Defect detection cooperative calculation load regulation and control system based on sneaker production line speed

Abstract

The invention relates to the technical field of intelligent manufacturing of sneakers, in particular to a sneaker production line speed-based defect detection cooperative computing load regulation and control system, which comprises a cooperative regulation and control management center, an interval division unit, a task generation unit, a migration acquisition unit, a node matching unit, an interval migration evaluation unit and a rear-end management unit, according to the method, full-link automatic processing from production line speed sensing to differential detection task packet generation is realized, task intelligent migration and optimal target node matching are carried out in combination with the load state of each node of the hybrid computing cluster, the condition of node overload or idle is avoided, and the load balancing efficiency is improved on the premise of ensuring the load balancing. The optimal target nodes are selected for different types of migration tasks, so that the migration cost is minimized, the continuity and the real-time performance of the detection tasks are ensured, the defect detection work is ensured to be stably carried out, meanwhile, the migration effects of different speed intervals are comprehensively evaluated, and the reliability and the effectiveness of system operation are further ensured.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology for athletic shoes, and in particular to a defect detection collaborative calculation load control system based on the speed of athletic shoe production lines. Background Technology

[0002] In the process of mass production of athletic shoes, defect detection is a key link to ensure product quality. With the popularization of automated production lines, machine vision-based defect detection technology has been widely used in athletic shoe production. It identifies problems such as scratches on the shoe upper, stitch defects, and poor fit of the sole through image acquisition and processing. Existing defect detection systems suffer from the following technical bottlenecks: Fixed detection tasks: Regardless of changes in production line speed, the same detection items and algorithm complexity are used, leading to resource waste at low speeds and detection delays at high speeds; Poor coordination among multiple detection nodes: Some nodes are overloaded while others are idle, affecting overall detection efficiency; Low specificity of migration effects: The method of obtaining the optimal target node is the same for different speed ranges, but it is difficult to provide targeted feedback and adjustments to the performance of migration after different speed ranges. To address the aforementioned technical shortcomings, a solution is proposed. Summary of the Invention

[0003] The purpose of this invention is to provide a collaborative computing load control system for defect detection based on the speed of a sports shoe production line, in order to solve the aforementioned technical defects. This invention achieves fully automated processing from production line speed perception to the generation of differentiated detection task packages. At the same time, it combines the load status of each node in the hybrid computing cluster to perform intelligent task migration and optimal target node matching, thereby minimizing migration costs and ensuring the continuity and real-time performance of detection tasks, thus guaranteeing the stable operation of defect detection.

[0004] The objective of this invention can be achieved through the following technical solution: a defect detection collaborative computing load control system based on the speed of a sports shoe production line, comprising a collaborative control management center, an interval division unit, a task generation unit, a migration acquisition unit, a node matching unit, an interval migration evaluation unit, and a backend management unit; The interval division unit is used to collect the real-time operating speed of the current sports shoe production line and perform production line speed division analysis. The obtained low-speed interval, medium-speed interval and high-speed interval are collectively referred to as speed interval. The task generation unit is used to perform interval and joint matching analysis on the retrieved list of detection items, and the resulting low-speed detection task package, medium-speed detection task package and high-speed detection task package are collectively referred to as detection task package; The migration acquisition unit is used to perform migration task management and analysis on the load status data of each collected node, and to include the remaining tasks in the nodes to be migrated into the migrateable task pool. The node matching unit is used to perform intelligent matching analysis on the load status data of regular nodes and set the candidate node corresponding to the maximum value in the expected load capacity index Q as the optimal target node. The interval migration evaluation unit is used to conduct a comprehensive evaluation and analysis of the migration effect data after the optimal target node migration in each speed interval, and to obtain the interval to be managed and the qualified interval.

[0005] Preferably, the analysis process for the interval division unit is as follows: Obtain the real-time operating speed of the current athletic shoe production line during its operating period; Call the preset speed range parameters: minimum value v1, maximum value v2; The real-time operating speed is processed for discrimination. If the real-time operating speed is less than the minimum value v1, it is determined to be in the low-speed range. If the minimum value v1 is less than or equal to the real-time operating speed and less than or equal to the maximum value v2, it is determined to be in the medium-speed range. If the real-time operating speed is greater than or equal to the maximum value v2, it is determined to be in the high-speed range.

[0006] Preferably, the analysis process of the task generation unit is as follows: Retrieve a list of preset test items from the database: basic test items include sole integrity and upper splicing alignment; enhanced test items include upper material texture analysis and stitch density distribution detection. At the same time, the pre-set proportional coefficients a1 and a2 corresponding to the enhanced detection items in the low-speed range and the medium-speed range are obtained respectively, where a1 > a2 > 0; Based on the shoe model parameters produced by the current sports shoe production line, the unnecessary inspection items for the shoe models produced by the current sports shoe production line are removed, and a subset of inspection items is constructed based on the retained inspection items; The real-time running speed is analyzed to obtain low-speed detection task packages, medium-speed detection task packages, and high-speed detection task packages.

[0007] Preferably, the analysis process of the migration acquisition unit is as follows: Real-time monitoring of the load status data of each node in the hybrid computing cluster (composed of edge computing nodes and cloud computing nodes), including real-time computing power utilization and task response latency; The load status data is preprocessed, and the preprocessed load status data is input into the pre-set load prediction model to obtain the predicted load value output by the pre-set load prediction model. The predicted load value is then processed to determine the node to be migrated and the regular node.

[0008] Preferably, a task list of nodes to be migrated is obtained, which includes detection task packages and remaining processing time after migration; Remove prohibited migration tasks and uneconomical migration tasks from the task list, and add the remaining tasks in the nodes to be migrated to the migration task pool. Migration is prohibited for tasks whose response latency is less than or equal to a preset response latency threshold. Migration is uneconomical for tasks whose remaining processing time after migration is less than a preset duration threshold.

[0009] Preferably, the analysis process of the node matching unit is as follows: obtain the remaining computing power, response latency, and network transmission latency from the load status data of each regular node; filter the regular nodes based on the remaining computing power, response latency, and network transmission latency; and select regular nodes that meet the following conditions: remaining computing power greater than a preset remaining computing power threshold, response latency less than a preset response latency threshold, and network transmission latency less than a preset network transmission latency threshold, and set them as migrateable nodes.

[0010] Preferably, the computing power score, response score, and network score are obtained within a preset range corresponding to the remaining computing power, response latency, and network transmission latency; A feasibility selection score is obtained by weighting and calculating the migrated nodes based on computing power score, response score, and network score. Based on the feasibility selection score, a secondary screening of migrateable nodes is performed to construct a candidate node pool for migrateable nodes whose feasibility selection scores are greater than or equal to a preset feasibility selection score threshold.

[0011] Preferably, the expected load capacity index Q of the hybrid computing cluster is calculated based on the candidate node pool after each candidate node receives tasks from the migrated node located in the migrateable task pool; The candidate node corresponding to the maximum value in the expected load capacity index Q is set as the optimal target node.

[0012] Preferably, the analysis process of the interval migration assessment unit is as follows: Based on different speed ranges, the migration effect data after the optimal target node migration is obtained. The migration effect data includes task performance indicators, resource utilization indicators, and migration cost indicators. The initial data is obtained by normalizing each indicator in the migration effect data; Input the task performance indicators, resource utilization indicators and migration cost indicators from the initial data into the pre-set task performance scoring model, resource utilization scoring model and migration cost scoring model respectively to obtain the output task performance score, resource utilization score and migration cost score. The migration effect coefficient is obtained by weighting and calculating the migration performance score, resource utilization score and migration cost score. The migration effect coefficient of each speed range is then processed to determine the range to be managed and the qualified range.

[0013] The beneficial effects of this invention are as follows: (1) This invention realizes the full-link automated processing from production line speed perception to the generation of differentiated detection task packages. It can ensure detection accuracy when the production line speed is slow, and it can not affect the production line operation efficiency when the production line speed is fast. At the same time, it combines the load status of each node of the hybrid computing cluster to perform intelligent task migration and match the optimal target node, avoiding the situation of node overload or idleness, so that computing resources are reasonably allocated and fully utilized, and the operation efficiency of the entire system is improved. (2) Under the premise of ensuring load balancing, this invention takes into account data transmission efficiency and task processing characteristics, selects the optimal target node for different types of migration tasks, thereby minimizing migration costs and ensuring the continuity and real-time performance of detection tasks, ensuring the stable progress of defect detection work. At the same time, it comprehensively evaluates the migration effect in different speed ranges, can promptly identify problems in the system operation process, and take corresponding measures to optimize and improve, further ensuring the reliability and effectiveness of system operation. Attached Figure Description

[0014] The invention will now be further described with reference to the accompanying drawings; Figure 1 This is a flowchart of the system of the present invention; Figure 2 This is a partial analytical diagram of Embodiment 1 of the present invention; Figure 3 This is a partial analysis diagram of Embodiment 2 of the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments; Example 1: Please see Figures 1 to 3As shown, the present invention is a defect detection collaborative computing load control system based on the speed of a sports shoe production line, including a collaborative control management center, an interval division unit, a task generation unit, a migration acquisition unit, a node matching unit, an interval migration evaluation unit, and a backend management unit. The coordinated control and management center has a two-way communication connection with the interval division unit, the interval division unit has a one-way communication connection with the task generation unit, the task generation unit has a one-way communication connection with the coordinated control and management center, the coordinated control and management center has a two-way communication connection with the migration acquisition unit, the migration acquisition unit has a one-way communication connection with the node matching unit, the node matching unit has a one-way communication connection with the coordinated control and management center, the coordinated control and management center has a one-way communication connection with the interval migration evaluation unit, and the interval migration evaluation unit has a one-way communication connection with the backend management unit. The interval division unit is used to collect the real-time operating speed of the current sports shoe production line and perform production line speed division analysis. The specific production line speed division analysis process is as follows: Obtain the real-time operating speed of the current athletic shoe production line during its operating period; Call the preset speed range parameters: minimum value v1, maximum value v2; The real-time running speed is judged and processed. If the real-time running speed is less than the minimum value v1, it is judged as a low-speed range. If the minimum value v1 is less than the real-time running speed and less than the maximum value v2, it is judged as a medium-speed range. If the real-time running speed is greater than the maximum value v2, it is judged as a high-speed range. The low-speed range, medium-speed range, and high-speed range are collectively referred to as speed ranges, and the speed ranges are sent to the coordinated control and management center for storage. The task generation unit is used to perform interval and detection item joint matching analysis on the retrieved preset list of detection items. The specific interval and detection item joint matching analysis process is as follows: Retrieve a pre-defined list of detection items from the database: Basic inspection items include mandatory items such as sole integrity, upper splicing alignment, and shoelace eyelet position deviation; Enhanced testing items include optional items such as upper material texture analysis, stitch density distribution detection, and shoe label printing clarity; At the same time, the pre-set proportional coefficients a1 and a2 corresponding to the enhanced detection items in the low-speed range and the medium-speed range are obtained respectively, where a1 > a2 > 0; Based on the shoe model parameters (size range, number of inspection areas per shoe, etc.) produced by the current sports shoe production line, the unnecessary inspection items for the current sports shoe production line are removed (e.g., if a specific shoe model does not require inspection of shoelace holes, the corresponding item is removed), and a subset of inspection items is constructed based on the retained inspection items. Analysis of real-time running speed: If the real-time operating speed of the current sports shoe production line is in the low-speed range, then obtain all basic detection items and enhanced detection items of the proportional coefficient a1 in the detection item subset, and construct a low-speed detection task package based on all basic detection items and enhanced detection items of the proportional coefficient a1. If the real-time operating speed during the current running period of the sports shoe production line is in the medium speed range, then obtain all basic detection items and enhanced detection items of the proportional coefficient a2 in the detection item subset, and construct a medium speed detection task package based on all basic detection items and enhanced detection items of the proportional coefficient a2. If the real-time operating speed of the current sports shoe production line is in the high-speed range, then obtain all basic inspection items in the inspection item subset and construct a high-speed inspection task package based on all basic inspection items; The obtained low-speed detection task packages, medium-speed detection task packages, and high-speed detection task packages are collectively referred to as detection task packages, and the detection task packages are sent to the collaborative control and management center for storage. Through the above process, the entire chain of automated processing, from production line speed perception to the generation of differentiated inspection task packages, is realized. The core logic is to dynamically adjust the number and type of inspection items based on the speed range, balancing inspection accuracy and production line operating efficiency.

[0017] Example 2: The migration acquisition unit is used to perform migration task management and analysis on the collected load status data of each node. The specific migration task management and analysis process is as follows: Real-time monitoring of the load status data of each node in the hybrid computing cluster (composed of edge computing nodes and cloud computing nodes), including real-time computing power utilization, task response latency, network transmission bandwidth, etc. For example: Real-time computing power utilization: GPU / CPU utilization is collected through the performance monitoring interface of the node operating system; Network transmission bandwidth: Real-time data transmission rate between the edge and the cloud is collected through the network interface controller (NIC), including uplink and downlink bandwidth; The load status data is preprocessed (e.g., cleaned, enhanced), and the preprocessed load status data is input into the pre-set load prediction model to obtain the predicted load value output by the pre-set load prediction model. The predicted load value is then judged. If the predicted load value is greater than or equal to the preset predicted load value threshold, the corresponding node is set as a node to be migrated. If the predicted load value is less than the preset predicted load value threshold, the corresponding node is set as a regular node. In the hybrid computing cluster of this invention, the comprehensive collection and accurate analysis of load status data is the foundation for realizing intelligent task scheduling. This data not only reflects the current operating status of each node, but also provides key basis for load prediction and task migration decisions. Obtain the task list of the nodes to be migrated. The task list includes the detection task package, the remaining processing time after migration, etc. Remove prohibited migration tasks (tasks with response latency less than or equal to the preset response latency threshold) and uneconomical migration tasks (tasks with remaining processing time less than the preset duration threshold after migration) from the task list. Add the remaining tasks in the nodes to be migrated to the migrateable task pool and send the migrateable task pool to the collaborative control and management center for storage. For example, in the task list of the node to be migrated, tasks that are prohibited from migration are removed: tasks whose response latency is less than or equal to the preset response latency threshold. For example, if the response latency is 5ms and the preset response latency threshold is 6.5ms, then 5ms < 6.5ms, and thus the task is removed from the task list. Migrating uneconomical tasks: Tasks whose remaining processing time after migration is less than a preset time threshold. For example, if the remaining processing time after migration is 2s and the preset time threshold is 3s, then 2s < 3s, and the task will be removed from the task list. The node matching unit is used to perform intelligent matching analysis on the load status data of regular nodes to facilitate node migration. The specific intelligent matching analysis process for node migration is as follows: The remaining computing power, response latency, and network transmission latency of each regular node are obtained. Based on the remaining computing power, response latency, and network transmission latency, regular nodes are filtered to select those that meet the following criteria: remaining computing power greater than a preset remaining computing power threshold, response latency less than a preset response latency threshold, and network transmission latency less than a preset network transmission latency threshold. These nodes are then set as migrateable nodes. The remaining computing power, response latency, and network transmission latency are respectively obtained within the preset intervals of computing power, response latency, and network transmission latency; The preset intervals corresponding to the remaining computing power are set in ascending order, and the computing power score is set in ascending order. The preset intervals corresponding to the response delay and network transmission delay are both set in ascending order, while the response score and network score are both set in descending order. For example, the lower the delay, the higher the score (e.g., a delay of 10ms gets 25 points, and a delay of 35ms gets 10 points). A feasibility selection score is obtained by weighting and calculating the migrated nodes based on computing power score, response score, and network score. Feasibility selection score = computing power score × β1 + response score × β2 + network score × β3, where β1, β2, and β3 are all greater than zero, and β1 + β2 + β3 = 1. For example, β1 = 0.6, β2 = 0.25, and β3 = 0.15. Based on the feasibility selection score, the migrateable nodes are screened a second time, and a candidate node pool is constructed for the migrateable nodes whose feasibility selection scores are greater than or equal to the preset feasibility selection score threshold. The expected load capacity index Q of the hybrid computing cluster is calculated based on the candidate node pool, after each candidate node receives tasks from the migrated node located in the migrateable task pool. Where Ci represents the computing power utilization rate of the i-th node, CB represents the load change caused by the migration task, CP represents the average load of the cluster after migration, n>0, and ± represents migration out and migration in. The candidate node corresponding to the maximum value in the expected load capacity index Q is set as the optimal target node; This invention can select the optimal target node for different types of migration tasks while ensuring load balancing, taking into account data transmission efficiency and task processing characteristics, thereby minimizing migration costs and ensuring the continuity and real-time performance of detection tasks.

[0018] Example 3: The interval migration evaluation unit is used to perform a comprehensive evaluation and analysis of the migration effect data after the optimal target node migration in each collected speed interval. The specific process of comprehensive evaluation and analysis of interval migration effect is as follows: Based on different speed ranges, the migration effect data after the optimal target node migration is obtained. The migration effect data includes task performance indicators (response latency, task completion rate, etc.), resource utilization indicators (resource utilization rate, node idle rate, etc.), and migration cost indicators (migration time, data transmission volume, etc.). The initial data is obtained by normalizing each indicator in the migration effect data; Input the task performance indicators, resource utilization indicators and migration cost indicators from the initial data into the pre-set task performance scoring model, resource utilization scoring model and migration cost scoring model respectively to obtain the output task performance score, resource utilization score and migration cost score. The migration effectiveness coefficient is obtained by weighting and calculating the score based on task performance, resource utilization, and migration cost. Among them, the migration effect coefficient = task performance score × weight 1 + resource utilization score × weight 2 + migration cost score × weight 3, where weight 1, weight 2 and weight 3 are all greater than zero; The migration effect coefficient of each speed range is judged and processed. The speed range with a migration effect coefficient less than the preset migration effect coefficient threshold is set as the range to be managed, and the speed range with a migration effect coefficient greater than or equal to the preset migration effect coefficient threshold is set as the qualified range. The backend management unit is used to respond to the range to be managed and the qualified range, and to optimize and adjust the acquisition and analysis of the optimal target node of the range to be managed in order to improve the reliability and effectiveness of the optimal target node of the range to be managed. In summary, this invention uses an interval division unit to perceive production line speed in real time and divide it into intervals. The task generation unit dynamically adjusts the number and type of detection items based on different intervals to achieve a balance between detection accuracy and production line efficiency. At the same time, the migration acquisition unit and node matching unit combine the load status of each node in the hybrid computing cluster to perform intelligent task migration and optimal target node matching, avoiding node overload or idleness and improving the utilization rate of computing resources. The interval migration evaluation unit comprehensively evaluates the migration effect of different speed intervals, and the backend management unit provides feedback and early warning adjustments for the intervals to be managed, further ensuring the reliability and effectiveness of system operation.

[0019] The threshold is set for comparative analysis of results to determine whether they are good or bad. The value of the threshold is determined by a combination of large-scale model analysis of sample data and human experience. It can also be adjusted appropriately based on seasonal or common-sense influencing factors. The size of the coefficient is a specific value obtained by quantifying each parameter to facilitate subsequent comparison. The size of the coefficient depends on the amount of sample data and the corresponding operating coefficient initially set by those skilled in the art for each set of sample data; as long as it does not affect the proportional relationship between the parameter and the quantified value.

[0020] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A defect detection collaborative calculation load control system based on the speed of a sports shoe production line, characterized in that, It includes a collaborative control and management center, a section division unit, a task generation unit, a migration acquisition unit, a node matching unit, a section migration evaluation unit, and a backend management unit; The interval division unit is used to collect the real-time operating speed of the current sports shoe production line and perform production line speed division analysis. The obtained low-speed interval, medium-speed interval and high-speed interval are collectively referred to as speed interval. The task generation unit is used to perform interval and joint matching analysis on the retrieved list of detection items, and the resulting low-speed detection task package, medium-speed detection task package and high-speed detection task package are collectively referred to as detection task package; The migration acquisition unit is used to perform migration task management and analysis on the load status data of each collected node, and to include the remaining tasks in the nodes to be migrated into the migrateable task pool. The node matching unit is used to perform intelligent matching analysis on the load status data of regular nodes and set the candidate node corresponding to the maximum value in the expected load capacity index Q as the optimal target node. The interval migration evaluation unit is used to conduct a comprehensive evaluation and analysis of the migration effect data after the optimal target node migration in each speed interval, and to obtain the interval to be managed and the qualified interval.

2. The defect detection collaborative calculation load control system based on the speed of a sports shoe production line according to claim 1, characterized in that, The analysis process for the interval division unit is as follows: Obtain the real-time operating speed of the current athletic shoe production line during its operating period; Call the preset speed range parameters: minimum value v1, maximum value v2; The real-time operating speed is processed for discrimination. If the real-time operating speed is less than the minimum value v1, it is determined to be in the low-speed range. If the minimum value v1 is less than or equal to the real-time operating speed and less than or equal to the maximum value v2, it is determined to be in the medium-speed range. If the real-time operating speed is greater than or equal to the maximum value v2, it is determined to be in the high-speed range.

3. The defect detection collaborative calculation load control system based on the speed of a sports shoe production line according to claim 1, characterized in that, The analysis process of the task generation unit is as follows: Retrieve a list of preset test items from the database: basic test items include sole integrity and upper splicing alignment; enhanced test items include upper material texture analysis and stitch density distribution detection. At the same time, the pre-set proportional coefficients a1 and a2 corresponding to the enhanced detection items in the low-speed range and the medium-speed range are obtained respectively, where a1 > a2 > 0; Based on the shoe model parameters produced by the current sports shoe production line, the unnecessary inspection items for the shoe models produced by the current sports shoe production line are removed, and a subset of inspection items is constructed based on the retained inspection items; The real-time running speed is analyzed to obtain low-speed detection task packages, medium-speed detection task packages, and high-speed detection task packages.

4. The defect detection collaborative calculation load control system based on the speed of a sports shoe production line according to claim 1, characterized in that, The analysis process of the migration acquisition unit is as follows: Real-time monitoring of the load status data of each node in the hybrid computing cluster (composed of edge computing nodes and cloud computing nodes), including real-time computing power utilization and task response latency; The load status data is preprocessed, and the preprocessed load status data is input into the pre-set load prediction model to obtain the predicted load value output by the pre-set load prediction model. The predicted load value is then processed to determine the node to be migrated and the regular node.

5. The defect detection collaborative calculation load control system based on the speed of a sports shoe production line according to claim 4, characterized in that, Obtain the task list of nodes to be migrated. The task list includes detection task packages and the remaining processing time after migration. Remove prohibited migration tasks and uneconomical migration tasks from the task list, and add the remaining tasks in the nodes to be migrated to the migration task pool. Migration is prohibited for tasks whose response latency is less than or equal to a preset response latency threshold. Migration is uneconomical for tasks whose remaining processing time after migration is less than a preset duration threshold.

6. The defect detection collaborative calculation load control system based on the speed of a sports shoe production line according to claim 1, characterized in that, The analysis process of the node matching unit is as follows: obtain the remaining computing power, response latency, and network transmission latency from the load status data of each regular node; filter the regular nodes based on the remaining computing power, response latency, and network transmission latency; select regular nodes that meet the following conditions: remaining computing power greater than a preset remaining computing power threshold, response latency less than a preset response latency threshold, and network transmission latency less than a preset network transmission latency threshold; and set them as migrateable nodes.

7. The defect detection collaborative calculation load control system based on the speed of a sports shoe production line according to claim 6, characterized in that, Obtain the computing power score, response score, and network score within the preset range corresponding to the remaining computing power, response latency, and network transmission latency; A feasibility selection score is obtained by weighting and calculating the migrated nodes based on computing power score, response score, and network score. Based on the feasibility selection score, a secondary screening of migrateable nodes is performed to construct a candidate node pool for migrateable nodes whose feasibility selection scores are greater than or equal to a preset feasibility selection score threshold.

8. The defect detection collaborative calculation load control system based on the speed of a sports shoe production line according to claim 7, characterized in that, The expected load capacity index Q of the hybrid computing cluster is calculated based on the candidate node pool, after each candidate node receives tasks from the migrated node located in the migrateable task pool. The candidate node corresponding to the maximum value in the expected load capacity index Q is set as the optimal target node.

9. The defect detection collaborative calculation load control system based on the speed of a sports shoe production line according to claim 1, characterized in that, The analysis process of the interval migration assessment unit is as follows: Based on different speed ranges, the migration effect data after the optimal target node migration is obtained. The migration effect data includes task performance indicators, resource utilization indicators, and migration cost indicators. The initial data is obtained by normalizing each indicator in the migration effect data; Input the task performance indicators, resource utilization indicators and migration cost indicators from the initial data into the pre-set task performance scoring model, resource utilization scoring model and migration cost scoring model respectively to obtain the output task performance score, resource utilization score and migration cost score. The migration effect coefficient is obtained by weighting and calculating the migration performance score, resource utilization score and migration cost score. The migration effect coefficient of each speed range is then processed to determine the range to be managed and the qualified range.

Images