An intelligent production management method, device, medium and product based on a large model

Abstract

The application provides a large model-based intelligent production management method, device, medium and product, and relates to the technical field of intelligent production. The method comprises the following steps: collecting real-time production data of a production line cluster and inputting the data into a preset model for joint analysis. Production bottleneck prediction signals and resource gap probability distributions output by the preset model are obtained, and a production bottleneck propagation path is determined accordingly. When an emergency production task needs to be inserted, conflict analysis is performed according to the production bottleneck prediction signals and the resource gap probability distributions, a production task adjustment instruction set is generated, and the production line cluster is adjusted according to the production task adjustment instruction set. The technical problem that in the related art, the multi-dimensional resource conflict caused by the sudden production bottleneck and the insertion of the emergency task in a large-scale complex production environment cannot be dynamically responded in real time is solved, and the effect of improving the real-time dynamic response capability of the multi-dimensional resource conflict caused by the sudden production bottleneck and the insertion of the emergency task in a large-scale complex production environment is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent production, and particularly relates to an intelligent production management method and device based on a large model, a medium and a product. BACKGROUND

[0002] With the rapid development of intelligent manufacturing, the market requirements for product delivery speed and quality are increasing, which puts forward more refined and intelligent requirements for the management of production line clusters.

[0003] In the related art, in the production line management process, production data is collected regularly, and based on statistical analysis methods such as arithmetic mean method and simple trend analysis, the production situation is evaluated. Specifically, this process usually includes the following steps: collecting the yield, equipment running time, raw material consumption quantity and other key data of each production node according to a fixed time interval (for example, every hour or every day); using data analysis software to perform arithmetic mean operation on these key data to calculate the average yield and other key indicators of each production node; and observing the trend of the data to roughly understand the overall situation of the production.

[0004] However, using the above method, since it relies on fixed-period static data analysis, it cannot capture real-time dynamic changes in the production process, and thus the multi-dimensional resource conflict caused by the insertion of sudden production bottlenecks and emergency tasks in a large-scale complex production environment cannot be dynamically responded in real time in the related art. SUMMARY

[0005] The present application provides an intelligent production management method and device based on a large model, a medium and a product, which are used to improve the real-time dynamic response capability to the multi-dimensional resource conflict caused by the insertion of sudden production bottlenecks and emergency tasks in a large-scale complex production environment.

[0006] In a first aspect, the present application provides a large model-based intelligent production management method, applied to the above-mentioned electronic device, which comprises: in the case of collecting real-time production data of the current production task of the production line cluster, inputting the real-time production data into a preset model to control the preset model to perform a joint analysis operation on the real-time production data; in the case of determining that the preset model has completed the joint analysis operation, obtaining the production bottleneck prediction signal and the resource gap probability distribution output by the preset model, wherein the production bottleneck prediction signal is the risk intensity of the first production node in the production line cluster in the first prediction period, the first prediction period is the bottleneck time range of the first production node predicted by the preset model to appear insufficient capacity, and the resource gap probability distribution is the probability distribution of the second production node in the production line cluster in the second prediction period to appear a multi-dimensional resource shortage event, and the second prediction period is the emergency time range of the second production node predicted by the preset model to appear a multi-dimensional resource shortage event; determining the production bottleneck propagation path according to the production bottleneck prediction signal, wherein the production bottleneck propagation path is a multi-stage influence path in the production line cluster triggered by the initial bottleneck node, and the initial bottleneck node includes the first production node and the second production node; in the case of detecting that the production line cluster needs to insert an emergency production task, performing a conflict analysis operation on the production bottleneck prediction signal, the production bottleneck propagation path, and the resource gap probability distribution on the emergency production task to generate a production task adjustment instruction set; and performing a production task adjustment operation on the production line cluster according to the production task adjustment instruction set.

[0007] By adopting the above technical solution, the real-time production data of the production line cluster is collected and input into the preset model for analysis, and the production bottleneck prediction signal and the resource gap probability distribution can be obtained. According to the prediction signal, the production bottleneck propagation path is determined, and when an emergency production task needs to be inserted, a conflict analysis is performed based on multiple information to generate a task adjustment instruction set, and the production task is adjusted accordingly. Thus, real-time and comprehensive monitoring and management of the production line cluster can be realized, production risks can be predicted in advance, when an emergency task is inserted, production can be adjusted through scientific analysis, production conflicts can be reduced, production processes can be optimized, production efficiency and resource utilization rate of the production line cluster can be greatly improved, and production can be ensured to proceed smoothly. Further, the technical problem that in related technologies, multi-dimensional resource conflicts caused by sudden production bottlenecks and emergency task insertion in large-scale complex production environments cannot be dynamically responded in real time is solved, and the technical effect of improving the real-time dynamic response capability to multi-dimensional resource conflicts caused by sudden production bottlenecks and emergency task insertion in large-scale complex production environments is achieved.

[0008] Optionally, in the case of collecting real-time production data of the current production task of the production line cluster, the real-time production data is input into the preset model to control the preset model to perform joint analysis operation on the real-time production data, specifically including: obtaining the equipment state change, material flow trajectory, order demand sequence and external market fluctuation parameter of each production node in the production line cluster through the distributed sensing device in real time, wherein the real-time production data includes the equipment state change, material flow trajectory, order demand sequence and external market fluctuation parameter, and each production node includes an initial bottleneck node; performing spatio-temporal feature fusion processing on the equipment state change, material flow trajectory, order demand sequence and external market fluctuation parameter to generate a multi-modal data set, wherein the multi-modal data set is a multi-dimensional data set with a unified time reference and a spatial identifier; inputting the multi-modal data set into the preset model to control the preset model to perform the following operations: the preset model performs abnormal detection analysis on the multi-modal data set to obtain an abnormal detection result when it is determined that the multi-modal data set is received; the preset model performs abnormal correlation analysis on the historical fault mode library and the abnormal detection result to generate a risk event correlation rule library; the preset model performs time series cross-validation analysis on the equipment state change and the order demand sequence to determine the time series coupling relationship of the equipment state change and the order demand sequence in the time series dimension; the preset model performs spatial topology mapping analysis on the equipment state change and the order demand sequence to determine the spatial coupling relationship of the equipment state change and the order demand sequence in the spatial dimension; the preset model performs confidence analysis on the production baseline data and the external market fluctuation parameter to determine the target confidence interval of the multi-dimensional resource shortage event, wherein the production baseline data is a reference data set for the production line cluster to maintain a resource supply-demand balance state; the preset model generates a production bottleneck prediction signal and a resource gap probability distribution according to the risk event correlation rule library, the time series coupling relationship, the spatial coupling relationship and the target confidence interval; and the preset model outputs the production bottleneck prediction signal and the resource gap probability distribution.

[0009] By adopting the above technical solution, multi-dimensional real-time production data is obtained through distributed sensing devices, a multi-modal data set is generated through spatio-temporal feature fusion, and then input into the preset model. The preset model performs a series of operations such as abnormal detection and correlation analysis on the data set to generate a production bottleneck prediction signal and a resource gap probability distribution. Thus, the collected data is comprehensive and accurate, the multi-modal data set provides high-quality data basis for model analysis, and the multi-dimensional analysis of the preset model makes the production bottleneck prediction signal and the resource gap probability distribution more accurate, providing a reliable basis for subsequent production decision-making.

[0010] Optionally, the device state changes, material flow trajectories, order demand sequences, and external market fluctuation parameters are spatiotemporal feature fusion processed to generate a multi-modal data set, specifically including: timestamp alignment processing of the device state changes, material flow trajectories, order demand sequences, and external market fluctuation parameters to generate original data streams with a unified time base; assigning spatial identifiers to the original data streams, wherein the spatial identifiers are used to identify the production device physical coordinates corresponding to the device state changes and the warehouse location and transportation path topological relationship corresponding to the material flow trajectories; extracting time series features of the device state changes, spatial distribution features of the material flow trajectories, demand fluctuation features of the order demand sequences, and market trend features of the external market fluctuation parameters from the original data streams with assigned spatial identifiers; feature fusion of the time series features, spatial distribution features, demand fluctuation features, and market trend features to generate a multi-modal data set.

[0011] By adopting the above technical solution, the device state changes and other data are timestamp aligned, spatial identifiers are assigned, data features are extracted, and a multi-modal data set is generated by fusion. Timestamp alignment ensures unified data time base, spatial identifiers clearly indicate data spatial position relationship, feature extraction obtains key information, and feature fusion integrates multi-source information. The generated multi-modal data set contains rich spatiotemporal features, can be better processed by a preset model, and improves model analysis accuracy and effectiveness.

[0012] Optionally, the preset model generates a production bottleneck prediction signal and a resource gap probability distribution according to a risk event association rule library, a time sequence coupling relationship, and a spatial coupling relationship target confidence interval, specifically including: the preset model performs fault mode matching on the abnormal detection result and the risk event association rule library to determine the association degree of the abnormal detection result and the fault mode in the historical fault mode library; the preset model determines the fault propagation weight of each production node according to the association degree; the preset model constructs a dynamic coupling model of the production line cluster according to the time sequence coupling relationship and the spatial coupling relationship; the preset model inputs the fault propagation weight into the dynamic coupling model to obtain the production capacity deviation of the first production node in the first prediction period and the supply-demand gap amount of the multi-dimensional resources of the second production node in the second prediction period output by the dynamic coupling model after forward propagation analysis of the fault propagation weight, wherein the production capacity deviation is the deviation value of the actual production capacity and the planned production capacity of the first production node in the first prediction period, and the supply-demand gap amount is the difference between the real-time inventory amount and the demand amount of the multi-dimensional resources of the second production node in the second prediction period; the preset model performs dynamic probability mapping of the production capacity deviation and the target confidence interval to generate a production bottleneck prediction signal; and the preset model performs multi-level probability calibration of the supply-demand gap amount and the target confidence interval to generate a resource gap probability distribution.

[0013] By adopting the technical scheme, the preset model matches the abnormality detection result with the risk event association rule library, determines the fault propagation weight, constructs a dynamic coupling model in combination with a time sequence and a space coupling relationship, inputs the fault propagation weight to obtain a production capacity deviation and a supply-demand gap, and then generates a production bottleneck prediction signal and a resource gap probability distribution through dynamic probability mapping and multi-level probability calibration. This process can mine deep correlations of data, construct a dynamic coupling model conforming to actual production, make the generated production bottleneck prediction signal and resource gap probability distribution more conform to the production status, and provide strong support for production decision-making.

[0014] Optionally, the production bottleneck propagation path is determined according to the production bottleneck prediction signal, and specifically includes: determining a time sequence association sequence of potential impact production nodes from the production line cluster according to the initial bottleneck node and the fault propagation weight, wherein the time sequence association sequence includes a causal relationship strength and an impact delay time between any two adjacent production nodes in the initial bottleneck node and the potential impact production nodes, the causal relationship strength is a fault propagation probability between any two adjacent production nodes, the impact delay time is a first shortest time interval required for fault propagation between any two adjacent production nodes, and the production bottleneck prediction signal includes the initial bottleneck node; a propagation delay matrix in a time sequence dimension is established according to the time sequence coupling relationship, and a space adjacency matrix in a space dimension is established according to the space coupling relationship, wherein the propagation delay matrix represents a symmetric matrix of fault propagation time delay between production nodes, a non-diagonal element value of the propagation delay matrix is a second shortest time interval required for fault propagation between production nodes, the second shortest time interval includes the first shortest time interval, the space adjacency matrix represents a topological matrix of physical connection relationships between production nodes, and a non-zero element value of the space adjacency matrix is a physical connection strength between production nodes; a plurality of propagation probabilities of a plurality of propagation paths of the initial bottleneck node are determined according to the propagation delay matrix and the space adjacency matrix; the plurality of propagation probabilities are compared with a preset propagation threshold to obtain a probability comparison result; a first propagation probability greater than the preset propagation threshold is determined from the plurality of propagation probabilities according to the probability comparison result; a first propagation path corresponding to the first propagation probability is determined from the plurality of propagation paths; and the first propagation path is determined as the production bottleneck propagation path.

[0015] By adopting the technical scheme, the time sequence association sequence of potential impact production nodes is determined according to the initial bottleneck node and the fault propagation weight, the propagation delay matrix and the space adjacency matrix are established in combination with the time sequence and the space coupling relationship, the propagation probabilities of the plurality of propagation paths are determined, and the production bottleneck propagation path is determined by comparison with the preset propagation threshold. The production bottleneck propagation path is determined from multiple dimensions by analyzing the relationship between production nodes, the accuracy and reliability of the determination of the production bottleneck propagation path are improved, and a clear path basis is provided for subsequent production decision-making.

[0016] Optionally, in the case of detecting that the production line cluster needs to insert an emergency production task, the production bottleneck prediction signal, the production bottleneck propagation path, and the resource gap probability distribution are subjected to a conflict analysis operation to generate a production task adjustment instruction set, specifically including: determining a time conflict area, a space conflict area, and a resource conflict area of the emergency production task and the current production task according to the production bottleneck prediction signal and the production bottleneck propagation path, wherein the time conflict area is an overlapping area of a required time period of the emergency production task and the first prediction time period and the second prediction time period, the space conflict area is a physical topology intersection area of a required device of the emergency production task and a device where the initial bottleneck node is located, and the resource conflict area is an intersection area of a resource type corresponding to a high confidence interval of the resource gap probability distribution and a required resource of the emergency production task, the target confidence interval including the high confidence interval; determining an adjustment priority index of the emergency production task according to the time conflict area, the space conflict area, and the resource conflict area; generating a resource allocation scheme of the emergency production task according to the target confidence interval and the adjustment priority index; and performing feasibility verification on the resource allocation scheme according to the time conflict area, the space conflict area, and the resource conflict area to generate the production task adjustment instruction set.

[0017] By adopting the above technical solution, when detecting that an emergency production task needs to be inserted, the time, space, and resource conflict areas are determined according to the production bottleneck prediction signal and the production bottleneck propagation path, the adjustment priority index is determined accordingly, the resource allocation scheme is generated in combination with the target confidence interval, and the feasibility verification is performed to generate the production task adjustment instruction set. This process comprehensively analyzes the conflict between the emergency production task and the current production, scientifically determines the adjustment strategy, and the generated production task adjustment instruction set can effectively guide the production adjustment, reduce the production conflict, and improve the production efficiency.

[0018] Optionally, the production task adjustment operation is performed on the production line cluster according to the production task adjustment instruction set, specifically including: determining the initial bottleneck node and the device scheduling parameter of the potentially affected production node according to the time rescheduling instruction, wherein the time rescheduling instruction is a time period optimization command generated according to the first prediction time period and the second prediction time period, the device scheduling parameter includes a device start-up time, a task execution time length and a buffer time interval, and the production task adjustment instruction set includes the time rescheduling instruction; reassigning the device runtime period of the initial bottleneck node and the potentially affected production node according to the device scheduling parameter and the time conflict area; determining the device layout parameter of the initial bottleneck node and the potentially affected production node according to the space optimization instruction, wherein the space optimization instruction is a space recombination command generated according to the physical topology intersection area, the device layout parameter includes a device spacing, a material transmission path and a cooperative working radius, and the production task adjustment instruction set includes the space optimization instruction; reconstructing the device workflow topology of the initial bottleneck node and the potentially affected production node according to the device layout parameter and the space conflict area; determining the resource hierarchical scheduling strategy of the initial bottleneck node and the potentially affected production node according to the resource allocation instruction, wherein the resource allocation instruction is a resource dynamic allocation command generated according to the target confidence interval, the resource hierarchical scheduling strategy includes a high confidence interval scheduling strategy, a medium confidence interval scheduling strategy and a low confidence interval scheduling strategy, and the production task adjustment instruction set includes the resource allocation instruction; determining the risk buffer scheme of the initial bottleneck node and the potentially affected production node according to the emergency preparation instruction, wherein the emergency preparation instruction is a risk prevention and control command generated according to the production bottleneck propagation path, the risk buffer scheme includes a backup device start-up plan, an emergency material reserve scheme and an exception handling plan, and the production task adjustment instruction set includes the emergency preparation instruction.

[0019] By adopting the above technical solutions, according to the production task adjustment instruction set, the device scheduling parameter is determined according to the time rescheduling instruction, the device runtime period is reassigned, the device layout parameter is determined according to the space optimization instruction, the device workflow topology is reconstructed, the resource hierarchical scheduling strategy is determined according to the resource allocation instruction, and the risk buffer scheme is determined according to the emergency preparation instruction. The production line cluster is comprehensively adjusted from time, space, resource and the like, the production task is ensured to be executed smoothly, and the ability of the production line to respond to unexpected situations is improved.

[0020] In a second aspect, an electronic device is provided, which includes one or more processors and a memory; the memory is coupled to the one or more processors, and is configured to store computer program codes, the computer program codes including computer instructions; the one or more processors invoke the computer instructions to enable the electronic device to perform the method described in the first aspect and any possible implementation manner of the first aspect.

[0021] In a third aspect, the embodiments of the present application provide a computer program product containing instructions, which, when executed on an electronic device, cause the electronic device to perform the method described in the first aspect and any possible implementation manner of the first aspect.

[0022] In a fourth aspect, the embodiments of the present application provide a computer-readable storage medium, including instructions, which, when executed on an electronic device, cause the electronic device to perform the method described in the first aspect and any possible implementation manner of the first aspect.

[0023] The one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0024] 1. The intelligent production management method based on a large model provided in the present application collects real-time production data of a production line cluster, inputs a preset model for analysis, and can obtain production bottleneck prediction signals and resource gap probability distribution. The production bottleneck propagation path is determined according to the prediction signals, and when an emergency production task needs to be inserted, conflict analysis is performed based on multiple aspects of information and a task adjustment instruction set is generated to adjust the production task. Thus, real-time and comprehensive monitoring and management of the production line cluster can be achieved, production risks can be predicted in advance, when an emergency task is inserted, production can be adjusted through scientific analysis, production conflicts can be reduced, production processes can be optimized, production efficiency and resource utilization rate of the production line cluster can be greatly improved, and production can be ensured to proceed smoothly.

[0025] 2. The intelligent production management method based on a large model provided in the present application acquires multi-dimensional real-time production data through distributed sensing equipment, generates a multi-modal data set through spatio-temporal feature fusion, and then inputs the preset model. The preset model performs a series of operations such as anomaly detection and correlation analysis on the data set to generate production bottleneck prediction signals and resource gap probability distribution. Thus, the collected data is comprehensive and accurate, the multi-modal data set provides a high-quality data basis for model analysis, and the multi-dimensional analysis of the preset model makes the production bottleneck prediction signals and resource gap probability distribution more accurate, providing a reliable basis for subsequent production decisions.

[0026] 3. The intelligent production management method based on a large model provided in the present application performs timestamp alignment on data such as device state changes, assigns spatial identifiers, extracts features of each data, and fuses to generate a multi-modal data set. Timestamp alignment ensures that the data time reference is unified, spatial identifiers clarify the spatial position relationship of the data, feature extraction obtains key information, and feature fusion integrates multi-source information. The generated multi-modal data set contains rich spatio-temporal features, can be better processed by the preset model, and improves the accuracy and effectiveness of model analysis. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1is a flowchart of a large model-based intelligent production management method in an embodiment of the present application.

[0028] Figure 2 is an entity device structure schematic diagram of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0029] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to be limiting of the present application. As used in the specification and the appended claims of the present application, the singular forms "a," "an" and "the" are intended to include both singular and plural forms, unless the context clearly indicates otherwise. It will be further understood that the terms "and / or" as used herein refers to any or all possible combinations of one or more of the associated listed items.

[0030] Hereinafter, the terms "first", "second" are only for the purpose of description, and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, the meaning of "a plurality of" is two or more, unless otherwise specified.

[0031] The present application provides a large model-based intelligent production management method, referring to Figure 1 , Figure 1 is a flowchart of a large model-based intelligent production management method in an embodiment of the present application, comprising the following steps:

[0032] Step S101, in the case of collecting real-time production data of the current production task of the production line cluster, inputting the real-time production data into the preset model to control the preset model to perform joint analysis operation on the real-time production data;

[0033] Step S102, in the case of determining that the preset model has completed the joint analysis operation, obtaining the production bottleneck prediction signal and the resource gap probability distribution output by the preset model, wherein the production bottleneck prediction signal is the risk intensity of the first production node in the production line cluster in the first prediction period, the first prediction period is the bottleneck time range of the first production node predicted by the preset model to appear insufficient capacity, and the resource gap probability distribution is the probability distribution of the second production node in the production line cluster in the second prediction period to appear multi-dimensional resource shortage event, and the second prediction period is the emergency time range of the second production node predicted by the preset model to appear multi-dimensional resource shortage event;

[0034] In step S103, a production bottleneck propagation path is determined according to the production bottleneck prediction signal, wherein the production bottleneck propagation path is a multi-stage influence path in the production line cluster triggered by the initial bottleneck node, and the initial bottleneck node includes the first production node and the second production node.

[0035] In step S104, in a case where it is detected that the production line cluster needs to insert an emergency production task, a conflict analysis operation is performed on the production bottleneck prediction signal, the production bottleneck propagation path, and the resource gap probability distribution to the emergency production task to generate a production task adjustment instruction set.

[0036] In step S105, a production task adjustment operation is performed on the production line cluster according to the production task adjustment instruction set.

[0037] In the above embodiment, the production line cluster represents a production system formed by combining a plurality of production lines, and is used for mass production of products. The real-time production data refers to data generated by the production line cluster in the current production task execution process, such as equipment status, material flow, etc. The preset model is a large model that is built in advance and used to analyze production data. The production bottleneck prediction signal is used to represent the risk intensity of the first production node in the production line cluster in the first prediction period. The resource gap probability distribution refers to the probability distribution of multi-dimensional resource shortage events in the second production node in the production line cluster in the second prediction period. The first production node and the second production node can be the same production node in the production line cluster, or two different production nodes in the production line cluster. The production bottleneck propagation path is a multi-stage influence path in the production line cluster triggered by the initial bottleneck node.

[0038] In the above embodiment, a large household appliance manufacturing factory has multiple coordinated household appliance production lines, forming a production line cluster, aiming to achieve large-scale household appliance production. Specifically, on each production line, various types of distributed sensing devices are densely deployed, such as temperature sensors, pressure sensors, displacement sensors, and RFID (Radio Frequency Identification) tag readers for material tracking. These sensors and tag readers continuously and automatically collect data such as the operating status of production equipment, the flow trajectory of materials, the order demand quantity at each time period, and transmit these real-time production data to edge computing nodes through the factory's internal industrial Ethernet. The edge computing nodes preliminarily filter and preprocess the collected data, remove abnormal and duplicate data, and integrate device status data, material flow data, and order demand data to form a multi-modal data set. The multi-modal data set is input into a preset model, which performs joint analysis on the multi-modal data set, including anomaly detection, time series and spatial correlation analysis, to determine potential risks in the production process. Through joint analysis, the preset model generates production bottleneck prediction signals and resource gap probability distributions. For example, the preset model predicts that a key assembly device on a production line may experience insufficient capacity within the next 4 hours (it can also be within the next 2 hours, 3 hours, 5 hours, etc., or within a certain time period in the future, etc.), and that a certain type of component required by another production line is at risk of shortage within the next 6 hours (it can also be within the next 3 hours, 4 hours, 5 hours, etc., or within a certain time period in the future, etc.).

[0039] In the above embodiment, based on the production bottleneck prediction signal, the model determines the production bottleneck propagation path based on the physical connection relationship between the production lines and the logical relationship of the production process. For example, the prediction of insufficient capacity of the key assembly device may lead to a backlog in the subsequent detection process, thereby affecting the output of the entire production line. When an urgent production task is detected, a conflict analysis mechanism is automatically triggered, i.e., based on the production bottleneck prediction signal, the production bottleneck propagation path, and the resource gap probability distribution, the conflict area between the urgent production task and the current production task in terms of time, space, and resources is determined. Further, a production task adjustment instruction set is generated, including adjusting the operation plan of the equipment, reallocating material resources, etc. The production task adjustment instruction set is issued to each production unit through the industrial internet, and the production line automatically adjusts the production task according to the instructions. For example, adjust the start time and running speed of the equipment, re-plan the material distribution path, etc., enable standby equipment or adjust the production process to ensure the smooth execution of the urgent production task while minimizing the impact on the original production task.

[0040] Through the above steps, the real-time production data of the production line cluster is collected, input into a preset model for analysis, and the production bottleneck prediction signal and resource gap probability distribution can be obtained. According to the prediction signal, the production bottleneck propagation path is determined, when an emergency production task needs to be inserted, a conflict analysis is performed based on multiple information and a task adjustment instruction set is generated, and the production task is adjusted accordingly. Thus, real-time and comprehensive monitoring and management of the production line cluster can be achieved, production risks can be predicted in advance, when an emergency task is inserted, the production can be adjusted through scientific analysis, production conflicts can be reduced, the production process can be optimized, the production efficiency and resource utilization rate of the production line cluster can be greatly improved, and the production can be ensured to proceed smoothly. Further, the technical problem that in the related art, in a large-scale complex production environment, the multi-dimensional resource conflict caused by the sudden production bottleneck and the insertion of the emergency task cannot be dynamically responded in real time is solved, and the technical effect that the real-time dynamic response capability to the multi-dimensional resource conflict caused by the sudden production bottleneck and the insertion of the emergency task in a large-scale complex production environment is improved is achieved.

[0041] The execution subject of the above steps can be an intelligent production management system with production risk prediction capability, or an intelligent production management platform with production risk prediction capability, or an intelligent production management device with production risk prediction capability, but is not limited thereto.

[0042] In an optional embodiment, in the case that real-time production data of a current production task of the production line cluster is collected, the real-time production data is input into the preset model to control the preset model to perform joint analysis operation on the real-time production data, specifically including: obtaining device state changes, material flow trajectories, order demand sequences and external market fluctuation parameters of each production node in the production line cluster in real time through distributed sensing devices, wherein the real-time production data includes the device state changes, the material flow trajectories, the order demand sequences and the external market fluctuation parameters, and each production node includes an initial bottleneck node; performing spatio-temporal feature fusion processing on the device state changes, the material flow trajectories, the order demand sequences and the external market fluctuation parameters to generate a multi-modal data set, wherein the multi-modal data set is a multi-dimensional data set with a unified time reference and a spatial identifier; inputting the multi-modal data set into the preset model to control the preset model to perform the following operations: the preset model performs abnormal detection analysis on the multi-modal data set to obtain an abnormal detection result in the case that it is determined that the multi-modal data set is received; the preset model performs abnormal association analysis on the historical fault mode library and the abnormal detection result to generate a risk event association rule library; the preset model performs time series cross-validation analysis on the device state changes and the order demand sequences to determine the time series coupling relationship of the device state changes and the order demand sequences in the time series dimension; the preset model performs spatial topology mapping analysis on the device state changes and the order demand sequences to determine the spatial coupling relationship of the device state changes and the order demand sequences in the spatial dimension; the preset model performs confidence analysis on the production baseline data and the external market fluctuation parameters to determine the target confidence interval of the multi-dimensional resource shortage event, wherein the production baseline data is a reference data set for the production line cluster to maintain a resource supply and demand balance state; the preset model generates a production bottleneck prediction signal and a resource gap probability distribution according to the risk event association rule library, the time series coupling relationship, the spatial coupling relationship and the target confidence interval; and the preset model outputs the production bottleneck prediction signal and the resource gap probability distribution.

[0043] In the above embodiments, the distributed sensing devices are devices distributed at various positions of the production line for collecting different types of data. The device state change represents a state change of the production device during operation, for example, starting, stopping, failure, etc. The material flow trajectory refers to the moving path of the material in the production process. The order demand sequence refers to a sequence of related information of the production order arranged in time sequence. The external market fluctuation parameter is a parameter reflecting the change of the external market, for example, raw material price fluctuation, market demand change, etc. The multi-modal data set is a multi-dimensional data set with unified time reference and spatial identifier. The anomaly detection analysis is an operation of detecting abnormal data in the multi-modal data set; the anomaly correlation analysis is an operation of correlating the anomaly detection result with the historical failure mode library; the time series cross-validation analysis is used to determine the coupling relationship of the device state change and the order demand sequence in the time series dimension; the spatial topology mapping analysis is used to determine the coupling relationship of the device state change and the order demand sequence in the spatial dimension; and the confidence analysis is an operation of analyzing the production baseline data and the external market fluctuation parameter to determine the target confidence interval of the multi-dimensional resource shortage event.

[0044] In the above embodiments, in a large household appliance manufacturing factory, multiple production lines cooperatively produce various household appliances, such as refrigerators, televisions, and washing machines, etc. In order to achieve fine management of the production process and obtain more accurate production analysis results, distributed sensing devices are widely deployed on the production line. Temperature sensors, vibration sensors, etc. are installed on each production device to monitor the running state of the device in real time, and once the temperature or vibration of the device is abnormal, it can be detected in time. RFID readers are installed on the material conveying belt to track the material flow trajectory and accurately record the movement information of the material from the raw material warehouse to each production link. At the same time, the order demand sequence of the manufacturing factory is obtained, and the raw material price fluctuation, market demand change, and other external market fluctuation parameters are collected through the market monitoring platform. The data collected by the distributed sensing devices are preliminarily fused and feature extracted. The device state change, material flow trajectory, etc. are time-stamped and aligned to make all data have a unified time reference. Then, the data with unified time reference are assigned spatial identifiers to clearly define the physical coordinates of the device and the storage location and transportation path topology relationship of the material. The time series features of the device state change, the spatial distribution features of the material flow trajectory, the demand fluctuation features of the order demand sequence, and the market trend features of the external market fluctuation parameter are extracted from the data with unified time reference and spatial identifier. The time series features, spatial distribution features, demand fluctuation features, and market trend features are fused to generate a multi-modal data set.

[0045] In the above embodiment, the generated multi-modal data set is transmitted to the preset model. After receiving the multi-modal data set, the preset model initiates a series of analysis operations. That is, abnormality detection analysis is performed to identify abnormal points in the data, such as sudden changes in equipment operating parameters or abnormal interruptions in material flow. The abnormality detection result is analyzed in association with the historical fault mode library to find out possible fault causes and risk events. At the same time, time series cross-validation analysis and spatial topology mapping analysis are performed on the equipment state change and order demand sequence to determine the coupling relationship between the two in the time dimension and the space dimension. In addition, confidence analysis is performed in combination with the pre-established production baseline data and the collected external market fluctuation parameters to determine the target confidence interval of the multi-dimensional resource shortage event. According to the risk event association rule library, the time series coupling relationship, the spatial coupling relationship, and the target confidence interval, a production bottleneck prediction signal and a resource gap probability distribution are generated. Through the implementation of the above steps, real-time and accurate management of the production line can be realized, and production bottlenecks and resource gaps can be predicted in a timely manner, providing strong data support for production decisions and effectively improving production efficiency and resource utilization.

[0046] In an optional embodiment, the equipment state change, the material flow trajectory, the order demand sequence, and the external market fluctuation parameter are subjected to spatio-temporal feature fusion processing to generate a multi-modal data set, specifically including: performing timestamp alignment processing on the equipment state change, the material flow trajectory, the order demand sequence, and the external market fluctuation parameter to generate original data streams with a unified time reference; assigning spatial identifiers to the original data streams, wherein the spatial identifiers are used to identify the topological relationship between the physical coordinates of the production equipment corresponding to the equipment state change and the storage location and transportation path corresponding to the material flow trajectory; extracting time series features of the equipment state change, spatial distribution features of the material flow trajectory, demand fluctuation features of the order demand sequence, and market trend features of the external market fluctuation parameter from the original data streams to which the spatial identifiers have been assigned; and performing feature fusion on the time series features, the spatial distribution features, the demand fluctuation features, and the market trend features to generate a multi-modal data set.

[0047] In the above embodiment, the timestamp alignment processing is an operation of making different source data have a unified time reference. The spatial identifier is used to identify the topological relationship between the physical coordinates of the production equipment corresponding to the equipment state change and the storage location and transportation path corresponding to the material flow trajectory. The time series feature refers to a data feature that changes over time, for example, the trend of the change of the equipment state over time. The spatial distribution feature represents the distribution characteristics of the data in space, for example, the distribution of materials in the warehouse. The demand fluctuation feature reflects the fluctuation of the order demand sequence; the market trend feature embodies the change trend of the external market fluctuation parameter; and the feature fusion is an operation of integrating multiple features.

[0048] In the above embodiment, it is assumed that the production line of a large food and beverage production factory covers multiple links such as raw material preprocessing, beverage filling, packaging, etc. In order to realize fine management and intelligent decision-making of the production process, different types of data acquisition devices are deployed at each production link. In the raw material storage area, a weight sensor with timestamp function is installed to monitor the inventory weight change of raw materials in real time; in the filling workshop, a flow sensor with timestamp function is used to record the filling speed and flow of beverages; in the packaging link, a high-definition camera is deployed to capture image data of the packaging process and add timestamps to the images. The data collected by these data acquisition devices includes device state changes (e.g., whether the sensor is working properly, etc.), material flow trajectories (e.g., the movement path of raw materials from the storage area to the production line, etc.), etc. The collected multi-source data is timestamped to ensure that all data has a unified time reference. For example, by synchronizing with the internal time server, the data collected by the weight sensor, flow sensor and camera is uniformly calibrated to generate a raw data stream with a unified time reference. Based on the digital map of the production workshop, a spatial identifier is assigned to the raw data stream.

[0049] In the above embodiment, in the raw material storage area, an RFID tag is attached to each storage shelf, and the data collected by the weight sensor is associated with the corresponding RFID tag to identify the storage location of the raw materials; in the filling and packaging link, a unique spatial code is assigned to each production device, and the data collected by the flow sensor and camera is bound to the corresponding device code to clearly identify the physical coordinates of the data corresponding device, and the transportation path topology relationship corresponding to the material flow trajectory is recorded. Time series features are extracted from device state data, such as sensor data fluctuations over time, to determine whether the device is abnormal; spatial distribution features are extracted from material flow data to analyze the distribution of raw materials in the workshop; demand fluctuation features are extracted from order data to understand market demand changes for different products; market trend features are extracted from external environment data, such as the impact of seasonal factors on beverage sales. Use data fusion algorithms to fuse the extracted time series features, spatial distribution features, demand fluctuation features and market trend features. Integrate different types of features into a data structure to generate a multi-modal data set. This multi-modal data set not only contains rich production information, but also has a unified time reference and spatial identifier, providing high-quality data support for subsequent preset model analysis. Through the above implementation steps, multi-source production data can be effectively integrated to provide a solid data foundation for intelligent analysis and decision-making of production management, and to optimize the production process and improve production capacity.

[0050] In an optional embodiment, the preset model generates the production bottleneck prediction signal and the resource gap probability distribution according to the risk event association rule library, the time sequence coupling relationship, and the spatial coupling relationship target confidence interval, specifically including: the preset model performs fault mode matching on the abnormal detection result and the risk event association rule library to determine the association degree of the abnormal detection result and the fault mode in the historical fault mode library; the preset model determines the fault propagation weight of each production node according to the association degree; the preset model constructs a dynamic coupling model of the production line cluster according to the time sequence coupling relationship and the spatial coupling relationship; the preset model inputs the fault propagation weight into the dynamic coupling model to obtain the output of the first production node in the first prediction period and the supply-demand gap amount of the multi-dimensional resource of the second production node in the second prediction period after the dynamic coupling model performs forward propagation analysis on the fault propagation weight, wherein the capacity deviation is the deviation value of the actual capacity and the planned capacity of the first production node in the first prediction period, and the supply-demand gap amount is the difference value between the real-time inventory amount and the demand amount of the multi-dimensional resource of the second production node in the second prediction period; the preset model performs dynamic probability mapping on the capacity deviation and the target confidence interval to generate the production bottleneck prediction signal; and the preset model performs multi-level probability calibration on the supply-demand gap amount and the target confidence interval to generate the resource gap probability distribution.

[0051] In the above embodiment, the fault mode matching is an operation of comparing the abnormal detection result with the fault mode in the historical fault mode library. The fault propagation weight represents the possibility of fault propagation between the production nodes. The dynamic coupling model is constructed according to the time sequence and spatial coupling relationship of the production line cluster, and is used for analyzing the fault propagation model. The capacity deviation is the deviation value of the actual capacity and the planned capacity of the first production node in the first prediction period. The supply-demand gap amount is the difference value between the real-time inventory amount and the demand amount of the multi-dimensional resource of the second production node in the second prediction period. The dynamic probability mapping is an operation of mapping the capacity deviation and the target confidence interval to generate the production bottleneck prediction signal. The multi-level probability calibration is an operation of calibrating the supply-demand gap amount and the target confidence interval to generate the resource gap probability distribution.

[0052] In the above embodiment, it is assumed that a certain automobile parts manufacturer needs to perform dynamic probability mapping and multi-level probability calibration after obtaining the first production node capacity deviation and the second production node supply-demand gap. The specific dynamic probability mapping is to collect production data in the past three years, including planned capacity, actual capacity of various automobile parts, and records of events affecting production such as equipment failure and raw material supply delay. These production data are classified according to production stages and part types to form a structured data set. For example, in the engine block production stage, the planned capacity of different batches is recorded as 1000 pieces per week, and the actual capacity of a certain batch is reduced to 800 pieces due to equipment maintenance. The number of production bottleneck events (such as production line stoppage and delivery delay) is counted for each production stage and part type under different capacity deviation. The frequency of production bottleneck events in different capacity deviation intervals is calculated to construct a capacity deviation-bottleneck event correlation table, for example, when the capacity deviation is in the 10%-15% interval, the frequency of production bottleneck events in the engine block production stage is 40%. In real-time production, the capacity deviation of the first production node is automatically obtained. By querying the capacity deviation-bottleneck event correlation table, the corresponding production bottleneck prediction signal, i.e. the probability of production bottleneck occurrence, is directly obtained, for example, the capacity deviation of the current engine block production is 12%, and by querying the correlation table, the probability of production bottleneck occurrence is 40%. It should be noted that the above examples of actual values are only exemplary embodiments, and the above examples of actual values are not limited to the above examples.

[0053] In the above embodiment, the specific multi-level probability calibration is to divide the automobile parts production resources into three levels according to the impact of resources on production and the difficulty of obtaining. Core parts, such as engines and transmissions, are listed as key resources; important parts, such as brake systems and suspension systems, are listed as important resources; and general parts, such as screws and nuts, are listed as general resources. For each level of resources, collect past supply-demand gap data and resource shortage event records. The number of resource shortage events of each level under different supply-demand gap is counted, the frequency is calculated, and a supply-demand gap-resource shortage correlation library is established, for example, for the key resource engine, when the supply-demand gap reaches 15%, the frequency of resource shortage events is 70%. The supply-demand gap of the second production node is obtained in real time, and the level of the resource is determined. By querying the supply-demand gap-resource shortage correlation library, the resource gap probability is calibrated to obtain the final resource gap probability distribution, for example, when the supply-demand gap of the engine is detected to be 15%, the resource shortage event probability is automatically calibrated to 70% through the supply-demand gap-resource shortage correlation library. It should be noted that the above examples of actual values are only exemplary embodiments, and the above examples of actual values are not limited to the above examples.

[0054] In an optional embodiment, the production bottleneck propagation path is determined according to the production bottleneck prediction signal, specifically comprising: determining a time sequence correlation sequence of the potential impact production nodes from the production line cluster according to the initial bottleneck node and the fault propagation weight, wherein the time sequence correlation sequence comprises a causal relationship strength and an impact delay time between any two adjacent production nodes in the initial bottleneck node and the potential impact production nodes, the causal relationship strength is a fault propagation probability between any two adjacent production nodes, the impact delay time is a first shortest time interval required for fault propagation between any two adjacent production nodes, and the production bottleneck prediction signal comprises the initial bottleneck node; establishing a propagation delay matrix in the time dimension according to the time sequence coupling relationship, and establishing a space adjacency matrix in the space dimension according to the space coupling relationship, wherein the propagation delay matrix represents a symmetric matrix of fault propagation time delay between the production nodes, a non-diagonal element value of the propagation delay matrix is a second shortest time interval required for fault propagation between the production nodes, the second shortest time interval comprises the first shortest time interval, the space adjacency matrix represents a topological matrix of the physical connection relationship between the production nodes, and a non-zero element value of the space adjacency matrix is a physical connection strength between the production nodes; determining a plurality of propagation probabilities of a plurality of propagation paths of the initial bottleneck node according to the propagation delay matrix and the space adjacency matrix; performing probability comparison on the plurality of propagation probabilities and the preset propagation threshold to obtain a probability comparison result; determining a first propagation probability greater than the preset propagation threshold from the plurality of propagation probabilities according to the probability comparison result; determining a first propagation path corresponding to the first propagation probability from the plurality of propagation paths; and determining the first propagation path as the production bottleneck propagation path.

[0055] In the above embodiment, the time sequence correlation sequence comprises a causal relationship strength and an impact delay time between any two adjacent production nodes in the initial bottleneck node and the potential impact production nodes. The causal relationship strength is a fault propagation probability between any two adjacent production nodes. The impact delay time is a first shortest time interval required for fault propagation between any two adjacent production nodes. The propagation delay matrix represents a symmetric matrix of fault propagation time delay between the production nodes. The space adjacency matrix represents a topological matrix of the physical connection relationship between the production nodes. The preset propagation threshold is a probability threshold value preset for screening the production bottleneck propagation path. The diagonal element of the propagation delay matrix represents the node itself, which is generally set to zero. The non-diagonal element represents the fault propagation delay time between the nodes. If there is no direct or indirect physical connection between two production nodes, the corresponding element value of the space adjacency matrix is zero, representing that the two nodes will not have fault propagation influence due to physical connection.

[0056] In the above embodiment, it is assumed that a certain automobile parts manufacturer has multiple production lines for producing key parts such as engines and transmissions, each production line is composed of multiple production nodes, and there is a close production link between the production nodes. Specifically, through long-term collection and analysis of past production data, a historical fault mode library of each production node is constructed, recording the frequency of faults, fault types, and the impact on adjacent nodes after the fault occurs. The running state of each production node and the physical connection relationship between nodes are collected in real time by distributed sensing equipment. It is assumed that at a certain time, it is detected that the milling station (i.e. the initial bottleneck node) of the engine cylinder body processing production line has failed, and it is analyzed that the fault propagation weight of the milling station is relatively high. According to the historical fault data and the fault propagation weight obtained by real-time analysis, the potentially affected production nodes are determined, for example, the subsequent drilling station, cleaning station, etc. By analyzing the causal relationship of fault propagation in the history of each node, the causal relationship strength between adjacent production nodes is determined. For example, after the milling station fails, the probability of the drilling station being affected is 80%, so the causal relationship strength is 0.8. At the same time, the shortest time required for the milling station fault to propagate to the drilling station in history is counted to determine the impact delay time, which is assumed to be 30 minutes. Thus, a time sequence association sequence including the initial bottleneck node and the potentially affected production nodes is constructed. It should be noted that the above example of actual values is only an exemplary embodiment, and the above example of actual values is not limited to the above example.

[0057] In the above embodiment, according to the time coupling relationship, the time delay data of the fault propagation between each production node is sorted into a propagation delay matrix. For example, the fault propagation delay time from the milling station to the drilling station is 30 minutes, and the fault propagation delay time to the cleaning station is 60 minutes. The corresponding positions in the propagation delay matrix are filled with these fault propagation delay times. According to the spatial coupling relationship, the physical connection relationship between each production node is sorted out to establish a spatial adjacency matrix. If the milling station and the drilling station are directly connected through a conveyor belt, the physical connection strength is set to 1; if the milling station and the cleaning station need to be connected through a transfer device, the physical connection strength is set to 0.5. The propagation probability of multiple propagation paths of the milling station (i.e. the initial bottleneck node) is calculated through the propagation delay matrix and the spatial adjacency matrix. Assuming that the preset propagation threshold is 60%, the calculated propagation probability is compared with the preset propagation threshold, and the first propagation probability greater than the preset propagation threshold is selected. For example, the propagation probability of the milling station-drilling station-cleaning station path is 70%, which is greater than the preset propagation threshold, so this path is determined as the production bottleneck propagation path. Through the above implementation steps, the production bottleneck propagation path can be accurately determined, which provides a strong basis for subsequent targeted production adjustment strategy, maximally reduces the impact of the production bottleneck on the overall production, and improves the production efficiency and resource utilization. It should be noted that the above example of actual values is only an example embodiment, and the example of actual values is not limited to the above example.

[0058] In an optional embodiment, in the case where it is detected that the production line cluster needs to insert an emergency production task, the production bottleneck prediction signal, the production bottleneck propagation path, and the resource gap probability distribution are subjected to a conflict analysis operation on the emergency production task to generate a production task adjustment instruction set, specifically including: determining a time conflict area, a space conflict area, and a resource conflict area of the emergency production task and the current production task according to the production bottleneck prediction signal and the production bottleneck propagation path, wherein the time conflict area is an overlapping area of the required time period of the emergency production task and the first predicted time period and the second predicted time period, the space conflict area is a physical topology intersection area of the required equipment of the emergency production task and the equipment where the initial bottleneck node is located, and the resource conflict area is an intersection area of the resource type corresponding to the high confidence interval of the resource gap probability distribution. The target confidence interval includes the high confidence interval; determining an adjustment priority index of the emergency production task according to the time conflict area, the space conflict area, and the resource conflict area; generating a resource allocation scheme of the emergency production task according to the target confidence interval and the adjustment priority index; and performing feasibility verification on the resource allocation scheme according to the time conflict area, the space conflict area, and the resource conflict area to generate the production task adjustment instruction set.

[0059] In the above embodiment, the time conflict region is an overlapping region of the required time period of the emergency production task and the first predicted time period and the second predicted time period; the space conflict region is a physical topology intersection region of the required equipment of the emergency production task and the equipment where the initial bottleneck node is located; the resource conflict region is an intersection region of the required resource of the emergency production task and the resource type corresponding to the high confidence interval of the resource gap probability distribution; the adjustment priority index is used to measure the order of the adjustment of the emergency production task; and the resource allocation scheme is a resource allocation plan formulated for the emergency production task.

[0060] In the above embodiment, a large clothing manufacturing factory is taken as an example, and its production line cluster covers multiple links such as cutting, sewing, ironing, etc. During the peak sales season, the clothing manufacturing factory suddenly receives an emergency clothing order that needs to be delivered within a week, which requires the insertion of an emergency production task in the existing production task. Specifically, after detecting the need to insert the emergency production task, the sewing link is found to be the initial bottleneck node according to the production bottleneck prediction signal and the production bottleneck propagation path. Some equipment in the sewing link frequently fails due to aging, and the production capacity is insufficient in the next two weeks. Some fabric inventory is in the high confidence interval of the resource gap probability distribution. The required time period of the emergency production task overlaps with the insufficient production capacity period of the sewing link, forming a time conflict region. Further analysis shows that the emergency task required part of the sewing equipment shares the same material transmission track with the sewing link fault equipment. During the maintenance period of the fault equipment, the transmission track will be occupied, which will reduce the transmission efficiency of the track and affect the material supply of the emergency task equipment, forming a space conflict region. In addition, the task required part of the fabric is consistent with the fabric type corresponding to the high confidence interval of the resource gap probability distribution, forming a resource conflict region.

[0061] In the above embodiment, considering the urgent delivery time of the emergency order and the key influence of the sewing link on the overall production progress, the time conflict priority of the emergency production task is set to high, and the space and resource conflict priorities are set to medium, and the adjustment priority index is comprehensively obtained. In combination with the target confidence interval, a resource allocation scheme is formulated. For the time conflict, the production process of the sewing link is optimized, and the parallel operation mode is adopted to simultaneously carry out the originally serial part of the process. At the same time, the sewing equipment is reasonably scheduled to reduce the idle time of the equipment and improve the overall production efficiency. For the space conflict, on the one hand, the fault equipment is repaired in priority to ensure the smoothness of the transmission track, and on the other hand, the material feeding sequence of the emergency task equipment is adjusted to improve the use efficiency of the transmission track. In the face of resource conflict, the inventory fabric is preferentially allocated, and the procurement department is notified to speed up the material supply.

[0062] In the above embodiment, the feasibility of the resource allocation scheme is verified from the dimensions of time conflict, space conflict, resource conflict, etc. In the time conflict dimension, the sequence and time consumption of each sewing link are analyzed, and a detailed process network diagram is constructed. Using historical production data and a preset model, the production process after changing the serial process to parallel operation is simulated. For example, the collar sewing process and the sleeve sewing process are analyzed. In the serial mode, these two processes need to be performed sequentially, which takes a long time. In the parallel operation mode, the two processes can be carried out simultaneously. Through simulation, the estimated time for completing the sewing task of the emergency order after parallel operation is obtained. Real-time state information of each device is collected, including the idle time and running time of the device. The comprehensive utilization rate of the device under different task allocation schemes is determined by a preset model. The device utilization rate before and after optimization is compared to evaluate the improvement effect of device scheduling optimization on production efficiency, and to determine whether the task can be completed within the one-week delivery deadline required by the emergency order.

[0063] In the above embodiment, in the space conflict dimension, a space topology graph containing device layout and material transmission track is constructed, and the positions of the devices, the directions of the transmission tracks, and the connection relationships between the devices and the tracks are accurately recorded. Using a preset model, the occupation of the transmission track by the faulty device during maintenance is simulated according to the fault type and maintenance time of the faulty device. On the basis of the space topology graph, the transmission process after adjusting the material feeding sequence is simulated. According to the type and quantity of the materials and the capacity limit of the transmission track, the transmission time and congestion probability of the materials on the track are determined. By comparing the transmission efficiency indicators before and after adjustment, it is determined whether the priority repair of the faulty device and the adjustment of the material feeding sequence can guarantee the material supply of the emergency task device.

[0064] In the above embodiment, in the resource conflict dimension, the real-time fabric inventory data is obtained by connecting with the inventory management database. According to the material list of the emergency order, it is analyzed whether the types and quantities of the inventory fabrics meet the production requirements. Through a preset model, the length of time that can support the production of the emergency order under the existing inventory condition is predicted. The preset model is integrated with the supplier management system to obtain the production plan and logistics information of the supplier. According to factors such as the production capacity of the supplier, the transportation distance and mode, etc., the time for the supplier to deliver the short supply of fabrics is predicted. It is determined whether the supplier can complete the replenishment within the time required for the production of the emergency order to ensure the timely supply of resources required for production.

[0065] In an optional embodiment, the production task adjustment operation is performed on the production line cluster according to the production task adjustment instruction set, specifically including: determining the initial bottleneck node and the device scheduling parameter of the potentially affected production node according to the time rescheduling instruction, wherein the time rescheduling instruction is a time period optimization command generated according to the first prediction period and the second prediction period, the device scheduling parameter includes device enabling time, task execution time and buffer time interval, and the production task adjustment instruction set includes the time rescheduling instruction; reassigning the device runtime period of the initial bottleneck node and the potentially affected production node according to the device scheduling parameter and the time conflict area; determining the device layout parameter of the initial bottleneck node and the potentially affected production node according to the space optimization instruction, wherein the space optimization instruction is a space recombination command generated according to the physical topology intersection area, the device layout parameter includes device spacing, material transmission path and cooperative working radius, and the production task adjustment instruction set includes the space optimization instruction; reconstructing the device workflow topology of the initial bottleneck node and the potentially affected production node according to the device layout parameter and the space conflict area; determining the resource hierarchical scheduling strategy of the initial bottleneck node and the potentially affected production node according to the resource allocation instruction, wherein the resource allocation instruction is a resource dynamic allocation command generated according to the target confidence interval, the resource hierarchical scheduling strategy includes a high confidence interval scheduling strategy, a medium confidence interval scheduling strategy and a low confidence interval scheduling strategy, and the production task adjustment instruction set includes the resource allocation instruction; determining the risk buffer scheme of the initial bottleneck node and the potentially affected production node according to the emergency preparation instruction, wherein the emergency preparation instruction is a risk prevention and control command generated according to the production bottleneck propagation path, and the risk buffer scheme includes a standby device startup plan, an emergency material reserve scheme and an exception handling plan, and the production task adjustment instruction set includes the emergency preparation instruction.

[0066] In the above embodiment, the time rescheduling instruction is a time period optimization command generated according to the first prediction period and the second prediction period; the device scheduling parameter includes device enabling time, task execution time and buffer time interval; the space optimization instruction is a space recombination command generated according to the physical topology intersection area; the device layout parameter includes device spacing, material transmission path and cooperative working radius; the resource allocation instruction is a resource dynamic allocation command generated according to the target confidence interval; the resource hierarchical scheduling strategy includes a high confidence interval scheduling strategy, a medium confidence interval scheduling strategy and a low confidence interval scheduling strategy; the emergency preparation instruction is a risk prevention and control command generated according to the production bottleneck propagation path; and the risk buffer scheme includes a standby device startup plan, an emergency material reserve scheme and an exception handling plan.

[0067] In the above embodiments, the device scheduling parameters are a set of parameters (e.g., device activation time, task execution duration, buffer time interval, device priority, task switching time, device maintenance period, etc.) that plan the order of task execution and time allocation of devices in the production process. By reasonably setting this set of parameters, the optimization of production resources is realized, the utilization rate of devices is improved, the production cost is reduced, and the production plan is ensured to be completed on time. The device activation time refers to the time point at which the device starts to participate in the production task. This parameter needs to be determined according to the production plan, the preparation of the device, and the connection requirements of upstream and downstream processes, to ensure that the device is put into use at the right time and reduce waiting time and resource waste. The task execution duration refers to the time required for the device to complete a specific production task, which is determined by factors such as task process requirements, device running speed, and operator proficiency, and is an important basis for arranging production progress and evaluating production efficiency. The buffer time interval refers to the time reserved between the completion of a production task by the device and the start of the next task. This period is used to deal with uncertain factors in the production process, such as device failure repair, material shortage replenishment, product quality detection, etc., to ensure the continuity and stability of production. The device priority gives each device a priority level. When arranging production tasks, high-priority devices will be assigned tasks first. For example, in electronic product manufacturing, high-precision chip mounters play a key role in product quality and can be assigned the highest priority to ensure that core processes proceed smoothly. The task switching time is the time required for the device to switch from one task to another, including the time spent on device debugging, tooling fixture replacement, program switching, etc. By accurately calculating the task switching time, the task order can be reasonably arranged during scheduling to reduce device idling, such as arranging similar process tasks to reduce device switching costs in automobile parts processing. The device maintenance period is the time interval for regular maintenance of the device. By including the device maintenance period in the scheduling parameters, maintenance tasks can be planned in advance to avoid conflicts between maintenance and production tasks, ensuring stable operation of the device, such as chemical production equipment, which is arranged for maintenance according to the maintenance period to prevent device failure from causing production accidents.

[0068] In the above embodiments, the device layout parameters are used to describe the arrangement relationship of the devices in the production space, including but not limited to device spacing, material transportation path, collaborative working radius, passage width, operation space area, device orientation, etc. Reasonable device layout parameters help to optimize production process, shorten material transportation distance, improve space utilization, and promote collaboration between devices and personnel. Device spacing refers to the physical distance between production devices. When determining the device spacing, the size of the device, the operation space requirement, the safety specification, and the convenience of material handling should be considered to avoid interference between devices and ensure the safety of the operators. The material transportation path refers to the route that the material takes in the production process, from the raw material input to the finished product output, through each production link. Scientific planning of the material transportation path can reduce transportation time and cost, reduce material loss, and improve production efficiency. The collaborative working radius refers to the spatial range within which multiple devices or personnel can efficiently collaborate to complete production tasks. Reasonable determination of the collaborative working radius helps to improve team collaboration efficiency, reduce communication cost, and improve overall production performance. The passage width is the width of the passage in the production area, which should be determined according to the size of the material handling equipment, the frequency of passage, and the personnel flow demand. Sufficient passage width can ensure smooth material handling and avoid congestion, such as large logistics warehouses, where wide passages facilitate the quick passage of forklifts and other handling equipment. The operation space area is the size of the space provided for the operator to operate the equipment. Factors such as device operation method and personnel activity range should be considered. Reasonable operation space can improve the work efficiency of the operator and reduce the risk of misoperation, such as in the machining area of a machine tool, where sufficient space is provided for the operator to perform operations such as feeding, tool changing, etc. The device orientation refers to the direction in which the device is placed in the production space. Correct device orientation helps to optimize the material transportation path and personnel operation process. For example, in assembly production lines, consistent device orientation can make material transportation smoother and improve assembly efficiency.

[0069] In the above embodiment, it is assumed that a certain automobile parts processing workshop is responsible for producing engine cylinder blocks, crankshafts and other key components, and the workshop has various production equipment such as numerical control machine tools, machining centers, cleaning equipment, etc. During the production process, the workshop receives a batch of urgent orders and needs to complete the processing of specific parts within a short time, requiring adjustment of the existing production tasks. Specifically, time rescheduling instructions are generated according to the first prediction period and the second prediction period. After analysis, the machining center is in a state of capacity shortage in the future period, and there is a time conflict with the urgent production task. According to the device priority, the machining center is set as a high-priority device, and the urgent production task is preferentially allocated. Considering the task switching time, similar process tasks are arranged in a centralized manner to reduce device debugging and program switching time. At the same time, combined with the device maintenance period, the maintenance task of the machining center is avoided during the execution of the urgent production task, and device scheduling parameters such as device activation time, task execution time and buffer time interval are determined. According to the determined device scheduling parameters, the running period of the machining center and other related devices is redistributed. The processing time of non-urgent tasks in the machining center is reduced to create a time window for the urgent production task and ensure the timely completion of the urgent task, thereby alleviating the time conflict. According to the physical topology intersection region, a space optimization instruction is generated. It is found that the devices required by the urgent production task and the existing faulty devices are in the same work area and share the material transfer channel. Considering parameters such as channel width, operating space area and device orientation, the device layout is optimized. The channel width is appropriately widened to ensure smooth passage of material handling equipment, sufficient operating space is left for operators, the device orientation is adjusted, the material transfer path is optimized, and the material handling time is reduced.

[0070] In the above embodiments, according to the optimized equipment layout parameters, the equipment workflow topology is reconstructed. The material transmission route between the equipment is re-planned to ensure that the materials can be efficiently and orderly transferred between the equipment, and the impact of space conflicts on production is reduced. Resource allocation instructions are generated according to the target confidence interval. According to the analysis, part of the raw material inventory is in the low confidence interval, which may affect the emergency production task. A resource hierarchical scheduling strategy is developed, and resources in the high confidence interval are preferentially allocated to the emergency production task. For resources in the low confidence interval, on the one hand, the supplier is coordinated to speed up the replenishment, and on the other hand, the production process is reasonably adjusted to reduce the dependence on such resources. Emergency preparation instructions are generated according to the production bottleneck propagation path. Considering the risks that may occur in the production process, such as equipment failure and raw material supply interruption, a risk buffer scheme is developed. A backup machining center is prepared to quickly switch production when the main machining center fails; an emergency material storage warehouse is established to ensure that production can be maintained for a certain period of time when the raw material supply is interrupted; and an abnormal handling plan is developed to clearly define the handling process and responsibility division in the event of production abnormalities. Through the implementation of the above steps, the workshop can effectively respond to emergency production tasks, complete order delivery on time under the premise of ensuring production quality, and improve the production management level and response ability of the workshop.

[0071] It should be noted that the above described embodiments are only part of the embodiments of the present application, not all. The present application will be specifically described below in conjunction with specific embodiments.

[0072] The embodiment of the present application provides an intelligent production management system based on a large model, which includes but is not limited to a central processor, a data collection module, a model analysis module and a plan adjustment module. The data collection module is responsible for collecting real-time data of the production line and transmitting it to the central processor. The model analysis module integrates a pre-trained large-scale machine learning model (corresponding to the above-mentioned preset model) for analyzing data and predicting potential problems in the production process. The plan adjustment module automatically adjusts the production plan according to the model analysis result and issues it to each production unit. It should be noted that the generation and training process of the preset model is prior art, which will not be described here.

[0073] Through the embodiment of the present application, the real-time production data of the production line cluster is collected, input into the preset model for analysis, and the production bottleneck prediction signal and resource gap probability distribution can be obtained. According to the prediction signal, the production bottleneck propagation path is determined, and when an emergency production task needs to be inserted, a task adjustment instruction set is generated based on multi-aspect information for conflict analysis and production task adjustment. Thus, real-time and comprehensive monitoring and management of the production line cluster can be realized, production risks can be predicted in advance, and when an emergency task is inserted, production can be adjusted through scientific analysis to reduce production conflicts, optimize production processes, greatly improve the production efficiency and resource utilization rate of the production line cluster, and ensure smooth production.

[0074] The electronic device in the embodiments of the present application will be described below from the perspective of hardware processing, with reference to Figure 2 , Figure 2 FIG. 1 is a schematic diagram of an entity device structure of the electronic device in the embodiments of the present application.

[0075] It should be noted that Figure 2 The structure of the electronic device shown is only an example and should not bring any limitation to the functions and use range of the embodiments of the present application.

[0076] As Figure 2 shown, the electronic device includes a central processing unit (CPU) 201, which can perform various appropriate actions and processes, such as executing the methods described in the above embodiments, according to programs stored in a read-only memory (ROM) 202 or loaded from a storage portion 208 into a random access memory (RAM) 203. In the RAM 203, various programs and data required for system operation are also stored.

[0077] The CPU 201, the ROM 202, and the RAM 203 are connected to each other through a bus 204. An input / output (I / O) interface 205 is also connected to the bus 204.

[0078] The following components are connected to the I / O interface 205: an input portion 206 including an audio input device, a button switch, and the like; an output portion 207 including a liquid crystal display (LCD), an audio output device, an indicator, and the like; a storage portion 208 including a hard disk and the like; and a communication portion 209 including a network interface card such as a LAN (Local Area Network) card, a modem, and the like. The communication portion 209 performs communication processing via a network such as the Internet. A drive 210 is also connected to the I / O interface 205 as needed. A removable medium 211 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, and the like is mounted on the drive 210 as needed, so that a computer program read therefrom is installed into the storage portion 208 as needed.

[0079] In particular, the processes described above with reference to the flowcharts can be implemented as a computer software program in accordance with embodiments of the present application. For example, embodiments of the present application include a computer program product comprising a computer program carried on a computer readable medium, the computer program comprising computer programs for executing the methods illustrated by the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network by the communication section 209, and / or installed from the detachable medium 211. When the computer program is executed by the central processing unit (CPU) 201, various functions defined in the present application are executed.

[0080] Note that specific examples of the computer readable storage medium can include, but are not limited to, an electric connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read only memory (ROM), an erasable programmable read only memory (EPROM), a flash memory, an optical fiber, a portable compact disc read only memory (CD-ROM), an optical storage device, a magnetic storage device, or any appropriate combination of the above. In the present application, the computer readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

[0081] The flowcharts and block diagrams in the drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowcharts or block diagrams can represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures.

[0082] In particular, the electronic device of the present embodiment includes a processor and a memory, and the memory stores a computer program, and the computer program is executed by the processor to implement the intelligent production management method based on a large model provided by the above-mentioned embodiments.

[0083] As another aspect, the present application also provides a computer-readable storage medium, which can be included in the electronic device described in the above embodiments, or can exist independently without being assembled into the electronic device. The storage medium carries one or more computer programs, which, when executed by a processor of the electronic device, enable the electronic device to implement the large model-based intelligent production management method provided in the above embodiments.

[0084] The above-described embodiments are merely used to illustrate the technical solutions of the present application, but not to limit the present application; even though the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some of the technical features can be replaced by equivalent features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0085] Those skilled in the art can understand that all or part of the processes in the above embodiments can be implemented by a computer program to instruct relevant hardware to complete, and the program can be stored in a computer-readable storage medium. When the program is executed, the processes of each method embodiment described above can be included. The foregoing storage medium includes ROM or random storage memory RAM, magnetic disk or optical disk, and various program code storage media.

Claims

1. A smart production management method based on a large model, characterized in that, include: When real-time production data of the current production task of the production line cluster is collected, the real-time production data is input into a preset model to control the preset model to perform joint analysis operations on the real-time production data; If the preset model has completed the joint analysis operation, the production bottleneck prediction signal and resource gap probability distribution output by the preset model are obtained. The production bottleneck prediction signal is the risk intensity of insufficient capacity of the first production node in the production line cluster during the first prediction period. The first prediction period is the bottleneck time range for insufficient capacity of the first production node predicted by the preset model. The resource gap probability distribution is the probability distribution of multi-dimensional resource shortage events of the second production node in the production line cluster during the second prediction period. The second prediction period is the emergency time range for multi-dimensional resource shortage events of the second production node predicted by the preset model. The production bottleneck propagation path is determined based on the production bottleneck prediction signal, wherein the production bottleneck propagation path is a multi-level influence path in the production line cluster that is triggered by an initial bottleneck node, and the initial bottleneck node includes the first production node and the second production node. If it is detected that the production line cluster needs to insert an emergency production task, a conflict analysis operation is performed on the emergency production task based on the production bottleneck prediction signal, the production bottleneck propagation path, and the resource gap probability distribution to generate a production task adjustment instruction set. The production task adjustment operation is performed on the production line cluster according to the production task adjustment instruction set.

2. The method according to claim 1, characterized in that, When real-time production data of the current production task of the production line cluster is collected, the real-time production data is input into a preset model to control the preset model to perform joint analysis operations on the real-time production data, specifically including: The real-time production data includes the changes in equipment status, material flow trajectory, order demand sequence, and external market fluctuation parameters of each production node in the production line cluster, obtained in real time through distributed sensing devices. The real-time production data includes the changes in equipment status, the material flow trajectory, the order demand sequence, and the external market fluctuation parameters. Each production node includes the initial bottleneck node. The changes in equipment status, the material flow trajectory, the order demand sequence, and the external market fluctuation parameters are subjected to spatiotemporal feature fusion processing to generate a multimodal dataset, wherein the multimodal dataset is a multi-dimensional data set with a unified time base and spatial identifier; The multimodal dataset is input into the preset model to control the preset model to perform the following operations: Upon determining that the multimodal dataset has been received, the preset model performs anomaly detection analysis on the multimodal dataset to obtain anomaly detection results. The preset model performs anomaly correlation analysis on the historical failure mode library and the anomaly detection results to generate a risk event correlation rule library; The preset model performs time-series cross-validation analysis on the equipment state changes and the order demand sequence to determine the time-series coupling relationship between the equipment state changes and the order demand sequence in the time-series dimension; The preset model performs spatial topology mapping analysis on the equipment state changes and the order demand sequence to determine the spatial coupling relationship between the equipment state changes and the order demand sequence in the spatial dimension; The preset model performs confidence analysis on the production baseline data and the external market fluctuation parameters to determine the target confidence interval of the multi-dimensional resource shortage event. The production baseline data is a pre-established reference dataset for maintaining the resource supply and demand balance of the production line cluster. The preset model generates the production bottleneck prediction signal and the resource gap probability distribution based on the risk event association rule base, the temporal coupling relationship, the spatial coupling relationship, and the target confidence interval. The preset model outputs the production bottleneck prediction signal and the resource gap probability distribution.

3. The method according to claim 2, characterized in that, The process of fusing spatiotemporal features of the equipment status changes, the material flow trajectory, the order demand sequence, and the external market fluctuation parameters to generate a multimodal dataset specifically includes: The equipment status changes, material flow trajectories, order demand sequences, and external market fluctuation parameters are timestamped to generate a raw data stream with a unified time base. Assign the spatial identifier to the original data stream, wherein the spatial identifier is used to identify the physical coordinates of the production equipment corresponding to the equipment state change and the topological relationship between the storage location and the transportation path corresponding to the material flow trajectory; Extract the time series features of the equipment status changes, the spatial distribution features of the material flow trajectory, the demand fluctuation features of the order demand sequence, and the market trend features of the external market fluctuation parameters from the raw data stream that has been assigned the spatial identifier; The time series features, spatial distribution features, demand fluctuation features, and market trend features are fused to generate the multimodal dataset.

4. The method according to claim 2, characterized in that, The preset model generates the production bottleneck prediction signal and the resource gap probability distribution based on the risk event association rule base, the temporal coupling relationship, the spatial coupling relationship, and the target confidence interval, specifically including: The preset model performs fault mode matching between the anomaly detection results and the risk event association rule base to determine the degree of association between the anomaly detection results and the fault modes in the historical fault mode base. The preset model determines the fault propagation weight of each production node based on the degree of correlation; The preset model constructs a dynamic coupling model of the production line cluster based on the temporal coupling relationship and the spatial coupling relationship; The preset model inputs the fault propagation weights into the dynamic coupling model to obtain the capacity deviation of the first production node in the first forecast period and the supply-demand gap of the second production node in the second forecast period after the dynamic coupling model performs forward propagation analysis on the fault propagation weights. The capacity deviation is the deviation between the actual capacity and the planned capacity of the first production node in the first forecast period, and the supply-demand gap is the difference between the real-time inventory and the demand of the multi-dimensional resources of the second production node in the second forecast period. The preset model dynamically maps the production capacity deviation to the target confidence interval to generate the production bottleneck prediction signal. The preset model performs multi-level probability calibration between the supply and demand gap and the target confidence interval to generate the resource gap probability distribution.

5. The method according to claim 4, characterized in that, The step of determining the production bottleneck propagation path based on the production bottleneck prediction signal specifically includes: Based on the initial bottleneck node and the fault propagation weight, a temporal correlation sequence that may affect production nodes is determined from the production line cluster. The temporal correlation sequence includes the causal relationship strength and the impact delay time between any two adjacent production nodes among the initial bottleneck node and the potentially affected production nodes. The causal relationship strength is the fault propagation probability between any two adjacent production nodes, and the impact delay time is the first shortest time interval required for fault propagation between any two adjacent production nodes. The production bottleneck prediction signal includes the initial bottleneck node. A propagation delay matrix in the temporal dimension is established based on the temporal coupling relationship, and a spatial adjacency matrix in the spatial dimension is established based on the spatial coupling relationship. The propagation delay matrix is ​​a symmetric square matrix representing the fault propagation time delay between each production node. The off-diagonal elements of the propagation delay matrix are the second shortest time interval required for fault propagation between each production node, and the second shortest time interval includes the first shortest time interval. The spatial adjacency matrix represents the topology matrix of the physical connection relationship between each production node, and the non-zero elements of the spatial adjacency matrix are the physical connection strength between each production node. Based on the propagation delay matrix and the spatial adjacency matrix, determine multiple propagation probabilities for multiple propagation paths of the initial bottleneck node; The multiple propagation probabilities are compared with a preset propagation threshold to obtain the probability comparison result; Based on the probability comparison results, a first propagation probability greater than the preset propagation threshold is determined from the plurality of propagation probabilities; From the plurality of propagation paths, determine the first propagation path corresponding to the first propagation probability; The first propagation path is determined as the production bottleneck propagation path.

6. The method according to any one of claims 1-5, characterized in that, When an emergency production task is detected to need to be inserted into the production line cluster, a conflict analysis operation is performed on the emergency production task based on the production bottleneck prediction signal, the production bottleneck propagation path, and the resource gap probability distribution to generate a production task adjustment instruction set, specifically including: Based on the production bottleneck prediction signal and the production bottleneck propagation path, the time conflict region, spatial conflict region, and resource conflict region between the emergency production task and the current production task are determined. The time conflict region is the overlapping region between the time period required by the emergency production task and the first and second predicted time periods. The spatial conflict region is the physical topology intersection region between the equipment required by the emergency production task and the equipment where the initial bottleneck node is located. The resource conflict region is the intersection region between the resources required by the emergency production task and the resource types corresponding to the high confidence interval of the resource gap probability distribution. The target confidence interval includes the high confidence interval. The adjustment priority index for the emergency production task is determined based on the time conflict area, the spatial conflict area, and the resource conflict area. A resource allocation plan for the emergency production task is generated based on the target confidence interval and the adjustment priority index. The feasibility of the resource allocation scheme is verified based on the time conflict region, the spatial conflict region, and the resource conflict region to generate the production task adjustment instruction set.

7. The method according to claim 6, characterized in that, The step of performing production task adjustment operations on the production line cluster according to the production task adjustment instruction set specifically includes: The equipment scheduling parameters of the initial bottleneck node and the potentially affected production node are determined according to the time rescheduling instruction. The time rescheduling instruction is a time period optimization command generated based on the first prediction period and the second prediction period. The equipment scheduling parameters include the equipment start time, task execution duration and buffer time interval. The production task adjustment instruction set includes the time rescheduling instruction. Based on the equipment scheduling parameters and the time conflict area, the equipment runtime segments of the initial bottleneck node and the potentially impactful production node are reallocated; The equipment layout parameters of the initial bottleneck node and the potentially affected production node are determined according to the space optimization instruction. The space optimization instruction is a space reorganization command generated based on the physical topology intersection area. The equipment layout parameters include equipment spacing, material transport path and collaborative working radius. The production task adjustment instruction set includes the space optimization instruction. Reconstruct the device workflow topology of the initial bottleneck node and the potentially impactful production node based on the device layout parameters and the spatial conflict area; The resource allocation instructions determine the resource hierarchical scheduling strategy for the initial bottleneck node and the potentially affected production node. The resource allocation instructions are dynamic resource allocation commands generated based on the target confidence interval. The resource hierarchical scheduling strategy includes a high confidence interval scheduling strategy, a medium confidence interval scheduling strategy, and a low confidence interval scheduling strategy. The production task adjustment instruction set includes the resource allocation instructions. The initial bottleneck node and the risk buffer scheme that may affect the production node are determined according to the emergency preparedness instructions. The emergency preparedness instructions are risk control commands generated according to the propagation path of the production bottleneck. The risk buffer scheme includes a backup equipment activation plan, an emergency material reserve plan, and an anomaly handling plan. The production task adjustment instruction set includes the emergency preparedness instructions.

8. An electronic device, characterized in that, The electronic device includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the electronic device to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that, When the instructions are executed on an electronic device, the electronic device causes the electronic device to perform the method as described in any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product is run on an electronic device, it causes the electronic device to perform the method as described in any one of claims 1-7.

Images