further includes: performing network parameter optimization with the time complexity difference between the first time complexity and the second time complexity as the optimization target; obtaining the time complexity difference between the load balancing decision link corresponding to the current candidate parameter variable and another load balancing decision link; updating the current candidate parameter variable through gradient until the absolute value of the time complexity difference is less than a preset tolerance threshold, and outputting the current parameter variable as a network optimization parameter.

[0016] Secondly, this application also provides a dynamic network load balancing system for multi-node intelligent fusion terminals, used to execute the dynamic network load balancing method for multi-node intelligent fusion terminals as described in the first aspect, comprising: a link extraction module, used to extract a first load balancing decision link and a second load balancing decision link of the dynamic network load balancing decision network, wherein the second load balancing decision link is connected to a feedback optimization link of the first load balancing decision link; a first time complexity calculation module, used to obtain a multi-source decision operation dataset based on the first load balancing decision link and calculate a first time complexity corresponding to the multi-source decision operation dataset; a second time complexity calculation module, used to obtain a feedback optimization decision operation dataset corresponding to the multi-source decision operation dataset based on the second load balancing decision link and calculate a second time complexity corresponding to the feedback optimization decision operation dataset; and a parameter optimization module, used to input the first time complexity and the second time complexity into a network parameter optimizer and perform parameter optimization on the dynamic network load balancing decision network according to the output first network optimization parameter or second network optimization parameter.

[0017] The technical solution provided in this application has at least the following technical effects or advantages: by achieving the technical goal of dynamically evaluating and optimizing the computational complexity in the multi-node load balancing decision-making process, it achieves the technical effects of improving scheduling response speed, optimizing resource utilization and enhancing the overall stability of network operation while ensuring balanced network task allocation.

[0018] The above description is merely an overview of the technical solution of this application. To enable a clearer understanding of the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the dynamic networking load balancing method for multi-node intelligent fusion terminals in this application.

[0021] Figure 2 This is a schematic diagram of the dynamic networking load balancing system of the multi-node intelligent fusion terminal of this application.

[0022] Figure labeling: Link extraction module 1, first time complexity calculation module 2, second time complexity calculation module 3, parameter optimization module 4. Detailed Implementation

[0023] This application provides a dynamic network load balancing method and system for multi-node intelligent fusion terminals. It addresses the technical problem in existing technologies where the lack of refined analysis and dynamic evaluation of the computational complexity and real-time scheduling characteristics of load balancing in multi-node, dynamically changing network environments leads to uneven resource allocation and task scheduling delays under high load conditions, further impacting network operating efficiency, service continuity, and overall system performance. The application achieves the technical goal of dynamically evaluating and optimizing the computational complexity in the multi-node load balancing decision-making process, thereby improving scheduling response speed, optimizing resource utilization, and enhancing overall network operational stability while ensuring balanced network task allocation.

[0024] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0025] Example 1, please refer to the appendix. Figure 1 This application provides a dynamic networking load balancing method for multi-node intelligent converged terminals, which is applied to a dynamic networking load balancing system for multi-node intelligent converged terminals. The method specifically includes the following steps:

[0026] Extract the first load balancing decision link and the second load balancing decision link of the dynamic network load balancing decision network, and the feedback optimization link that connects the second load balancing decision link to the first load balancing decision link.

[0027] Specifically, extracting the first and second load balancing decision links of the dynamic network load balancing decision network represents identifying and dividing decision-making paths with different functions within the dynamic network load balancing decision network. The dynamic network load balancing decision network represents the decision-making structure used to perform node connection management, task allocation control, and network resource scheduling. It constitutes a dynamically changing load balancing processing system through multiple decision nodes, link connections, and data interaction mechanisms within the network. Extraction means identifying and separating different processing flows within the dynamic network load balancing decision network based on preset link partitioning rules, data flow relationships, or functional execution logic. The first load balancing decision link represents the processing path used to execute the initial load balancing decision, mainly undertaking node scheduling, resource allocation, and task load distribution functions. The second load balancing decision link represents the processing path used to perform decision result feedback analysis and optimization correction, mainly undertaking load balancing result verification, impact assessment, and optimization adjustment functions. The first and second load balancing decision links correspond to different stages of the decision-making process, thus forming a multi-stage link processing structure in the dynamic network load balancing process. The feedback optimization link, connecting the first and second load balancing decision links, represents an established processing path for transmitting decision results and feedback information between them. This feedback optimization link serves as a data interaction channel for decision result feedback, optimization analysis, and parameter correction. By establishing a data mapping relationship between the links, the second load balancing decision link can perform subsequent optimization processing based on the decision results output by the first load balancing decision link. The connection indicates that the decision information output by the first load balancing decision link is transmitted to the second load balancing decision link according to preset data flow rules, thus achieving continuous processing between different decision stages. Feedback optimization involves performing impact analysis, status verification, and optimization correction processing on the initial load balancing decision results to improve the stability and adaptability of load balancing decision results in dynamic networking environments. By establishing the association between the first and second load balancing decision links through the feedback optimization link, collaborative processing and closed-loop optimization control of links can be achieved during the dynamic network load balancing decision process.

[0028] Based on the first load balancing decision link, a multi-source decision operation dataset based on a multi-node intelligent fusion terminal is obtained, and the first time complexity corresponding to the multi-source decision operation dataset is calculated.

[0029] Furthermore, this application also includes: obtaining the dataset size parameter of the multi-source decision-making running dataset, wherein the multi-source decision-making running dataset includes the resource running status, task allocation information, network topology status information, and real-time load distribution of each node; processing the decision algorithm type of the multi-source decision-making running dataset according to the first load balancing decision link; and calculating the time complexity of the dataset size parameter and the decision algorithm type using a time complexity calculation function to obtain a first time complexity.

[0030] Furthermore, this application also includes: obtaining the number of k key execution steps of the decision algorithm type, wherein the key execution steps include at least node comparison steps, task and node matching calculations, queue operation processing, and iteration cost calculation steps; extracting the number of participating nodes, the number of tasks to be assigned, the feature vector dimension, and the number of constraints of the dataset size parameter; and calculating the functional relationship between the number of the k key execution steps and the corresponding dataset size parameter based on the time complexity calculation function to obtain the first time complexity.

[0031] Furthermore, this application also includes: mapping the number of the k key execution steps to function expressions with the number of participating nodes, the number of tasks to be assigned, the dimension of feature vectors, and the number of constraints as independent variables; summing the k function expressions corresponding to the number of the k key execution steps to obtain a comprehensive function expression of the total number of execution steps; and performing a time complexity function O(f(n)) transformation on the comprehensive function expression to obtain a first time complexity.

[0032] Specifically, the acquisition of a multi-source decision-making operation dataset based on a multi-node intelligent fusion terminal, according to the first load balancing decision-making link, means collecting and summarizing multi-dimensional operational information generated by the multi-node intelligent fusion terminal during dynamic network operation, following the load balancing decision-making process corresponding to the first load balancing decision-making link. The first load balancing decision-making link represents the decision-making path used to perform initial load balancing analysis and task scheduling. The multi-node intelligent fusion terminal represents multiple network node devices with data processing, communication interaction, and task collaboration capabilities. The multi-source decision-making operation dataset represents a comprehensive set of operational data from different nodes, different operating states, and different network layers, used to characterize the operating state and load changes of each node in a dynamic network environment.

[0033] Furthermore, the dataset size parameter of the multi-source decision-making operation dataset is obtained, representing a statistical analysis of the data size characteristics corresponding to the multi-source decision-making operation dataset. The dataset size parameter represents the data parameters used to characterize the overall computational scale of the multi-source decision-making operation dataset, including the number of nodes participating in data processing, the number of tasks, the number of feature dimensions, and the number of constraints, which are used to reflect the data processing scale and computational complexity corresponding to the first load balancing decision link in the process of executing load balancing decisions.

[0034] The multi-source decision-making runtime dataset includes the resource operation status, task allocation information, network topology status information, and real-time load distribution for each node, indicating that the dataset is composed of different types of runtime status data. Resource operation status represents the resource utilization of a node at its current operational stage, including CPU utilization, memory utilization, network bandwidth usage, and storage I / O status, reflecting the node's resource consumption level. Task allocation information represents the resource requirements and task attributes of the tasks to be scheduled, including task resource requirements, task priority, and task type, characterizing the task scheduling requirements. Network topology status information represents the connection relationships and link status information between nodes in a dynamic networking environment, including link connection status, link transmission delay, and link bandwidth status, characterizing the network communication structure. Real-time load distribution represents the current task load status of different nodes, including the number of tasks per node or the node load index, reflecting the load differences between nodes in a dynamic networking environment.

[0035] Furthermore, based on the decision algorithm type used by the first load balancing decision link to process the multi-source decision-making dataset, the algorithm processing category adopted by the first load balancing decision link in executing load balancing decisions is determined. The decision algorithm type represents the type of algorithm model used to perform node resource allocation, task scheduling, and load balancing analysis, including scheduling algorithms based on sorting rules, optimization algorithms based on search mechanisms, or dynamic scheduling algorithms based on iterative analysis. It is used to characterize the corresponding operation rules and processing methods of the first load balancing decision link when processing the multi-source decision-making dataset.

[0036] Furthermore, the number of execution steps for the k key execution steps of the decision algorithm type is obtained, representing a step-by-step statistical analysis of the core computational processes of the decision algorithm in the load balancing decision-making process. The decision algorithm type represents the algorithmic processing method adopted by the first load balancing decision link in the process of resource scheduling and task allocation. The number of execution steps for the k key execution steps represents the number of executions corresponding to different core computational steps during the operation of the decision algorithm, used to characterize the resource consumption and computational scale changes of different computational stages during algorithm processing. Key execution steps represent the core computational processes that have a major impact on the overall time complexity. By statistically analyzing the number of executions corresponding to the key execution steps, the computational growth trend of the decision algorithm under different data scale conditions can be reflected.

[0037] The node comparison step represents the process by which the decision-making algorithm sorts and compares the running status, resource occupancy status, or load status of multiple nodes. The number of executions of the node comparison step characterizes the computational scale in the node selection and sorting process. The number of task-node matching calculations represents the number of times the resource matching relationship between the task to be assigned and the candidate nodes is calculated and analyzed, reflecting the complexity of resource matching in the task scheduling process. The queue operation processing represents the process of inserting, deleting, updating, and sorting tasks, nodes, or scheduling cache queues. The number of executions of queue operation processing reflects the data management overhead in the scheduling process. The iterative cost calculation step represents the process by which the decision-making algorithm performs repetitive calculations during multiple rounds of scheduling analysis or dynamic optimization. The number of executions of the iterative cost calculation step reflects the computational growth during the algorithm's iterative analysis phase.

[0038] Furthermore, the dataset size parameters—number of participating nodes, number of tasks to be assigned, feature vector dimension, and number of constraints—are extracted to represent the data size variables used for time complexity analysis. The number of participating nodes represents the total number of nodes involved in the dynamic network load balancing decision-making process, characterizing the node scheduling scope and resource scheduling scale. The number of tasks to be assigned represents the total number of tasks requiring resource scheduling and task allocation processing, characterizing the task processing scale. The feature vector dimension represents the number of feature parameters used to describe node status, task status, or network status, characterizing the feature computation scale during the decision analysis process. The number of constraints represents the number of scheduling restrictions that need to be met during the load balancing decision-making process, including resource restrictions, network transmission restrictions, or task priority restrictions, characterizing the constraint analysis complexity during the execution of the decision algorithm.

[0039] Furthermore, the number of execution steps for each of the k key execution steps is mapped to a functional expression with the number of participating nodes, the number of tasks to be assigned, the feature vector dimension, and the number of constraints as independent variables. This represents a functional description of the quantitative relationship between the number of operations for different key execution steps and the data scale variable. The number of execution steps for each of the k key execution steps represents the number of executions of the decision algorithm at different core operation stages, including the statistical results of the number of operations for node comparison, task matching, queue processing, and iterative calculation. The number of participating nodes represents the scale of nodes participating in the dynamic network load balancing decision-making process; the number of tasks to be assigned represents the scale of tasks requiring resource scheduling and task allocation; the feature vector dimension represents the scale of feature parameters used to describe node and task states; and the number of constraints represents the scale of constraints that need to be satisfied during the load balancing decision-making process. The independent variables represent the data scale variable that affects the result of the function change, and the functional expression represents the mathematical relationship between the number of execution steps for key execution steps and the data scale, used to describe the growth trend of the corresponding operation steps under different data scale conditions.

[0040] Furthermore, the k function expressions corresponding to the number of times each of the k key execution steps is summed to obtain a comprehensive function expression for the total number of operations. This comprehensive function expression represents the unified accumulation of the functional relationships corresponding to multiple key execution steps, thus yielding the overall result of the total number of operations for the decision-making algorithm. The k function expressions represent the growth functions of operations corresponding to different key execution steps. By summing these multiple function expressions, the comprehensive computational scale of the decision-making algorithm during the complete execution process can be obtained. The comprehensive function expression for the total number of operations represents the overall functional relationship formed by the number of operations of multiple key execution steps, and is used to characterize the growth trend of the total computational scale of the decision-making algorithm under different data scale conditions.

[0041] Furthermore, the time complexity of the synthesized function expression is transformed using the O(f(n)) function to obtain the first time complexity, representing an asymptotic complexity analysis of the synthesized function expression. The O(f(n)) time complexity transformation converts the synthesized function expression into a Big O complexity form used to describe the growth trend of the algorithm. The dominant scale parameter represents the scale variable with the greatest impact on the overall computational scale among multiple scale parameters, and variable n represents a single variable used to uniformly describe the trend of data scale changes. Representing the other scale parameters besides the dominant scale parameter as functional relationships with the dominant scale parameter and substituting them into the synthesized function expression represents a unified variableization of the multivariate functional relationships, thus obtaining the total number of operations function T(n) with variable n as the single independent variable. The total number of operations function T(n) represents the unified functional relationship between the overall number of operations of the decision algorithm and the change in data scale.

[0042] Furthermore, identifying the term in the total number of operations function T(n) that grows fastest as the variable n increases infinitely and retaining it as the highest-order term represents selecting the computational term with the greatest impact on the overall growth trend from the total number of operations function T(n). The highest-order term represents the function term with the fastest growth rate under continuously expanding data scale, used to characterize the main source of growth in algorithm complexity. Ignoring the constant coefficient of the highest-order term and deleting all other terms in the total number of operations function T(n) except for the highest-order term represents simplifying the non-dominant growth part, thus retaining the core function part that determines the growth trend of algorithm complexity. The Big O time complexity function O(f(n)) represents the complexity function form used to describe the asymptotic growth trend of the algorithm, used to characterize the computational complexity of the first load balancing decision link under the condition of continuously increasing data scale. The first time complexity represents the final complexity analysis result corresponding to the first load balancing decision link, used to reflect the overall computational growth trend and resource consumption scale in the load balancing decision process.

[0043] Based on the second load balancing decision link, obtain the feedback optimization decision running dataset corresponding to the multi-source decision running dataset, and calculate the second time complexity corresponding to the feedback optimization decision running dataset.

[0044] Furthermore, this application also includes: calculating the second time complexity corresponding to the feedback optimization decision running dataset using the time complexity calculation function, and calculating the third time complexity corresponding to the previous multi-source decision running dataset using the time complexity calculation function; wherein, the feedback optimization decision running dataset includes a set of feedback correction nodes participating in the feedback optimization decision running, a set of feedback correction tasks, feedback path status information, and feedback optimization constraint information.

[0045] Furthermore, this application also includes: extracting the dataset size parameter of the feedback optimization decision running dataset; obtaining the feedback optimization algorithm type of the second load balancing decision link processing the feedback optimization decision running dataset; obtaining the number of m key feedback execution steps corresponding to the feedback optimization algorithm type, wherein the key feedback execution steps include at least a feedback path backtracking step, a local redistribution calculation step, a feedback constraint verification step, and a feedback iterative update step; calculating the functional relationship between the number of the m key feedback execution steps and the corresponding dataset size parameter based on the time complexity calculation function to obtain the first time complexity.

[0046] Specifically, the feedback optimization decision-making dataset, obtained based on the second load balancing decision-making link and corresponding to the multi-source decision-making dataset, represents the reprocessing of the results of the multi-source decision-making dataset after the initial load balancing decision execution, according to the feedback optimization processing flow corresponding to the second load balancing decision-making link, thereby forming a dataset for feedback optimization analysis. The second load balancing decision-making link represents the decision-making processing path used for impact assessment and correction analysis of the initial load balancing decision results. The multi-source decision-making dataset represents a comprehensive set of operational data composed of the operating states of multiple nodes, task allocation states, and network topology states in a dynamic network environment. The feedback optimization decision-making dataset represents a dataset formed on the basis of the initial decision execution, used to describe the feedback correction process and reflect the state changes and optimization requirements of the load balancing decision results after execution.

[0047] Furthermore, the second time complexity corresponding to the feedback optimization decision-making runtime dataset is calculated using a time complexity calculation function. This represents the computational complexity of the second load balancing decision-making link during the feedback correction process, based on the scale characteristics of the feedback optimization decision-making runtime dataset and the processing method of the feedback optimization algorithm. The feedback optimization decision-making runtime dataset represents a dedicated dataset used for performing feedback correction analysis, characterizing the system state changes after the load balancing decision is executed. The time complexity calculation function represents a function structure used to model the computational growth trend during algorithm execution. By correlating the scale parameter corresponding to the feedback optimization decision-making runtime dataset with the feedback optimization algorithm type, the complexity change relationship during the feedback optimization process can be obtained. The second time complexity represents the complexity result corresponding to the second load balancing decision-making link during the feedback optimization process, characterizing the computational resource consumption level and computational growth trend in the feedback correction stage.

[0048] Simultaneously, a third time complexity is calculated using a time complexity calculation function corresponding to the previous multi-source decision-making dataset. This represents the complexity analysis of the initial load balancing decision execution process corresponding to the previous time step or round of multi-source decision-making dataset. The previous multi-source decision-making dataset represents the set of historical decision-making data prior to the current feedback optimization process, reflecting the system's operating state in the previous round of load balancing decision-making. The third time complexity represents the complexity result of the previous round of multi-source decision-making dataset in executing the initial load balancing decision process, characterizing the changes in computational scale corresponding to historical decision-making processes.

[0049] The feedback optimization decision-making runtime dataset includes a set of feedback correction nodes participating in the feedback optimization decision-making process, a set of feedback correction tasks, feedback path status information, and feedback optimization constraint information. This indicates that the dataset is composed of multiple key elements used for feedback correction processing. The feedback correction node set represents the set of nodes participating in feedback optimization processing, used to perform the redistribution and correction calculations of load balancing results. The feedback correction task set represents the set of tasks that need to be readjusted or redistributed during the feedback optimization process, used to characterize the task adjustment requirements in the feedback phase. The feedback path status information represents the runtime status information of the data backhaul path and processing path during the feedback optimization process, used to describe the flow characteristics of feedback data in the network. The feedback optimization constraint information represents the set of constraints that need to be met during the feedback optimization process, used to constrain the resource allocation and scheduling rules during the feedback correction process.

[0050] Furthermore, the dataset size parameter of the feedback optimization decision-making operation dataset is extracted to represent the statistical analysis of the data scale characteristics corresponding to the feedback optimization decision-making operation dataset. The feedback optimization decision-making operation dataset represents the data set used to perform feedback correction analysis and optimization processing. The dataset size parameter represents the data variables used to describe the overall processing scale of the feedback optimization decision-making operation dataset, including the number of nodes participating in feedback optimization processing, the number of feedback correction tasks, the number of feedback state feature dimensions, and the number of feedback constraints, which are used to characterize the data processing scale and computational analysis scale corresponding to the feedback optimization stage.

[0051] Furthermore, the feedback optimization algorithm type for the second load balancing decision link's feedback optimization decision execution dataset is obtained, indicating the algorithm processing method adopted by the second load balancing decision link during the feedback correction and optimization analysis process. The second load balancing decision link represents the decision path used to perform feedback correction analysis, local optimization adjustment, and decision result correction processing. The feedback optimization algorithm type represents the algorithm category used to process the feedback optimization decision execution dataset, including feedback optimization algorithms based on local search, feedback analysis algorithms based on path correction, or feedback adjustment algorithms based on dynamic iteration, used to characterize the corresponding operational rules and data processing mechanisms in the feedback optimization stage.

[0052] Furthermore, the number of execution steps for the m key feedback steps corresponding to the feedback optimization algorithm type is obtained, representing a statistical analysis of the core computational steps in the feedback correction process. The number of execution steps for the m key feedback steps represents the statistical results of the execution count of the feedback optimization algorithm in different feedback processing stages, reflecting the changes in computational scale and computational resource consumption during the feedback optimization process. The key feedback execution steps represent the core computational processes that have a major impact on the complexity of feedback optimization. By statistically analyzing the execution counts of different key feedback execution steps, the computational growth trend of the feedback optimization algorithm under varying data scale conditions can be characterized.

[0053] The feedback path backtracking step involves analyzing historical decision paths or data transmission paths during the feedback optimization process to determine path status information affecting the optimization results. The local reallocation calculation step involves rescheduling and reallocating resources or tasks from certain nodes to perform local load balancing optimization analysis. The feedback constraint verification step verifies resource, network, or scheduling rules during the optimization process to ensure the optimization results meet preset constraints. The feedback iterative update step involves multiple rounds of dynamic updates and repeated corrections to the feedback correction results during the optimization analysis, improving the stability and adaptability of the optimization results.

[0054] Furthermore, based on the time complexity calculation function, the functional relationship between the number of execution steps of the m key feedback steps and the corresponding dataset size parameter is calculated to obtain the first time complexity. This represents the analysis of the computational complexity in the feedback optimization process based on the quantitative change relationship between the execution number of the key feedback steps and the data size variable in the feedback optimization stage. The time complexity calculation function represents a complexity analysis function used to describe the relationship between the number of algorithm executions and the data size variable. By correlating the number of key feedback execution steps with the dataset size parameter corresponding to the feedback optimization stage, the complexity change relationship of the feedback optimization algorithm under different scale conditions can be obtained. The functional relationship represents the quantitative change relationship between the execution number of key feedback steps and the data size variable, used to characterize the growth trend of the computational scale of the feedback optimization algorithm as the data size changes. The first time complexity represents the complexity analysis result corresponding to the feedback optimization stage, used to reflect the overall computational resource consumption level and complexity change trend of the second load balancing decision link in the process of executing feedback correction.

[0055] The first time complexity and the second time complexity are input into the network parameter optimizer, and the parameters of the dynamic network load balancing decision network are optimized according to the output first network optimization parameters or second network optimization parameters.

[0056] Furthermore, this application also includes: if the first time complexity is greater than the second time complexity, performing network parameter optimization with the first load balancing decision link as the input parameter variable and the time complexity difference between the first time complexity and the second time complexity as the optimization objective, and outputting first network optimization parameters; if the first time complexity is less than the second time complexity, performing network parameter optimization with the second load balancing decision link as the input parameter variable and the time complexity difference between the first time complexity and the second time complexity as the optimization objective, and outputting second network optimization parameters; if the first time complexity is equal to the second time complexity, not activating the network parameter optimizer.

[0057] Furthermore, this application also includes: performing network parameter optimization with the time complexity difference between the first time complexity and the second time complexity as the optimization target; obtaining the time complexity difference between the load balancing decision link corresponding to the current candidate parameter variable and another load balancing decision link; updating the current candidate parameter variable through gradient until the absolute value of the time complexity difference is less than a preset tolerance threshold, and outputting the current parameter variable as a network optimization parameter.

[0058] Furthermore, this application also includes: obtaining the previous multi-source decision-making operation dataset based on the first load balancing decision link and the multi-node intelligent fusion terminal, calculating the third time complexity corresponding to the previous multi-source decision-making operation dataset; inputting the first time complexity, the second time complexity and the third time complexity into the network parameter optimizer to re-output the first network optimization parameter or the second network optimization parameter.

[0059] Specifically, if the first time complexity is greater than the second time complexity, network parameter optimization is performed using the first load balancing decision link as input parameter variables and the time complexity difference between the first and second time complexities as the optimization objective. The output is the first network optimization parameter, indicating that when the computational complexity corresponding to the initial load balancing decision stage is higher than the computational complexity corresponding to the feedback optimization stage, optimization and adjustment processing is performed on the network parameters corresponding to the first load balancing decision link. The first time complexity represents the complexity analysis result of the first load balancing decision link during the initial load balancing decision process, and the second time complexity represents the complexity analysis result of the second load balancing decision link during the feedback optimization process. The first load balancing decision link represents the decision path used to perform node resource scheduling and initial task allocation processing. The parameter variables represent the set of parameters used to influence the computational scale and scheduling strategy of the first load balancing decision link, including node selection parameters, task allocation parameters, iteration control parameters, or constraint processing parameters. The time complexity difference represents the complexity difference between the first and second time complexities, used to characterize the degree of difference in computational scale between the two decision stages. Network parameter optimization refers to performing adjustment analysis on the parameter variables in the decision-making process to reduce the degree of computational imbalance corresponding to differences in complexity. The first network optimization parameters represent the first load balancing decision-making process parameter results obtained after complexity optimization, which are used to adjust the computational complexity corresponding to the initial load balancing decision-making stage.

[0060] Furthermore, if the first time complexity is less than the second time complexity, network parameter optimization is performed using the second load balancing decision link as input parameter variables and the time complexity difference between the first and second time complexities as the optimization objective. The output is the second network optimization parameter, indicating that when the computational complexity corresponding to the feedback optimization stage is higher than that corresponding to the initial load balancing decision stage, optimization and adjustment processing is performed on the network parameters corresponding to the second load balancing decision link. The second load balancing decision link represents the decision path used to perform feedback correction analysis and local optimization processing. The parameter variables represent the set of parameters used to influence the computational scale and feedback correction strategy in the feedback optimization process, including feedback path analysis parameters, local redistribution parameters, feedback constraint processing parameters, and feedback iteration control parameters. The time complexity difference serves as the optimization objective in the network parameter optimization process, constraining the range of complexity differences between the feedback optimization stage and the initial decision stage. The second network optimization parameter represents the result of the second load balancing decision link parameters obtained after complexity balancing optimization processing, used to reduce the computational complexity corresponding to the feedback optimization stage, thereby improving the real-time processing capability in the dynamic network load balancing feedback process. Specifically, obtaining the time complexity difference between the load balancing decision path corresponding to the current candidate parameter variable and another load balancing decision path represents the calculation and analysis of the complexity change of the corresponding load balancing decision path for the parameter variable currently participating in the optimization analysis. The current candidate parameter variable represents the set of parameters participating in the complexity adjustment analysis during the current round of parameter optimization, including node scheduling parameters, task allocation parameters, feedback correction parameters, or iterative control parameters. The load balancing decision path represents the decision path used to perform load balancing decision processing or feedback optimization processing. Another load balancing decision path represents another decision path that did not participate in the current parameter optimization processing. The time complexity difference represents the complexity difference between the corresponding load balancing decision path and another load balancing decision path under the influence of the current candidate parameter variable, reflecting the degree of influence of the current parameter variable on the complexity balance state.

[0061] Furthermore, by updating the current candidate parameter variables through gradient updates until the absolute value of the time complexity difference is less than a preset tolerance threshold, the current parameter variables are output as network optimization parameters. This indicates that a gradient iterative approach is used to continuously adjust and analyze the current candidate parameter variables. Gradient updates represent incremental adjustments to the parameter variables based on the direction of change in the time complexity difference, used to gradually reduce the complexity difference between the two load balancing decision links. The absolute value of the time complexity difference represents the magnitude of the complexity deviation between the two decision stages, used to reflect the degree of complexity balance. The preset tolerance threshold represents the maximum allowable range of complexity difference, used to constrain the error range of the complexity balance target. When the absolute value of the time complexity difference is less than the preset tolerance threshold, it indicates that the preset complexity balance condition has been reached between the two load balancing decision links. The network optimization parameters represent the parameter optimization results obtained after gradient update processing, used to adjust the computational complexity and link processing efficiency in the dynamic network load balancing decision process.

[0062] Furthermore, if the first time complexity equals the second time complexity, the network parameter optimizer is not activated. This means that when the computational complexity of the first and second load balancing decision links is balanced, no additional parameter optimization is performed. The network parameter optimizer is an optimization module used to perform parameter adjustment analysis based on the difference in time complexity. When the first and second time complexities are equal, it indicates that the computational resource consumption levels corresponding to the initial load balancing decision stage and the feedback optimization stage are consistent. The complexity balance can meet the collaborative processing requirements in the dynamic network load balancing process. Therefore, network parameter optimization is stopped, thereby avoiding unnecessary parameter updates that could affect the stability of the dynamic network decision.

[0063] The dynamic network load balancing decision network is optimized based on either the first or second network optimization parameters. This means that the decision processing parameters in the dynamic network load balancing decision network are updated based on the parameter adjustment results obtained from the network parameter optimization. The first network optimization parameter represents the complexity optimization result corresponding to the first load balancing decision link, and is mainly used to adjust the node scheduling rules, task allocation rules, or iterative control rules in the initial load balancing decision stage. The second network optimization parameter represents the complexity optimization result corresponding to the second load balancing decision link, and is mainly used to adjust the feedback correction rules, local optimization rules, or feedback iteration rules in the feedback optimization stage. The dynamic network load balancing decision network represents a dynamic decision processing network used to perform node resource scheduling, task load allocation, and feedback optimization. Parameter optimization means adjusting the computational control parameters in the dynamic network load balancing decision network based on the complexity analysis results, thereby improving the complexity balance and link collaborative processing capabilities between different decision stages.

[0064] Furthermore, outputting the first or second network optimization parameters also includes obtaining the previous multi-source decision-making runtime dataset based on the multi-node intelligent fusion terminal using the first load balancing decision link, and calculating the third time complexity corresponding to the previous multi-source decision-making runtime dataset. This indicates that, based on the current complexity optimization analysis, the complexity analysis results corresponding to the historical decision-making stage are introduced for joint optimization processing. The previous multi-source decision-making runtime dataset represents the historical runtime data set corresponding to the previous round of dynamic network load balancing decision-making, used to reflect the node running status, task allocation status, and network load changes in the historical stage. The third time complexity represents the complexity analysis result of the previous multi-source decision-making runtime dataset during the execution of the first load balancing decision link processing, used to characterize the computational complexity and complexity change trend corresponding to the historical decision-making stage. By introducing the third time complexity, the complexity changes of the dynamic network load balancing decision network at different time stages can be continuously analyzed, thereby improving the dynamic adaptability in the complexity optimization process.

[0065] Furthermore, inputting the first, second, and third time complexities into the network parameter optimizer and re-outputting the first or second network optimization parameters indicates that the complexity of the current initial decision stage, the complexity of the current feedback optimization stage, and the complexity of the historical decision stage are used together as the basis for complexity optimization analysis, and the network parameter optimization process is re-iterated and analyzed. The network parameter optimizer represents the optimization processing module used to perform complexity balancing optimization and parameter adjustment analysis. By jointly analyzing the changing relationships between the first, second, and third time complexities, more stable complexity change trend analysis results can be obtained. Re-outputting the first or second network optimization parameters indicates that the parameter optimization results are regenerated based on the multi-stage complexity joint analysis results, which are used to further adjust the parameter configuration state in the dynamic network load balancing decision network, thereby reducing the degree of complexity fluctuation and dynamic operation deviation between different decision stages.

[0066] In summary, the dynamic networking load balancing method for multi-node intelligent fusion terminals provided in this application has the following technical effects: by achieving the technical goal of dynamically evaluating and optimizing the computational complexity in the multi-node load balancing decision-making process, it achieves the technical effects of improving scheduling response speed, optimizing resource utilization, and enhancing the overall network operation stability while ensuring balanced network task allocation.

[0067] Example 2: Based on the same inventive concept as the dynamic networking load balancing method for multi-node intelligent fusion terminals in the foregoing examples, this application also provides a dynamic networking load balancing system for multi-node intelligent fusion terminals. Please refer to the appendix. Figure 2The system includes: a link extraction module 1, used to extract a first load balancing decision link and a second load balancing decision link of the dynamic network load balancing decision network, and a feedback optimization link connected to the first load balancing decision link; a first time complexity calculation module 2, used to obtain a multi-source decision operation dataset based on a multi-node intelligent fusion terminal based on the first load balancing decision link, and calculate the first time complexity corresponding to the multi-source decision operation dataset; a second time complexity calculation module 3, used to obtain a feedback optimization decision operation dataset corresponding to the multi-source decision operation dataset based on the second load balancing decision link, and calculate the second time complexity corresponding to the feedback optimization decision operation dataset; and a parameter optimization module 4, used to input the first time complexity and the second time complexity into a network parameter optimizer, and optimize the parameters of the dynamic network load balancing decision network according to the output first network optimization parameters or second network optimization parameters.

[0068] Furthermore, the dynamic networking load balancing system of the multi-node intelligent fusion terminal is also used to: obtain the previous multi-source decision operation dataset based on the first load balancing decision link, calculate the third time complexity corresponding to the previous multi-source decision operation dataset; input the first time complexity, the second time complexity and the third time complexity into the network parameter optimizer to re-output the first network optimization parameter or the second network optimization parameter.

[0069] Furthermore, the dynamic networking load balancing system of the multi-node intelligent fusion terminal is also used to: obtain the dataset size parameter of the multi-source decision-making operation dataset, wherein the multi-source decision-making operation dataset includes the resource operation status, task allocation information, network topology status information, and real-time load distribution of each node; process the decision algorithm type of the multi-source decision-making operation dataset according to the first load balancing decision link; and calculate the time complexity of the dataset size parameter and the decision algorithm type using a time complexity calculation function to obtain a first time complexity.

[0070] Furthermore, the dynamic networking load balancing system of the multi-node intelligent fusion terminal is also used to: obtain the number of k key execution steps of the decision algorithm type, wherein the key execution steps include at least node comparison steps, task and node matching calculations, queue operation processing, and iteration cost calculation steps; extract the number of participating nodes, the number of tasks to be assigned, the feature vector dimension, and the number of constraints of the dataset size parameter; and calculate the functional relationship between the number of the k key execution steps and the corresponding dataset size parameter based on the time complexity calculation function to obtain the first time complexity.

[0071] Furthermore, the dynamic networking load balancing system of the multi-node intelligent fusion terminal is also used to: map the number of the k key execution steps into function expressions with the number of participating nodes, the number of tasks to be assigned, the dimension of the feature vector, and the number of constraints as independent variables; sum the k function expressions corresponding to the number of the k key execution steps to obtain a comprehensive function expression of the total number of execution steps; and perform a time complexity function O(f(n)) transformation on the comprehensive function expression to obtain the first time complexity.

[0072] Furthermore, the dynamic networking load balancing system of the multi-node intelligent fusion terminal is also used to: calculate the second time complexity corresponding to the feedback optimization decision running dataset using the time complexity calculation function, and calculate the third time complexity corresponding to the previous multi-source decision running dataset using the time complexity calculation function; wherein, the feedback optimization decision running dataset includes a set of feedback correction nodes participating in the feedback optimization decision running, a set of feedback correction tasks, feedback path status information, and feedback optimization constraint information.

[0073] Furthermore, the dynamic networking load balancing system of the multi-node intelligent fusion terminal is also used for: extracting the dataset size parameter of the feedback optimization decision running dataset; obtaining the feedback optimization algorithm type of the second load balancing decision link processing the feedback optimization decision running dataset; obtaining the number of m key feedback execution steps corresponding to the feedback optimization algorithm type, wherein the key feedback execution steps include at least a feedback path backtracking step, a local redistribution calculation step, a feedback constraint verification step, and a feedback iteration update step; and calculating the functional relationship between the number of the m key feedback execution steps and the corresponding dataset size parameter based on the time complexity calculation function to obtain the first time complexity.

[0074] Furthermore, the dynamic network load balancing system of the multi-node intelligent fusion terminal is also used for: if the first time complexity is greater than the second time complexity, performing network parameter optimization with the first load balancing decision link as the input parameter variable and the time complexity difference between the first time complexity and the second time complexity as the optimization target, and outputting the first network optimization parameter; if the first time complexity is less than the second time complexity, performing network parameter optimization with the second load balancing decision link as the input parameter variable and the time complexity difference between the first time complexity and the second time complexity as the optimization target, and outputting the second network optimization parameter; if the first time complexity is equal to the second time complexity, not activating the network parameter optimizer.

[0075] Furthermore, the dynamic networking load balancing system of the multi-node intelligent fusion terminal is also used to: perform network parameter optimization with the time complexity difference between the first time complexity and the second time complexity as the optimization target; obtain the time complexity difference between the load balancing decision link corresponding to the current candidate parameter variable and another load balancing decision link; update the current candidate parameter variable through gradient until the absolute value of the time complexity difference is less than a preset tolerance threshold, and output the current parameter variable as a network optimization parameter.

[0076] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The dynamic networking load balancing method and specific examples of the multi-node intelligent fusion terminal in the foregoing embodiment one are also applicable to the dynamic networking load balancing system of the multi-node intelligent fusion terminal in this embodiment. Through the foregoing detailed description of the dynamic networking load balancing method of the multi-node intelligent fusion terminal, those skilled in the art can clearly understand the dynamic networking load balancing system of the multi-node intelligent fusion terminal in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0077] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0078] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A dynamic networking load balancing method for multi-node intelligent fusion terminals, characterized in that, The method includes: Extract the first load balancing decision link and the second load balancing decision link of the dynamic network load balancing decision network, and the feedback optimization link that connects the second load balancing decision link to the first load balancing decision link. Based on the first load balancing decision link, a multi-source decision operation dataset based on a multi-node intelligent fusion terminal is obtained, and the first time complexity corresponding to the multi-source decision operation dataset is calculated. Based on the second load balancing decision link, obtain the feedback optimization decision running dataset corresponding to the multi-source decision running dataset, and calculate the second time complexity corresponding to the feedback optimization decision running dataset; The first time complexity and the second time complexity are input into the network parameter optimizer, and the parameters of the dynamic network load balancing decision network are optimized according to the output first network optimization parameters or second network optimization parameters.

2. The dynamic networking load balancing method for multi-node intelligent fusion terminals as described in claim 1, characterized in that, The method for outputting the first or second network optimization parameters also includes: Based on the first load balancing decision link, obtain the previous multi-source decision running dataset based on the multi-node intelligent fusion terminal, and calculate the third time complexity corresponding to the previous multi-source decision running dataset. The first time complexity, the second time complexity, and the third time complexity are input into the network parameter optimizer to re-output the first network optimization parameters or the second network optimization parameters.

3. The dynamic networking load balancing method for multi-node intelligent fusion terminals as described in claim 1, characterized in that, The method for calculating the first time complexity of the multi-source decision-making dataset includes: Obtain the dataset size parameter of the multi-source decision-making operation dataset, which includes the resource operation status, task allocation information, network topology status information, and real-time load distribution of each node; The decision algorithm type for processing the multi-source decision running dataset is based on the first load balancing decision link; The time complexity is calculated using a time complexity calculation function to determine the time complexity of the dataset size parameter and the decision algorithm type, thus obtaining the first time complexity.

4. The dynamic networking load balancing method for multi-node intelligent fusion terminals as described in claim 3, characterized in that, The method involves calculating the time complexity of the dataset size parameter and the decision algorithm type using a time complexity calculation function to obtain a first time complexity. Obtain the number of k key execution steps of the decision algorithm type, wherein the key execution steps include at least the node comparison step, the number of task and node matching calculations, queue operation processing, and the iteration cost calculation step; Extract the following parameters from the dataset size parameters: number of participating nodes, number of tasks to be assigned, feature vector dimension, and number of constraints. The first time complexity is obtained by calculating the functional relationship between the number of the k key execution steps and the corresponding dataset size parameter based on the time complexity calculation function.

5. The dynamic networking load balancing method for multi-node intelligent fusion terminals as described in claim 4, characterized in that, The first time complexity is obtained by calculating the functional relationship between the number of execution steps of the k key execution steps and the corresponding dataset size parameter based on the time complexity calculation function. The method includes: The number of the k key execution steps is mapped to a function expression with the number of participating nodes, the number of tasks to be assigned, the dimension of the feature vector, and the number of constraints as independent variables; Summing the k function expressions corresponding to the number of times the k key execution steps are performed, a comprehensive function expression for the total number of times is obtained. The comprehensive function expression is then transformed with a time complexity function of O(f(n)) to obtain the first time complexity.

6. The dynamic networking load balancing method for multi-node intelligent fusion terminals as described in claim 3, characterized in that, The second time complexity corresponding to the feedback optimization decision running dataset is calculated using the time complexity calculation function, and the third time complexity corresponding to the previous multi-source decision running dataset is calculated using the time complexity calculation function. The feedback optimization decision-making operation dataset includes a set of feedback correction nodes participating in the feedback optimization decision-making operation, a set of feedback correction tasks, feedback path status information, and feedback optimization constraint information.

7. The dynamic networking load balancing method for multi-node intelligent fusion terminals as described in claim 6, characterized in that, The method for calculating the second time complexity of the feedback optimization decision-making dataset using the time complexity calculation function includes: Extract the dataset size parameter of the feedback optimization decision-making dataset; Obtain the feedback optimization algorithm type of the feedback optimization decision running dataset processed by the second load balancing decision link; Obtain the number of m key feedback execution steps corresponding to the feedback optimization algorithm type, wherein the key feedback execution steps include at least the feedback path backtracking step, the local redistribution calculation step, the feedback constraint verification step, and the feedback iterative update step; The first time complexity is obtained by calculating the functional relationship between the number of execution steps of the m feedback key steps and the corresponding dataset size parameter based on the time complexity calculation function.

8. The dynamic networking load balancing method for multi-node intelligent fusion terminals as described in claim 1, characterized in that, The method of inputting the first time complexity and the second time complexity into the network parameter optimizer includes: If the first time complexity is greater than the second time complexity, network parameter optimization is performed with the first load balancing decision link as the input parameter variable and the time complexity difference between the first time complexity and the second time complexity as the optimization target, and the first network optimization parameters are output. If the first time complexity is less than the second time complexity, take the second load balancing decision link as the input parameter variable, take the time complexity difference between the first time complexity and the second time complexity as the optimization target, perform network parameter optimization, and output the second network optimization parameters. If the first time complexity is equal to the second time complexity, the network parameter optimizer is not activated.

9. The dynamic networking load balancing method for multi-node intelligent fusion terminals as described in claim 8, characterized in that, Network parameter optimization is performed with the time complexity difference between the first time complexity and the second time complexity as the optimization objective; The time complexity of the load balancing decision link corresponding to the current candidate parameter variable is compared with the time complexity difference between another load balancing decision link. The current candidate parameter variable is updated by gradient until the absolute value of the time complexity difference is less than a preset tolerance threshold, at which point the current parameter variable is output as the network optimization parameter.

10. A dynamic networking load balancing system for multi-node intelligent fusion terminals, characterized in that, The steps for implementing the dynamic networking load balancing method for a multi-node intelligent fusion terminal according to any one of claims 1 to 9 include: The link extraction module is used to extract the first load balancing decision link and the second load balancing decision link of the dynamic network load balancing decision network, and the feedback optimization link connected to the first load balancing decision link by the second load balancing decision link. The first time complexity calculation module is used to obtain a multi-source decision operation dataset based on a multi-node intelligent fusion terminal based on the first load balancing decision link, and calculate the first time complexity corresponding to the multi-source decision operation dataset. The second time complexity calculation module is used to obtain the feedback optimization decision running dataset corresponding to the multi-source decision running dataset based on the second load balancing decision link, and calculate the second time complexity corresponding to the feedback optimization decision running dataset. The parameter optimization module is used to input the first time complexity and the second time complexity into the network parameter optimizer, and optimize the parameters of the dynamic network load balancing decision network according to the output first network optimization parameters or second network optimization parameters.

Images