Distributed Quality Inspection Task Dynamic Scheduling Method Based on Internet and Artificial Intelligence

Abstract

This application relates to the technical field of quality inspection tasks, and provides a dynamic scheduling method for distributed quality inspection tasks based on the Internet and artificial intelligence. The dynamic scheduling method for distributed quality inspection tasks based on the Internet and artificial intelligence includes: Step S1: Preprocess the core features through a normalization algorithm and a one-hot encoding algorithm; Step S2: Calculate the probability P1 of the quality inspector matching the task and the local matching probability P2, and calculate the final matching probability corresponding to P1 and P2; Step S3: Set a genetic algorithm; Step S4: Based on the final matching probability in Step S2 and the genetic algorithm in Step S3, assign the quality inspection task to the quality inspector, and the quality inspector performs quality inspection processing on the quality inspection task to generate a quality inspection task result; Step S5: Based on the real busy index and the personnel ability portrait model, perform feedback adjustment on the assignment schemes in Steps S2 and S3. This application realizes precise matching in the scenario of quality inspection task assignment for structured data extraction, and enhances the interpretability of the system.

Description

Technical Field

[0001] This application relates to the technical field of quality inspection tasks, and specifically relates to a dynamic scheduling method for distributed quality inspection tasks based on the Internet and artificial intelligence. Background Art

[0002] In the scenario of quality inspection task allocation for structured data extraction, the current mainstream methods mainly include two modes: static automatic allocation based on the idle status / level of personnel and self-claiming in the task pool. However, both of them have significant drawbacks: 1. Common defects: Both methods lack a closed-loop feedback mechanism and cannot dynamically optimize the allocation strategy according to the historical task execution quality data, resulting in the repeated occurrence of the same type of errors; 2. Limitations of static allocation: The automatic allocation mode relying on fixed rules, due to the lack of quantitative evaluation of the matching relationship between personnel skills and task complexity, is prone to ability mismatch and low resource utilization; 3. Deficiencies of the claiming mode: The self-claiming mechanism is prone to uneven task allocation (such as complex tasks remaining unclaimed), and additional manpower is required to verify the task integrity, resulting in an increase in management costs. Summary of the Invention

[0003] In order to help solve the above technical problems, this application provides a dynamic scheduling method for distributed quality inspection tasks based on the Internet and artificial intelligence, and adopts the following technical solutions:

[0004] A dynamic scheduling method for distributed quality inspection tasks based on the Internet and artificial intelligence, wherein the dynamic scheduling method for distributed quality inspection tasks based on the Internet and artificial intelligence includes:

[0005] Step S1: Screen the input features through the LASSO regression algorithm to obtain the core features, and preprocess the core features through the normalization algorithm and the one-hot encoding algorithm;

[0006] Step S2: Calculate the matching probability P1 between the quality inspector and the task and the local matching probability P2 according to the core features after Step S1 through the logistic regression model and the KNN model, and calculate the final matching probability corresponding to P1 and P2 through the SHAP algorithm;

[0007] Step S3: Set the genetic algorithm, the genetic algorithm includes an objective function, the core features include skill matching degree, urgency, and load pressure, and the objective function is Maximize∑(the weight of the first objective function • skill matching degree + the weight of the second objective function • urgency - the weight of the third objective function • load pressure);

[0008] Step S4: Allocate the quality inspection task to the quality inspector based on the final matching probability in Step S2 and the genetic algorithm in Step S3, and the quality inspector conducts quality inspection processing on the quality inspection task to generate the quality inspection task result;

[0009] Step S5: Based on the real busy-idle index and the personnel ability portrait model, feedback adjustment is performed on the allocation schemes in steps S2 and S3, including:

[0010] Calculate the real busy-idle index based on the analysis of the quality inspector's terminal behavior. The terminal behavior analysis includes active duration and task switching frequency. The real busy-idle index = the first real busy-idle index weight × (active duration / total working duration) + the second real busy-idle index weight × (1 - task switching frequency / 10). The real busy-idle index is used to represent the work load situation.

[0011] Set up the personnel ability portrait model, including introducing comprehensive skill tags, historical task quality scores, and real-time monitoring response speed and matching them with core features. The comprehensive skill tags, historical task quality scores, and real-time monitoring response speed are automatically updated through the quality inspection task results.

[0012] Based on the real busy-idle index, the personnel ability portrait model, the final matching probability in step S2, and the genetic algorithm in step S3, assign the quality inspection tasks to quality inspectors. The quality inspectors process the quality inspection tasks, generate quality inspection task results, and conduct result acceptance on the quality inspection task results to obtain task acceptance results.

[0013] Preferably, the distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence further includes step S6: Based on the task acceptance results, dynamically adjust the target function weight parameters in step S3 through the RL reinforcement learning algorithm.

[0014] The RL reinforcement learning algorithm includes state design, action design, and reward design.

[0015] The state design includes:

[0016] The current task queue, including the list of quality inspection tasks to be assigned and the core features corresponding to each quality inspection task.

[0017] Personnel load, including the current workload of each quality inspector. The current workload includes the number of assigned quality inspection tasks and the estimated completion time.

[0018] Skill matching degree distribution, indicating the skill matching degree between quality inspectors and quality inspection tasks.

[0019] The action design includes adjusting the target function weight.

[0020] The reward design includes calculating the reward according to the task acceptance results. The task acceptance results include task completion timeliness scores and error rates. The reward = task completion timeliness score × the first reward weight + (1 - error rate) × the second reward weight.

[0021] Preferably, the distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence further includes step S7: Based on the task acceptance result, if the error rate > the preset error rate threshold, then reduce the score of the comprehensive skill label of the personnel ability portrait model corresponding to this quality inspector in step S5.

[0022] Preferably, the distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence further includes step S8: Set up a task board, which is used to display the task progress and the personnel load heat map in real time, generate a reason explanation for the allocation of each task, and the task board includes:

[0023] A global view, which is used to display the load of each work station through a real-time heat map,

[0024] A detailed panel design, which is used to click on the quality inspection task to view the allocation reason.

[0025] Preferably, step S1 includes:

[0026] Take the task characteristics, personnel characteristics and real-time status as input features, screen the input features through LASSO regression to obtain core features, the core features include continuous features and categorical features, convert the continuous features with different dimensions to the same dimension through the normalization algorithm, and convert the categorical features to numerical features through the one-hot encoding algorithm.

[0027] Preferably, step S2 includes:

[0028] Input the preprocessed core features into the logistic regression model, output the probability P1 of the quality inspector matching the task, find n historical quality inspection tasks that match the current core features based on the historical core features through the KNN model, calculate the local matching probability P2 of the quality inspector in the n historical quality inspection tasks, take the probability P1 of the quality inspector matching the task and the local matching probability P2 of the quality inspector as inputs, calculate the contribution degrees of P1 and P2 to the final prediction result through the SHAP algorithm, and assign weights according to the contribution degrees, and generate the final matching probability according to the weights respectively, where n is a positive integer.

[0029] Preferably, step S3 includes:

[0030] Gene coding design: Each chromosome represents an allocation plan,

[0031] Genetic operation design:

[0032] Crossover operation, including selecting two parent plans and exchanging part of the task allocation segments of the two parent plans,

[0033] Mutation operation, including randomly replacing the quality inspector of a certain quality inspection task,

[0034] Constraints, including the requirement that the number of tasks processed by the same quality inspector at the same time ≤ the threshold of the number of tasks processed at the same time, and tasks with an urgency greater than the urgency threshold are assigned within a preset time.

[0035] Preferably, step S5 includes: the weight of the first true busy - idle index is 0.6, the weight of the second true busy - idle index is 0.4, and the total working hours is 8 hours.

[0036] Preferably, step S7 includes: based on the task acceptance result, if task rework is triggered due to skill mismatch, then trigger the correction of the allocation strategy; if task rework is triggered due to defects in the detection process, then notify the process engineer to optimize the SOP.

[0037] Preferably, step S8 includes: when a quality inspector continuously processes a urgent tasks, it is automatically marked as having an overload risk; when a sudden batch of quality inspection tasks enters, the skill tags that need to be preferentially assigned are highlighted, where a is a positive integer.

[0038] In summary, the method of the present invention realizes precise matching in the scenario of structured data extraction quality inspection task allocation through a three - layer architecture of data - driven modeling, multi - objective dynamic optimization, and human - machine collaborative decision - making. At the same time, it reduces mismatching and resource waste, can flexibly respond to complex dynamic environments, and enhances the system interpretability and user trust. This technical effect is universal in fields such as data quality inspection and data operation and maintenance management, and provides a standardized and extensible solution for the multi - node task allocation problem. Brief Description of the Drawings

[0039] Figure 1 It is a schematic block diagram of an embodiment of a distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence of the present application. Detailed Embodiment

[0040] The following further illustrates the present application with reference to the drawings. The structure and principle of the present application are very clear to those skilled in the art. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0041] Figure 1 It is a schematic block diagram of an embodiment of a distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence of the present application. Figure 1"Task reception" in it corresponds to step S1, "algorithm" corresponds to steps S2 and S3, "task distribution", "task processing" and "task result" correspond to step S4, "personnel ability portrait", "workload situation" and "result acceptance" correspond to step S5, the connection between "result acceptance" and "algorithm" corresponds to step S6, the connection between "result acceptance" and "personnel ability portrait" corresponds to step S7, "human load heat map" and "task progress display" correspond to step S8.

[0042] Reinforcement Learning (RL) is a paradigm of machine learning. Its core idea is that through trial-and-error and reward mechanisms, an agent can gradually learn and optimize its decision-making strategy in the interaction with the environment to achieve specific goals.

[0043] Genetic Algorithm (GA) is an optimization algorithm that simulates natural selection and genetic mechanisms. It searches for the optimal solution or approximate optimal solution in the solution space by simulating operations such as natural selection, crossover, and mutation.

[0044] A distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence in this application includes:

[0045] Step S1: Screen the input features through the LASSO regression algorithm to obtain the core features, and preprocess the core features through the normalization algorithm and one-hot encoding algorithm. In step S1, task features, personnel features, and real-time status are used as input features. The input features are screened through LASSO regression to obtain the core features. The core features include continuous features and categorical features. The continuous features with different dimensions are converted to the same dimension through the normalization algorithm, and the categorical features are converted to numerical features through the one-hot encoding algorithm.

[0046] Step S2: Calculate the probability P1 of the quality inspector matching the task and the local matching probability P2 according to the core features after step S1 through the logistic regression model and the KNN model, and calculate the final matching probability corresponding to P1 and P2 through the SHAP algorithm. In step S2, the preprocessed core features are input into the logistic regression model, and the probability P1 of the quality inspector matching the task is output. The KNN model finds n historical quality inspection tasks that match the current core features based on the historical core features, and calculates the local matching probability P2 of the quality inspector in the n historical quality inspection tasks. The probability P1 of the quality inspector matching the task and the local matching probability P2 of the quality inspector are used as inputs, and the contribution degrees of P1 and P2 to the final prediction result are calculated through the SHAP algorithm, and weights are assigned according to the contribution degrees, and the final matching probability is generated respectively according to the weights. n is a positive integer.

[0047] Step S3: Set up the genetic algorithm. The genetic algorithm includes an objective function, and the core features include skill matching degree, urgency, and load pressure. The objective function is Maximize∑(the weight of the first objective function • skill matching degree + the weight of the second objective function • urgency - the weight of the third objective function • load pressure). In step S3, it also includes gene coding design, genetic operation design, and constraint conditions. Gene coding design: Each chromosome represents an allocation plan. Genetic operation design includes crossover operation and mutation operation. The crossover operation includes selecting two parental plans and exchanging partial task allocation segments of the two parental plans. The mutation operation includes randomly replacing the quality inspector of a certain quality inspection task. The constraint conditions include requiring that the number of tasks processed by the same quality inspector at the same time ≤ the threshold of the number of tasks processed at the same time, and tasks with an urgency greater than the urgency threshold are allocated within a preset time.

[0048] Step S4: Allocate the quality inspection tasks to the quality inspectors based on the final matching probability in step S2 and the genetic algorithm in step S3. The quality inspectors conduct quality inspection on the quality inspection tasks and generate quality inspection task results.

[0049] Step S5: Conduct feedback adjustment on the allocation plans in steps S2 and S3 based on the real busy and idle index and the personnel ability portrait model, including:

[0050] Calculate the real busy and idle index based on the analysis of the quality inspector's terminal behavior. The terminal behavior analysis includes active duration and task switching frequency. The real busy and idle index = the weight of the first real busy and idle index × (active duration / total working duration) + the weight of the second real busy and idle index × (1 - task switching frequency / 10). The real busy and idle index is used to represent the working load situation.

[0051] Set up the personnel ability portrait model, including introducing comprehensive skill tags, historical task quality scores, and real-time monitoring response speed and matching them with the core features. The comprehensive skill tags, historical task quality scores, and real-time monitoring response speed are automatically updated through the quality inspection task results.

[0052] Allocate the quality inspection tasks to the quality inspectors based on the real busy and idle index, the personnel ability portrait model, the final matching probability in step S2, and the genetic algorithm in step S3. The quality inspectors process the quality inspection tasks, generate quality inspection task results, and conduct result acceptance on the quality inspection task results to obtain task acceptance results. In step S5, the weight of the first real busy and idle index is 0.6, the weight of the second real busy and idle index is 0.4, and the total working duration is 8 hours.

[0053] Step S6: Based on the task acceptance result, dynamically adjust the weight parameters of the objective function in Step S3 through the RL reinforcement learning algorithm. The RL reinforcement learning algorithm includes state design, action design, and reward design. The state design includes the current queue and personnel load. The current task queue includes the list of quality inspection tasks to be assigned and the core features corresponding to each quality inspection task. The personnel load includes the current workload of each quality inspector. The current workload includes the number of assigned quality inspection tasks and the estimated completion time. The skill matching degree distribution represents the skill matching degree between the quality inspector and the quality inspection task. The action design includes adjusting the weight of the objective function. The reward design includes calculating the reward based on the task acceptance result. The task acceptance result includes the task completion timeliness score and the error rate. Reward = task completion timeliness score × first reward weight + (1 - error rate) × second reward weight.

[0054] Step S7: Based on the task acceptance result, if the error rate > the preset error rate threshold, then reduce the score of the comprehensive skill label of the personnel ability portrait model in Step S5 corresponding to this quality inspector.

[0055] In Step S7, based on the task acceptance result, if task rework is triggered due to skill mismatch, then trigger the correction of the allocation strategy. If task rework is triggered due to defects in the detection process, then notify the process engineer to optimize the SOP. SOP refers to Standard Operating Procedure, that is, the standard operation procedure or standard operating procedure. It is a detailed guidance document used to specify the standard steps, methods, and requirements for carrying out specific activities within an organization.

[0056] Step S8: Set up a task dashboard. The task dashboard is used to display the task progress and the personnel load heat map in real time, and generate a reason explanation for the allocation of each task. The task dashboard includes a global view and a detailed panel design. The global view is used to display the load of each work station through a real-time heat map. The detailed panel design is used to view the allocation reason by clicking on the quality inspection task.

[0057] In Step S8, when a certain quality inspector continuously processes a emergency tasks, it is automatically marked as having an overload risk. When a sudden batch of quality inspection tasks enters, the skill labels that need to be prioritized for allocation are highlighted. a is a positive integer.

[0058] Specifically, the scenario background: A structured data extraction quality inspection system needs to complete multi-link quality inspections of more than 2,000 tasks per day. The following are the existing pain points:

[0059] 1. Uneven allocation: Highly skilled quality inspectors are frequently assigned simple tasks, resulting in the backlog of complex tasks.

[0060] 2. Delay in emergency tasks: The quality inspection response to sudden emergency tasks is slow, affecting the delivery timeliness.

[0061] 3. Load imbalance: Some quality inspectors are overloaded with work, while others are idle for a long time due to skill mismatches.

[0062] 4. High error rate: Manual allocation relies on experience, and skill mismatches result in a missed inspection rate as high as 8%.

[0063] This application includes:

[0064] 1. Data preprocessing is a crucial step in machine learning tasks. It involves cleaning, transforming, and normalizing raw data to provide high-quality data for subsequent model training. In the above solution, the data preprocessing in step 1 mainly includes feature selection using LASSO regression, normalization of continuous features, and one-hot encoding of categorical features.

[0065] This application uses LASSO regression for feature selection:

[0066] LASSO (Least Absolute Shrinkage and Selection Operator) regression is a linear regression method. By adding an L1 regularization term to the loss function, it can make the coefficients of some unimportant features become 0, thus achieving the purpose of feature selection.

[0067] In this solution, task features (such as complexity, urgency, business type), personnel features (such as skill tags, historical quality inspection accuracy, average response speed), and real-time status (such as the number of tasks assigned on the same day, terminal operation logs) are used as input features. Through LASSO regression, the most critical features for task allocation can be screened out, such as task complexity, urgency, personnel skill matching degree, historical accuracy, etc. Through feature selection, the interference of noise features can be reduced, the complexity of the model can be lowered, and the training efficiency and generalization ability of the model can be improved.

[0068] This application normalizes continuous features:

[0069] Normalization is to transform features with different dimensions into the same dimension so that they are comparable. For continuous features, such as task complexity and urgency, their value ranges may be different, and directly inputting them into the model may affect the performance of the model. Using the method of linear transformation, the value ranges of continuous features are scaled to between [0,1]. Normalization can make the model more stable and improve the convergence speed and accuracy of the model.

[0070] This application performs one-hot encoding on categorical features:

[0071] Categorical features, such as business type (interim announcement, prospectus, interest rate bond), are non-numerical features and cannot be directly input into the model. One-hot encoding is a method of converting categorical features into numerical features. For each categorical feature, a binary vector equal to the number of its categories is created. For example, for the business type feature, a 3-dimensional binary vector can be created, corresponding to interim announcement, prospectus, and interest rate bond respectively. If the business type of a certain task is an interim announcement, the corresponding vector representation is [1, 0, 0]. One-hot encoding can enable categorical features to be correctly processed by the model and improve the accuracy of the model.

[0072] Step 1 compresses the feature dimension from 15 dimensions to 8 dimensions, reduces noise interference, and improves the model training efficiency.

[0073] 2. Logistic regression is a widely used linear model suitable for handling linearly separable tasks. In this solution, logistic regression is used to predict whether a quality inspector is suitable for handling a specific type of task.

[0074] The input features include preprocessed task features (such as complexity, urgency) and personnel features (such as skill match degree, historical accuracy rate). The model output is the probability P1 of the quality inspector matching the task, indicating the suitability of the quality inspector for handling this task. KNN (K-Nearest Neighbors) is a non-parametric model used to capture the local similarity between data points. In this solution, KNN is used to find the past tasks most similar to the current task requirements, thereby evaluating the local matching ability of the quality inspector. Select k = 5, that is, find the 5 past tasks most similar to the current task requirements based on historical data. Calculate the local matching probability P2 of the quality inspector in these 5 tasks, indicating the performance of the quality inspector when handling similar tasks. SHAP (SHapley Additive exPlanations) is a method for explaining the predictions of machine learning models. In this solution, the SHAP explanation layer is used to fuse the prediction results of logistic regression and KNN to generate an interpretable weight assignment. Take the probability P1 of the quality inspector matching the task and the local matching probability P2 as inputs, and calculate the contribution degree of each feature to the final prediction result through the SHAP method. Assign weights according to the contribution degree, such as 60% contribution from logistic regression and 40% contribution from KNN, to generate the final matching probability.

[0075] The personnel ability profile is used to comprehensively describe the ability characteristics of quality inspectors, including dimensions such as skill tags, historical task quality scores, response speed, etc. This application sets comprehensive skill tags, such as "Text Parsing Expert", "Numeric Verification Expert", etc., which are automatically updated according to the historical performance of quality inspectors. The historical task quality score is introduced to reflect the performance of quality inspectors in past tasks, and real-time status characteristics such as response speed are monitored in real time to update the personnel ability profile. The real busy-idle index is used to reflect the real-time working status of quality inspectors and avoid overloading or idleness. This application calculates the busy-idle index through terminal behavior analysis, such as active duration, task switching frequency, etc. Busy-idle index = 0.6×(active duration / 8 hours) + 0.4×(1 - task switching frequency / 10), and the higher the value, the busier it indicates.

[0076] Step 2 improves the task matching accuracy from 70% to 92%, and the rework rate caused by mismatching decreases by 40%.

[0077] 3. Through the genetic algorithm, gradually generate a better task allocation plan to maximize resource utilization, improve the response speed of urgent tasks, and ensure the rationality of task allocation.

[0078] Objective function: Maximize ∑(w1•skill matching degree + w2•urgency - w3•load pressure).

[0079] In this application, the objective function is specifically formulated as Maximize ∑(0.5•skill matching degree + 0.3•urgency - 0.2•load pressure). Here, w1, w2, and w3 represent the weights of skill matching degree, urgency, and load pressure respectively. By adjusting these weights, different aspects of the allocation plan can be optimized. The higher the skill matching degree, the more the task matches the skills of the quality inspector; the higher the urgency, the faster the task needs to be processed; and the smaller the load pressure, the smaller the current workload of the quality inspector.

[0080] Coding method: Each chromosome represents an allocation plan, that is, a mapping relationship from a task to a quality inspector. For example, [Task 1 → Person A, Task 2 → Person C,...] represents a possible allocation plan.

[0081] Crossover: Select two parent plans and exchange some of their task allocation segments.

[0082] Mutation: Randomly replace the assigned person for a certain task, and preferably select a quality inspector with a high skill matching degree and a low load.

[0083] Constraint conditions: The number of tasks processed by the same quality inspector at the same time ≤ 3: Avoid overloading the quality inspector and ensure the rationality of task allocation. Tasks with an urgency greater than 3 must be allocated within 30 minutes: Ensure that urgent tasks can be processed in a timely manner.

[0084] Step 3 increases the resource utilization rate from 65% to 85% and shortens the average response time for urgent tasks to 15 minutes.

[0085] 4. Use the RL reinforcement learning method to dynamically adjust the weight parameters of the allocation strategy.

[0086] Status: Current task queue: including the list of tasks to be allocated, and the characteristics of each task (such as complexity, urgency, business type, etc.). Personnel load: the current workload of each quality inspector, including the number of tasks already allocated, the estimated completion time, etc. Skill matching degree distribution: the degree of skill matching between quality inspectors and tasks, calculated based on the personnel ability profile and skill tags.

[0087] Action: Adjust the weights of the objective function: Dynamically adjust the weight parameters (such as w1, w2, w3) in the objective function according to the current status to optimize the allocation strategy.

[0088] Reward: Calculate the reward based on the task completion timeliness and error rate to encourage the model to make better allocation decisions. The reward function can be designed as: Reward = Completion timeliness score × 0.7 + (1 - Error rate) × 0.3, where the completion timeliness score and error rate are calculated based on the actual task acceptance results.

[0089] Closed-loop feedback mechanism: Quality feedback: After task acceptance, if the error rate exceeds the set threshold (such as 5%), then reduce the score of the corresponding skill tag of this quality inspector to reflect their performance in actual tasks. Abnormal backtracking: Automatically analyze tasks with frequent rework, identify the root cause and take corresponding measures. If rework is caused by skill mismatch, such as assigning a non-"temporary announcement expert" to handle a difficult temporary announcement, then trigger the correction of the allocation strategy and optimize the calculation of the skill matching degree. If rework is caused by defects in the detection process, such as the lack of specific instruments or steps, then notify the process engineer to optimize the standard operating procedure (SOP).

[0090] Step 4 shortens the system's adaptive adjustment cycle from 2 weeks to 2 days, and the accuracy rate of identifying process defects reaches 90%.

[0091] 5. Task dashboard design:

[0092] Global view: A real-time heat map shows the load of each work station (red for high load, green for idle).

[0093] Details panel: Click on a task to view the allocation reason (such as "Select Engineer Zhang: Skill matching degree 95% + Historical accuracy rate 98%").

[0094] Early warning system:

[0095] When a quality inspector continuously processes 3 urgent tasks, it is automatically marked as "overload risk", and it is recommended that the manager intervene.

[0096] When sudden batch tasks enter, the skill tags that need to be preferentially assigned are prominently displayed on the kanban (such as "urgently need temporary announcement experts").

[0097] Effect: The efficiency of manager intervention is increased by 50%, and the employee satisfaction with the fairness of assignment rises from 60% to 88%.

[0098] In summary, through the implementation of the technical solution of this application in a structured data extraction quality detection system, the following have been achieved:

[0099] 1. Precise dynamic matching: multi-model fusion + GA optimization to reduce skill mismatches and load imbalances.

[0100] 2. Agile response: RL dynamic parameter adjustment + visual kanban to quickly respond to emergency orders and abnormal events.

[0101] 3. Continuous self-optimization: a closed-loop feedback mechanism promotes system iteration and gradually approaches the Pareto optimum.

Claims

1. A distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence, characterized in that: The distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence includes: Step S1: The input features are screened by the LASSO regression algorithm to obtain the core features, and the core features are preprocessed by the normalization algorithm and the one-hot encoding algorithm; Step S2: Calculate the probability P1 of the quality inspector matching with the task and the local matching probability P2 based on the core features in step S1 through the logistic regression model and the KNN model, and calculate the final matching probabilities corresponding to P1 and P2 through the SHAP algorithm; Step S3: Setting a genetic algorithm, the genetic algorithm includes an objective function, the core features include skill matching, urgency and load pressure, and the objective function is Maximize∑(first objective function weight • skill matching + second objective function weight • urgency − third objective function weight • load pressure); Step S4: Based on the final matching probability of step S2 and the genetic algorithm of step S3, the quality inspection task is assigned to the quality inspector, and the quality inspector performs quality inspection on the quality inspection task and generates the quality inspection task result; Step S5: Feedback adjustment is performed on the allocation schemes of steps S2 and S3 based on the real busy-idle index and the personnel capability profile model, including: The real busy-idle index is calculated based on the terminal behavior analysis of the quality inspector. The terminal behavior analysis includes active time and task switching frequency. The real busy-idle index = the first real busy-idle index weight × (active time / total working time) + the second real busy-idle index weight × (1-task switching frequency / 10). The real busy-idle index is used to indicate the workload. Set up a personnel capability portrait model, including introducing comprehensive skill labels, historical task quality scores, and real-time monitoring response speed and matching them with core features. Comprehensive skill labels, historical task quality scores, and real-time monitoring response speed are automatically updated through quality inspection task results. Based on the real busy-idle index, the personnel capability portrait model, the final matching probability of step S2 and the genetic algorithm of step S3, the quality inspection task is assigned to the quality inspector. The quality inspector processes the quality inspection task, generates the quality inspection task result, and performs acceptance on the quality inspection task result to obtain the task acceptance result.

2. The method for dynamic scheduling of distributed quality inspection tasks based on the Internet and artificial intelligence according to claim 1 is characterized in that: The distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence also includes step S6: based on the task acceptance result, dynamically adjusting the objective function weight parameter of step S3 through the RL reinforcement learning algorithm, RL reinforcement learning algorithm includes state design, action design and reward design. State design includes: The current task queue, including the list of quality inspection tasks to be assigned and the core features corresponding to each quality inspection task; Personnel load, including the current workload of each quality inspector. The current workload includes the number of assigned quality inspection tasks and the estimated completion time. Skill matching distribution, which indicates the matching degree between the quality inspectors and the quality inspection tasks. Action design involves adjusting the objective function weights, The reward design includes calculating the reward based on the task acceptance results. The task acceptance results include the task completion time score and the error rate. The reward = task completion time score × the first reward weight + (1-error rate) × the second reward weight.

3. The distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence according to claim 2 is characterized in that: The distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence also includes step S7: based on the task acceptance result, if the error rate is greater than the preset error rate threshold, the score of the comprehensive skill label of the personnel capability portrait model in step S5 corresponding to the quality inspector is reduced.

4. The distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence according to claim 3 is characterized in that: The distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence also includes step S8: setting a task board, which is used to display the task progress and personnel load heat map in real time, and generate a reason explanation for the allocation of each task. The task board includes: Global view, used to display the load of each workstation through real-time heat map, The details panel is designed to click on the quality inspection task to view the assignment reason.

5. The distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence according to claim 1 is characterized in that: The step S1 comprises: Task characteristics, personnel characteristics and real-time status are taken as input features. LASSO regression is used to screen the input features to obtain core features. The core features include continuous features and categorical features. The normalization algorithm is used to convert continuous features of different dimensions to the same dimension, and the one-hot encoding algorithm is used to convert categorical features into numerical features.

6. The distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence according to claim 1 is characterized in that: The step S2 comprises: The preprocessed core features are input into the logistic regression model, and the probability P1 of the quality inspector matching the task is output. The KNN model is used to find n historical quality inspection tasks that match the current core features based on the historical core features, and the local matching probability P2 of the quality inspector in the n historical quality inspection tasks is calculated. The probability P1 of the quality inspector matching the task and the local matching probability P2 of the quality inspector are used as input, and the contribution of P1 and P2 to the final prediction result are calculated through the SHAP algorithm, and weights are assigned according to the contribution. The final matching probabilities are generated according to the weights, and n is a positive integer.

7. The distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence according to claim 1 is characterized in that: The step S3 comprises: Genetic coding design: Each chromosome represents a distribution scheme, Genetic manipulation design: The crossover operation includes selecting two parent solutions and exchanging some of the task allocation segments of the two parent solutions. Mutation operations include randomly replacing the inspectors of a certain quality inspection task. The constraints include requiring the same quality inspector to handle the same number of tasks at the same time ≤ the threshold of the number of tasks to be handled at the same time, and tasks with an urgency greater than the threshold of the urgency are allocated within a preset time.

8. The distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence according to claim 1 is characterized in that: The step S5 includes: the weight of the first real busy-idle index is 0.6, the weight of the second real busy-idle index is 0.4, and the total working time is 8 hours.

9. The distributed quality inspection task dynamic scheduling method based on the Internet and artificial intelligence according to claim 3 is characterized in that: The step S7 includes: based on the task acceptance result, if the task rework is triggered due to skill mismatch, triggering the allocation strategy correction; if the task rework is triggered due to detection process defects, notifying the process engineer to optimize the SOP.

10. The method for dynamic scheduling of distributed quality inspection tasks based on the Internet and artificial intelligence according to claim 4 is characterized in that: The step S8 includes: when a quality inspector continuously handles a urgent tasks, he is automatically marked as having an overload risk; when a sudden batch quality inspection task comes in, the skill label that needs to be assigned first is highlighted, and a is a positive integer.

Images