An AI agent-based production management information processing method, device and medium

Abstract

The application discloses a production management information processing method and device based on an AI agent, and a medium, relates to the technical field of production management information processing, and comprises the following steps: according to a standardized production event stream and a data quality meta stream, performing hierarchical management and state aggregation on the standardized production event stream, and generating a credible production state atlas; performing local query on order, process, equipment, material and personnel associated data corresponding to an influence subgraph in the credible production state atlas according to the influence subgraph by an inquiry planning AI agent, generating a local query result set, and performing incremental recalculation and synchronous updating by a recalculation AI agent to generate a local recalculation result package; and performing execution condition judgment on the local recalculation result package by a decision arrangement AI agent, generating a processing scheme, and performing arrangement to generate an execution instruction package. The application enhances the focus of state change identification and the order of result organization.

Description

Technical Field

[0001] This invention relates to the field of production management information processing technology, and in particular to a production management information processing method, equipment and medium based on AI intelligent agents. Background Technology

[0002] With the continuous development of intelligent manufacturing and industrial digitalization, production management is gradually shifting from processing by a single business system to collaborative processing of multi-source business data such as orders, processes, equipment, materials and personnel. Related technologies have expanded from traditional planning management and process recording to event flow analysis, master data governance, state graph construction, and task identification and scheduling based on artificial intelligence.

[0003] When faced with continuous changes in multi-source heterogeneous data, existing production management information processing technologies typically focus on global correlation analysis and overall result updates. They struggle to quickly define the scope of influence around state change nodes and form a data processing link that can be directly used for local queries and incremental recalculation. As a result, it is difficult to coordinate and guarantee the correlation granularity, response efficiency, and state consistency in real-time processing. The core issue is the lack of mechanisms for identifying and recalculating the scope of local influence in dynamic change scenarios. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a production management information processing method based on AI intelligent agents to solve the problem of difficulty in accurately identifying and recalculating the local impact range in dynamic scenarios of multi-source heterogeneous production data.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] Firstly, this invention provides a production management information processing method based on an AI intelligent agent, comprising: collecting real-time business data of production management, and performing time alignment, master data mapping, duplicate record elimination, and quality status calculation to generate a standardized production event flow and a data quality metaflow; performing hierarchical governance and state aggregation on the standardized production event flow based on the standardized production event flow and the data quality metaflow to generate a reliable production state map; and having a task parsing AI intelligent agent identify task types and state change ranges based on the reliable production state map, and track upstream and downstream influences along the business relationships between orders, processes, equipment, materials, and personnel. The system generates an impact subgraph based on the impact path. The query planning AI agent performs local queries on the order, process, equipment, material, and personnel data corresponding to the impact subgraph in the trusted production status graph, generating a local query result set. The recalculation AI agent then performs incremental recalculation and synchronous updates, generating a local recalculation result package. The decision orchestration AI agent determines the execution conditions of the local recalculation result package, generates a processing plan, and orchestrates it, generating an execution instruction package. Finally, the execution feedback AI agent performs business dissemination, feedback consistency verification, and update event write-back processing on the execution instruction package, generating a feedback correction result package.

[0008] As a preferred embodiment of the AI-based intelligent agent-based production management information processing method of the present invention, the specific steps for generating standardized production event streams and data quality metadata streams are as follows:

[0009] Collect real-time production management business data from the order management terminal, production scheduling terminal, warehouse management terminal, equipment acquisition terminal, quality record terminal and manual entry terminal, and add source identifier, business object identifier and collection time identifier to the real-time production management business data to generate the original production management dataset;

[0010] The original production management dataset is processed for time alignment, master data matching, and object mapping to generate a mapped production management dataset.

[0011] The mapping production management dataset is subjected to duplicate record elimination, data quality inspection, field standardization, event structure transformation, and quality status association to generate a standardized production event stream and data quality meta-stream.

[0012] As a preferred embodiment of the AI-based intelligent agent-based production management information processing method of the present invention, the specific steps for generating a reliable production status map are as follows:

[0013] Based on the standardized production event flow and data quality metaflow, establish the correlation between events and quality status, and classify them according to preset quality rules to generate a hierarchical production event set;

[0014] Implement hierarchical governance for the hierarchical production event set, and extract the status information and relationships of order events, process events, equipment events, material events, and personnel events to generate a status-related production event set;

[0015] Based on the state-related production event set, merge the state changes of the same order, the same process, the same equipment, the same material, and the same person at different times, and write the order, process, equipment, material, and person into the graph nodes, and write the business dependencies and state transit relationships into the graph edges to generate a reliable production state graph.

[0016] As a preferred embodiment of the AI-based intelligent agent-based production management information processing method of the present invention, the specific steps for generating the influence subgraph are as follows:

[0017] Extract task trigger information and determine task type from the trusted production status map, identify the status change information corresponding to new tasks, abnormal tasks, query tasks and adjustment tasks, and generate a task type determination result set.

[0018] The task type determination result set is used to identify the range of state changes and track upstream and downstream relationships. The order nodes, process nodes, equipment nodes, material nodes and personnel nodes that have changed in the trusted production state map are located, and the upstream and downstream impact paths are tracked along business dependencies and state transmission relationships to generate an impact path set.

[0019] Based on the set of affected paths, redundant paths are pruned, the scope of influence is limited, and the corresponding graph nodes and edges are extracted to generate an affected subgraph.

[0020] As a preferred embodiment of the AI-based intelligent agent-based production management information processing method of the present invention, the specific steps for generating a local query result set are as follows:

[0021] Extract the order, process, equipment, material, and personnel nodes to be queried, as well as their corresponding business relationships, from the influence subgraph to generate a local query scope set;

[0022] Map and match the local query range set with the corresponding order, process, equipment, material and personnel association data in the trusted production status map to locate the data range corresponding to the local query range set and generate a local query dataset;

[0023] Perform a partial query on the dataset to be queried, extract the associated status data of the current task according to the dimensions of order, process, equipment, material and personnel, and generate a partial query result set.

[0024] As a preferred embodiment of the AI-based intelligent agent-based production management information processing method of the present invention, the specific steps for generating the local recalculation result package are as follows:

[0025] The current state in the local query result set is compared with the historical state, and the key state items that have changed are identified to generate a set of difference state results.

[0026] Based on the difference status result set, recalculate the affected order progress, process connection, equipment load, material availability and personnel allocation relationship, and generate an incremental recalculation result set;

[0027] Write the incremental recalculation result set into the data structure corresponding to the local query result set to generate the local recalculation result package.

[0028] In a preferred embodiment of the AI-based intelligent agent-based production management information processing method of the present invention, the specific steps for generating the execution instruction package are as follows:

[0029] The system parses the results of the partial recalculation and determines the execution conditions, extracts order change information, process change information, equipment change information, material change information and personnel change information, and identifies the execution conditions corresponding to the current task according to the preset execution rules, generating an execution condition determination result set.

[0030] Based on the result set of execution conditions, corresponding adjustment paths are generated for the associated objects of orders, processes, equipment, materials and personnel, and resource occupation conflicts, business timing conflicts and status transmission conflicts are verified to generate a set of post-verification processing solutions.

[0031] Based on the urgency of the task, the scope of the status change, and the intensity of the business impact, the set of post-verification processing solutions is prioritized and arranged. The arranged processing content is then converted into executable business operation instructions to generate an execution instruction package.

[0032] In a preferred embodiment of the AI-based intelligent agent-based production management information processing method of the present invention, the specific steps for generating the feedback correction result package are as follows:

[0033] The execution instruction package is split and executed, and the business execution feedback set is generated.

[0034] The business execution feedback set is aggregated and its consistency is verified to generate a consistency verification result set.

[0035] Update events are extracted from the consistency check result set, and the update events and consistency check results are encapsulated to generate a feedback correction result package.

[0036] In a second aspect, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the production management information processing method based on AI intelligent agents as described in the first aspect of the present invention.

[0037] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the production management information processing method based on AI intelligent agents as described in the first aspect of the present invention.

[0038] The beneficial effects of this invention are as follows: by performing local queries on the corresponding related data in the reliable production state map based on the influence subgraph and generating a local recalculation result package, a data processing scope with clear boundaries can be constructed around the nodes and related links involved in the state change. This allows the extraction of related data, identification of differences, incremental recalculation, and aggregation of results to be continuously connected along a unified processing path. It can integrate the scattered state relationships between orders, processes, equipment, materials, and personnel into a structured data foundation that can directly participate in the recalculation, thereby enhancing the focus of state change identification and the orderliness of result organization. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a flowchart of a production management information processing method based on AI intelligent agents.

[0041] Figure 2 A flowchart for generating a trusted production state map.

[0042] Figure 3 The flowchart that influences the generation of subgraphs.

[0043] Figure 4 A flowchart for generating feedback correction result packages. Detailed Implementation

[0044] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0045] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0046] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0047] Reference Figures 1-4 This is one embodiment of the present invention, which provides a production management information processing method based on an AI intelligent agent, including the following steps:

[0048] S1. Collect real-time business data for production management, and perform time alignment, master data mapping, duplicate record elimination, and quality status calculation to generate standardized production event streams and data quality meta-streams.

[0049] S1.1 Collect real-time production management business data from the order management terminal, production scheduling terminal, warehouse management terminal, equipment acquisition terminal, quality record terminal and manual entry terminal, and add source identifier, business object identifier and collection time identifier to the real-time production management business data to generate the original production management dataset.

[0050] Specifically, real-time production management data is continuously received from the order management, production scheduling, warehouse management, equipment acquisition, quality recording, and manual data entry terminals. Following unified access rules, each piece of order, process, equipment, material, personnel, and time information within the real-time production management data is parsed. A corresponding source identifier, business object identifier, and acquisition time identifier are written for each piece of real-time production management data. Based on the source identifier, real-time production management data from different sources are categorized; based on the business object identifier, related real-time production management data for the same business object are matched; and based on the acquisition time identifier, the formation sequence of the real-time production management data is recorded. The real-time production management data for which source identifiers, business object identifiers, and acquisition time identifiers have been supplemented are collected and organized to generate the original production management dataset.

[0051] It should also be noted that unified access rules are a set of rules used to standardize the processing of real-time production management business data from different sources according to a unified data format, field standards, identification methods, and receiving order.

[0052] S1.2 Perform time alignment, master data matching, and object mapping on the original production management dataset to generate a mapped production management dataset.

[0053] Specifically, the records in the original production management dataset are sorted and organized according to their collection time identifiers. Time information from different sources is corrected based on a unified time benchmark to obtain a processing result with a unified temporal relationship after time alignment. Order information, process information, equipment information, material information, and personnel information are extracted from the time-aligned processing result and compared item by item with the corresponding information in the preset master data set to identify the correspondence of the same business object in different records. After master data matching, the master data correspondence result is obtained. Based on the master data correspondence result, object mapping processing is performed on different records in the original production management dataset related to the same business object. Business objects from different sources, with different names, or with different codes are uniformly mapped to the same business object identifier, while maintaining the continuity and consistency of the association relationship between each record. The results after time alignment, master data matching, and object mapping processing are integrated to generate a mapped production management dataset.

[0054] It should also be noted that the preset master data set is pre-collected and organized from the basic information of business objects in the order management terminal, production scheduling terminal, warehouse management terminal, equipment acquisition terminal, quality record terminal and manual entry terminal.

[0055] S1.3 Perform duplicate record elimination, data quality inspection, field standardization, event structure transformation, and quality status association on the mapped production management dataset to generate a standardized production event stream and data quality meta-stream.

[0056] Specifically, each record in the mapped production management dataset is compared according to the business object identifier, collection time identifier, and record content. Records with duplicate content or pointing to the same business object and expressing the same information are identified, and valid records are retained to complete the duplicate record elimination. After the duplicate record elimination is completed, the records are checked in turn for field integrity, value consistency, time validity, and status anomalies to obtain the corresponding data quality inspection results. According to the unified field standard, the order information, process information, equipment information, material information, personnel information, and status information in the records are uniformly processed in terms of field names, field order, and field value format to form field standardization results. Based on the field standardization results, each record is converted into a unified event structure containing event type, business object identifier, status content, and time information to form event structure conversion results. The quality status in the data quality inspection results is written into the event record corresponding to the event structure conversion results, so that each event record is associated with the corresponding quality status, generating a standardized production event flow and data quality metaflow.

[0057] S2. Based on the standardized production event flow and data quality metaflow, perform hierarchical governance and state aggregation on the standardized production event flow to generate a trusted production state graph.

[0058] S2.1. Based on the standardized production event flow and data quality metaflow, establish the correlation between events and quality status, classify them according to preset quality rules, and generate a hierarchical production event set.

[0059] Specifically, each event record in the standardized production event stream is matched item by item with the corresponding quality status in the data quality metastream according to the business object identifier, event type, and time information. This ensures that each event record in the standardized production event stream obtains corresponding quality status information, forming a correlation between events and quality statuses. Based on preset quality rules, the associated event records are classified according to the quality levels corresponding to completeness, consistency, timeliness, and abnormal states. Event records with different quality conditions are divided into high-confidence, medium-confidence, and low-confidence levels. Event records that meet the completeness, consistency, timeliness, and abnormal state judgment results are classified into the high-confidence level. Event records with partial missing information, deviations, or timeliness fluctuations but still usable after correction are classified into the medium-confidence level. Event records with abnormal conflicts, missing key fields, or that do not meet processing requirements are classified into the low-confidence level. The event records in the same quality level are then collected and organized according to the classification results, maintaining the continuity and consistency of the correlation between each event record and its corresponding quality status, generating a hierarchical production event set.

[0060] It should also be noted that the preset quality rules are the availability requirements for subsequent state aggregation, impact analysis, and result processing of real-time business data in production management in terms of completeness, consistency, timeliness, and abnormal states.

[0061] S2.2 Implement hierarchical governance for the hierarchical production event set, and extract the status information and relationships of order events, process events, equipment events, material events and personnel events to generate a status-related production event set.

[0062] Specifically, the event records corresponding to different quality levels in the hierarchical production event set are processed separately. Event records classified to the high-confidence level are directly written into the status extraction process. Event records with missing fields or expression deviations are supplemented, corrected, and standardized according to their corresponding quality status. Event records with abnormal conflicts are isolated, marked, and their relationships are removed, resulting in a rectified hierarchical production event set. From the rectified hierarchical production event set, the status content, time information, and business object identifiers corresponding to order events, process events, equipment events, material events, and personnel events are extracted, forming the status results for order events, process events, equipment events, material events, and personnel events, respectively. Based on the business object identifiers, time information, and business relationships between events, the status results for order events, process events, equipment events, material events, and personnel events are correlated and organized to form a set of events with continuously corresponding status information and relationships, generating a status-associated production event set.

[0063] S2.3. Based on the state-related production event set, merge the state changes of the same order, the same process, the same equipment, the same material, and the same person at different times, and write the order, process, equipment, material, and person into the graph nodes, and write the business dependencies and state transit relationships into the graph edges to generate a reliable production state graph.

[0064] Specifically, based on the order events, process events, equipment events, material events, and personnel events recorded in the state-related production event set, the state content corresponding to the same order, process, equipment, material, and personnel at different times is sequentially organized according to business object identifiers and time information. Continuous state changes are merged according to chronological order, forming order state change results, process state change results, equipment state change results, material state change results, and personnel state change results. These results are then written into graph nodes. Based on the established business relationships in the state-related production event set, the corresponding business dependencies and state propagation relationships between orders, processes, equipment, materials, and personnel are organized and written into graph edges. The correspondence between graph nodes and edges is uniformly organized to ensure that the state changes of the same order, process, equipment, material, and personnel are continuously associated with business dependencies and state propagation relationships, generating a reliable production state graph.

[0065] It should be noted that coupling standardized production event flows with data quality metaflows does not simply involve routine aggregation or static modeling of production events. Instead, events are first stratified and governed based on quality information such as completeness, consistency, timeliness, and abnormal states. Then, the governed order events, process events, equipment events, material events, and personnel events are aggregated to construct a reliable production state graph with quality constraint semantics. This allows for the initial screening of event reliability, followed by the organization of state relationships based on reliable events. This ensures that the state connections between orders, processes, equipment, materials, and personnel have clear quality boundaries and semantic consistency, thereby providing a structurally stable and reliable data foundation for subsequent impact path tracing, local queries, and incremental recalculation.

[0066] S3. The task analysis AI agent identifies the task type and state change range based on the trusted production state map, and traces the upstream and downstream impact paths along the business relationships between orders, processes, equipment, materials and personnel to generate an impact sub-graph.

[0067] S3.1 Extract task trigger information and determine task type from the trusted production status map, identify the status change information corresponding to new tasks, abnormal tasks, query tasks and adjustment tasks, and generate a task type determination result set.

[0068] Specifically, the system reads the graph node status of orders, processes, equipment, materials, and personnel from the trusted production state graph, as well as the graph edge changes corresponding to business dependencies and state propagation relationships. It then extracts task trigger information that triggers the current processing flow according to the chronological order of state changes and the business object identifier, forming state change information for new tasks, abnormal tasks, query tasks, and adjustment tasks. Based on the change objects, change directions, change frequencies, and relationships corresponding to these state change information, the task trigger information is categorized and classified into the corresponding task types: new tasks, abnormal tasks, query tasks, and adjustment tasks. Finally, the task types are matched one-to-one with the corresponding state change information, ensuring a continuous association between each task type and its corresponding state change information, generating a task type determination result set.

[0069] S3.2. Identify the range of state changes and track upstream and downstream relationships in the task type determination result set, locate the order nodes, process nodes, equipment nodes, material nodes and personnel nodes that have changed in the trusted production state map, and track the upstream and downstream impact paths along business dependencies and state transmission relationships to generate an impact path set.

[0070] Specifically, based on the status change information corresponding to each task type in the task type determination result set, the order nodes, process nodes, equipment nodes, material nodes, and personnel nodes corresponding to the status change information are first identified in the trusted production state graph. The scope of status change is then defined based on the business object identifier, time information, and direction of change involved in the status change information, resulting in a set of nodes that have changed. Starting from the set of nodes that have changed, the propagation paths are traced upstream and downstream layer by layer along the business dependencies and status transmission relationships in the trusted production state graph. The propagation paths that are continuously associated with the changed order nodes, process nodes, equipment nodes, material nodes, and personnel nodes are extracted sequentially to form the upstream and downstream impact path results. The upstream and downstream impact path results are then collected and organized according to task type, scope of status change, and node association order, so that each path maintains a continuous correspondence with the corresponding order node, process node, equipment node, material node, and personnel node, generating an impact path set.

[0071] S3.3. Based on the set of affected paths, prune redundant paths, limit the scope of influence, and extract the corresponding graph nodes and edges to generate an affected subgraph.

[0072] Specifically, based on the starting node, transmission node, and termination node corresponding to each path in the set of impact paths, the comprehensive impact value of each path is calculated, as expressed by:

[0073] ;

[0074] in, Indicates the overall value of the impact path. This represents the sum of the intensity of state changes at nodes undergoing state transitions. This indicates the depth of the propagation hierarchy affecting the path. This indicates the strength of resource associations between orders, processes, equipment, materials, and personnel. The weights represent the intensity of the state change. This indicates the weight corresponding to the depth of the transmission hierarchy. The weight representing the strength of resource association, and , , All coefficients are positive zero and satisfy α + β + γ = 1. The weights α, β, and γ are determined based on the contribution of A, B, and C to the actual business impact in historical task processing. A, B, and C are normalized and combined with the order delay, process change range, equipment load fluctuation, material coordination impact, and personnel adjustment range corresponding to the historical task to form a historical impact calibration value. The weights are then determined according to the contribution ratio of A, B, and C to the historical impact calibration value. For example, α is 0.40 to 0.50, β is 0.15 to 0.25, and γ is 0.30 to 0.40, preferably α = 0.45, β = 0.20, and γ = 0.35. Among them, the sum of the change intensity of state change nodes has the strongest direct representation effect on the path impact, so α has the highest value. The resource association strength reflects the coupling propagation ability between multiple types of business objects, so γ has the second highest value. The propagation level depth is mainly used to characterize the impact propagation distance, so β ​​has a relatively low value.

[0075] Next, the comprehensive value of the impact path is compared with the preset path threshold. Paths with a comprehensive value of impact path greater than or equal to the preset path threshold and directly related to the task type and state change range are retained. Paths that are repeated, have the same transmission direction, and have the same impact result are pruned to obtain the path results after limiting the impact range. The corresponding order nodes, process nodes, equipment nodes, material nodes, and personnel nodes are extracted from the path results after limiting the impact range. At the same time, the graph edges corresponding to the business dependencies and state transmission relationships connecting the order nodes, process nodes, equipment nodes, material nodes, and personnel nodes are extracted. The extracted order nodes, process nodes, equipment nodes, material nodes, and personnel nodes are associated and organized with the corresponding graph edges to generate an impact subgraph.

[0076] It should also be noted that the preset path threshold is determined based on the distribution of the comprehensive value of the impact path in historical task processing, the actual impact of state change propagation on subsequent order progress and resource coordination, and the stability requirements of the results after impact path pruning. For example, in specific implementation, the sum of the change intensity of state change nodes, the depth of the transmission hierarchy of the impact path, and the strength of resource association are normalized, and the weight coefficients are controlled within the range of a sum of 1, so that the comprehensive value P of the impact path falls between 0 and 1. At this time, the preset path threshold can be set to 0.4 to 0.7. When the impact propagation range is large and it is necessary to improve the retention accuracy, the preset path threshold can be taken as 0.6 to 0.7. When it is necessary to expand the impact identification range, the preset path threshold can be taken as 0.4 to 0.6.

[0077] S4. The query planning AI agent performs local queries on the order, process, equipment, material and personnel related data in the trusted production status map corresponding to the influence subgraph based on the influence subgraph, generates a local query result set, and the recalculation AI agent performs incremental recalculation and synchronous update to generate a local recalculation result package.

[0078] S4.1 Extract the order, process, equipment, material and personnel nodes to be queried and their corresponding business relationships from the influence subgraph to generate a local query range set.

[0079] Specifically, based on the order nodes, process nodes, equipment nodes, material nodes, and personnel nodes in the influence subgraph, the business object identifiers, status information, and connection relationship information corresponding to each order node, process node, equipment node, material node, and personnel node are read item by item. Then, the order nodes, process nodes, equipment nodes, material nodes, and personnel nodes corresponding to the task type are filtered according to the task type and the range of status changes. The existing connection relationships between the filtered order nodes, process nodes, equipment nodes, material nodes, and personnel nodes are organized, and the business association relationships corresponding to the order nodes, process nodes, equipment nodes, material nodes, and personnel nodes are extracted. Finally, the order nodes, process nodes, equipment nodes, material nodes, and personnel nodes are aggregated with the corresponding business association relationships to generate a local query range set.

[0080] S4.2 Map and match the local query range set with the corresponding order, process, equipment, material and personnel association data in the trusted production status map, locate the data range corresponding to the local query range set, and generate a local query dataset.

[0081] Specifically, based on the order nodes, process nodes, equipment nodes, material nodes, and personnel nodes recorded in the local query scope set, and their corresponding business relationships, the system searches for the order, process, equipment, material, and personnel-related data corresponding to the order nodes, process nodes, equipment nodes, material nodes, and personnel nodes in the trusted production status graph, and performs mapping and matching according to business object identifiers, status information, and connection relationships. Based on the mapping and matching results, the system filters out the order, process, equipment, material, and personnel-related data that are consistent with the local query scope set, thus clarifying the data range corresponding to the local query scope set. The filtered order, process, equipment, material, and personnel-related data are then collected and organized to generate a local query dataset.

[0082] S4.3 Perform a partial query on the local dataset to be queried, extract the associated status data of the current task according to the order dimension, process dimension, equipment dimension, material dimension and personnel dimension, and generate a local query result set.

[0083] Specifically, based on the order, process, equipment, material, and personnel association data centrally recorded in the local query dataset, the status information corresponding to each order, process, equipment, material, and personnel is extracted according to the order dimension, process dimension, process dimension, material dimension, and personnel dimension. The extracted results from these dimensions are then correlated with the business relationships corresponding to the current task, ensuring a continuous correspondence between the status data of orders, processes, equipment, materials, and personnel. Finally, the correlated and correlated results are aggregated to generate a local query result set.

[0084] It should also be noted that associated status data refers to the set of status information that connects orders, processes, equipment, materials, and personnel under the business relationship corresponding to the current task and can reflect the status transmission relationship.

[0085] S4.4 Compare the current state with the historical state in the local query result set, identify the key state items that have changed, and generate a difference state result set.

[0086] Specifically, based on the order dimension extraction results, process dimension extraction results, equipment dimension extraction results, material dimension extraction results, and personnel dimension extraction results recorded in the local query results set, the corresponding current status information is read, and the historical status information corresponding to the order dimension extraction results, process dimension extraction results, equipment dimension extraction results, material dimension extraction results, and personnel dimension extraction results is retrieved. The current status information and historical status information are compared item by item according to business object identifier, status content, and time order to identify records where the status value has changed, the status order has changed, or the status relationship has changed. The identified status change records are collected and organized according to the business relationship between orders, processes, equipment, materials, and personnel to generate a set of differential status results.

[0087] S4.5 Based on the difference status result set, recalculate the affected order progress, process connection, equipment load, material availability and personnel allocation relationship, and generate an incremental recalculation result set.

[0088] Specifically, based on the status change content recorded in the differential status results, the associated records of changed orders, processes, equipment, materials, and personnel are traced back and located according to business object identification, time sequence, and business relationships to determine the affected order progress, process connection, equipment load, material availability, and personnel allocation relationships. Using the change point corresponding to the status change record as the recalculation starting point, the start time, completion time, and processing sequence of the affected processes are updated sequentially, and the completed and remaining progress of orders are simultaneously corrected based on the updated process time results. Based on the updated process occupancy time and task allocation results, the task occupancy and idle time of the corresponding equipment during the changed period are recalculated to obtain the updated equipment load. Based on the updated process material requirements, existing inventory, in-transit replenishment, and allocated quantity, the material surplus for the corresponding period is recalculated to obtain the updated material availability. Based on the updated process execution time, job requirements, and personnel occupancy status, the allocation relationship of the corresponding personnel during the changed period is adjusted to obtain the updated personnel allocation relationship. The updated order progress, process connection, equipment load, material availability, and personnel allocation relationships are collected and organized according to business object identification and time sequence to generate an incremental recalculation result set.

[0089] S4.6 Write the incremental recalculation result set into the data structure corresponding to the local query result set to generate the local recalculation result package.

[0090] Specifically, based on the recalculation results recorded in the incremental recalculation result set regarding order progress, process connection, equipment load, material availability, and personnel allocation, each item is matched with the corresponding data content in the local query result set according to the business object identifier, time sequence, and status category. The recalculation results corresponding to the incremental recalculation result set are written into the data structure corresponding to the local query result set, maintaining the continuous correspondence between business relationships of orders, processes, equipment, materials, and personnel. The written data content is then collected and organized to form a unified result structure between the recalculated order progress, process connection, equipment load, material availability, and personnel allocation relationships and other related status data in the local query result set, generating a local recalculation result package.

[0091] It should be noted that an impact subgraph is generated based on the set of impact paths. Then, a local query is performed only on the order, process, equipment, material, and personnel related data in the trusted production state map that correspond to the impact subgraph. The difference states are identified around the local query result set, and only the affected order progress, process connection, equipment load, material availability, and personnel allocation relationship are recalculated to form a local recalculation result package. This constrains the state changes within a clear data range, keeps the query boundary, recalculation range, and update path consistent, and enhances the pertinence of state propagation analysis, the coherence of the data processing link, and the input stability of subsequent processing solutions.

[0092] S5. The decision-making and orchestration AI agent determines the execution conditions of the local recalculation result package, generates a processing plan, and orchestrates it to generate an execution instruction package.

[0093] S5.1. Perform result parsing and execution condition determination on the partial recalculation result package, extract order change information, process change information, equipment change information, material change information and personnel change information, and identify the execution conditions corresponding to the current task according to the preset execution rules, and generate an execution condition determination result set.

[0094] Specifically, based on the recalculation results recorded in the partial recalculation result package regarding order progress, process connection, equipment load, material availability, and personnel allocation, the partial recalculation result package is parsed according to business object identification, time sequence, and status category. Order change information, process change information, equipment change information, material change information, and personnel change information are extracted respectively. These information are then matched with the business objectives corresponding to the current task. Each of these information is then evaluated according to preset execution rules to identify the execution conditions corresponding to the current task. Finally, the order change information, process change information, equipment change information, material change information, and personnel change information are associated and aggregated with the identified execution conditions to generate an execution condition judgment result set.

[0095] It should also be noted that the preset execution rules are determined based on order delivery requirements, process sequence constraints, equipment availability, material supply status, personnel job matching relationships, and the corresponding timeliness and resource coordination requirements in production management objectives.

[0096] S5.2 Based on the result set of execution conditions, generate corresponding adjustment paths for the associated objects of orders, processes, equipment, materials and personnel, and verify resource occupation conflicts, business timing conflicts and status transmission conflicts, and generate a set of post-verification processing solutions.

[0097] Specifically, based on the order change information, process change information, equipment change information, material change information, personnel change information, and execution conditions recorded in the execution condition judgment results, corresponding adjustment paths are generated for the associated objects of orders, processes, equipment, materials, and personnel, ensuring that the business relationships of the associated objects of orders, processes, equipment, materials, and personnel remain continuously corresponding in the corresponding adjustment paths; each of the associated objects of orders, processes, equipment, materials, and personnel is checked for resource occupation conflicts, business timing conflicts, and status transmission conflicts according to the corresponding adjustment paths, and corresponding adjustment paths that do not meet the execution conditions are eliminated or corrected; the corresponding adjustment paths that pass the verification are collected and organized according to business object identifiers and relationships to generate a set of post-verification processing solutions.

[0098] It should also be noted that resource occupancy conflict refers to the situation where the same equipment, material, or personnel is repeatedly occupied by multiple related objects within the same time frame, or the occupancy exceeds the available range; business sequence conflict refers to the situation where the processing order of orders, processes, equipment, materials, and personnel does not meet order delivery requirements, process sequence constraints, or time requirements; state transmission conflict refers to the situation where, after the state change of an upstream related object is transmitted to a downstream related object, there is a logical inconsistency, business semantic contradiction, or break in the transmission relationship between the preceding and following states; for example, when two corresponding adjustment paths arrange the same equipment to execute different processes at the same time, it is judged as a resource occupancy conflict; when the preceding process has not been completed but the subsequent process has been scheduled for execution, it is judged as a business sequence conflict; when the equipment state has changed to shutdown but the corresponding process remains in the execution state, it is judged as a state transmission conflict.

[0099] S5.3. Prioritize and arrange the post-verification processing scheme set according to the urgency of the task, the range of status changes, and the intensity of business impact. Then, convert the arranged processing content into executable business operation instructions and generate an execution instruction package.

[0100] Specifically, based on the task type, status change content, and business relevance of each corresponding adjustment path in the post-verification processing scheme set, the corresponding adjustment paths are sorted according to task urgency, status change range, and business impact intensity. The sorted corresponding adjustment paths are then arranged into schemes according to their priority. The processing content corresponding to orders, processes, equipment, materials, and personnel is extracted from the arranged corresponding adjustment paths, and the processing content corresponding to orders, processes, equipment, materials, and personnel is converted into executable business operation instructions with business object identifiers, processing actions, execution order, and time requirements. The converted executable business operation instructions are collected and organized according to the arrangement order to generate an execution instruction package.

[0101] S6. The execution feedback AI agent performs business distribution, feedback consistency verification, and update event write-back processing on the execution instruction package, and generates a feedback correction result package.

[0102] S6.1. Split the execution instruction package and execute the business deployment, generating a business execution feedback set.

[0103] Specifically, based on the executable business operation instructions corresponding to orders, processes, equipment, materials, and personnel recorded in the execution instruction package, the executable business operation instructions are broken down according to the business object identifier, processing action, execution order, and time requirements to form issuance instructions corresponding to orders, processes, equipment, materials, and personnel. The issuance instructions corresponding to orders, processes, equipment, materials, and personnel are sent to the corresponding business processing links according to the execution order, and the receiving status, processing status, and return status of each issuance instruction are recorded. The receiving status, processing status, and return status of each issuance instruction are collected and organized according to the business object identifier and execution order to generate a business execution feedback set.

[0104] S6.2 Perform feedback aggregation and feedback consistency verification on the business execution feedback set, and generate a consistency verification result set.

[0105] Specifically, based on the received, processing, and return statuses recorded in the business execution feedback set, the received, processing, and return statuses are aggregated according to the business object identifier, execution order, and processing action, ensuring continuous correspondence between the feedback content for orders, processes, equipment, materials, and personnel. The aggregated received, processing, and return statuses are then compared item by item with the corresponding executable business operation instructions in the execution instruction package. A correspondence between feedback records and executable business operation instructions is established based on the business object identifier. Next, the actual processing actions recorded in the received, processing, and return statuses are compared to those recorded in the executable business operation instructions to see if they match. The execution order is also compared to see if the actual execution sequence recorded in the received, processing, and return statuses matches the execution order recorded in the executable business operation instructions. Finally, the expected processing content is compared to see if the actual processing result recorded in the return status matches the expected processing content recorded in the executable business operation instructions. The results of the processing action comparison, execution order comparison, and expected processing content comparison are then summarized and organized to generate a consistency verification result set.

[0106] S6.3 Extract update events from the consistency check result set, and encapsulate the update events and consistency check results to generate a feedback correction result package.

[0107] Specifically, based on the comparison results of processing actions, execution order, and expected processing content recorded in the consistency verification result set, records that have passed verification and whose status has changed are selected. The corresponding business object identifier, status change content, and time information are extracted from the records that have passed verification and whose status has changed to form update events. The update events are written back to the corresponding event records according to the business object identifier, status change content, and time information to ensure that the update events and the corresponding verification status in the consistency verification result set are continuously correlated. The update events, processing action comparison results, execution order comparison results, and expected processing content comparison results are collected and organized, and the processing results are encapsulated according to the business object identifier and time order to generate a feedback correction result package.

[0108] It should be noted that the task parsing AI agent, query planning AI agent, recalculation AI agent, decision orchestration AI agent, and execution feedback AI agent are all task-oriented AI agents deployed on computer devices. Each AI agent is formed by corresponding task prompt templates, rule constraint sets, business field mapping relationships, and tool interface configurations. Among them, the task parsing AI agent is used to identify task types and state change ranges based on the trusted production state map; the query planning AI agent is used to generate a local query range set based on the influence subgraph; the recalculation AI agent is used to generate an incremental recalculation result set based on the local query result set and the difference state result set; the decision orchestration AI agent is used to determine execution conditions, check conflicts, and orchestrate schemes based on the local recalculation result package and generate an execution instruction package; and the execution feedback AI agent is used to check feedback consistency based on the business execution feedback set and generate a feedback correction result package. Each AI agent uses the output of the previous AI agent as the input of the next AI agent for sequential collaborative processing, thereby forming a closed-loop processing link in which multiple AI agents participate collaboratively.

[0109] This embodiment also provides a computer device applicable to the production management information processing method based on AI intelligent agents, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the production management information processing method based on AI intelligent agents as proposed in the above embodiment.

[0110] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0111] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the production management information processing method based on AI intelligent agents as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0112] In summary, this invention, by performing local queries on corresponding related data in the reliable production state map based on the influence subgraph and generating a local recalculation result package, can construct a data processing scope with clear boundaries around the nodes and related links involved in state changes. This allows related data extraction, difference identification, incremental recalculation, and result aggregation to be continuously connected along a unified processing path. It can integrate the scattered state relationships between orders, processes, equipment, materials, and personnel into a structured data foundation that can directly participate in recalculation, enhancing the focus of state change identification and the orderliness of result organization.

[0113] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A production management information processing method based on AI intelligent agents, characterized in that: include, Collect real-time business data for production management, and perform time alignment, master data mapping, duplicate record elimination, and quality status calculation to generate standardized production event streams and data quality meta-streams; Based on the standardized production event flow and data quality metaflow, the standardized production event flow is subjected to hierarchical governance and state aggregation to generate a reliable production state graph. The task-analysis AI agent identifies task types and the range of state changes based on a reliable production state map, and traces the upstream and downstream impact paths along the business relationships between orders, processes, equipment, materials and personnel, generating an impact sub-map; The query planning AI agent performs local queries on the order, process, equipment, material and personnel related data in the trusted production status map corresponding to the influence subgraph based on the influence subgraph, generates a local query result set, and the recalculation AI agent performs incremental recalculation and synchronous update to generate a local recalculation result package. The decision-making and orchestration AI agent determines the execution conditions of the local recalculation result package, generates a processing plan, and orchestrates it to generate an execution instruction package. The execution feedback AI agent performs business distribution, feedback consistency verification, and update event write-back processing on the execution instruction package, generating a feedback correction result package.

2. The production management information processing method based on AI intelligent agents as described in claim 1, characterized in that: The specific steps for generating standardized production event streams and data quality metadata streams are as follows. Collect real-time production management business data from the order management terminal, production scheduling terminal, warehouse management terminal, equipment acquisition terminal, quality record terminal and manual entry terminal, and add source identifier, business object identifier and collection time identifier to the real-time production management business data to generate the original production management dataset; The original production management dataset is processed for time alignment, master data matching, and object mapping to generate a mapped production management dataset. The mapping production management dataset is subjected to duplicate record elimination, data quality inspection, field standardization, event structure transformation, and quality status association to generate a standardized production event stream and data quality meta-stream.

3. The production management information processing method based on AI intelligent agents as described in claim 2, characterized in that: The specific steps for generating the trusted production status map are as follows: Based on the standardized production event flow and data quality metaflow, establish the correlation between events and quality status, and classify them according to preset quality rules to generate a hierarchical production event set; Implement hierarchical governance for the hierarchical production event set, and extract the status information and relationships of order events, process events, equipment events, material events, and personnel events to generate a status-related production event set; Based on the state-related production event set, merge the state changes of the same order, the same process, the same equipment, the same material, and the same person at different times, and write the order, process, equipment, material, and person into the graph nodes, and write the business dependencies and state transit relationships into the graph edges to generate a reliable production state graph.

4. The production management information processing method based on AI intelligent agents as described in claim 3, characterized in that: The specific steps for generating the influence subgraph are as follows: Extract task trigger information and determine task type from the trusted production status map, identify the status change information corresponding to new tasks, abnormal tasks, query tasks and adjustment tasks, and generate a task type determination result set. The task type determination result set is used to identify the range of state changes and track upstream and downstream relationships. The order nodes, process nodes, equipment nodes, material nodes and personnel nodes that have changed in the trusted production state map are located, and the upstream and downstream impact paths are tracked along business dependencies and state transmission relationships to generate an impact path set. Based on the set of affected paths, redundant paths are pruned, the scope of influence is limited, and the corresponding graph nodes and edges are extracted to generate an affected subgraph.

5. The production management information processing method based on AI intelligent agents as described in claim 4, characterized in that: The specific steps for generating the local query result set are as follows: Extract the order, process, equipment, material, and personnel nodes to be queried, as well as their corresponding business relationships, from the influence subgraph to generate a local query scope set; Map and match the local query range set with the corresponding order, process, equipment, material and personnel association data in the trusted production status map to locate the data range corresponding to the local query range set and generate a local query dataset; Perform a partial query on the dataset to be queried, extract the associated status data of the current task according to the dimensions of order, process, equipment, material and personnel, and generate a partial query result set.

6. The production management information processing method based on AI intelligent agents as described in claim 5, characterized in that: The specific steps for generating the local recalculation result package are as follows: The current state in the local query result set is compared with the historical state, and the key state items that have changed are identified to generate a set of difference state results. Based on the difference status result set, recalculate the affected order progress, process connection, equipment load, material availability and personnel allocation relationship, and generate an incremental recalculation result set; Write the incremental recalculation result set into the data structure corresponding to the local query result set to generate the local recalculation result package.

7. The production management information processing method based on AI intelligent agents as described in claim 6, characterized in that: The specific steps for generating the execution instruction package are as follows: The system parses the results of the partial recalculation and determines the execution conditions, extracts order change information, process change information, equipment change information, material change information and personnel change information, and identifies the execution conditions corresponding to the current task according to the preset execution rules, generating an execution condition determination result set. Based on the result set of execution conditions, corresponding adjustment paths are generated for the associated objects of orders, processes, equipment, materials and personnel, and resource occupation conflicts, business timing conflicts and status transmission conflicts are verified to generate a set of post-verification processing solutions. Based on the urgency of the task, the scope of the status change, and the intensity of the business impact, the set of post-verification processing solutions is prioritized and arranged. The arranged processing content is then converted into executable business operation instructions to generate an execution instruction package.

8. The production management information processing method based on AI intelligent agents as described in claim 7, characterized in that: The specific steps for generating the feedback correction result package are as follows: The execution instruction package is split and executed, and the business execution feedback set is generated. The business execution feedback set is aggregated and its consistency is verified to generate a consistency verification result set. Update events are extracted from the consistency check result set, and the update events and consistency check results are encapsulated to generate a feedback correction result package.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the production management information processing method based on AI intelligent agents as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the production management information processing method based on AI intelligent agents as described in any one of claims 1 to 8.

Images