AutoGen-based intelligent manufacturing process coordination agent method and system

Abstract

The invention provides a method and a system for coordinating and agenting an intelligent manufacturing process based on AutoGen. The intelligent manufacturing process coordination agent method based on AutoGen comprises the following steps: defining roles of Agents, and configuring and initializing a system unit corresponding to each role Agent; monitoring the operation data of the system unit corresponding to each role Agent in real time, and judging whether abnormity exists or not according to the operation data of the system unit corresponding to each role Agent; and performing performance evaluation on the system unit corresponding to each role Agent according to a preset time period, and sending an operation and maintenance prompt to a worker terminal when the performance of the system unit corresponding to each role Agent is abnormal. The system comprises modules corresponding to the steps of the method.

Description

Technical Field

[0001] The invention proposes a method and a system based on AutoGen intelligent manufacturing process coordination agent, belonging to the technical field of process coordination control. Background Art

[0002] In a manufacturing plant with highly automated assembly lines, precise coordination of various production links is required to optimize production efficiency, reduce waste, and ensure product quality. Commonly used technologies include SCADA (Supervisory Control and Data Acquisition System), which is used to monitor industrial processes in real time; MES (Manufacturing Execution System), which manages the production activities of the factory; automated robots, which users perform specific production tasks such as welding, assembly or packaging; regular quality inspections, etc. There are several technical defects and problems in the existing technology:

[0003] (1) Lack of flexibility: Fixed automated processes are difficult to adapt to rapid changes in production needs.

[0004] (2) Manual intervention: Production line adjustments and problem solving usually require manual operation.

[0005] (3) Data silos: Data between different systems and devices are not interoperable, making comprehensive analysis difficult.

[0006] (4) Long response time: It takes time to respond to and resolve quality issues or production bottlenecks. Summary of the invention

[0007] The present invention provides a method and system based on AutoGen intelligent manufacturing process coordination agent to solve the above technical problems existing in the prior art. The technical solutions adopted are as follows:

[0008] The method based on AutoGen intelligent manufacturing process coordination agent includes:

[0009] Define the roles of Agents, and configure and initialize the system units corresponding to each role Agent;

[0010] Monitor the operation data of the system unit corresponding to each role Agent in real time, and determine whether there is an abnormality according to the operation data of the system unit corresponding to each role Agent;

[0011] The performance of the system unit corresponding to each role Agent is evaluated according to a preset time period, and it is determined whether to send an operation and maintenance prompt to the staff terminal according to the performance evaluation result of the system unit corresponding to each role Agent.

[0012] Furthermore, define the roles of the Agent, and configure and initialize the system units corresponding to each role Agent, including:

[0013] Define the role of the Agent;

[0014] Using the API provided by AutoGen to create a system unit corresponding to each role Agent; wherein the system unit corresponding to each role Agent includes a production monitoring agent, a quality control agent, a predictive maintenance agent, a supply chain coordination agent, and a production scheduling agent;

[0015] Configure system parameters for the system unit corresponding to each role Agent to clarify the task responsibilities of the system unit corresponding to each role Agent;

[0016] According to the task responsibilities corresponding to the system units corresponding to each role Agent, an LLM model corresponding to the task responsibilities is configured; wherein the LLM model is a model that has been trained and tested and stored in a database.

[0017] Furthermore, real-time monitoring of the operating data of the system unit corresponding to each role Agent and determining whether there is an abnormality according to the operating data of the system unit corresponding to each role Agent include:

[0018] Real-time monitoring of the operating data of the system unit corresponding to each role Agent;

[0019] Whether the operation data of the system unit corresponding to each role Agent is abnormal is determined according to the invalid data acquired in each unit time of the system unit corresponding to each role Agent.

[0020] Further, determining whether there is an abnormality in the operation data of the system unit corresponding to each role Agent according to the invalid data acquired per unit time of the system unit corresponding to each role Agent includes:

[0021] Extracting the amount of invalid data corresponding to each unit time acquired by the system unit corresponding to each role Agent;

[0022] Extracting the data type corresponding to the invalid data acquired per unit time by the system unit corresponding to each role Agent and the weight value corresponding to the data type;

[0023] The evaluation parameter of the invalid data influence degree is obtained according to the data amount corresponding to the invalid data and the corresponding weight value; wherein the evaluation parameter of the invalid data influence degree is obtained by the following formula:

[0024]

[0025] Where I represents the evaluation parameter of the impact of invalid data; M represents the number of unit time that has passed; k hj Indicates the weight and value of the type of invalid data generated in the jth unit time; k z represents the weights and values ​​corresponding to all data types; H represents the evaluation parameter variable, and the evaluation parameter variable is obtained by the following formula:

[0026]

[0027] Among them, H represents the evaluation parameter; n j V represents the number of data types with invalid data corresponding to the jth unit time; i k represents the amount of invalid data corresponding to the jth unit time of the i-th data type; i Indicates the weight value corresponding to the i-th data type; D i Represents the diversity index corresponding to the i-th data type, which is used to indicate the dispersion or frequency of this type of data in the overall data; i represents the attenuation coefficient corresponding to the diversity index of the i-th data type and the i-th data type; n z Indicates the total number of data types;

[0028] When the evaluation parameter of the impact degree of the invalid data exceeds a preset impact degree evaluation parameter threshold, it is determined that there is an abnormality in the operation of the system unit corresponding to the role Agent corresponding to which the evaluation parameter of the impact degree of the invalid data exceeds the preset impact degree evaluation parameter threshold.

[0029] Further, a performance evaluation is performed on the system unit corresponding to each role Agent according to a preset time period, and according to the performance evaluation result of the system unit corresponding to each role Agent, it is determined whether to send an operation and maintenance prompt to the staff terminal, including:

[0030] Real-time monitoring of the operating parameters of the system unit corresponding to each role Agent in each of the preset time periods, wherein the operating parameters include task completion rate, task start response time and task execution time;

[0031] The unit operation evaluation coefficient corresponding to the system unit corresponding to each role Agent is obtained according to the operation parameters of the system unit corresponding to each role Agent; wherein the unit operation evaluation coefficient is obtained by the following formula:

[0032]

[0033] Where S represents the unit operation evaluation coefficient; n represents the number of time periods; P i represents the task completion rate corresponding to the i-th time period; T zi represents the median value of the task start response time corresponding to the number of left and right task starts in the i-th time period; T pi represents the average task start response time corresponding to the number of task starts in the i-th time period; T maxi represents the maximum task start response time corresponding to the number of task starts in the i-th time period; T b represents the standard deviation of the task start response time corresponding to the number of task starts in the i-th time period; m i represents the number of tasks in the i-th time period; T sj represents the task execution time of the jth task corresponding to the jth time period; T yj represents the reference duration of the jth task corresponding to the jth time period; M b represents the standard deviation of the number of tasks corresponding to n time periods; M max Indicates the maximum number of tasks in a single time period that appear in n time periods;

[0034] When the unit operation evaluation coefficient corresponding to the system unit corresponding to each role Agent is lower than the corresponding evaluation coefficient threshold, the correlation coefficient between the system unit corresponding to the role Agent lower than the corresponding evaluation coefficient threshold and the system units corresponding to other role Agents not lower than the corresponding evaluation coefficient threshold is obtained; wherein the correlation coefficient is obtained by the following formula:

[0035]

[0036] Where f represents the correlation coefficient; k represents the number of system units corresponding to other role agents that are not lower than the corresponding evaluation coefficient threshold; S i represents the unit operation evaluation coefficient of the system unit corresponding to the i-th other role agent that is not lower than its corresponding evaluation coefficient threshold; S yi represents the evaluation coefficient threshold corresponding to the system unit of the i-th other role agent whose evaluation coefficient threshold is not lower than its corresponding evaluation coefficient threshold; S d Indicates the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is lower than its corresponding evaluation coefficient threshold; S cmax It represents the maximum value of the difference between the unit operation evaluation coefficients of the system units corresponding to other role agents that are not lower than their corresponding evaluation coefficient thresholds; S z Indicates that the middle value of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than the corresponding evaluation coefficient threshold; Smax and S min Indicates the maximum and minimum values ​​of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than its corresponding evaluation coefficient threshold;

[0037] The comprehensive correlation coefficient is obtained by using the correlation coefficient of the system unit corresponding to the role agent whose evaluation coefficient is lower than the corresponding evaluation coefficient threshold and the system unit corresponding to other role agents whose evaluation coefficient is not lower than the corresponding evaluation coefficient threshold; wherein the comprehensive correlation coefficient is obtained by the following formula:

[0038]

[0039] Where CI represents the comprehensive correlation coefficient; r represents the number of system units corresponding to the role agent that is lower than its corresponding evaluation coefficient threshold; f i It represents the correlation coefficient of the system unit corresponding to the i-th role agent that is lower than its corresponding evaluation coefficient threshold;

[0040] When the comprehensive correlation coefficient exceeds a preset coefficient threshold, it is determined that a system operation and maintenance prompt needs to be sent to the staff terminal.

[0041] The system based on AutoGen intelligent manufacturing process coordination agent includes:

[0042] Define the initialization module, which is used to define the role of the Agent and configure and initialize the system unit corresponding to each role Agent;

[0043] A real-time monitoring module is used to monitor the operation data of the system unit corresponding to each role Agent in real time, and determine whether there is an abnormality according to the operation data of the system unit corresponding to each role Agent;

[0044] The performance evaluation module is used to perform performance evaluation on the system unit corresponding to each role Agent according to a preset time period, and determine whether to send an operation and maintenance prompt to the staff terminal according to the performance evaluation result of the system unit corresponding to each role Agent.

[0045] Furthermore, the definition initialization module includes:

[0046] Definition module, used to define the role of Agent;

[0047] A creation module, used to create a system unit corresponding to each role Agent using the API provided by AutoGen; wherein the system unit corresponding to each role Agent includes a production monitoring agent, a quality control agent, a predictive maintenance agent, a supply chain coordination agent, and a production scheduling agent;

[0048] The parameter configuration module is used to configure the system parameters for the system unit corresponding to each role Agent, so as to clarify the task responsibilities corresponding to the system unit corresponding to each role Agent;

[0049] The model configuration module is used to configure the LLM model corresponding to the task responsibilities according to the task responsibilities corresponding to the system units corresponding to each role Agent; wherein the LLM model is a model that has been trained and tested and stored in the database.

[0050] Furthermore, the real-time monitoring module includes:

[0051] Data monitoring module, used to monitor the operation data of the system unit corresponding to each role Agent in real time;

[0052] The abnormality judgment module is used to determine whether there is an abnormality in the operation data of the system unit corresponding to each role Agent according to the invalid data obtained in each unit time of the system unit corresponding to each role Agent.

[0053] Further, determining whether there is an abnormality in the operation data of the system unit corresponding to each role Agent according to the invalid data acquired per unit time of the system unit corresponding to each role Agent includes:

[0054] A first data information extraction module is used to extract the data volume corresponding to the invalid data acquired per unit time of the system unit corresponding to each role Agent;

[0055] A second data information extraction module is used to extract the data type corresponding to the invalid data acquired per unit time by the system unit corresponding to each role Agent and the weight value corresponding to the data type;

[0056] The evaluation parameter acquisition module is used to obtain the evaluation parameter of the invalid data impact degree according to the data volume corresponding to the invalid data and its corresponding weight value; wherein the evaluation parameter of the invalid data impact degree is obtained by the following formula:

[0057]

[0058] Where I represents the evaluation parameter of the impact of invalid data; M represents the number of unit time that has passed; khj Indicates the weight and value of the type of invalid data generated in the jth unit time; k z represents the weights and values ​​corresponding to all data types; H represents the evaluation parameter variable, and the evaluation parameter variable is obtained by the following formula:

[0059]

[0060] Among them, H represents the evaluation parameter; n j V represents the number of data types with invalid data corresponding to the jth unit time; i k represents the amount of invalid data corresponding to the jth unit time of the i-th data type; i Indicates the weight value corresponding to the i-th data type; D i Represents the diversity index corresponding to the i-th data type, which is used to indicate the dispersion or frequency of this type of data in the overall data; i represents the attenuation coefficient corresponding to the diversity index of the i-th data type and the i-th data type; n z Indicates the total number of data types;

[0061] The abnormality determination module is used to determine that when the evaluation parameter of the impact degree of the invalid data exceeds the preset impact degree evaluation parameter threshold, there is an abnormality in the operation of the system unit corresponding to the role Agent corresponding to the evaluation parameter of the impact degree of the invalid data exceeding the preset impact degree evaluation parameter threshold.

[0062] Furthermore, the performance evaluation module includes:

[0063] An operation parameter monitoring module is used to monitor in real time the operation parameters of the system unit corresponding to each role Agent in each of the preset time periods, wherein the operation parameters include task completion rate, task start response time and task execution time;

[0064] The unit operation evaluation coefficient acquisition module is used to acquire the unit operation evaluation coefficient corresponding to the system unit corresponding to each role Agent according to the operation parameters of the system unit corresponding to each role Agent; wherein the unit operation evaluation coefficient is acquired by the following formula:

[0065]

[0066] Where S represents the unit operation evaluation coefficient; n represents the number of time periods; P i represents the task completion rate corresponding to the i-th time period; T zi represents the median value of the task start response time corresponding to the number of left and right task starts in the i-th time period; T pirepresents the average task start response time corresponding to the number of task starts in the i-th time period; T maxi represents the maximum task start response time corresponding to the number of task starts in the i-th time period; T b represents the standard deviation of the task start response time corresponding to the number of task starts in the i-th time period; m i represents the number of tasks in the i-th time period; T sj represents the task execution time of the jth task corresponding to the jth time period; T yj represents the reference duration of the jth task corresponding to the jth time period; M b represents the standard deviation of the number of tasks corresponding to n time periods; M max Indicates the maximum number of tasks in a single time period that appear in n time periods;

[0067] The correlation coefficient acquisition module is used to obtain the correlation coefficient between the system unit corresponding to the role agent below the corresponding evaluation coefficient threshold and the system units corresponding to other role agents not below the corresponding evaluation coefficient threshold when the unit operation evaluation coefficient corresponding to the system unit corresponding to each role agent is lower than the corresponding evaluation coefficient threshold; wherein the correlation coefficient is obtained by the following formula:

[0068]

[0069] Where f represents the correlation coefficient; k represents the number of system units corresponding to other role agents that are not lower than the corresponding evaluation coefficient threshold; S i represents the unit operation evaluation coefficient of the system unit corresponding to the i-th other role agent that is not lower than its corresponding evaluation coefficient threshold; S yi represents the evaluation coefficient threshold corresponding to the system unit of the i-th other role agent whose evaluation coefficient threshold is not lower than its corresponding evaluation coefficient threshold; S d Indicates the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is lower than its corresponding evaluation coefficient threshold; S cmax It represents the maximum value of the difference between the unit operation evaluation coefficients of the system units corresponding to other role agents that are not lower than the corresponding evaluation coefficient threshold; S z Indicates that the middle value of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than the corresponding evaluation coefficient threshold; S max and S min Indicates the maximum and minimum values ​​of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than its corresponding evaluation coefficient threshold;

[0070] The comprehensive correlation coefficient acquisition module is used to obtain the comprehensive correlation coefficient by using the correlation coefficients of the system units corresponding to the role Agent whose evaluation coefficient is lower than the corresponding evaluation coefficient threshold and the system units corresponding to other role Agents whose evaluation coefficient is not lower than the corresponding evaluation coefficient threshold; wherein the comprehensive correlation coefficient is obtained by the following formula:

[0071]

[0072] Where CI represents the comprehensive correlation coefficient; r represents the number of system units corresponding to the role agent that is lower than its corresponding evaluation coefficient threshold; f i It represents the correlation coefficient of the system unit corresponding to the i-th role agent that is lower than its corresponding evaluation coefficient threshold;

[0073] The operation and maintenance prompt module is used to determine the need to send a system operation and maintenance prompt to the staff terminal when the comprehensive correlation coefficient exceeds a preset coefficient threshold.

[0074] Beneficial effects of the present invention:

[0075] The method and system based on the AutoGen intelligent manufacturing process coordination agent proposed in the present invention realizes real-time monitoring, predictive analysis, adaptive decision-making and integrated collaboration of the production process by integrating the multi-agent system under the AutoGen architecture in the key modules. The agent can process a large amount of production data in real time, provide automated decision support based on real-time data and historical trends, break the data island, and realize comprehensive collaboration between different systems. This improvement in flexibility and automation level enables the agent to quickly respond to changes in production demand, reduce human errors and production interruptions, and provide a more intuitive and friendly user experience. The self-learning and optimization capabilities of the agent enable it to continuously improve its performance over time, and the fast exception response mechanism further reduces production delays. In addition, the customized solution of the agent can meet the unique needs of specific factories. Compared with traditional systems, the agent provides a more intelligent and automated advanced manufacturing process management method. At the same time, through real-time monitoring and performance evaluation, bottlenecks and problems in the production process can be discovered and solved in a timely manner to ensure the smooth progress of the production process, thereby improving overall production efficiency. Automated monitoring and abnormality judgment reduce the frequency and intensity of manual inspections and reduce labor costs. At the same time, potential problems are discovered and handled in a timely manner, avoiding greater losses caused by the expansion of faults. Through strict monitoring and performance evaluation of the production process, each production link is ensured to meet quality requirements, thereby improving the quality of the final product and customer satisfaction. The agent-based intelligent manufacturing process coordination agent method is highly flexible and scalable. As production needs change, the role and configuration of the agent can be easily adjusted to adapt to new production scenarios and needs. The application of this method has promoted the in-depth development of intelligent manufacturing and provided strong support for the intelligent transformation of enterprises. By introducing intelligent agent technology, enterprises can manage and optimize production processes more efficiently and improve overall competitiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 is a flow chart of the method of the present invention;

[0077] Figure 2 The system block diagram of the system of the present invention. DETAILED DESCRIPTION

[0078] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0079] The embodiment of the present invention proposes a method based on AutoGen intelligent manufacturing process coordination agent, such as Figure 1 As shown, the method based on AutoGen intelligent manufacturing process coordination agent includes:

[0080] S1. Define the roles of Agents, and configure and initialize the system units corresponding to each role Agent;

[0081] S2, real-time monitoring of the operating data of the system unit corresponding to each role Agent, and determining whether there is an abnormality according to the operating data of the system unit corresponding to each role Agent;

[0082] S3. Perform performance evaluation on the system unit corresponding to each role Agent according to a preset time period, and determine whether to send an operation and maintenance prompt to the staff terminal according to the performance evaluation result of the system unit corresponding to each role Agent.

[0083] The working principle of the above technical solution is as follows: AutoGen is a multi-agent framework that supports the development of LLM applications using multiple agents that can communicate with each other to solve tasks. Agents in AutoGen are customizable, conversational, and seamlessly allow human participation. They can run LLMs in various modes that use a combination of human input and tools. In the key technical modules, agents with different roles are integrated, including production monitoring agents to monitor the status of the production line in real time (machine operation, production speed, etc.); quality control agents to be responsible for product quality inspection and real-time analysis of product quality data; predictive maintenance agents to predict equipment failures and arrange maintenance work in advance; supply chain coordination agents to manage raw material supply and inventory and respond to supply chain changes; production scheduling agents to intelligently arrange production plans based on order requirements and production capacity, etc.

[0084] The above technical solution of this embodiment realizes real-time monitoring, predictive analysis, adaptive decision-making and integrated collaboration of the production process by integrating a multi-agent system under the AutoGen architecture in key modules. Agent can process a large amount of production data in real time, provide automated decision support based on real-time data and historical trends, break the data island, and realize comprehensive collaboration between different systems. This improvement in flexibility and automation level enables Agent to respond quickly to changes in production demand, reduce human errors and production interruptions, and provide a more intuitive and user-friendly experience. Agent's self-learning and optimization capabilities enable it to continuously improve performance over time, and the rapid exception response mechanism further reduces production delays. In addition, Agent's customized solutions can meet the unique needs of specific factories. Compared with traditional systems, Agent provides a more intelligent and automated way to manage advanced manufacturing processes.

[0085] In the above technical solution of this embodiment, first, the role played by each Agent in the intelligent manufacturing process is clarified, and these roles may include raw material supply Agent, production and processing Agent, quality inspection Agent, logistics and transportation Agent, etc. The system units corresponding to each Agent are configured and initialized to ensure that they can operate according to the preset rules and parameters.

[0086] The system collects and monitors the operating data of each Agent's corresponding system unit in real time. These data may include equipment status, production efficiency, energy consumption, product quality, etc. By analyzing these data, the system can determine whether there are abnormal conditions, such as equipment failure, reduced production efficiency, substandard product quality, etc.

[0087] The system evaluates the performance of each system unit corresponding to each Agent at a preset time period (such as daily, weekly or monthly). The evaluation content may include production efficiency, energy efficiency, failure rate and other aspects. According to the evaluation results, if the performance of a system unit is lower than the preset standard or there is a potential problem, the system will send an operation and maintenance prompt to the staff terminal so that timely measures can be taken for maintenance or optimization.

[0088] The effect of the above technical solution is: through real-time monitoring and performance evaluation, bottlenecks and problems in the production process can be discovered and solved in time, ensuring the smooth progress of the production process, thereby improving the overall production efficiency. Automated monitoring and abnormality judgment reduce the frequency and intensity of manual inspections and reduce labor costs. At the same time, potential problems are discovered and handled in a timely manner to avoid greater losses caused by the expansion of faults. Through strict monitoring and performance evaluation of the production process, it is ensured that each production link meets the quality requirements, thereby improving the quality of the final product and customer satisfaction. The agent-based intelligent manufacturing process coordination agent method is highly flexible and scalable. As production needs change, the role and configuration of the agent can be easily adjusted to adapt to new production scenarios and needs. The application of this method has promoted the in-depth development of intelligent manufacturing and provided strong support for the intelligent transformation of enterprises. By introducing intelligent agent technology, enterprises can manage and optimize production processes more efficiently and improve overall competitiveness.

[0089] An embodiment of the present invention defines the role of an Agent, and configures and initializes the system unit corresponding to each role Agent, including:

[0090] S101. Define the role of Agent;

[0091] S102, using the API provided by AutoGen to create a system unit corresponding to each role Agent; wherein the system unit corresponding to each role Agent includes a production monitoring agent, a quality control agent, a predictive maintenance agent, a supply chain coordination agent, and a production scheduling agent;

[0092] S103, configuring system parameters for the system unit corresponding to each role Agent, so as to clarify the task responsibilities corresponding to the system unit corresponding to each role Agent;

[0093] S104, configuring an LLM model corresponding to the task responsibilities according to the task responsibilities corresponding to the system units corresponding to each role Agent; wherein the LLM model is a model that has been trained and tested and stored in a database.

[0094] The working principle of the above technical solution is as follows: In the intelligent manufacturing process, the roles of various agents are first clarified, such as production monitoring agent, quality control agent, predictive maintenance agent, supply chain coordination agent and production scheduling agent, etc. These roles correspond to different links and responsibilities in the intelligent manufacturing process.

[0095] Using the API provided by AutoGen, corresponding system units are created for each defined Agent role. These system units are physical or logical entities that Agents use to perform tasks in the intelligent manufacturing process, and are responsible for collecting data, performing operations, transmitting information, etc.

[0096] For each system unit corresponding to the Agent role, detailed configuration of system parameters is performed. These parameters define the specific task responsibilities of each system unit, including which data to monitor, which operations to perform, and which other system units to interact with. Through parameter configuration, each system unit can be ensured to operate according to the preset rules and processes.

[0097] According to the task responsibilities of the system unit corresponding to each role Agent, the corresponding LLM (Large Language Model) model is configured. These LLM models are trained and tested models stored in the database, which can understand and process natural language input and generate corresponding output. By integrating LLM models, the system can process and analyze data more intelligently and make more accurate decisions and predictions.

[0098] The effect of the above technical solution is: by introducing Agent and LLM model, each link in the intelligent manufacturing process can be automated and intelligent. Agent can autonomously perform tasks, monitor operating status, make decisions, etc., while LLM model can provide more accurate data analysis and prediction, thereby improving the intelligence level of the entire process. The collaborative work between various Agent roles can optimize resource allocation and ensure that each link in the production process can operate efficiently. At the same time, through real-time monitoring and data analysis, the system can promptly discover and solve potential problems, avoid production interruptions and waste, and thus improve production efficiency.

[0099] The quality control agent can monitor product quality data in real time to ensure that the product meets standards and requirements. By integrating the LLM model for data analysis, the system can more accurately identify potential quality problems and take appropriate measures to improve them. This helps to improve product quality and customer satisfaction. The predictive maintenance agent can predict equipment failures and maintenance needs in advance, thereby avoiding production interruptions and losses caused by sudden equipment downtime. By maintaining and servicing equipment in a timely manner, the system can reduce operation and maintenance costs and risks and improve the reliability and stability of equipment. The technical solution adopts a modular and configurable design approach, which enables the system to be flexibly adjusted according to different production needs and scenarios. At the same time, with the continuous development and advancement of technology, the system can easily integrate new agent roles and LLM models to adapt to more complex production processes and higher intelligence requirements.

[0100] In one embodiment of the present invention, the operation data of the system unit corresponding to each role Agent is monitored in real time, and whether an abnormality exists according to the operation data of the system unit corresponding to each role Agent is determined, including:

[0101] S201, real-time monitoring of the operating data of the system unit corresponding to each role Agent;

[0102] S202: Determine whether there is an abnormality in the operation data of the system unit corresponding to each role Agent according to the invalid data acquired in each unit time of the system unit corresponding to each role Agent.

[0103] Wherein, determining whether the operation data of the system unit corresponding to each role Agent is abnormal according to the invalid data acquired per unit time of the system unit corresponding to each role Agent includes:

[0104] S2021, extracting the amount of invalid data corresponding to each unit time acquired by the system unit corresponding to each role Agent;

[0105] S2022, extracting the data type corresponding to the invalid data acquired per unit time by the system unit corresponding to each role Agent and the weight value corresponding to the data type;

[0106] S2023. Obtain an evaluation parameter of the invalid data impact degree according to the data volume corresponding to the invalid data and its corresponding weight value; wherein the evaluation parameter of the invalid data impact degree is obtained by the following formula:

[0107]

[0108] Where I represents the evaluation parameter of the impact of invalid data; M represents the number of unit time that has passed; k hj Indicates the weight and value of the type of invalid data generated in the jth unit time; k z represents the weights and values ​​corresponding to all data types; H represents the evaluation parameter variable, and the evaluation parameter variable is obtained by the following formula:

[0109]

[0110] Among them, H represents the evaluation parameter; n j V represents the number of data types with invalid data corresponding to the jth unit time; i k represents the amount of invalid data corresponding to the jth unit time of the i-th data type; i Indicates the weight value corresponding to the i-th data type; D i Represents the diversity index corresponding to the i-th data type, which is used to indicate the dispersion or frequency of this type of data in the overall data; i represents the attenuation coefficient corresponding to the diversity index of the i-th data type and the i-th data type; n z Indicates the total number of data types;

[0111] S2024. When the evaluation parameter of the impact degree of the invalid data exceeds a preset impact degree evaluation parameter threshold, it is determined that there is an abnormality in the operation of the system unit corresponding to the role Agent corresponding to which the evaluation parameter of the impact degree of the invalid data exceeds the preset impact degree evaluation parameter threshold.

[0112] The working principle of the above technical solution is as follows: the solution first monitors the operating data of the system unit corresponding to each role Agent in real time. This is a key step to ensure system stability and timely discover potential problems. In each unit time, the system extracts and counts the invalid data obtained by the system unit corresponding to each role Agent. These invalid data may be generated due to various reasons, such as data transmission errors, sensor failures, etc. By analyzing the data volume, data type and corresponding weight value of the invalid data, the system calculates the impact evaluation parameter (I) of the invalid data. This parameter comprehensively considers the number, type, weight and dispersion or frequency of occurrence of the data type of the invalid data in the overall data. Finally, the system compares the calculated invalid data impact evaluation parameter with the preset threshold. If the evaluation parameter exceeds the threshold, it is determined that the corresponding role Agent system unit is abnormal in operation.

[0113] The effect of the above technical solution is: through real-time monitoring and abnormal judgment, the system can timely discover and deal with potential operation problems, thereby avoiding the expansion of problems and improving the stability and reliability of the overall system. The solution can monitor and judge the system unit corresponding to each role Agent, making the problem location more accurate and facilitating the rapid adoption of targeted solutions. By evaluating the impact of invalid data, the system can allocate resources more reasonably, give priority to problems with greater impact, and improve resource utilization efficiency. Timely discovery and resolution of system problems can reduce user inconvenience and losses caused by system anomalies and improve user experience and satisfaction. For system units involving sensitive data, timely discovery and processing of invalid data helps protect data security and prevent data leakage or abuse.

[0114] In summary, this technical solution achieves accurate monitoring of the operating status of system units and abnormality judgment through real-time monitoring, invalid data identification and statistics, invalid data impact assessment, and abnormality judgment. It is of great significance for improving system stability, optimizing resource allocation, enhancing user experience, and enhancing data security.

[0115] In one embodiment of the present invention, a performance evaluation is performed on a system unit corresponding to each role Agent according to a preset time period, and a determination is made whether to send an operation and maintenance prompt to a staff terminal according to the performance evaluation result of the system unit corresponding to each role Agent, including:

[0116] S301, real-time monitoring of the operating parameters of the system unit corresponding to each role Agent in each preset time period, wherein the operating parameters include task completion rate, task start response time and task execution time;

[0117] S302, obtaining a unit operation evaluation coefficient corresponding to the system unit corresponding to each role Agent according to the operation parameters of the system unit corresponding to each role Agent; wherein the unit operation evaluation coefficient is obtained by the following formula:

[0118]

[0119] Where S represents the unit operation evaluation coefficient; n represents the number of time periods; P i represents the task completion rate corresponding to the i-th time period; T zi represents the median value of the task start response time corresponding to the number of left and right task starts in the i-th time period; T pi represents the average task start response time corresponding to the number of task starts in the i-th time period; T maxi represents the maximum task start response time corresponding to the number of task starts in the i-th time period; T b represents the standard deviation of the task start response time corresponding to the number of task starts in the i-th time period; m i represents the number of tasks in the i-th time period; T sj represents the task execution time of the jth task corresponding to the jth time period; T yj represents the reference duration of the jth task corresponding to the jth time period; M b represents the standard deviation of the number of tasks corresponding to n time periods; M max Indicates the maximum number of tasks in a single time period that appear in n time periods;

[0120] S303, when the unit operation evaluation coefficient corresponding to the system unit corresponding to each role Agent is lower than the corresponding evaluation coefficient threshold, the correlation coefficient between the system unit corresponding to the role Agent lower than the corresponding evaluation coefficient threshold and the system units corresponding to other role Agents not lower than the corresponding evaluation coefficient threshold is obtained; wherein the correlation coefficient is obtained by the following formula:

[0121]

[0122] Where f represents the correlation coefficient; k represents the number of system units corresponding to other role agents that are not lower than the corresponding evaluation coefficient threshold; S i represents the unit operation evaluation coefficient of the system unit corresponding to the i-th other role agent that is not lower than its corresponding evaluation coefficient threshold; S yi represents the evaluation coefficient threshold corresponding to the system unit of the i-th other role agent whose evaluation coefficient threshold is not lower than its corresponding evaluation coefficient threshold; S dIndicates the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is lower than its corresponding evaluation coefficient threshold; S cmax It represents the maximum value of the difference between the unit operation evaluation coefficients of the system units corresponding to other role agents that are not lower than the corresponding evaluation coefficient threshold; S z Indicates that the middle value of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than the corresponding evaluation coefficient threshold; S max and S min Indicates the maximum and minimum values ​​of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than its corresponding evaluation coefficient threshold;

[0123] S304, obtaining a comprehensive correlation coefficient by using the correlation coefficients of the system units corresponding to the role Agents whose evaluation coefficients are lower than the corresponding evaluation coefficient thresholds and the system units corresponding to other role Agents whose evaluation coefficients are not lower than the corresponding evaluation coefficient thresholds; wherein the comprehensive correlation coefficient is obtained by the following formula:

[0124]

[0125] Where CI represents the comprehensive correlation coefficient; r represents the number of system units corresponding to the role agent that is lower than its corresponding evaluation coefficient threshold; f i It represents the correlation coefficient of the system unit corresponding to the i-th role agent that is lower than its corresponding evaluation coefficient threshold;

[0126] S305: When the comprehensive correlation coefficient exceeds a preset coefficient threshold, it is determined that a system operation and maintenance prompt needs to be sent to the staff terminal.

[0127] The working principle of the above technical solution is as follows: the solution first monitors the operating parameters of the system unit corresponding to each role Agent in real time according to a preset time period. These parameters include task completion rate, task start response time and task execution time, which are important indicators for evaluating the performance of the system unit. According to the collected operating parameters, a specific formula is used to calculate the unit operation evaluation coefficient (S) of the system unit corresponding to each role Agent. This coefficient comprehensively considers multiple statistics of task completion rate, task start response time (such as median value, average value, maximum value, standard deviation, etc.) and the difference between task execution time and reference time, so as to comprehensively evaluate the performance of the system unit.

[0128] When the unit operation evaluation coefficient of a system unit is lower than its corresponding evaluation coefficient threshold, it is considered that the system unit may have performance problems. At this time, the correlation coefficient (f) between the system unit and other system units that are not lower than the threshold is further calculated to evaluate the degree of mutual influence between them. Then, the comprehensive correlation coefficient (CI) is calculated using the correlation coefficients of all system units below the threshold. This coefficient reflects the comprehensive correlation degree between low-performance system unit groups.

[0129] When the comprehensive correlation coefficient exceeds the preset coefficient threshold, it indicates that there is a strong mutual influence between these low-performance system units, which may need to be processed in a coordinated manner. At this time, the system determines that it is necessary to send a system operation and maintenance prompt to the staff terminal so that timely measures can be taken for intervention and optimization.

[0130] The effect of the above technical solution is: by comprehensively considering multiple operating parameters and statistics, the solution can achieve a refined evaluation of the performance of the system units, which helps to more accurately identify potential problems. The combination of real-time monitoring and periodic evaluation can promptly detect performance anomalies of system units, providing the possibility for rapid response and problem solving. By calculating the correlation coefficient and the comprehensive correlation coefficient, the solution can reveal the mutual influence relationship between low-performance system units, and provide strong support for collaborative processing and optimization. When it is necessary to send an operation and maintenance prompt, the system can accurately locate the group of system units that need attention, avoid blind troubleshooting and waste of resources, and thus improve operation and maintenance efficiency. Timely detection and handling of performance problems of system units can help reduce system failures and downtime, and enhance the stability and reliability of the overall system.

[0131] In summary, this technical solution provides strong support for system operation and maintenance through refined performance evaluation, timely anomaly detection, correlation analysis, and efficient operation and maintenance prompt mechanism, which helps to improve system performance and stability.

[0132] The embodiment of the present invention proposes a system based on AutoGen intelligent manufacturing process coordination agent, such as Figure 2 As shown, the system based on AutoGen intelligent manufacturing process coordination agent includes:

[0133] Define the initialization module, which is used to define the role of the Agent and configure and initialize the system unit corresponding to each role Agent;

[0134] A real-time monitoring module is used to monitor the operation data of the system unit corresponding to each role Agent in real time, and determine whether there is an abnormality according to the operation data of the system unit corresponding to each role Agent;

[0135] The performance evaluation module is used to perform performance evaluation on the system unit corresponding to each role Agent according to a preset time period, and determine whether to send an operation and maintenance prompt to the staff terminal according to the performance evaluation result of the system unit corresponding to each role Agent.

[0136] The working principle of the above technical solution is as follows: AutoGen is a multi-agent framework that supports the development of LLM applications using multiple agents that can communicate with each other to solve tasks. Agents in AutoGen are customizable, conversational, and seamlessly allow human participation. They can run LLMs in various modes that use a combination of human input and tools. In the key technical modules, agents with different roles are integrated, including production monitoring agents to monitor the status of the production line in real time (machine operation, production speed, etc.); quality control agents to be responsible for product quality inspection and real-time analysis of product quality data; predictive maintenance agents to predict equipment failures and arrange maintenance work in advance; supply chain coordination agents to manage raw material supply and inventory and respond to supply chain changes; production scheduling agents to intelligently arrange production plans based on order requirements and production capacity, etc.

[0137] The above technical solution of this embodiment realizes real-time monitoring, predictive analysis, adaptive decision-making and integrated collaboration of the production process by integrating a multi-agent system under the AutoGen architecture in key modules. Agent can process a large amount of production data in real time, provide automated decision support based on real-time data and historical trends, break the data island, and realize comprehensive collaboration between different systems. This improvement in flexibility and automation level enables Agent to respond quickly to changes in production demand, reduce human errors and production interruptions, and provide a more intuitive and user-friendly experience. Agent's self-learning and optimization capabilities enable it to continuously improve performance over time, and the rapid exception response mechanism further reduces production delays. In addition, Agent's customized solutions can meet the unique needs of specific factories. Compared with traditional systems, Agent provides a more intelligent and automated way to manage advanced manufacturing processes.

[0138] In the above technical solution of this embodiment, first, the role played by each Agent in the intelligent manufacturing process is clarified, and these roles may include raw material supply Agent, production and processing Agent, quality inspection Agent, logistics and transportation Agent, etc. The system units corresponding to each Agent are configured and initialized to ensure that they can operate according to the preset rules and parameters.

[0139] The system collects and monitors the operating data of each Agent's corresponding system unit in real time. These data may include equipment status, production efficiency, energy consumption, product quality, etc. By analyzing these data, the system can determine whether there are abnormal conditions, such as equipment failure, reduced production efficiency, substandard product quality, etc.

[0140] The system evaluates the performance of each system unit corresponding to each Agent at a preset time period (such as daily, weekly or monthly). The evaluation content may include production efficiency, energy efficiency, failure rate and other aspects. According to the evaluation results, if the performance of a system unit is lower than the preset standard or there is a potential problem, the system will send an operation and maintenance prompt to the staff terminal so that timely measures can be taken for maintenance or optimization.

[0141] The effect of the above technical solution is: through real-time monitoring and performance evaluation, bottlenecks and problems in the production process can be discovered and solved in time, ensuring the smooth progress of the production process, thereby improving the overall production efficiency. Automated monitoring and abnormality judgment reduce the frequency and intensity of manual inspections and reduce labor costs. At the same time, potential problems are discovered and handled in a timely manner to avoid greater losses caused by the expansion of faults. Through strict monitoring and performance evaluation of the production process, it is ensured that each production link meets the quality requirements, thereby improving the quality of the final product and customer satisfaction. The agent-based intelligent manufacturing process coordination agent method is highly flexible and scalable. As production needs change, the role and configuration of the agent can be easily adjusted to adapt to new production scenarios and needs. The application of this method has promoted the in-depth development of intelligent manufacturing and provided strong support for the intelligent transformation of enterprises. By introducing intelligent agent technology, enterprises can manage and optimize production processes more efficiently and improve overall competitiveness.

[0142] In one embodiment of the present invention, the definition initialization module includes:

[0143] Definition module, used to define the role of Agent;

[0144] A creation module, used to create a system unit corresponding to each role Agent using the API provided by AutoGen; wherein the system unit corresponding to each role Agent includes a production monitoring agent, a quality control agent, a predictive maintenance agent, a supply chain coordination agent, and a production scheduling agent;

[0145] The parameter configuration module is used to configure the system parameters for the system unit corresponding to each role Agent, so as to clarify the task responsibilities corresponding to the system unit corresponding to each role Agent;

[0146] The model configuration module is used to configure the LLM model corresponding to the task responsibilities according to the task responsibilities corresponding to the system units corresponding to each role Agent; wherein the LLM model is a model that has been trained and tested and stored in the database.

[0147] The working principle of the above technical solution is as follows: In the intelligent manufacturing process, the roles of various agents are first clarified, such as production monitoring agent, quality control agent, predictive maintenance agent, supply chain coordination agent and production scheduling agent, etc. These roles correspond to different links and responsibilities in the intelligent manufacturing process.

[0148] Using the API provided by AutoGen, corresponding system units are created for each defined Agent role. These system units are physical or logical entities that Agents use to perform tasks in the intelligent manufacturing process, and are responsible for collecting data, performing operations, transmitting information, etc.

[0149] For each system unit corresponding to the Agent role, detailed configuration of system parameters is performed. These parameters define the specific task responsibilities of each system unit, including which data to monitor, which operations to perform, and which other system units to interact with. Through parameter configuration, each system unit can be ensured to operate according to the preset rules and processes.

[0150] According to the task responsibilities of the system unit corresponding to each role Agent, the corresponding LLM (Large Language Model) model is configured. These LLM models are trained and tested models stored in the database, which can understand and process natural language input and generate corresponding output. By integrating LLM models, the system can process and analyze data more intelligently and make more accurate decisions and predictions.

[0151] The effect of the above technical solution is: by introducing Agent and LLM model, each link in the intelligent manufacturing process can be automated and intelligent. Agent can autonomously perform tasks, monitor operating status, make decisions, etc., while LLM model can provide more accurate data analysis and prediction, thereby improving the intelligence level of the entire process. The collaborative work between various Agent roles can optimize resource allocation and ensure that each link in the production process can operate efficiently. At the same time, through real-time monitoring and data analysis, the system can promptly discover and solve potential problems, avoid production interruptions and waste, and thus improve production efficiency.

[0152] The quality control agent can monitor product quality data in real time to ensure that the product meets standards and requirements. By integrating the LLM model for data analysis, the system can more accurately identify potential quality problems and take appropriate measures to improve them. This helps to improve product quality and customer satisfaction. The predictive maintenance agent can predict equipment failures and maintenance needs in advance, thereby avoiding production interruptions and losses caused by sudden equipment downtime. By maintaining and servicing equipment in a timely manner, the system can reduce operation and maintenance costs and risks and improve the reliability and stability of equipment. The technical solution adopts a modular and configurable design approach, which enables the system to be flexibly adjusted according to different production needs and scenarios. At the same time, with the continuous development and advancement of technology, the system can easily integrate new agent roles and LLM models to adapt to more complex production processes and higher intelligence requirements.

[0153] In one embodiment of the present invention, the real-time monitoring module includes:

[0154] Data monitoring module, used to monitor the operation data of the system unit corresponding to each role Agent in real time;

[0155] The abnormality judgment module is used to determine whether there is an abnormality in the operation data of the system unit corresponding to each role Agent according to the invalid data obtained in each unit time of the system unit corresponding to each role Agent.

[0156] Wherein, determining whether the operation data of the system unit corresponding to each role Agent is abnormal according to the invalid data acquired per unit time of the system unit corresponding to each role Agent includes:

[0157] A first data information extraction module is used to extract the data volume corresponding to the invalid data acquired per unit time of the system unit corresponding to each role Agent;

[0158] A second data information extraction module is used to extract the data type corresponding to the invalid data acquired per unit time by the system unit corresponding to each role Agent and the weight value corresponding to the data type;

[0159] The evaluation parameter acquisition module is used to obtain the evaluation parameter of the invalid data impact degree according to the data volume corresponding to the invalid data and its corresponding weight value; wherein the evaluation parameter of the invalid data impact degree is obtained by the following formula:

[0160]

[0161] Where I represents the evaluation parameter of the impact of invalid data; M represents the number of unit time that has passed; k hjIndicates the weight and value of the type of invalid data generated in the jth unit time; k z represents the weights and values ​​corresponding to all data types; H represents the evaluation parameter variable, and the evaluation parameter variable is obtained by the following formula:

[0162]

[0163] Among them, H represents the evaluation parameter; n j V represents the number of data types with invalid data corresponding to the jth unit time; i k represents the amount of invalid data corresponding to the jth unit time of the i-th data type; i Indicates the weight value corresponding to the i-th data type; D i Represents the diversity index corresponding to the i-th data type, which is used to indicate the dispersion or frequency of this type of data in the overall data; i represents the attenuation coefficient corresponding to the diversity index of the i-th data type and the i-th data type; n z Indicates the total number of data types;

[0164] The abnormality determination module is used to determine that when the evaluation parameter of the impact degree of the invalid data exceeds the preset impact degree evaluation parameter threshold, there is an abnormality in the operation of the system unit corresponding to the role Agent corresponding to the evaluation parameter of the impact degree of the invalid data exceeding the preset impact degree evaluation parameter threshold.

[0165] The working principle of the above technical solution is as follows: the solution first monitors the operating data of the system unit corresponding to each role Agent in real time. This is a key step to ensure system stability and timely discover potential problems. In each unit time, the system extracts and counts the invalid data obtained by the system unit corresponding to each role Agent. These invalid data may be generated due to various reasons, such as data transmission errors, sensor failures, etc. By analyzing the data volume, data type and corresponding weight value of the invalid data, the system calculates the impact evaluation parameter (I) of the invalid data. This parameter comprehensively considers the number, type, weight and dispersion or frequency of occurrence of the data type of the invalid data in the overall data. Finally, the system compares the calculated invalid data impact evaluation parameter with the preset threshold. If the evaluation parameter exceeds the threshold, it is determined that the corresponding role Agent system unit is abnormal in operation.

[0166] The effect of the above technical solution is: through real-time monitoring and abnormal judgment, the system can timely discover and deal with potential operation problems, thereby avoiding the expansion of problems and improving the stability and reliability of the overall system. The solution can monitor and judge the system unit corresponding to each role Agent, making the problem location more accurate and facilitating the rapid adoption of targeted solutions. By evaluating the impact of invalid data, the system can allocate resources more reasonably, give priority to problems with greater impact, and improve resource utilization efficiency. Timely discovery and resolution of system problems can reduce user inconvenience and losses caused by system anomalies and improve user experience and satisfaction. For system units involving sensitive data, timely discovery and processing of invalid data helps protect data security and prevent data leakage or abuse.

[0167] In summary, this technical solution achieves accurate monitoring of the operating status of system units and abnormality judgment through real-time monitoring, invalid data identification and statistics, invalid data impact assessment, and abnormality judgment. It is of great significance for improving system stability, optimizing resource allocation, enhancing user experience, and enhancing data security.

[0168] In one embodiment of the present invention, the performance evaluation module includes:

[0169] An operation parameter monitoring module is used to monitor in real time the operation parameters of the system unit corresponding to each role Agent in each of the preset time periods, wherein the operation parameters include task completion rate, task start response time and task execution time;

[0170] The unit operation evaluation coefficient acquisition module is used to acquire the unit operation evaluation coefficient corresponding to the system unit corresponding to each role Agent according to the operation parameters of the system unit corresponding to each role Agent; wherein the unit operation evaluation coefficient is acquired by the following formula:

[0171]

[0172] Where S represents the unit operation evaluation coefficient; n represents the number of time periods; P i represents the task completion rate corresponding to the i-th time period; T zi represents the median value of the task start response time corresponding to the number of left and right task starts in the i-th time period; T pi represents the average task start response time corresponding to the number of task starts in the i-th time period; T maxi represents the maximum task start response time corresponding to the number of task starts in the i-th time period; T b represents the standard deviation of the task start response time corresponding to the number of task starts in the i-th time period; m irepresents the number of tasks in the i-th time period; T sj represents the task execution time of the jth task corresponding to the jth time period; T yj represents the reference duration of the jth task corresponding to the jth time period; M b represents the standard deviation of the number of tasks corresponding to n time periods; M max Indicates the maximum number of tasks in a single time period that appear in n time periods;

[0173] The correlation coefficient acquisition module is used to obtain the correlation coefficient between the system unit corresponding to the role agent below the corresponding evaluation coefficient threshold and the system units corresponding to other role agents not below the corresponding evaluation coefficient threshold when the unit operation evaluation coefficient corresponding to the system unit corresponding to each role agent is lower than the corresponding evaluation coefficient threshold; wherein the correlation coefficient is obtained by the following formula:

[0174]

[0175] Where f represents the correlation coefficient; k represents the number of system units corresponding to other role agents that are not lower than the corresponding evaluation coefficient threshold; S i represents the unit operation evaluation coefficient of the system unit corresponding to the i-th other role agent that is not lower than its corresponding evaluation coefficient threshold; S yi represents the evaluation coefficient threshold corresponding to the system unit of the i-th other role agent whose evaluation coefficient threshold is not lower than its corresponding evaluation coefficient threshold; S d Indicates the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is lower than its corresponding evaluation coefficient threshold; S cmax It represents the maximum value of the difference between the unit operation evaluation coefficients of the system units corresponding to other role agents that are not lower than their corresponding evaluation coefficient thresholds; S z Indicates that the middle value of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than the corresponding evaluation coefficient threshold; S max and S min Indicates the maximum and minimum values ​​of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than its corresponding evaluation coefficient threshold;

[0176] The comprehensive correlation coefficient acquisition module is used to obtain the comprehensive correlation coefficient by using the correlation coefficients of the system units corresponding to the role Agent whose evaluation coefficient is lower than the corresponding evaluation coefficient threshold and the system units corresponding to other role Agents whose evaluation coefficient is not lower than the corresponding evaluation coefficient threshold; wherein the comprehensive correlation coefficient is obtained by the following formula:

[0177]

[0178] Where CI represents the comprehensive correlation coefficient; r represents the number of system units corresponding to the role agent that is lower than its corresponding evaluation coefficient threshold; f i It represents the correlation coefficient of the system unit corresponding to the i-th role agent that is lower than its corresponding evaluation coefficient threshold;

[0179] The operation and maintenance prompt module is used to determine the need to send a system operation and maintenance prompt to the staff terminal when the comprehensive correlation coefficient exceeds a preset coefficient threshold.

[0180] The working principle of the above technical solution is as follows: the solution first monitors the operating parameters of the system unit corresponding to each role Agent in real time according to a preset time period. These parameters include task completion rate, task start response time and task execution time, which are important indicators for evaluating the performance of the system unit. According to the collected operating parameters, a specific formula is used to calculate the unit operation evaluation coefficient (S) of the system unit corresponding to each role Agent. This coefficient comprehensively considers multiple statistics of task completion rate, task start response time (such as median value, average value, maximum value, standard deviation, etc.) and the difference between task execution time and reference time, so as to comprehensively evaluate the performance of the system unit.

[0181] When the unit operation evaluation coefficient of a system unit is lower than its corresponding evaluation coefficient threshold, it is considered that the system unit may have performance problems. At this time, the correlation coefficient (f) between the system unit and other system units that are not lower than the threshold is further calculated to evaluate the degree of mutual influence between them. Then, the comprehensive correlation coefficient (CI) is calculated using the correlation coefficients of all system units below the threshold. This coefficient reflects the comprehensive correlation degree between low-performance system unit groups.

[0182] When the comprehensive correlation coefficient exceeds the preset coefficient threshold, it indicates that there is a strong mutual influence between these low-performance system units, which may need to be processed in a coordinated manner. At this time, the system determines that it is necessary to send a system operation and maintenance prompt to the staff terminal so that timely measures can be taken for intervention and optimization.

[0183] The effect of the above technical solution is: by comprehensively considering multiple operating parameters and statistics, the solution can achieve a refined evaluation of the performance of the system units, which helps to more accurately identify potential problems. The combination of real-time monitoring and periodic evaluation can promptly detect performance anomalies of system units, providing the possibility for rapid response and problem solving. By calculating the correlation coefficient and the comprehensive correlation coefficient, the solution can reveal the mutual influence relationship between low-performance system units, and provide strong support for collaborative processing and optimization. When it is necessary to send an operation and maintenance prompt, the system can accurately locate the group of system units that need attention, avoid blind troubleshooting and waste of resources, and thus improve operation and maintenance efficiency. Timely detection and handling of performance problems of system units can help reduce system failures and downtime, and enhance the stability and reliability of the overall system.

[0184] In summary, this technical solution provides strong support for system operation and maintenance through refined performance evaluation, timely anomaly detection, correlation analysis, and efficient operation and maintenance prompt mechanism, which helps to improve system performance and stability.

[0185] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method based on AutoGen intelligent manufacturing process coordination agent, characterized in that: The method based on AutoGen intelligent manufacturing process coordination agent includes: Define the roles of Agents, and configure and initialize the system units corresponding to each role Agent; Monitor the operation data of the system unit corresponding to each role Agent in real time, and determine whether there is an abnormality according to the operation data of the system unit corresponding to each role Agent; The performance of the system unit corresponding to each role Agent is evaluated according to a preset time period, and it is determined whether to send an operation and maintenance prompt to the staff terminal according to the performance evaluation result of the system unit corresponding to each role Agent.

2. The method based on AutoGen intelligent manufacturing process coordination agent according to claim 1 is characterized in that: Define the roles of the Agent and configure and initialize the system units corresponding to each role Agent, including: Define the role of the Agent; Using the API provided by AutoGen to create a system unit corresponding to each role Agent; wherein the system unit corresponding to each role Agent includes a production monitoring agent, a quality control agent, a predictive maintenance agent, a supply chain coordination agent, and a production scheduling agent; Configure system parameters for the system unit corresponding to each role Agent to clarify the task responsibilities of the system unit corresponding to each role Agent; According to the task responsibilities corresponding to the system units corresponding to each role Agent, an LLM model corresponding to the task responsibilities is configured.

3. The method based on AutoGen intelligent manufacturing process coordination agent according to claim 1 is characterized in that: Real-time monitoring of the operating data of the system unit corresponding to each role Agent, and determining whether there is an abnormality according to the operating data of the system unit corresponding to each role Agent, including: Real-time monitoring of the operating data of the system unit corresponding to each role Agent; Whether the operation data of the system unit corresponding to each role Agent is abnormal is determined according to the invalid data acquired in each unit time of the system unit corresponding to each role Agent.

4. The method based on AutoGen intelligent manufacturing process coordination agent according to claim 3 is characterized in that: Determining whether there is an abnormality in the operation data of the system unit corresponding to each role Agent according to the invalid data acquired per unit time of the system unit corresponding to each role Agent includes: Extracting the amount of invalid data corresponding to each unit time acquired by the system unit corresponding to each role Agent; Extracting the data type corresponding to the invalid data acquired per unit time by the system unit corresponding to each role Agent and the weight value corresponding to the data type; The evaluation parameter of the invalid data influence degree is obtained according to the data amount corresponding to the invalid data and the corresponding weight value; wherein the evaluation parameter of the invalid data influence degree is obtained by the following formula: Where I represents the evaluation parameter of the impact of invalid data; M represents the number of unit time that has passed; k hj Indicates the weight and value of the type of invalid data generated in the jth unit time; k z represents the weights and values ​​corresponding to all data types; H represents the evaluation parameter variable, and the evaluation parameter variable is obtained by the following formula: Among them, H represents the evaluation parameter; n j V represents the number of data types with invalid data corresponding to the jth unit time; i k represents the amount of invalid data corresponding to the jth unit time of the i-th data type; i Indicates the weight value corresponding to the i-th data type; D i Represents the diversity index corresponding to the i-th data type, which is used to indicate the dispersion or frequency of this type of data in the overall data; i represents the attenuation coefficient corresponding to the diversity index of the i-th data type and the i-th data type; n z Indicates the total number of data types; When the evaluation parameter of the impact degree of the invalid data exceeds a preset impact degree evaluation parameter threshold, it is determined that there is an abnormality in the operation of the system unit corresponding to the role Agent corresponding to which the evaluation parameter of the impact degree of the invalid data exceeds the preset impact degree evaluation parameter threshold.

5. The method based on AutoGen intelligent manufacturing process coordination agent according to claim 1 is characterized in that: The system unit corresponding to each role Agent is evaluated for performance according to a preset time period, and whether to send an operation and maintenance prompt to the staff terminal is determined according to the performance evaluation result of the system unit corresponding to each role Agent, including: Real-time monitoring of the operating parameters of the system unit corresponding to each role Agent in each of the preset time periods, wherein the operating parameters include task completion rate, task start response time and task execution time; The unit operation evaluation coefficient corresponding to the system unit corresponding to each role Agent is obtained according to the operation parameters of the system unit corresponding to each role Agent; wherein the unit operation evaluation coefficient is obtained by the following formula: Where S represents the unit operation evaluation coefficient; n represents the number of time periods; P i represents the task completion rate corresponding to the i-th time period; T zi represents the median value of the task start response time corresponding to the number of left and right task starts in the i-th time period; T pi represents the average task start response time corresponding to the number of task starts in the i-th time period; T maxi represents the maximum task start response time corresponding to the number of task starts in the i-th time period; T b represents the standard deviation of the task start response time corresponding to the number of task starts in the i-th time period; m i represents the number of tasks in the i-th time period; T sj represents the task execution time of the jth task corresponding to the jth time period; T yj represents the reference duration of the jth task corresponding to the jth time period; M b represents the standard deviation of the number of tasks corresponding to n time periods; M max Indicates the maximum number of tasks in a single time period that appear in n time periods; When the unit operation evaluation coefficient corresponding to the system unit corresponding to each role Agent is lower than the corresponding evaluation coefficient threshold, the correlation coefficient between the system unit corresponding to the role Agent lower than the corresponding evaluation coefficient threshold and the system units corresponding to other role Agents not lower than the corresponding evaluation coefficient threshold is obtained; wherein the correlation coefficient is obtained by the following formula: Where f represents the correlation coefficient; k represents the number of system units corresponding to other role agents that are not lower than the corresponding evaluation coefficient threshold; S i represents the unit operation evaluation coefficient of the system unit corresponding to the i-th other role agent that is not lower than its corresponding evaluation coefficient threshold; S yi represents the evaluation coefficient threshold corresponding to the system unit of the i-th other role agent whose evaluation coefficient threshold is not lower than its corresponding evaluation coefficient threshold; S d Indicates the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is lower than its corresponding evaluation coefficient threshold; S cmax It represents the maximum value of the difference between the unit operation evaluation coefficients of the system units corresponding to other role agents that are not lower than their corresponding evaluation coefficient thresholds; S z Indicates that the middle value of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than the corresponding evaluation coefficient threshold; S max and S min Indicates the maximum and minimum values ​​of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than its corresponding evaluation coefficient threshold; The comprehensive correlation coefficient is obtained by using the correlation coefficient of the system unit corresponding to the role agent whose evaluation coefficient is lower than the corresponding evaluation coefficient threshold and the system unit corresponding to other role agents whose evaluation coefficient is not lower than the corresponding evaluation coefficient threshold; wherein the comprehensive correlation coefficient is obtained by the following formula: Where CI represents the comprehensive correlation coefficient; r represents the number of system units corresponding to the role agent that is lower than its corresponding evaluation coefficient threshold; f i It represents the correlation coefficient of the system unit corresponding to the i-th role agent that is lower than its corresponding evaluation coefficient threshold; When the comprehensive correlation coefficient exceeds a preset coefficient threshold, it is determined that a system operation and maintenance prompt needs to be sent to the staff terminal.

6. The system based on AutoGen intelligent manufacturing process coordination agent is characterized by: The system based on AutoGen intelligent manufacturing process coordination agent includes: Define the initialization module, which is used to define the role of the Agent and configure and initialize the system unit corresponding to each role Agent; A real-time monitoring module is used to monitor the operation data of the system unit corresponding to each role Agent in real time, and determine whether there is an abnormality according to the operation data of the system unit corresponding to each role Agent; The performance evaluation module is used to perform performance evaluation on the system unit corresponding to each role Agent according to a preset time period, and determine whether to send an operation and maintenance prompt to the staff terminal according to the performance evaluation result of the system unit corresponding to each role Agent.

7. The system based on AutoGen intelligent manufacturing process coordination agent according to claim 6 is characterized in that: The definition initialization module includes: Definition module, used to define the role of Agent; A creation module, used to create a system unit corresponding to each role Agent using the API provided by AutoGen; wherein the system unit corresponding to each role Agent includes a production monitoring agent, a quality control agent, a predictive maintenance agent, a supply chain coordination agent, and a production scheduling agent; The parameter configuration module is used to configure the system parameters for the system unit corresponding to each role Agent, so as to clarify the task responsibilities corresponding to the system unit corresponding to each role Agent; The model configuration module is used to configure the LLM model corresponding to the task responsibility according to the task responsibility corresponding to the system unit corresponding to each role Agent.

8. The system based on AutoGen intelligent manufacturing process coordination agent according to claim 6 is characterized in that: The real-time monitoring module comprises: Data monitoring module, used to monitor the operation data of the system unit corresponding to each role Agent in real time; The abnormality judgment module is used to determine whether there is an abnormality in the operation data of the system unit corresponding to each role Agent according to the invalid data obtained in each unit time of the system unit corresponding to each role Agent.

9. The system based on AutoGen intelligent manufacturing process coordination agent according to claim 8 is characterized in that: Determining whether there is an abnormality in the operation data of the system unit corresponding to each role Agent according to the invalid data acquired per unit time of the system unit corresponding to each role Agent includes: A first data information extraction module is used to extract the data volume corresponding to the invalid data acquired per unit time of the system unit corresponding to each role Agent; A second data information extraction module is used to extract the data type corresponding to the invalid data acquired per unit time by the system unit corresponding to each role Agent and the weight value corresponding to the data type; The evaluation parameter acquisition module is used to obtain the evaluation parameter of the invalid data impact degree according to the data volume corresponding to the invalid data and its corresponding weight value; wherein the evaluation parameter of the invalid data impact degree is obtained by the following formula: Where I represents the evaluation parameter of the impact of invalid data; M represents the number of unit time that has passed; k hj Indicates the weight and value of the type of invalid data generated in the jth unit time; k z represents the weights and values ​​corresponding to all data types; H represents the evaluation parameter variable, and the evaluation parameter variable is obtained by the following formula: Among them, H represents the evaluation parameter; n j V represents the number of data types with invalid data corresponding to the jth unit time; i k represents the amount of invalid data corresponding to the jth unit time of the i-th data type; i Indicates the weight value corresponding to the i-th data type; D i Represents the diversity index corresponding to the i-th data type, which is used to indicate the dispersion or frequency of this type of data in the overall data; i represents the attenuation coefficient corresponding to the diversity index of the i-th data type and the i-th data type; n z Indicates the total number of data types; The abnormality determination module is used to determine that when the evaluation parameter of the impact degree of the invalid data exceeds the preset impact degree evaluation parameter threshold, there is an abnormality in the operation of the system unit corresponding to the role Agent corresponding to the evaluation parameter of the impact degree of the invalid data exceeding the preset impact degree evaluation parameter threshold.

10. The system based on AutoGen intelligent manufacturing process coordination agent according to claim 6 is characterized in that: The performance evaluation module includes: An operation parameter monitoring module is used to monitor in real time the operation parameters of the system unit corresponding to each role Agent in each of the preset time periods, wherein the operation parameters include task completion rate, task start response time and task execution time; The unit operation evaluation coefficient acquisition module is used to acquire the unit operation evaluation coefficient corresponding to the system unit corresponding to each role Agent according to the operation parameters of the system unit corresponding to each role Agent; wherein the unit operation evaluation coefficient is acquired by the following formula: Where S represents the unit operation evaluation coefficient; n represents the number of time periods; P i represents the task completion rate corresponding to the i-th time period; T zi represents the median value of the task start response time corresponding to the number of left and right task starts in the i-th time period; T pi represents the average task start response time corresponding to the number of task starts in the i-th time period; T maxi represents the maximum task start response time corresponding to the number of task starts in the i-th time period; T b represents the standard deviation of the task start response time corresponding to the number of task starts in the i-th time period; m i represents the number of tasks in the i-th time period; T sj represents the task execution time of the jth task corresponding to the jth time period; T yj represents the reference duration of the jth task corresponding to the jth time period; M b represents the standard deviation of the number of tasks corresponding to n time periods; M max Indicates the maximum number of tasks in a single time period that appear in n time periods; The correlation coefficient acquisition module is used to obtain the correlation coefficient between the system unit corresponding to the role agent below the corresponding evaluation coefficient threshold and the system units corresponding to other role agents not below the corresponding evaluation coefficient threshold when the unit operation evaluation coefficient corresponding to the system unit corresponding to each role agent is lower than the corresponding evaluation coefficient threshold; wherein the correlation coefficient is obtained by the following formula: Where f represents the correlation coefficient; k represents the number of system units corresponding to other role agents that are not lower than the corresponding evaluation coefficient threshold; S i represents the unit operation evaluation coefficient of the system unit corresponding to the i-th other role agent that is not lower than its corresponding evaluation coefficient threshold; S yi represents the evaluation coefficient threshold corresponding to the system unit of the i-th other role agent whose evaluation coefficient threshold is not lower than its corresponding evaluation coefficient threshold; S d Indicates the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is lower than its corresponding evaluation coefficient threshold; S cmax It represents the maximum value of the difference between the unit operation evaluation coefficients of the system units corresponding to other role agents that are not lower than their corresponding evaluation coefficient thresholds; S z Indicates that the middle value of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than the corresponding evaluation coefficient threshold; S max and S min Indicates the maximum and minimum values ​​of the unit operation evaluation coefficient of the system unit corresponding to the role Agent that is not lower than its corresponding evaluation coefficient threshold; The comprehensive correlation coefficient acquisition module is used to obtain the comprehensive correlation coefficient by using the correlation coefficients of the system units corresponding to the role Agent whose evaluation coefficient is lower than the corresponding evaluation coefficient threshold and the system units corresponding to other role Agents whose evaluation coefficient is not lower than the corresponding evaluation coefficient threshold; wherein the comprehensive correlation coefficient is obtained by the following formula: Where CI represents the comprehensive correlation coefficient; r represents the number of system units corresponding to the role agent that is lower than its corresponding evaluation coefficient threshold; f i It represents the correlation coefficient of the system unit corresponding to the i-th role agent that is lower than its corresponding evaluation coefficient threshold; The operation and maintenance prompt module is used to determine the need to send a system operation and maintenance prompt to the staff terminal when the comprehensive correlation coefficient exceeds a preset coefficient threshold.

Images