Production line operation monitoring system based on digital twinning technology

Abstract

The invention discloses a production line operation monitoring system based on a digital twin technology, and the system comprises a data collection module which is used for collecting the data of each task in a production line; the scheduling unit construction module is used for mapping each task instance in the virtual task relationship into a cat-shaped scheduling unit and initializing a role field based on a preset threshold value; the scheduling game engine module is used for executing a combat cat algorithm on the game model and generating a scheduling scheme set comprising a plurality of candidate scheduling schemes; the simulation evaluation module is used for executing simulation evaluation on each candidate scheduling scheme in the digital twin environment to obtain a corresponding task completion rate, a resource utilization rate and a system load balance degree; and the strategy management module is used for sorting the candidate scheduling schemes according to the simulation evaluation result and screening the current optimal scheduling scheme. According to the invention, intelligent production task scheduling, real-time response and strategy adaptive optimization are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of production line operation monitoring, and in particular to a production line operation monitoring system based on digital twin technology. Background Art

[0002] Currently, the manufacturing industry is accelerating its transformation towards intelligent and digital directions. The real-time monitoring and optimal scheduling of production lines have become key links in achieving flexible manufacturing and efficient operation and maintenance. To meet the requirements of refined management and dynamic control, digital twin technology has been widely introduced into the manufacturing field to construct a virtual mapping of physical production systems, enabling state visualization, behavior prediction, and system collaboration driven by data. Existing production line monitoring systems mainly rely on traditional data collection and rule control models. Although they can achieve basic equipment status monitoring and alarm functions, in the face of scenarios such as complex multi-task scheduling, multi-resource constraints, and emergency handling, their response capabilities are insufficient, the scheduling schemes are difficult to adapt dynamically, and there is still much room for improvement in the overall operation efficiency and robustness of the system.

[0003] Traditional scheduling methods mostly adopt fixed priority rules or heuristic algorithms such as genetic algorithms and particle swarm algorithms. Although they have certain optimization capabilities, in the actual production environment with frequent changes in task priorities, intense resource competition, and variable dynamic constraint conditions, problems such as scheduling deadlocks, inability to coordinate resource conflicts, and delays in emergency handling often occur. At the same time, most systems lack the ability to model the interdependencies between tasks and are difficult to make reasonable scheduling decisions based on the actual urgency of tasks and the status of system resources. In addition, existing methods often use static evaluation indicators in simulation evaluation and scheduling strategy selection, lacking the dynamic optimization ability driven by real-time operation state feedback, resulting in the disconnection between scheduling results and actual production and the lack of pertinence and adaptability in strategy selection.

[0004] In the face of the above problems, this paper proposes a production line operation monitoring system based on digital twin technology. The system constructs a virtual mapping of production tasks and resource status, combines a multi-round game evolution process, and establishes a dynamic scheduling and evaluation closed-loop mechanism. A swarm intelligence-like modeling mechanism is introduced into the system. Tasks are mapped to cat-type scheduling units, which are given behavioral roles and scheduling attributes. Through the battle cat algorithm, multi-round simulation games are carried out to realize the evolutionary optimization of scheduling strategies. The system not only considers key factors such as task priorities, resource request quantities, and execution status, but also constructs a complete behavioral response mechanism for complex situations such as support behaviors, role switching, and resource release. Based on the candidate strategies output by the scheduling game process, the system uses the digital twin platform for simulation evaluation and selects the scheduling scheme with the optimal resource utilization rate, task completion rate, and system load balance, thereby significantly improving the adaptability and real-time performance of the scheduling strategy.

[0005] Therefore, how to provide a production line operation monitoring system based on digital twin technology is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0006] An object of the present invention is to propose a production line operation monitoring system based on digital twin technology, which uses the Battle Cat algorithm to perform intelligent modeling and game evolution on the multi-task scheduling process, and combines the digital twin platform to realize virtual simulation and dynamic evaluation of task status, resource allocation and policy results. The present invention fully integrates digital twin modeling, intelligent agent role division, construction of non-linear priority functions and game-based resource scheduling methods, details the construction mechanism of cat-type scheduling units, the generation process of multi-round evolution strategies and the strategy optimization method based on simulation feedback, and has the advantages of fast scheduling response, high resource utilization rate and strong strategy adaptability.

[0007] A production line operation monitoring system based on digital twin technology according to an embodiment of the present invention includes: A data acquisition module for collecting task identifiers, resource requirements, estimated execution times and dependencies of each task in the production line; A scheduling unit construction module for mapping each task instance to a cat-type scheduling unit, the cat-type scheduling unit including task attributes, resource request amounts, priority values and role fields, and initializing the role fields based on a preset threshold; A scheduling game engine module for executing the Battle Cat algorithm on a task scheduling game model composed of cat-type scheduling units, performing multi-round task scheduling evolution according to the behavior rules of the role fields, and generating a scheduling plan set including multiple candidate scheduling plans; A simulation evaluation module for performing simulation evaluation on each candidate scheduling plan in the digital twin environment to obtain the corresponding task completion rate, resource utilization rate and system load balance degree; A policy management module for sorting the candidate scheduling plans according to the simulation evaluation results and screening the current optimal scheduling plan.

[0008] Optionally, the modules are implemented by the following method: S1. Collect task data in the current production line, including task identifiers, resource requirements, estimated execution times and dependencies, and establish corresponding virtual task relationships based on digital twin technology; S2. Map each task instance in the virtual task relationship to a cat-type scheduling unit, the cat-type scheduling unit including task attributes, resource request amounts, priority values and role fields; S3. Initialize the role fields of the cat-type scheduling units according to the task data, and set them to attack type, support type or defense type; S4. Build a task scheduling game model based on all cat - type scheduling units, and set role - behavior rules, including that the attack - type preferentially acquires resources, the support - type coordinates conflicts, and the defense - type undertakes delayed and redundant tasks; S5. Run the battle cat algorithm, simulate the game process among cat - type scheduling units, perform several rounds of resource competition and strategy evolution, and output several candidate scheduling schemes; S6. Conduct simulation evaluation on the candidate scheduling schemes, calculate the task completion rate, resource utilization rate, and system load balance degree, and select the scheme with the best score as the optimal scheduling scheme; S7. Apply the optimal scheduling scheme to the production line operation monitoring system, and during the operation, trigger the reset of the role field according to emergencies, and re - execute steps S4 to S6 to achieve adaptive optimization.

[0009] Optionally, the specific content of S2 includes: S21. Traverse each task instance in the virtual task relationship, and extract the resource requirements of the task instance as a resource request volume vector , where represents the demand of the th task for the th type of resource, is the number of resource types, and the resource request volume vector is used as the resource request volume field of the cat - type scheduling unit ; S22. Extract the task priority value from the task instance, which reflects the urgency of the task in the overall scheduling, and use it as the priority value field of the cat - type scheduling unit ; S23. Define as the role field corresponding to the i - th task, generate the cat - type scheduling unit , and add it to the scheduling set , where represents the total number of scheduling tasks.

[0010] Optionally, the specific content of S3 includes: S31. Set the task priority threshold , and the total resource request threshold ; S32. For any cat - type scheduling unit in the generated scheduling set , extract the task priority value and the resource request volume vector , and calculate the total resource request volume ; S33. According to the comparison result of and , set the cat - type scheduling unit Role field Value, and the classification rules are as follows: If and , then assign the value , indicating an attacking role; If and , then assign the value , indicating a supporting role; If , then assign the value , indicating a defensive role; S34. Write the updated cat-type scheduling unit into the scheduling set , completing the initialization operation of the role fields of all task instances.

[0011] Optionally, the S4 specifically includes: S41. On the basis of the scheduling set , construct a task scheduling game model , where represents the set of schedulable resources, represents the set of scheduling strategies; S42. Define the number of resource request failures in the previous round of the cat-type scheduling unit , and the value is derived from the number of unmet resource types in the requests initiated by the previous round of combat cat algorithm game to the set of schedulable resources ; S43. Determine the scheduling behavior rules according to the value of the role field , and the rules are as follows: If , for an attacking unit, when this unit requests resources, the task priority value is the main order, and the allocated resource request amount is , where is the minimum value function, is the schedulable amount of the th type of resource; If , for a defensive unit, it is executed after other units' scheduling is completed, and the resource request amount is , where is the buffer coefficient set by the system; If , for a supporting unit, perform a support operation according to the resource conflict information in the game process. The support operation includes the following three types of behaviors: Behavior 1: Release the part of the resources occupied by itself that conflict with the attacking unit, that is, if there is And if the attacking unit has a shortage of Class resources, then let , where is the maximum value function, is the release step size; Behavior 2: Adjust the scheduling order, delay the execution round of the current task, and at the same time adjust the scheduling priority value to , where is the delay penalty parameter; Behavior 3: If the number of failed requests for Class resources by the attacking unit is greater than or equal to the set threshold , then the supporting unit gives priority to ensuring the reallocation of resources by releasing the corresponding resources or delaying its own scheduling order; S44. Define the scheduling priority function of the cat - type scheduling unit ; S45. Sort the cat - type scheduling units in the current round according to the scheduling priority function , and use the sorting result as the input to the task - scheduling game model to complete the task scheduling for this round.

[0012] Optionally, the specific steps of S5 include: S51. Based on the task - scheduling game model , with the set of cat - type scheduling units as the game participants, the set of resources as the schedulable resource pool, and the set of strategies including all combinations of scheduling orders and resource allocation paths; S52. Set the maximum number of evolution rounds , and initialize the scheduling priority function for the 0th round; S53. For each round of game iteration , perform the following operations: S531. Sort the set of cat - type scheduling units in descending order according to the scheduling priority function for the current round to construct a scheduling queue; S532. Traverse each cat - type scheduling unit in the queue in turn, and execute the corresponding behavior rules according to the role field ; S533. Record the resource allocation status of the cat - type scheduling unit in this round, calculate the number of unmet resource items and update the resource request failure count , and update the execution status field to , representing waiting, in - execution, and completed respectively; After completing a round of scheduling game, recalculate the next-round scheduling priority function based on the updated value, and enter the next round of iteration until the iteration process is terminated when any of the following conditions is met: All cat-type scheduling units satisfy ; The evolution round reaches the maximum value ; S55. Record and output the task execution trajectory, resource allocation path, and game state formed in each round of scheduling results as a set of scheduling schemes for candidate scheduling schemes , where each contains the scheduling results of one round, .

[0013] Optionally, the set of scheduling schemes in S55 specifically includes the following policy combinations based on conditional judgments: The first policy: When the attack-type unit satisfies the priority value , and there exists a resource type that satisfies the requested quantity , then select the scheduling unit with the priority value and the role field , and update the current-round resource request quantity to ; The second policy: If there exists an attack-type unit that satisfies the same resource type requested, and , is the priority similarity threshold, then allocate the resource in a time-slicing rotation manner within the scheduling period , the execution time slice is , and alternately allocate it to and ; The third policy: If the support-type unit has a resource request item , and at the same time there is an attack-type unit that satisfies and requests the same resource , then update the requested quantity of the th type of resource for the th time; The fourth policy: When the total number of tasks in the current round is greater than the preset value of the total number of tasks, and the total available resource is less than the preset value of the available resource, then for all scheduling units , update the scheduling priority function; Fifth strategy: If there is a scheduling unit that satisfies in two consecutive rounds and , then update the priority.

[0014] Optionally, the S6 simulation evaluation specifically includes: executing in a simulation platform, calculating the corresponding task completion rate, resource utilization rate, and system load balance degree, and using the combination of the three indicators as the evaluation basis to form the performance evaluation result of the scheduling strategy.

[0015] The beneficial effects of the present invention are: (1) By constructing a cat-shaped scheduling unit and introducing attack, support, and defense role behavior mechanisms, the present invention enables production tasks to have differentiated response strategies under different resource tension levels and priority conditions, improves the scheduling and coordination ability between complex tasks, and effectively solves the problems of scheduling deadlocks and response delays in traditional systems.

[0016] (2) The present invention uses the battle cat algorithm to carry out multiple rounds of game evolution in a digital twin environment to generate multiple candidate scheduling schemes, and introduces a non-linear priority function and a failure feedback mechanism to dynamically adjust the strategy selection logic. The system can continuously optimize the scheduling strategy in the case of resource state changes, increasing task failure rates, etc., and has the ability of self-learning and self-adjustment.

[0017] (3) By simulating and evaluating the scheduling strategy in a digital twin platform and evaluating and screening based on three types of indicators: task completion rate, resource utilization rate, and load balance degree, it is ensured that the finally executed scheduling scheme has verifiability and high execution efficiency, overcomes the problem of the disconnection between the strategy and the real production line in traditional methods, and significantly improves the operation efficiency and stability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention, and do not constitute a limitation to the present invention. In the drawings: Figure 1 is a flowchart of a production line operation monitoring system based on digital twin technology proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] Now, the present invention will be further described in detail with reference to the drawings. These drawings are all simplified schematic diagrams, only showing the basic structure of the present invention in a schematic way, so they only show the components related to the present invention.

[0020] Refer to Figure 1 , a production line operation monitoring system based on digital twin technology, includes: A data collection module, which is used to collect the task identifier, resource requirements, estimated execution time, and dependencies of each task in the production line; A scheduling unit construction module, which is used to map each task instance to a cat-shaped scheduling unit. The cat-shaped scheduling unit includes task attributes, resource request amounts, priority values, and role fields, and initializes the role fields based on a preset threshold; A scheduling game engine module, which is used to execute the battle cat algorithm on the task scheduling game model composed of cat-shaped scheduling units, and perform multi-round task scheduling evolution according to the role field behavior rules to generate a scheduling scheme set containing multiple candidate scheduling schemes; A simulation evaluation module, which is used to perform simulation evaluation on each candidate scheduling scheme in the digital twin environment to obtain the corresponding task completion rate, resource utilization rate, and system load balance degree; A strategy management module, which is used to sort the candidate scheduling schemes according to the simulation evaluation results and screen the current optimal scheduling scheme.

[0021] The system structure provided by the present invention covers all process modules from data collection, unit modeling, scheduling game, simulation evaluation to strategy management. The logic between the modules is clear and the functions are in a closed loop. Compared with the problems of the separation of scheduling strategies and control mechanisms and the lack of integration of ontology modeling and evaluation in traditional systems, the present invention integrates scheduling logic and feedback mechanisms in a unified architecture through modular design, realizes the integrated operation of the generation, evaluation, and execution of scheduling strategies, and improves the overall real-time performance, scalability, and intelligent decision-making ability of the system.

[0022] In this embodiment, the modules are implemented through the following methods: S1. Collect task data in the current production line, including task identifiers, resource requirements, estimated execution time, and dependencies, and establish corresponding virtual task relationships based on digital twin technology. Specifically, construct task models in the virtual environment that correspond one-to-one with physical tasks. This virtual task model not only restores the running characteristics of tasks but also establishes logical associations such as sequential constraints, resource competition, and execution priorities between tasks; S2. Map each task instance in the virtual task relationship to a cat-shaped scheduling unit. The cat-shaped scheduling unit includes task attributes, resource request amounts, priority values, and role fields; S3. Initialize the role fields of the cat-shaped scheduling unit according to the task data and set them to attack type, support type, or defense type; S4. Construct a task scheduling game model based on all cat-shaped scheduling units and set role behavior rules, including that the attack type preferentially obtains resources, the support type coordinates conflicts, and the defense type undertakes delayed and redundant tasks; S5. Run the Battle Cat algorithm to simulate the game process among cat-type scheduling units, execute several rounds of resource competition and strategy evolution, and output several candidate scheduling schemes; S6. Conduct simulation evaluation on the candidate scheduling schemes, calculate the task completion rate, resource utilization rate, and system load balance degree, and select the scheme with the best score as the optimal scheduling scheme; S7. Apply the optimal scheduling scheme to the production line operation monitoring system. During the operation process, trigger the reset of the role field according to emergencies, and re-execute steps S4 to S6 to achieve adaptive optimization. Specifically, when detecting emergencies such as task failures or resource faults, the system will trigger the reset of the role field of the cat-type scheduling unit, and reassign it as an attack-type, support-type, or defense-type role according to the current task urgency and resource status to adapt to the new scheduling strategy requirements.

[0023] The overall method flow provided by the present invention realizes a complete closed-loop from task data collection, cat-type unit construction, role initialization, to game scheduling driven by the Battle Cat algorithm, simulation evaluation, and strategy feedback optimization. Compared with the existing production line monitoring methods that only rely on static rules or single-round optimization strategies, the present invention constructs a dynamic evolution mechanism and role behavior differentiation strategies, enabling the scheduling system to have the ability to handle resource conflicts, task delays, and multi-objective trade-offs, improving the response efficiency and scheduling flexibility in complex environments, and having a significant improvement in system intelligence and adaptability.

[0024] In this embodiment, the specific steps of S2 include: S21. Traverse each task instance in the virtual task relationship, and extract the resource requirements of the task instance as a resource request volume vector , where represents the demand of the th task for the th type of resource, is the number of resource types, and the resource request volume vector is used as the resource request volume field of the cat-type scheduling unit ; S22. Extract the task priority value from the task instance, which reflects the urgency of the task in the overall scheduling, and use it as the priority value field of the cat-type scheduling unit ; S23. Define as the role field corresponding to the i-th task, generate the cat-type scheduling unit , and add it to the scheduling set , where represents the total number of scheduling tasks.

[0025] By introducing the modeling method of the cat-type scheduling unit, the present invention maps traditional task data into scheduling agent individuals with behavioral characteristics, and uses the resource request volume, priority value, and role field as a unified scheduling parameter structure, laying the foundation for multi-agent game modeling. Compared with the traditional scheduling representation methods based on task queues or graph structures, the present invention introduces role behavior variables at the modeling stage, providing structural support for subsequent scheduling behavior games and resource coordination, and enhancing the expression ability of the scheduling model and the flexibility of the behavior-driven logic.

[0026] In this embodiment, S3 specifically includes: S31. Set the task priority threshold , and the total resource request threshold ; S32. For any cat-type scheduling unit in the generated scheduling set , extract the task priority value and the resource request volume vector , and calculate the total resource request volume: ; Wherein, is the number of resource types; S33. According to the comparison result of and , set the value of the role field of the cat-type scheduling unit , and the classification rules are as follows: If and , then assign the value , indicating an attack-type role; If and , then assign the value , indicating a support-type role; If , then assign the value , indicating a defense-type role; S34. Write the updated cat-type scheduling unit into the scheduling set , and complete the initialization operation of the role fields of all task instances.

[0027] In the initialization process of the role field of the present invention, a combined condition judgment of the task priority value and the total amount of resource requests is adopted, and the scheduling tasks are clearly classified into three types of roles: attack type, support type, and defense type, realizing the pre-classification and structural division of the behavior rules in the game scheduling. Compared with the existing method that only uses fixed priority sorting for decision-making, the present invention, through the structured role division mechanism, embeds the differences in scheduling game behaviors in advance, provides a basis for the adaptive coordination of scheduling rules, and improves the system's processing ability for sudden scenarios and the flexibility of strategy adjustment.

[0028] In this embodiment, the specific content of S4 includes: S41. On the basis of the scheduling set , construct a task scheduling game model , where represents the set of schedulable resources, and represents the set of scheduling strategies; S42. Define the number of resource request failures of the previous round of the cat-type scheduling unit , and the value is derived from the number of resource types that were not satisfied in the requests initiated by the previous round of the battle cat algorithm game to the set of schedulable resources , that is: ; Among them, is the number of the -th type of resource requested by the cat-type scheduling unit , represents the remaining available amount of the -th type of resource in the -th round of the game, is an indicator function, and the rule is that if the condition is true, the value is 1, otherwise it is 0; S43. According to the value of the role field , determine the scheduling behavior rules, and the rules are as follows: If , it is an attack-type unit, then when this unit requests resources, the task priority value is used as the main order, and the allocated resource request amount is , where is the minimum value function, and is the current remaining available amount of the -th type of resource; If , it is a defense-type unit, then it is executed after the scheduling of other units, and the resource request amount is , where is the buffer coefficient set by the system; If , as a support unit, performs support operations according to the resource conflict information in the game process. The support operations include the following three types of behaviors: Behavior 1: Release the part of the resources occupied by itself that conflict with the attack unit, that is, if there is and the attack unit has a shortage of the type of resource, then let , where is the maximum value function, is the release step size; Behavior 2: Adjust the scheduling order, delay the execution round of the current task, and at the same time adjust the scheduling priority value to , where is the delay penalty parameter; Behavior 3: If the number of request failures of the attack unit for the type of resource is greater than or equal to the set threshold , then the support unit gives priority to the reallocation of resources by releasing the corresponding resources or delaying its own scheduling order; S44. Define the scheduling priority function of the cat-type scheduling unit , and the calculation method is: ; where is the logarithmic function with base 2, is the hyperbolic tangent function; This scheduling priority function comprehensively considers the urgency of the task itself, the resource consumption intensity, and the historical scheduling failure situation, and constructs a non-linear scoring mechanism using logarithmic and hyperbolic tangent functions, improving the rationality of the scheduling order and the system's adaptive regulation ability for resource conflicts and failed tasks; S45. Sort the cat-type scheduling units in the current round according to the scheduling priority function , and use the sorting result as the input to the task scheduling game model to complete the task scheduling for this round.

[0029] The present invention formulates clear behavior rules for various scheduling roles, and establishes a non-linear scheduling scoring mechanism by comprehensively considering task priorities, resource consumption, and failure feedback information through the scheduling priority function. Different from the static weighting or linear evaluation methods in the prior art, the function constructed by the present invention introduces a feedback term for the number of failed resource types, improving the self-adaptability of the scheduling order in multi-round evolution, effectively solving the problem that high-priority tasks are repeatedly postponed, and enhancing the strategy difference and strategy quality in resource contention scheduling.

[0030] In this embodiment, the specific content of S5 includes: S51. Based on the task scheduling game model Take the set of cat - type scheduling units as the game participants, the resource set as the schedulable resource pool, and the strategy set contains all combinations of scheduling orders and resource allocation paths; S52. Set the maximum number of evolution rounds and initialize the scheduling priority function for the 0th round ; S53. For each round of game iteration , perform the following operations: S531. Sort the set of cat - type scheduling units in descending order according to the scheduling priority function of the current round to construct a scheduling queue; S532. Traverse each cat - type scheduling unit in the queue in turn , and execute the corresponding behavior rules according to the role field ; S533. Record the resource allocation status of the cat - type scheduling unit in this round, calculate the number of unmet resource items and update the resource request failure count , and update the execution status field to , which represent waiting, in - execution, and completed respectively; S54. After completing one round of scheduling game, recalculate the scheduling priority function for the next round based on the updated value, enter the next round of iteration, and terminate the iteration process until any of the following conditions is met: All cat - type scheduling units meet ; The number of evolution rounds reaches the maximum value ; ; S55. Record and output the task execution trajectory, resource configuration path, and game state formed in each round of scheduling results as candidate scheduling schemes , where each contains the scheduling results of one round .

[0031] In the process of evolving the battle cat algorithm, the present invention introduces a multi - round simulation mechanism, and dynamically adjusts the scheduling priority according to the resource allocation results, role behaviors, and failure times in each round, enabling the scheduling unit to have the ability of self - learning during the evolution process. Compared with the traditional single - step decision - making scheduling method, the evolved multi - round scheduling mechanism adopted by the present invention can continuously optimize the scheduling strategy in a dynamic production environment, improve the system's adaptability and repair ability to complex scenarios such as task failure feedback and resource tension degree, and has high practicality and robustness.

[0032] In this embodiment, the scheduling scheme set in S55 specifically includes the following policy combinations based on conditional judgments: The first policy: When the attack unit satisfies the priority value , and there exists a resource type that satisfies the request volume , then select the scheduling unit with the priority value and the role field , update the current round of resource request volume to ; ; The second policy: If there exists an attack unit that satisfies the same resource type of the request , and , is the priority similarity threshold, then allocate the resource in a round-robin manner within the scheduling period , the execution time slice is , and alternately allocate it to and ; The third policy: If the support unit has a resource request item , and at the same time there is an attack unit that satisfies and requests the same resource , then update the request volume of for the th type of resource to: ; Among them, is the resource request volume of the th round for the th type of resource, is the resource request volume of the th round for the th type of resource, is the resource release ratio coefficient; This formula is used in the resource competition conflict scenario. The support unit actively releases part of the resource request to cede key resources to the attack unit with more failure times or higher priority. By proportionally reducing its own resource demand, resource cession and scheduling buffer are achieved, and the coordination of the overall system and the scheduling success rate of key tasks are improved. The formula controls the release intensity through the parameter to balance local sacrifice and global optimality; The fourth policy: When the total number of tasks in the current round is greater than the preset value of the total number of tasks, and the total available resource is less than the preset value of the available resource, then for all scheduling units , update the scheduling priority function to: ; Among them, is the round-robin scheduling priority function, is the round-robin scheduling priority function, is the delay penalty constant; This formula delays the scheduling time of defensive scheduling units by reducing their priorities, giving priority to ensuring the execution of more urgent or critical tasks. When the number of tasks or resource pressure exceeds the threshold, this strategy helps to relieve scheduling congestion and release resources to high-priority tasks. The degree of priority reduction is controlled by , reflecting the system's scheduling tilt and dynamic adjustment ability for tasks with different roles, thus enhancing the overall scheduling stability and efficiency; The fifth strategy: If there is a scheduling unit that satisfies and in two consecutive rounds, then update the priority to: ; Among them is the failure compensation factor.

[0033] This formula implements a dynamic priority repair mechanism based on failure feedback. When a task is not successfully scheduled in the current round, the system will automatically increase its priority according to the degree of failure, so that it can obtain a more favorable execution opportunity in the next round of scheduling. By introducing the linkage between the number of failures and the linear compensation term, the system can effectively avoid the problem of tasks being suppressed for a long time, improve task fairness and the overall scheduling success rate, and at the same time enhance the robustness and self-healing ability of the scheduling system.

[0034] In the present invention, the candidate scheduling scheme is defined as a set of scheduling policy rules with specific triggering conditions and behavioral logics, clarifying the decision-making paths for resource competition, role coordination, and scheduling failure handling among tasks. Compared with the existing scheme that only outputs a list of scheduling results, the rule-based policy set expressed by the "if-then" structure in the present invention enhances the comprehensibility, executability, and systematic expression ability of the scheduling scheme, provides a clear basis for the automatic decision-making of the scheduling engine, and significantly improves the maintainability and scenario expansion ability of the system.

[0035] In this embodiment, the S6 simulation evaluation specifically includes: executing in a simulation platform to calculate the corresponding task completion rate, resource utilization rate, and system load balance degree; among them, the task completion rate is obtained by dividing the number of cat-type scheduling units with the status of "completed" within a certain time period by the total number of tasks in this cycle; the system records the allocation and occupation of resource types in each round of game, and the resource utilization rate is obtained by dividing the cumulative resource usage by the total available resources; the system load balance degree is the standard deviation of each resource utilization rate or task distribution, and the smaller the value, the more uniform the system load distribution and the more balanced the scheduling strategy; the ternary index combination is used as the evaluation basis to form the evaluation result of the scheduling strategy performance.

[0036] Example 1: To verify the feasibility of the present invention in implementation, the present invention is applied to the automatic assembly production line of an intelligent equipment manufacturing enterprise. This production line is mainly responsible for the precision assembly of various electronic control modules, involving multiple processes such as chip mounting, electrical testing, plug-in, and assembly. The average number of tasks to be processed per day is about 380. The production line is equipped with 18 types of schedulable resources, including chip mounters, plug-in machines, test benches, assembly stations, intelligent AGV cars, manual work teams, tool groups, and electrical energy supply units, etc.

[0037] In actual operation, the original scheduling system of this enterprise is based on a static priority strategy and uses a single weighted algorithm for task sorting. It cannot dynamically handle problems such as the insertion of sudden tasks, frequent resource conflicts, and off-peak scheduling of night shift production capacity. There are often phenomena such as task congestion and too long waiting time on the critical path, resulting in an average equipment utilization rate of less than 65% and a task delay rate as high as 18%. To solve this series of scheduling bottleneck problems, the enterprise introduces the production line operation monitoring system based on digital twin technology proposed by the present invention and deploys it on a typical combined production line of electrical testing and assembly for a 30-day comparative experiment.

[0038] The system first uses the data acquisition module of the present invention to collect the real-time execution logs of all tasks on this production line for 30 days, record elements such as task identifiers, resource request volumes, estimated execution durations, and task dependencies, and construct corresponding virtual task models. After the model is generated, the system uses cat-type scheduling units for task modeling and initializes the roles according to the priority values and resource consumption, automatically dividing them into three types of roles: attack type, support type, and defense type. Subsequently, the system enters the multi-round scheduling game stage of the battle cat algorithm, dynamically evolves the task priority order based on the priority function, and outputs candidate scheduling plans.

[0039] Before deployment, the scheduling plan is verified and evaluated through a digital twin simulation platform. For each round of the plan, the task completion rate, resource utilization rate, and system load balance degree are calculated. After the evaluation, the system selects the scheduling plan with the best performance and directly issues it to the MES system to update and control the production line task sequence and resource allocation instructions. During the scheduling execution process, the system adjusts the task scheduling strategy in real time according to the task execution status and resource feedback. If a task fails due to equipment failure or the occupied work station, the system can automatically re-evaluate the scheduling priority and re-adjust the order according to the game feedback mechanism to ensure that the overall production capacity target is not affected.

[0040] Through a 30-day system comparison test, it is found that the present invention shows significant superiority in the actual production line. Compared with the original system, the average task waiting time has decreased from 21.3 minutes to 9.6 minutes, the resource utilization rate has increased from 63.5% to 82.1%, and the task delay rate has decreased from 18.2% to 4.7%. In two sudden task insertion events, the system of the present invention can complete the scheduling diagram update and realize the automatic strategy switch within 10 seconds, successfully ensuring the on-time delivery of critical work orders and avoiding losses of about 30,000 yuan in man-hour waste and production line shutdown.

[0041] By deploying the system of the present invention, the enterprise has realized the field application of the multi-role game scheduling strategy in the industrial-level production line for the first time, and shows excellent scalability and stability in dimensions such as intelligent production scheduling, task plug-and-play scheduling, and adaptive collaborative regulation, verifying the feasibility, robustness, and system economy of the present invention in industrial scenarios with multi-tasks and high complexity.

[0042] Table 1: Performance comparison table of the game-based scheduling system based on digital twin and traditional system First of all, in terms of task scheduling efficiency, the system of the present invention reduces the average task waiting time from 21.3 minutes to 9.6 minutes, and the shortening amplitude exceeds half, significantly improving the response speed of tasks in the queue. At the same time, the average task completion time is also shortened from 45.8 minutes to 31.2 minutes, indicating that the execution cycle of the overall task is compressed, effectively alleviating the idle waiting between processes and the resource switching delay.

[0043] In terms of resource utilization rate, the system of the present invention has achieved a significant increase from 63.5% to 82.1%, indicating that the system can allocate and schedule various production resources more efficiently, avoiding the phenomenon of equipment idleness and repeated waiting. This increase also directly drives the production capacity performance. The maximum number of output tasks per day has increased from 386 to 447, an increase of nearly 16%, broadening the production line throughput capacity without increasing hardware investment.

[0044] In terms of task delay rate, the system of the present invention effectively reduces the delay ratio from 18.2% to 4.7%, reducing the risks of task delivery delay and rework, and further improving the delivery reliability and customer satisfaction. At the same time, the response time of the scheduling strategy is shortened from 65 seconds to 9.8 seconds, and the average repair time of rescheduling is reduced from 77 seconds to 11.2 seconds, indicating that the system has extremely strong rapid response and strategy recovery capabilities in the face of disturbances such as sudden task changes or equipment anomalies.

[0045] More notably, the present invention shows outstanding performance in the dimension of "success rate of adapting to emergencies". The success rate has been significantly increased from 42.3% of the original system to 91.6%, indicating that the scheduling game mechanism after introducing the Battle Cat algorithm demonstrates extremely high robustness and adaptive capabilities under typical challenges such as complex task conflicts and high resource contention.

[0046] Generally speaking, the present invention not only solves the problems existing in the original system, such as static scheduling rigidity, resource coordination imbalance, and non-adjustable strategy superiority, but also constructs a complete feedback loop from modeling to evaluation and then to strategy update, demonstrating extremely strong practicality, scalability, and economic benefits in the actual industrial production environment. While maintaining scheduling intelligence, the system significantly improves the comprehensive operation efficiency of the production line and has broad application and promotion value.

[0047] The above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent substitutions or changes should be covered within the protection scope of the present invention.

Claims

1. A production line operation monitoring system based on digital twin technology, characterized in that, Including: A data acquisition module, which is used to acquire the task identifier, resource requirements, estimated execution time, and dependencies of each task in the production line; A scheduling unit construction module, which is used to map each task instance to a cat-shaped scheduling unit. The cat-shaped scheduling unit includes task attributes, resource request quantities, priority values, and role fields, and initializes the role fields based on a preset threshold; A scheduling game engine module, which is used to execute the battle cat algorithm on the task scheduling game model composed of cat-shaped scheduling units. The battle cat algorithm is a virtual task scheduling method that simulates the behaviors of attack, support, and defense roles, and performs multi-round task scheduling evolution according to the role field behavior rules to generate a set of scheduling plans containing multiple candidate scheduling plans; A simulation evaluation module, which is used to perform simulation evaluation on each candidate scheduling plan in the digital twin environment to obtain the corresponding task completion rate, resource utilization rate, and system load balance degree; A strategy management module, which is used to sort the candidate scheduling plans according to the simulation evaluation results and screen the current optimal scheduling plan.

2. The production line operation monitoring system based on digital twin technology according to claim 1, wherein The modules are implemented through the following methods: S1. Collect task data in the current production line, including task identifiers, resource requirements, estimated execution time, and dependencies, and establish corresponding virtual task relationships based on digital twin technology; S2. Map each task instance in the virtual task relationship to a cat-shaped scheduling unit, and the cat-shaped scheduling unit includes task attributes, resource request quantities, priority values, and role fields; S3. Initialize the role fields of the cat-shaped scheduling units according to the task data, and set them to attack, support, or defense types; S4. Build a task scheduling game model based on all cat-shaped scheduling units, and set role behavior rules, including that the attack type preferentially obtains resources, the support type coordinates conflicts, and the defense type undertakes delayed and redundant tasks; S5. Run the battle cat algorithm. The battle cat algorithm is a virtual task scheduling method that simulates the behaviors of attack, support, and defense roles, simulates the game process between cat-shaped scheduling units, performs several rounds of resource competition and strategy evolution, and outputs several candidate scheduling plans; S6. Perform simulation evaluation on the candidate scheduling plans, calculate the task completion rate, resource utilization rate, and system load balance degree, and select the plan with the best score as the optimal scheduling plan; S7. Apply the optimal scheduling plan to the production line operation monitoring system, trigger the reset of the role fields according to emergencies during the operation process, and re-execute steps S4 to S6 to achieve adaptive optimization.

3. The production line operation monitoring system based on digital twin technology according to claim 2, characterized in that, The specific content of S2 includes: S21. Traverse each task instance in the virtual task relationship, and extract the resource requirements of the task instance as a resource request volume vector , where represents the demand of the -th task for the -th type of resource, is the number of resource types, and the resource request volume vector serves as the resource request volume field of the cat-type scheduling unit ; S22. Extract the task priority value from the task instance , which reflects the urgency of the task in the overall scheduling and serves as the priority value field of the cat-shaped scheduling unit ; S23. Define as the role field corresponding to the i-th task, generate a cat-type scheduling unit , and add it to the scheduling set , where represents the total number of scheduling tasks.

4. The production line operation monitoring system based on digital twin technology according to claim 3, characterized in that, The specific content of S3 includes: S31. Set the task priority threshold , and the total resource request threshold ; S32. For any cat-type scheduling unit in the generated scheduling set , extract the task priority value and the resource request volume vector , and calculate the total resource request volume ; S33. Set the role field value of the cat - type scheduling unit according to the comparison result with , and the classification rules are as follows: ​ If and , then assign the value , indicating an attacking character; If and , then assign the value , indicating a support character; If , then assign the value , indicating a defensive character; S34. Write the updated cat-type scheduling unit to the scheduling set , and complete the initialization operation of the role fields of all task instances.

5. The production line operation monitoring system based on digital twin technology according to claim 4, characterized in that The specific content of S4 includes: S41. On the basis of the scheduling set , construct a task scheduling game model , where represents the set of schedulable resources, represents the set of scheduling strategies; S42. Define the cat-type scheduling unit The number of resource request failures in the previous round , and the value is obtained from the previous round of battle cat algorithm game to the schedulable resource set Among the requests initiated, the number of resource types that are not satisfied; S43. Determine the scheduling behavior rules according to the value of the role field as follows: If is an attack unit, when this unit requests resources, the task priority value is the main order, and the allocated resource request amount is , where is the minimum value function is the schedulable amount of the type of resource; If is a defensive unit, it will be executed after the scheduling of other units is completed, and the resource request volume is , where is the buffer coefficient set by the system; If is a support unit, it performs a support operation according to the resource conflict information in the game process. The support operation includes the following three types of behaviors: Behavior 1: Release the part of the resources occupied by itself that conflicts with the attack units. That is, if there is and the attack units have a shortage of the type of resources, then let , where is the maximum value function, is the release step size; Action 2: Adjust the scheduling order, delay the execution round of the current task, and at the same time adjust the scheduling priority value to , where is the delay penalty parameter; Behavior three: If the number of failed requests of the attack type unit for the type resources is greater than or equal to the set threshold , the support type unit preferentially guarantees the reallocation of resources by releasing the corresponding resources or delaying its own scheduling order; S44. Define the scheduling priority function of the cat-type scheduling unit of ; S45. According to the scheduling priority function Sort the cat-type scheduling units in the current round, and use the sorting result as the input to be passed to the task scheduling game model to complete the task scheduling for this round.

6. The production line operation monitoring system based on digital twin technology according to claim 5, wherein The specific content of S5 includes: S51. Based on the task scheduling game model , taking the set of cat-type scheduling units as the game participants, and the resource set as the schedulable resource pool, and the strategy set includes all combinations of scheduling orders and resource allocation paths; S52. Set the maximum number of evolution rounds , and initialize the scheduling priority function for the 0th round ; S53. For each round of game iteration , perform the following operations: S531. For the set of cat-type scheduling units Sort in descending order according to the scheduling priority function of the current round to construct a scheduling queue; S532. Traverse each cat-shaped scheduling unit in the queue in sequence , and execute the corresponding behavior rules according to the role field ; S533. Record the cat-type scheduling unit In the resource allocation status of this round, calculate the number of unmet resource items and update the number of failed resource requests , update the execution status field to , which respectively represent waiting, in execution, and completed; S54. After completing a round of scheduling game, recalculate the next-round scheduling priority function based on the updated value, and enter the next round of iteration until the iteration process is terminated when any of the following conditions is met: ​ All cat-type scheduling units satisfy ; The evolution round reaches the maximum value ; S55. Record and output the task execution trajectory, resource allocation path, and game state formed in each round of scheduling results as a set of scheduling plans for candidate scheduling plans , where each contains the scheduling results of one round, .

7. A production line operation monitoring system based on digital twin technology according to claim 6, characterized in that, The specific content of the scheduling plan set in S55 specifically includes the following policy combinations based on conditional judgments: The first strategy: when the attacking unit meets the priority value , and there exists a resource type that meets the requested quantity , then select the scheduling unit with the priority value and the role field , update the current round of resource request quantity to ;​ Second strategy: If there are attack units with the same resource type that meets the request , and , is the priority similarity threshold, then the resource will be allocated in a round-robin manner within the scheduling period , and the execution time slice is , and it will be alternately allocated to and ; The third strategy: If the support unit has a resource request item , and at the same time there is an attack unit that satisfies and requests the same resource , then update the request volume for the type of resource; Fourth strategy: If the total number of tasks in the current round is greater than the preset value of the total number of tasks, and the total available resources are less than the preset value of the available resources, then for all scheduling units , update the scheduling priority function; Fifth strategy: If there is a scheduling unit that satisfies in two consecutive rounds and , then update the priority.

8. An on-line production line operation monitoring system based on digital twin technology according to claim 7, characterized in that, The specific content of the simulation evaluation includes: executing in a simulation platform, calculating the corresponding task completion rate, resource utilization rate, and system load balance degree, and weighted integrating the corresponding task completion rate, resource utilization rate, and system load balance degree as the evaluation basis to form the scheduling strategy performance evaluation result.

Images