Multi-agent collaborative production optimization method and system based on industrial large model

Abstract

The invention discloses a multi-agent collaborative production optimization method and system based on an industrial large model, and relates to the technical field of production control, and the method comprises the steps: connecting an industrial knowledge base platform, carrying out the platform retrieval and interpretation reconstruction of a target production scene, and determining a production line interaction chain; aiming at a production line interaction chain and a driving agent construction interface, calling and splicing optimization of a code fragment library are carried out, lightweight construction of agents is carried out, multi-agent integration is carried out by introducing an intelligent interaction contract, and an industrial large model is established; and deploying the industrial large model in an industrial production system, establishing interaction with production line equipment, triggering the industrial large model along with the start of a target production scene, and performing automatic production management of the production line equipment through multi-agent cooperative control driven by an intelligent interaction contract. The technical problem that in the prior art, an industrial production system lacks intelligent and multi-dimensional collaborative optimization is solved, and the technical effects of improving the production efficiency, the resource utilization rate and collaborative work of production line equipment are achieved.

Description

Technical Field

[0001] The present invention relates to the field of production control technology, and in particular to a multi-agent collaborative production optimization method and system based on an industrial large model. Background Art

[0002] Industrial production processes typically rely on traditional manual scheduling or single-device control, lacking intelligent optimization mechanisms. The scheduling and management of production line equipment is mostly static, making it difficult to cope with complex production environments and emergencies, resulting in uneven resource allocation and low production efficiency. Traditional methods also fail to achieve effective collaboration between production line equipment, resulting in poor coordination between different production nodes and the inability to achieve full automation and real-time response. These issues limit the flexibility and efficiency of production processes and fail to meet the modern industrial demand for rapid response, precise control, and resource optimization. Summary of the Invention

[0003] This application provides a multi-agent collaborative production optimization method and system based on an industrial large model, which is used to solve the technical problem of the lack of intelligent and multi-dimensional collaborative optimization of industrial production systems in the existing technology.

[0004] In view of the above problems, this application provides a multi-agent collaborative production optimization method and system based on industrial large models.

[0005] In a first aspect, the present application provides a multi-agent collaborative production optimization method based on an industrial large model, the method comprising:

[0006] Connect to the industrial knowledge base platform, perform platform retrieval and interpretation reconstruction of the target production scenario, and determine the production line interaction chain, wherein each production node is marked with an interactive collaboration condition; for the production line interaction chain, drive the intelligent agent to build an interface, perform call and splicing optimization of the code fragment library, perform lightweight construction of the intelligent agent, and integrate multiple intelligent agents by introducing intelligent interaction contracts to establish an industrial large model, wherein the intelligent interaction contract includes spatial dimension interaction and time dimension interaction; deploy the industrial large model in the industrial production system, establish interaction with the production line equipment, and trigger the industrial large model with the start of the target production scenario, and use multi-agent collaborative control driven by the intelligent interaction contract to perform automated production management of the production line equipment.

[0007] The second aspect of the present application provides a multi-agent collaborative production optimization system based on an industrial large model, the system comprising:

[0008] An interaction chain determination module is used to connect to the industrial knowledge base platform, perform platform retrieval and interpretation reconstruction of the target production scenario, and determine the production line interaction chain, wherein each production node is marked with an interaction and collaboration condition; an industrial large model establishment module is used to drive the intelligent agent to build an interface for the production line interaction chain, perform call and splicing optimization of the code fragment library, perform lightweight construction of the intelligent agent, integrate multiple intelligent agents by introducing intelligent interaction contracts, and establish an industrial large model, wherein the intelligent interaction contract includes spatial dimension interaction and time dimension interaction; a collaborative control module is used to deploy the industrial large model in the industrial production system, establish interaction with the production line equipment, and trigger the industrial large model with the start of the target production scenario, and perform automated production management of the production line equipment through multi-agent collaborative control driven by the intelligent interaction contract.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] This application connects to the industrial knowledge base platform, performs platform retrieval and interpretation reconstruction on the target production scenario, and determines the production line interaction chain, wherein each production node is marked with an interactive collaboration condition; for the production line interaction chain, drives the intelligent agent to build an interface, performs call and splicing optimization of the code fragment library, performs lightweight construction of the intelligent agent, integrates multiple intelligent agents by introducing intelligent interaction contracts, and establishes an industrial large model, wherein the intelligent interaction contract includes spatial dimension interaction and time dimension interaction; deploys the industrial large model in the industrial production system, establishes interaction with the production line equipment, and triggers the industrial large model with the start of the target production scenario, and performs automated production management of the production line equipment with multi-agent collaborative control driven by the intelligent interaction contract. The present invention solves the technical problem of the lack of intelligent and multi-dimensional collaborative optimization of industrial production systems in the prior art, and achieves the technical effect of improving production efficiency, resource utilization and collaborative work of production line equipment by introducing industrial large models and multi-agent collaborative control. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A schematic diagram of the process flow of a multi-agent collaborative production optimization method based on an industrial large model provided in an embodiment of the present application;

[0013] Figure 2 Schematic diagram of the structure of a multi-agent collaborative production optimization system based on an industrial large model provided in an embodiment of the present application.

[0014] Description of the accompanying drawings: interactive chain determination module 11, industrial large model establishment module 12, collaborative control module 13. DETAILED DESCRIPTION

[0015] This application provides a multi-agent collaborative production optimization method and system based on industrial large models, aiming to solve the technical problem of the lack of intelligent and multi-dimensional collaborative optimization of industrial production systems in the existing technology. By introducing industrial large models and multi-agent collaborative control, the technical effect of improving production efficiency, resource utilization and collaborative work of production line equipment is achieved.

[0016] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0017] It should be noted that any variations of the terms "include" and "have" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0018] Example 1, as Figure 1 As shown, the present application provides a multi-agent collaborative production optimization method based on an industrial large model, the method comprising:

[0019] Step S100: Connect to the industrial knowledge base platform, perform platform retrieval and interpretation reconstruction on the target production scenario, and determine the production line interaction chain, where each production node is marked with an interactive collaboration condition.

[0020] In an embodiment of the present application, after connecting to the industrial knowledge base platform, index entries are first determined based on the target production scenario. The index entries include scenario intrinsic entries and extended entries. The scenario intrinsic entries contain the core feature information of the production scenario, while the extended entries provide additional contextual information. Then, based on these index entries, a targeted search is performed in the industrial knowledge base platform to screen out the scenario knowledge set related to the target production scenario. Among them, the industrial knowledge base platform is a database that integrates a large amount of industrial production data, technical standards, operating procedures and optimization strategies to support production optimization and intelligent decision-making.

[0021] By interpreting and reconstructing the scenario knowledge set, the extracted production information is converted into practical production factors, thereby determining the production line interaction chain. This interaction chain defines the collaborative relationship between production nodes, and each production node is identified with specific interaction and coordination conditions to ensure that each node in the production process can operate efficiently and collaboratively.

[0022] Furthermore, in the method provided in the embodiment of the application, platform retrieval and interpretation and reconstruction of the target production scenario are performed to determine the production line interaction chain, and the method further includes:

[0023] According to the target production scenario, index entries are determined, wherein the index entries include scenario intrinsic entries and extended entries; using the index entries, a targeted search is performed in the industrial knowledge base platform to determine the scenario knowledge set; and by interpreting and reconstructing the scenario knowledge set, the production line interaction chain is determined.

[0024] In an embodiment of the present application, based on the target production scenario, an index entry is first determined, and the index entry consists of a scenario intrinsic entry and an extended entry. The scenario intrinsic entry contains the core elements of the production scenario, such as equipment configuration, process flow, production tasks, etc., which represent the basic characteristics of the target production scenario; while the extended entry includes supplementary information, such as historical data, external environmental factors (such as market demand, climate conditions, etc.), to provide more background support for production decisions. After determining the index entries, a targeted search is performed in the industrial knowledge base platform based on these entries. The industrial knowledge base platform is a database that contains a large amount of industry knowledge, production processes, equipment information and optimization strategies. Through targeted retrieval, a scenario knowledge set that is highly relevant to the target production scenario is screened out.

[0025] Next, the scenario knowledge set is interpreted and reconstructed to determine the production line interaction chain. Specifically, first, for the target production scenario, production line elements are extracted and key elements related to the production process are identified. Then, based on the association characteristics between production line elements, cascade nodes are deployed, and each node is associated with at least one production line element. Next, through element attribution clustering, the scenario knowledge set is divided into different node knowledge sets. For each node knowledge set, the production driving conditions and interactive coordination conditions are determined, and then the cascade nodes are initialized to generate the production line interaction chain. The interactive coordination conditions include interaction requirements in the spatial and temporal dimensions to ensure the collaborative work of each node in the production process.

[0026] Furthermore, in the method provided in the embodiment of the application, the production line interaction chain is determined by interpreting and reconstructing the scenario knowledge set, and further includes:

[0027] For the target production scenario, production line elements are extracted; based on the production line elements, cascade nodes are deployed with element association characteristics, wherein each node contains at least one production line element; the scenario knowledge set is divided based on element affiliation clustering to determine the node knowledge set; for the node knowledge set, production driving conditions and interactive coordination conditions are determined, the cascade nodes are initialized, and the production line interaction chain is generated, wherein the interactive coordination conditions include spatial dimension interaction and temporal dimension interaction.

[0028] In the embodiments of this application, production line elements are first extracted based on the target production scenario. These elements include key elements directly related to the production process, such as equipment type, process steps, and production tasks. For example, in an automated assembly line, production line elements include the equipment configuration of each workstation, the process requirements for each workstation, and the operation schedule.

[0029] Next, cascade nodes are deployed based on the relationships between production line elements to ensure that production tasks at each node are executed in the correct order and time. For example, if a task at a node depends on the completion of a previous node's task, or requires the output of a previous process as input, the cascade nodes are deployed to associate each node with at least one production line element based on these relationships. This ensures the correct execution order of production tasks through the deployment of cascade nodes.

[0030] After deploying the cascade nodes, the scenario knowledge set is divided through factor-attribution clustering to determine the node knowledge set. Factor-attribution clustering groups related production data and knowledge based on similarities between nodes, ensuring that each node integrates all relevant production factors. For example, if a node is related to factors such as equipment operation and temperature control, the node knowledge set for that node will include information such as equipment configuration, operating conditions, and temperature control requirements.

[0031] Next, for each node knowledge set, production-driving conditions and interaction-coordination conditions are determined. Production-driving conditions include key parameters that influence node initiation and execution, such as equipment status and production task priority. Interaction-coordination conditions define the collaboration rules between nodes, such as temporal and spatial dependencies. Spatial interaction requires nodes to rationally allocate resources, equipment, and workstations in space, while temporal interaction ensures that production tasks are executed in the predetermined order and timing. For example, if a node depends on the completion of a previous node before it can begin execution, temporal interaction defines this temporal dependency.

[0032] Finally, based on production-driven conditions and interactive coordination conditions, the cascaded nodes are initialized to generate a production line interaction chain. This production line interaction chain ensures that the nodes collaborate within predetermined spatial and temporal conditions, thereby achieving efficient production process coordination. For example, after the completion of one node, subsequent nodes can be started promptly according to the scheduled schedule, avoiding delays and conflicts in the production process.

[0033] Step S200: For the production line interaction chain, drive the intelligent agent to build an interface, perform call and splicing optimization of the code fragment library, perform lightweight construction of the intelligent agent, integrate multiple intelligent agents by introducing intelligent interaction contracts, and establish an industrial large model, wherein the intelligent interaction contract includes spatial dimension interaction and time dimension interaction.

[0034] In an embodiment of the present application, for the production line interaction chain, the intelligent agent is first driven to build an interface, and then spliced ​​by calling relevant code snippets in the code fragment library. The code fragment library contains the control logic and scheduling rules commonly used in industrial production processes, and each code snippet represents a functional module in the production task. In this process, according to the requirements of the target production scenario, fragments related to the production task are selected from the code fragment library and spliced ​​into a control program that meets the requirements of the production task. The splicing of these code snippets provides each intelligent agent with the functional modules required to perform the specified task. For example, if the task involves equipment monitoring and temperature control, the code snippets related to these functions are extracted and spliced ​​to form an intelligent agent for performing temperature control and monitoring.

[0035] Next, we perform lightweight agent construction, ensuring that the constructed agents can operate efficiently in a production environment and avoid consuming excessive computing resources. This process ensures that each agent possesses only the essential functionality required to complete its assigned task. For example, when multiple agents work together, each will only perform a specific production task, avoiding duplication of functionality or waste of resources, allowing each agent to respond quickly and execute tasks efficiently.

[0036] Subsequently, multiple agents were integrated by introducing smart interaction contracts to ensure their collaborative work during the production process. Smart interaction contracts are protocols that define the interaction rules between agents, ensuring their orderly collaboration across space and time. Smart interaction contracts encompass both spatial and temporal interactions. Spatial interaction involves how agents coordinate and share spatial resources, such as workbenches and equipment locations. Through spatial interaction, multiple agents can rationally arrange and share physical space, ensuring optimal allocation of equipment resources. Temporal interaction defines the temporal coordination rules for agents in production scheduling, including the order in which tasks are executed and temporal dependencies. This allows each agent's tasks to be rationally arranged according to temporal scheduling rules, ensuring that production tasks are completed on time and without conflict.

[0037] Finally, after the above steps, an industrial large model is constructed, which enables multiple intelligent agents to efficiently perform tasks under a coordinated framework and collaboratively control equipment and resources on the production line.

[0038] Furthermore, the method provided in the application embodiment also includes:

[0039] The code fragment library stores code fragment templates in the field of industrial production, and the code fragment library is interconnected with the intelligent body construction interface; by identifying the production line interaction chain, a construction instruction is generated; according to the construction instruction, the intelligent body construction interface is driven to perform element matching based on the code fragment library, and the matching fragment template is returned; according to the matching fragment template, the intelligent body is constructed.

[0040] In an embodiment of the present application, the code fragment library contains a plurality of code fragment templates related to industrial production. These templates represent common control logic and scheduling rules in the production process, covering functional modules such as equipment control, task scheduling, and data acquisition. Each code fragment can execute a certain function independently, and after splicing multiple code fragments, complex production control tasks can be realized. Access and call relevant templates in the code fragment library through the intelligent agent construction interface. The intelligent agent construction interface is a standardized API that allows external systems or applications to access the code fragment library, thereby realizing real-time acquisition of required code snippets. The interface communicates with the library through a network protocol to ensure that each request can efficiently return relevant code snippets to meet the needs of production tasks.

[0041] The process of identifying the production line interaction chain begins by analyzing the various production links and their dependencies within the target production scenario. These dependencies include the timing requirements and resource sharing needs of each production node, forming a production line interaction chain that describes the sequence and coordination of production tasks. Based on this information, corresponding build instructions are generated, indicating the specific tasks and functions to be completed at each production node. This ensures that each task is executed in the specified order and is coordinated with other production links.

[0042] Based on the build instructions, the agent builds an interface to match elements in a code fragment library. Based on the requirements in the build instructions, such as specific production functions or equipment control requirements, the interface selects code fragments that match the task requirements. These code fragments may include functional modules such as equipment status monitoring, temperature control, and resource scheduling. During the matching process, the interface extracts the corresponding code fragments from the library based on the instruction information, ensuring that the selected template meets the specific requirements of the production task.

[0043] After matching is complete, matching fragment templates are returned. These templates contain the specific code modules required for the production task, ready for agent construction. Matching fragment templates are essentially a collection of selected and organized code fragments that are used to generate a complete agent control program. Each agent program contains the specific code required to execute the production task and is deployed to the corresponding production node.

[0044] Finally, based on the matching fragment templates, agents are constructed. Specifically, the production drive logic is first defined based on production driving conditions and interactive coordination conditions. This logic is conditionally mined with a preset degree of difference to obtain the required control logic. Next, the first production drive logic is extracted, and based on this logic, the first matching fragment template is determined. This is then logically oriented and spliced ​​to construct the first agent. Subsequently, multiple agents based on different production drive logics are constructed sequentially and integrated to form a single agent set.

[0045] Furthermore, in the method provided in the embodiment of the application, constructing an intelligent agent based on the matching fragment template further includes:

[0046] Based on the production driving conditions and the interactive coordination conditions, a production driving logic is defined, wherein the production driving logic is obtained by conditional mining with a preset difference degree; the first production driving logic is extracted, the first matching fragment template is determined, and logic-oriented adjustment optimization and splicing are performed to construct a first intelligent agent, wherein the first production driving logic is any one of the production driving logics; the construction of the Nth intelligent agent based on the Nth production driving logic is completed and integrated as an intelligent agent set.

[0047] In an embodiment of the present application, the production driving logic is first defined based on the production driving conditions and interactive collaboration conditions. The production driving conditions include key parameters that affect task execution, such as the equipment operating status, the sequence of production steps, the type of materials required, and the production speed requirements; while the interactive collaboration conditions involve the collaboration requirements between production nodes, including factors such as the execution time of the task, resource sharing, and coordination. When defining the production driving logic, these conditions jointly determine the task execution rules and timing arrangements for each production link. When defining the production driving logic, conditional mining is performed using a preset difference degree. The difference degree is a standard for measuring the difference in requirements between different production tasks, and it is adjusted according to changes in conditions in the production link. For example, some tasks may need to be completed on specific equipment, or the execution of a certain process may be closely related to the completion status of the previous link. These differences will affect the scheduling method of the task. By comparing the difference degrees between tasks, the control logic that meets the requirements of each task is automatically adjusted and mined to ensure that each task in the production process can be smoothly executed under appropriate conditions.

[0048] Next, the first production-driving logic is extracted. This logic defines the control requirements for a specific task or process, such as temperature control, equipment startup and shutdown, and task scheduling. Based on the first production-driving logic, the first matching fragment template is selected and determined from the code fragment library. These templates are predefined code segments representing specific functional modules, such as equipment status detection, production task scheduling, and temperature control. By selecting a fragment template from the code fragment library that matches the production task, the correct execution and control of the task is ensured. At this point, the matching fragment template contains all the control logic and parameters required for execution.

[0049] Matching fragment templates are then adjusted and combined through logic-directed adjustment and splicing to ensure they accurately execute the production-driven logic. For example, certain tasks may require real-time monitoring of equipment and adjustments based on its status. Through logic-directed adjustment, fragment templates are optimized to ensure that each code snippet accurately responds to task requirements during execution, avoiding redundant or unnecessary operations. These adjusted templates are then spliced ​​into a complete control program, forming the first intelligent agent. The first intelligent agent executes tasks based on the first production-driven logic, such as controlling the start and stop of equipment and adjusting the production rhythm, to ensure successful task completion.

[0050] Finally, based on the Nth production-driven logic, the Nth agent is constructed. Similar to the first agent, the Nth agent selects appropriate code snippets, splices them together, and adjusts them based on its corresponding production-driven logic and task requirements. Each agent's control logic is independent and may have different execution processes depending on the task requirements. For example, some tasks may involve the coordinated operation of multiple devices, while others may only involve the control of a single device. By adjusting the production-driven logic, each agent is ensured to operate efficiently in the target production scenario.

[0051] Ultimately, all agents are constructed through the above process and integrated into an agent ensemble, which coordinates and works together to complete complex production tasks.

[0052] Furthermore, the method provided in the application embodiment also includes:

[0053] According to the production driving conditions, a node mapping between the set of intelligent agents and the production line interaction chain is established, wherein one intelligent agent is associated with at least one node.

[0054] In this embodiment, when mapping the nodes of the agent collection to the production line interaction chain, we first analyze the nodes in the production line interaction chain. These nodes represent various aspects of the production process, such as equipment operation, production task scheduling, or quality inspection. Each node defines the specific requirements of the production task, the execution conditions, and the coordination relationship with other nodes.

[0055] Next, the responsibilities and tasks of each agent are determined based on production-driven conditions. These production-driven conditions include task execution order, equipment operating status, material flow, production priority, and other factors, which determine the functions and behaviors of each agent. Based on these conditions, each agent is responsible for one or more tasks related to a node in the production line interaction chain.

[0056] Once the task requirements of each agent are determined, they can be mapped to the corresponding nodes in the production line interaction chain. Each agent performs the production tasks of its associated node, ensuring that task dependencies and collaborations between nodes are met. For example, a production node might require equipment monitoring and scheduling. In this case, these tasks can be mapped to a single agent for management.

[0057] Through this mapping method, each agent in the agent set is mapped to one or more nodes in the production line interaction chain, thereby achieving efficient coordination and control of the production process.

[0058] Furthermore, the method provided in the embodiment of the application introduces a smart interactive contract, and further includes:

[0059] Based on the production-driven conditions, a first interaction contract is determined, wherein the first interaction contract is intelligent body collaboration in the spatial dimension; based on the interaction collaboration conditions, a second interaction contract is determined, wherein the second interaction contract is intelligent body collaboration linked to the time dimension; and the first interaction contract and the second interaction contract are integrated to determine the intelligent interaction contract.

[0060] In an embodiment of the present application, based on the production driving conditions, a first interaction contract is first determined, which defines the collaboration of intelligent agents in the spatial dimension. Production driving conditions refer to key factors that affect production tasks, such as equipment status, the order of production links, resource requirements, process requirements, etc. These conditions determine how each intelligent agent collaborates and allocates resources in the production process. In the collaboration of intelligent agents in the spatial dimension, the purpose of the first interaction contract is to ensure that multiple intelligent agents can spatially share and coordinate the use of physical resources, such as workbenches, equipment, tools, etc. For example, in an automated production line, multiple intelligent agents may need to use the same equipment in the same time period. The first interaction contract ensures the rational use of equipment by stipulating the timing of equipment use, the order of task execution, and the priority, and avoids production delays or idleness caused by resource conflicts or unreasonable allocation. Specifically, the first interaction contract ensures efficient collaboration of each node in the production process and optimizes resource allocation by controlling how each intelligent agent uses these resources in space.

[0061] Next, based on the interactive coordination conditions, a second interactive contract is determined, which involves agent coordination in the time dimension. The interactive coordination conditions describe the timing and dependencies between production tasks and define the coordination rules between different production nodes. Within the time dimension of agent coordination, the second interactive contract ensures that each production node executes tasks in the correct chronological order. For example, within a production process, a task in a certain production link must wait until the previous link is completed before it can begin. By defining this second interactive contract, each task in the production schedule is guaranteed to start and end at the appropriate time, thus avoiding production delays caused by time conflicts.

[0062] Finally, the first interactive contract is merged with the second interactive contract to form the final intelligent interactive contract. This step combines the coordination requirements of the spatial and temporal dimensions, ensuring that multiple agents can collaborate efficiently in resource utilization and execute on demand in time scheduling. By coordinating these two dimensions, the intelligent interactive contract ensures that each node in the production process can execute efficiently within the specified time and optimizes resource sharing and collaboration between various production nodes. For example, multiple agents may need to share certain equipment and workstations within the same time period. The intelligent interactive contract not only coordinates the order in which these devices are used, but also ensures that tasks are executed in sequence according to time requirements, thereby avoiding production bottlenecks.

[0063] Step S300: Deploy the industrial big model in the industrial production system, establish interaction with the production line equipment, and trigger the industrial big model with the start of the target production scenario to perform automated production management of the production line equipment through multi-agent collaborative control driven by intelligent interactive contracts.

[0064] In this embodiment, the industrial master model is deployed within the industrial production system, establishing an interactive relationship with the production line equipment. When the target production scenario begins, the industrial master model runs synchronously, monitoring and managing the production process in real time. Intelligent interactive contracts serve as the driving force, ensuring that multiple agents can coordinate and cooperate when executing tasks through multi-agent collaborative control, optimizing production scheduling and resource allocation.

[0065] In this process, smart interaction contracts drive different agents to execute tasks within predetermined time and space conditions, based on the needs of each node in the production line's interactive chain. Each agent is responsible for a specific production link. Through collaborative control, efficient execution of production tasks is achieved, ensuring the rational utilization and efficient operation of production line equipment. Smart interaction contracts define the task collaboration rules between each agent, ensuring optimized resource sharing, equipment utilization, and task sequencing.

[0066] Ultimately, the industrial large model and the intelligent interactive contract work together to realize the automated production management of production line equipment.

[0067] Furthermore, in the method provided in the embodiment of the application, the automated production management of production line equipment is performed by multi-agent collaborative control driven by intelligent interactive contracts, and further includes:

[0068] As the industrial production of the target production scenario begins, the industrial large model built into the industrial production system runs synchronously; multi-agent collaborative drive management is carried out through the intelligent interactive contract, and the automated production management of the target production scenario is executed in response to the production line equipment of each link node in the production line interaction chain.

[0069] In this embodiment of the present application, upon the initiation of the target production scenario, the industrial master model built into the industrial production system begins operating simultaneously. As an intelligent management platform integrating production data, equipment status, process flow, and production scheduling, the industrial master model receives and analyzes various information from the production site in real time, providing real-time support for production decision-making. With the start of the target production scenario, the industrial master model automatically schedules production tasks and resources based on the set production goals and task requirements.

[0070] Through intelligent interaction contracts, multi-agent collaborative control begins to drive task execution within the production process. Intelligent interaction contracts are protocols that coordinate and manage collaboration between different agents. They define how tasks are scheduled, resources are allocated, and time is coordinated across different production links. Based on the task requirements of each node in the production line's interactive chain, intelligent interaction contracts ensure that tasks at each node are executed on time and that equipment and resources are effectively utilized, thereby achieving optimal production scheduling and task execution.

[0071] In response to the needs of each node in the production line interaction chain, the intelligent interaction contract guides each agent to perform operations related to its task, ensuring coordination and connection between production links, and ultimately realizing automated production management in the target production scenario.

[0072] Furthermore, in the method provided in the embodiment of the application, after performing automated production management of the production line equipment, the method further includes:

[0073] According to the production line interaction chain, a directional monitoring device is deployed on the production line side, wherein the directional monitoring device is set with a critical threshold; with the critical threshold as a constraint, the data threshold judgment under the monitoring of the directional monitoring device is performed, the out-of-limit monitoring data is located and the intelligent body traceability is performed, and directional feedback control is performed.

[0074] In the embodiments of this application, directional monitoring devices are deployed along the production line, based on the production line interaction chain. These devices monitor key equipment or process parameters during the production process, such as temperature, pressure, and vibration. Each directional monitoring device has a set critical threshold, which serves as the threshold for the device's monitoring data and determines when the data is considered to be outside the normal range. The critical threshold is set based on the device's normal operating range and process requirements, ensuring that abnormalities can be detected promptly during the production process.

[0075] Once the directional monitoring equipment begins operation, it performs a data threshold determination on the real-time data collected based on the set critical thresholds. This process compares the real-time monitoring data with the preset critical thresholds. If a parameter exceeds the threshold, the data is marked as out-of-limit monitoring data and an alarm mechanism is triggered. For example, if the device temperature exceeds the set safety range, the monitoring equipment will immediately identify the anomaly and record it as out-of-limit monitoring data.

[0076] Next, intelligent traceability begins tracing the out-of-limit monitoring data to determine the source of the anomaly. By analyzing the correlation between the out-of-limit monitoring data and relevant production processes, intelligent traceability identifies the cause of the anomaly. This process involves tracing the data's source nodes and operational history, using intelligent traceability technology to quickly locate potential faults. For example, if the equipment temperature is too high, intelligent traceability may reveal that it is caused by a cooling system failure or incorrect process parameter settings.

[0077] Finally, based on the agent's traceability analysis results, targeted feedback control is implemented. Based on the traceability analysis results, targeted feedback control automatically adjusts relevant equipment or process parameters to restore the production system to normal operation. For example, if the temperature anomaly is caused by a cooling system failure, targeted feedback control will adjust the cooling system's operating parameters or activate backup cooling equipment to ensure the equipment returns to a safe operating state.

[0078] In the embodiments of the present application, in summary, the embodiments of the present application have at least the following technical effects:

[0079] This application connects to the industrial knowledge base platform, performs platform retrieval and interpretation reconstruction on the target production scenario, and determines the production line interaction chain, wherein each production node is marked with an interactive collaboration condition; for the production line interaction chain, drives the intelligent agent to build an interface, performs call and splicing optimization of the code fragment library, performs lightweight construction of the intelligent agent, integrates multiple intelligent agents by introducing intelligent interaction contracts, and establishes an industrial large model, wherein the intelligent interaction contract includes spatial dimension interaction and time dimension interaction; deploys the industrial large model in the industrial production system, establishes interaction with the production line equipment, and triggers the industrial large model with the start of the target production scenario, and performs automated production management of the production line equipment with multi-agent collaborative control driven by the intelligent interaction contract. The present invention solves the technical problem of the lack of intelligent and multi-dimensional collaborative optimization of industrial production systems in the prior art, and achieves the technical effect of improving production efficiency, resource utilization and collaborative work of production line equipment by introducing industrial large models and multi-agent collaborative control.

[0080] The second embodiment is based on the same inventive concept as the multi-agent collaborative production optimization method based on the industrial large model in the above embodiment. Figure 2 As shown, the present application provides a multi-agent collaborative production optimization system based on an industrial large model. The system and method embodiments in the present application are based on the same inventive concept. The system includes:

[0081] The interaction chain determination module 11 is used to connect to the industrial knowledge base platform, perform platform retrieval and interpretation reconstruction of the target production scenario, and determine the production line interaction chain, wherein each production node is marked with an interaction and coordination condition; the industrial large model establishment module 12 is used to drive the intelligent agent to build an interface for the production line interaction chain, perform call and splicing optimization of the code fragment library, perform lightweight construction of the intelligent agent, integrate multiple intelligent agents by introducing intelligent interaction contracts, and establish an industrial large model, wherein the intelligent interaction contract includes spatial dimension interaction and time dimension interaction; the collaborative control module 13 is used to deploy the industrial large model in the industrial production system, establish interaction with the production line equipment, and trigger the industrial large model with the start of the target production scenario, and perform automated production management of the production line equipment through multi-agent collaborative control driven by the intelligent interaction contract.

[0082] Furthermore, the system is also used to implement the following functions:

[0083] According to the target production scenario, index entries are determined, wherein the index entries include scenario intrinsic entries and extended entries; using the index entries, a targeted search is performed in the industrial knowledge base platform to determine the scenario knowledge set; and by interpreting and reconstructing the scenario knowledge set, the production line interaction chain is determined.

[0084] Furthermore, the system is also used to implement the following functions:

[0085] For the target production scenario, production line elements are extracted; based on the production line elements, cascade nodes are deployed with element association characteristics, wherein each node contains at least one production line element; the scenario knowledge set is divided based on element affiliation clustering to determine the node knowledge set; for the node knowledge set, production driving conditions and interactive coordination conditions are determined, the cascade nodes are initialized, and the production line interaction chain is generated, wherein the interactive coordination conditions include spatial dimension interaction and temporal dimension interaction.

[0086] Furthermore, the system is also used to implement the following functions:

[0087] By identifying the production line interaction chain, a construction instruction is generated; according to the construction instruction, the intelligent agent construction interface is driven to perform element matching based on the code fragment library, and a matching fragment template is returned; according to the matching fragment template, the intelligent agent is constructed.

[0088] Furthermore, the system is also used to implement the following functions:

[0089] Based on the production driving conditions and the interactive coordination conditions, a production driving logic is defined, wherein the production driving logic is obtained by conditional mining with a preset difference degree; the first production driving logic is extracted, the first matching fragment template is determined, and logic-oriented adjustment optimization and splicing are performed to construct a first intelligent agent, wherein the first production driving logic is any one of the production driving logics; the construction of the Nth intelligent agent based on the Nth production driving logic is completed and integrated as an intelligent agent set.

[0090] Furthermore, the system is also used to implement the following functions:

[0091] According to the production driving conditions, a node mapping between the set of intelligent agents and the production line interaction chain is established, wherein one intelligent agent is associated with at least one node.

[0092] Furthermore, the system is also used to implement the following functions:

[0093] Based on the production-driven conditions, a first interaction contract is determined, wherein the first interaction contract is intelligent body collaboration in the spatial dimension; based on the interaction collaboration conditions, a second interaction contract is determined, wherein the second interaction contract is intelligent body collaboration linked to the time dimension; and the first interaction contract and the second interaction contract are integrated to determine the intelligent interaction contract.

[0094] Furthermore, the system is also used to implement the following functions:

[0095] As the industrial production of the target production scenario begins, the industrial large model built into the industrial production system runs synchronously; multi-agent collaborative drive management is carried out through the intelligent interactive contract, and the automated production management of the target production scenario is executed in response to the production line equipment of each link node in the production line interaction chain.

[0096] Furthermore, the system is also used to implement the following functions:

[0097] According to the production line interaction chain, a directional monitoring device is deployed on the production line side, wherein the directional monitoring device is set with a critical threshold; with the critical threshold as a constraint, the data threshold judgment under the monitoring of the directional monitoring device is performed, the out-of-limit monitoring data is located and the intelligent body traceability is performed, and directional feedback control is performed.

[0098] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0099] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

[0100] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A multi-agent collaborative production optimization method based on an industrial large model, characterized by: The method comprises: Connect to the industrial knowledge base platform to perform platform retrieval and interpretation reconstruction of the target production scenario, and determine the production line interaction chain, where each production node is marked with interactive coordination conditions; For the production line interaction chain, drive the intelligent agent to build an interface, perform call and splicing optimization of the code fragment library, perform lightweight construction of the intelligent agent, integrate multiple intelligent agents by introducing intelligent interaction contracts, and establish an industrial large model, wherein the intelligent interaction contract includes spatial dimension interaction and temporal dimension interaction; The industrial big model is deployed in the industrial production system, and interaction is established with the production line equipment. With the start of the target production scenario, the industrial big model is triggered, and the automated production management of the production line equipment is carried out through multi-agent collaborative control driven by intelligent interactive contracts.

2. The multi-agent collaborative production optimization method based on industrial large model according to claim 1 is characterized in that: Perform platform retrieval and interpretation reconstruction of the target production scenario to determine the production line interaction chain, including: Determine index entries according to the target production scenario, wherein the index entries include scenario intrinsic entries and extended entries; Using the index entries, a targeted search is performed in the industrial knowledge base platform to determine a scenario knowledge set; The production line interaction chain is determined by interpreting and reconstructing the scenario knowledge set.

3. The multi-agent collaborative production optimization method based on industrial large model according to claim 2 is characterized in that: By interpreting and reconstructing the scenario knowledge set, the production line interaction chain is determined, including: Extracting production line elements for the target production scenario; Deploy cascade nodes based on the production line elements and element association characteristics, wherein each node contains at least one production line element; The scene knowledge set is divided according to element attribution clustering to determine the node knowledge set; For the node knowledge set, production driving conditions and interactive coordination conditions are determined, the cascade nodes are initialized, and the production line interactive chain is generated, wherein the interactive coordination conditions include spatial dimension interaction and temporal dimension interaction.

4. The multi-agent collaborative production optimization method based on industrial large model according to claim 3 is characterized in that: The code fragment library stores code fragment templates in the field of industrial production, and the code fragment library is interconnected with the intelligent agent construction interface; Generate build instructions by identifying the production line interaction chain; According to the construction instruction, the agent construction interface is driven to perform element matching based on the code fragment library and return a matching fragment template; An intelligent agent is constructed according to the matching fragment template.

5. The multi-agent collaborative production optimization method based on industrial large model according to claim 4 is characterized in that: According to the matching fragment template, constructing an intelligent agent includes: Defining a production driving logic according to the production driving condition and the interactive coordination condition, wherein the production driving logic is obtained by conditional mining with a preset difference degree; Extracting a first production driving logic, determining a first matching segment template, and performing logic-oriented adjustment optimization and splicing to construct a first intelligent agent, wherein the first production driving logic is any one of the production driving logics; Complete the construction of the Nth intelligent agent based on the Nth production driving logic and integrate it as an intelligent agent collection.

6. The multi-agent collaborative production optimization method based on industrial large model according to claim 5 is characterized in that: According to the production driving conditions, a node mapping between the set of intelligent agents and the production line interaction chain is established, wherein one intelligent agent is associated with at least one node.

7. The multi-agent collaborative production optimization method based on industrial large model according to claim 6 is characterized in that: Introducing smart interactive contracts, including: Determining a first interaction contract based on the production driving condition, wherein the first interaction contract is intelligent agent collaboration in a spatial dimension; Determining a second interaction contract based on the interaction collaboration condition, wherein the second interaction contract is an agent collaboration linked to the time dimension; The first interaction contract and the second interaction contract are integrated to determine the smart interaction contract.

8. The multi-agent collaborative production optimization method based on industrial large model according to claim 1 is characterized in that: Automated production management of production line equipment is achieved through multi-agent collaborative control driven by intelligent interactive contracts, including: As the industrial production of the target production scenario begins, the industrial large model built into the industrial production system runs synchronously; Multi-agent collaborative drive management is carried out through the intelligent interactive contract, and the automated production management of the target production scenario is executed in response to the production line equipment of each link node in the production line interactive chain.

9. The multi-agent collaborative production optimization method based on industrial large model according to claim 1 is characterized in that: After the automated production management of production line equipment is carried out, it includes: Deploy a directional monitoring device on the production line side according to the production line interaction chain, wherein the directional monitoring device is set with a critical threshold; With the critical threshold as a constraint, the data threshold judgment under the monitoring of the directional monitoring equipment is performed, the out-of-limit monitoring data is located and the intelligent body is traced, and directional feedback control is performed.

10. A multi-agent collaborative production optimization system based on an industrial large model, characterized by: The system is used to execute the multi-agent collaborative production optimization method based on an industrial large model as described in any one of claims 1 to 9, and the system includes: The interaction chain determination module is used to connect to the industrial knowledge base platform, perform platform retrieval and interpretation and reconstruction of the target production scenario, and determine the production line interaction chain, in which each production node is marked with interaction and coordination conditions; The industrial large-scale model building module is used to drive the intelligent agent to build an interface for the production line interaction chain, call and splice the code fragment library, optimize the lightweight construction of the intelligent agent, integrate multiple intelligent agents by introducing intelligent interaction contracts, and build an industrial large-scale model, wherein the intelligent interaction contract includes spatial dimension interaction and temporal dimension interaction; The collaborative control module is used to deploy the industrial big model in the industrial production system, establish interaction with the production line equipment, trigger the industrial big model with the start of the target production scenario, and perform automated production management of the production line equipment through multi-agent collaborative control driven by intelligent interactive contracts.

Images