A Modeling Method and System for Industrial Chain Maps Based on Deep Learning

Abstract

The present invention provides a method and system for modeling an industrial chain map based on deep learning, which relates to the technical field of industrial chain analysis and modeling. The method includes: collecting multi-source heterogeneous data in the industrial chain, including demand information, resource status, process characteristics, and operation and maintenance data, and preprocessing the multi-source heterogeneous data to form a standardized data set; performing semantic vectorization processing on the text-type data in the standardized data set, and performing node embedding on the multi-relationship graph to generate a low-dimensional representation of entities that integrates text and structural features. By integrating multi-source heterogeneous data, fusing text and structural features, constructing an industrial chain knowledge graph, and dynamically optimizing, the present invention realizes the precise adjustment of production line configuration and task allocation strategies. At the same time, combined with the risk propagation model, it effectively predicts industrial trends, analyzes risk paths, and identifies resource bottlenecks, providing a scientific basis for industrial chain decision-making and improving the operation efficiency and stability of the industrial chain.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial chain analysis and modeling, and particularly to a method and system for modeling an industrial chain map based on deep learning. Background Art

[0002] Most traditional methods for modeling industrial chain maps highly rely on expert experience and manual analysis, which results in strong subjectivity in the model construction process and is difficult to comprehensively and objectively reflect the actual situation of the industrial chain.

[0003] For example, in the automotive manufacturing industrial chain, traditional methods may only rely on the intuitive judgments and experience summaries of industry experts for links such as component suppliers, vehicle manufacturers, and distributors to construct the industrial chain map. However, this method based on expert experience may overlook some emerging enterprises or potential important suppliers, resulting in an incomplete or biased map.

[0004] Secondly, with the rapid development of information technology, the amount of data related to the industrial chain has increased explosively. Traditional methods for modeling industrial chain maps are inefficient in processing large-scale data and are difficult to adapt to the rapidly changing market environment. In the e-commerce industrial chain, there are massive amounts of transaction data, user behavior data, logistics data, etc. Traditional methods may use manual sorting or simple data analysis tools to process this data, resulting in slow processing speed and low accuracy. For example, during shopping festivals such as "Double Eleven", the transaction volume surges, and traditional methods are difficult to update the industrial chain map in real time to reflect the latest market conditions, which may cause enterprises to miss market opportunities or make wrong decisions.

[0005] In addition, in the intelligent manufacturing industrial chain, technologies such as equipment networking and real-time data acquisition have caused a sharp increase in the amount of data. Traditional methods may face problems such as storage difficulties and insufficient computing resources when processing this data. For example, an intelligent manufacturing enterprise hopes to construct a comprehensive industrial chain map including equipment status, production processes, supply chain information, etc., but traditional methods are difficult to efficiently process such a large amount of data, resulting in a long map construction cycle and untimely updates, which affects the enterprise's ability to monitor the production process in real time and make optimized decisions. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method and system for modeling an industrial chain map based on deep learning to improve the intelligent and automated level of industrial chain management.

[0007] To solve the above technical problem, the technical solution of the present invention is as follows:

[0008] In the first aspect, a method for modeling an industrial chain map based on deep learning, the method includes:

[0009] Collect multi-source heterogeneous data in the acquisition industrial chain, including demand information, resource status, process characteristics, and operation and maintenance data, and preprocess the multi-source heterogeneous data to form a standardized data set;

[0010] Perform semantic vectorization processing on the text data in the standardized data set, and perform node embedding on the multi-relationship graph to generate a low-dimensional representation of entities that integrates text and structural features;

[0011] Construct an industrial chain ontology based on the low-dimensional representation of entities, including structured data, semi-structured text, and unstructured building information models, and construct an industrial chain knowledge graph based on the industrial chain ontology;

[0012] According to the industrial chain knowledge graph, through a two-stage deep Q-network, dynamically optimize the production line configuration and task allocation strategy, and update the optimization results to the node attributes and relationship weights of the knowledge graph to generate an optimized dynamic industrial chain knowledge graph;

[0013] According to the optimized dynamic industrial chain knowledge graph, calculate the node importance weights through a graph attention network, and combine a risk propagation model to conduct industrial trend prediction, risk propagation path analysis, and resource bottleneck identification, and finally generate decision-making suggestions.

[0014] Further, perform semantic vectorization processing on the text data in the standardized data set, and perform node embedding on the multi-relationship graph to generate a low-dimensional representation of entities that integrates text and structural features, including:

[0015] Perform word segmentation and part-of-speech tagging on the text data in the standardized data set, and extract the preliminary semantic features of the text, including word frequency vectors and co-occurrence matrices; extract entities from the standardized data set, including enterprises, equipment, processes, and association relationships, including supply-demand relationships and collaboration relationships, and construct the initial structure of the multi-relationship graph;

[0016] Map the word frequency vectors and co-occurrence matrices to the corresponding nodes of the multi-relationship graph, and add preliminary semantic representations to each node;

[0017] Calculate the structural features of each node in the multi-relationship graph, including degree centrality and local clustering coefficient features, and splice and combine the structural features and semantic features to form a composite feature vector containing text and structural information;

[0018] Perform a linear projection operation on the composite feature vector, and perform a non-linear transformation on the projected features to generate a low-dimensional representation vector of the entity.

[0019] Further, construct an industrial chain ontology based on the low-dimensional representation of entities, including structured data, semi-structured text, and unstructured building information models, and construct an industrial chain knowledge graph based on the industrial chain ontology, including:

[0020] Define the ontology model of the industrial chain based on the low-dimensional representation of entities, including structured data, semi-structured text, and unstructured building information models;

[0021] Map the entities, relationships, and attributes in the ontology model to the nodes, edges, and attribute fields of the graph database to generate an industrial chain knowledge graph containing nodes, edges, and relationship weights.

[0022] Furthermore, in the industrial chain knowledge graph, nodes represent entities in the industrial chain, including enterprises, equipment, and processes, and edges represent the relationships between entities, including supply and demand, collaboration, and dependence.

[0023] Furthermore, according to the industrial chain knowledge graph, through a two-stage deep Q-network, dynamically optimize the production line configuration and task allocation strategy, and update the optimization results to the node attributes and relationship weights of the knowledge graph to generate an optimized dynamic industrial chain knowledge graph, including:

[0024] According to the current state of the industrial chain knowledge graph, use the two-stage deep Q-network to simulate the combination of production line configuration adjustment and task allocation strategy;

[0025] Use the two-stage deep Q-network to dynamically adjust and optimize the production line configuration and task allocation strategy in the knowledge graph environment, calculate the target Q-value and update the parameters during the iterative learning process to determine the final production line configuration and task allocation strategy combination;

[0026] According to the final production line configuration and task allocation strategy combination, update the attributes of the equipment nodes in the knowledge graph;

[0027] After the attributes of the equipment nodes are updated, re-evaluate and adjust the relationship weights between the nodes in the knowledge graph according to the new task allocation strategy;

[0028] Integrate all updated node attributes and relationship weights to generate an optimized dynamic industrial chain knowledge graph.

[0029] Furthermore, use the two-stage deep Q-network to dynamically adjust and optimize the production line configuration and task allocation strategy in the knowledge graph environment, calculate the target Q-value and update the parameters during the iterative learning process to determine the final production line configuration and task allocation strategy combination, including:

[0030] Calculate the difference ratio of the efficiency between the new state and the old state, the cost difference between the old state and the new state, and the difference ratio of the resource utilization rate between the new state and the old state to obtain the immediate reward;

[0031] Predict the next state based on the current state, and select the action that maximizes the predicted Q-value of the current Q-network as the final action in the corresponding predicted state;

[0032] Input the future state and the final action into the target Q-network to obtain the Q-value prediction of the future state;

[0033] Fuse the immediate reward and the Q-value prediction of the future state to obtain the target Q-value.

[0034] Furthermore, according to the optimized dynamic industrial chain knowledge graph, calculate the node importance weights through the graph attention network, and combine with the risk propagation model for industrial trend prediction, risk propagation path analysis and resource bottleneck identification, and finally generate decision-making suggestions, including:

[0035] Extract node features, edge features and topological structure information from the optimized dynamic industrial chain knowledge graph;

[0036] Use the graph attention network to process the node features, edge features and topological structure information, automatically learn the association strength between nodes, and assign an importance weight to each node;

[0037] Combine the node importance weights with the predefined risk propagation model to evaluate the risk propagation path and potential impact in the industrial chain, and obtain the risk propagation analysis result;

[0038] According to the risk propagation analysis result, predict the future development trend of the industrial chain, including identifying key nodes and potential risk points in the industrial chain, and obtain the industrial trend prediction result;

[0039] Analyze the node importance weights, risk propagation model and industrial trend prediction result, identify key risk propagation nodes and paths, as well as resource bottleneck nodes and potential resource shortage problems in the industrial chain, and generate decision-making suggestions, including adjustment of production line configuration, optimization of task allocation, formulation of risk prevention and control measures, planning of resource allocation and scheduling.

[0040] In the second aspect, a deep learning-based industrial chain map modeling system includes:

[0041] A data processing module for collecting multi-source heterogeneous data from the industrial chain, including demand information, resource status, process characteristics and operation and maintenance data, and preprocessing the multi-source heterogeneous data to form a standardized data set;

[0042] A knowledge fusion module for performing semantic vectorization processing on the text data in the standardized data set and performing node embedding on the multi-relationship graph to generate a low-dimensional representation of entities that integrates text and structural features;

[0043] A map construction module for constructing an industrial chain knowledge graph using the low-dimensional representation of entities that integrates text and structural features, where nodes represent different entities in the industrial chain and edges represent relationships between entities;

[0044] A reinforcement learning module, which is used to dynamically optimize the production line configuration and task assignment strategy through a two-stage deep Q-network, and generate an optimized dynamic industrial chain knowledge graph;

[0045] A decision support module, which is used to perform industrial trend prediction, risk propagation path analysis and resource bottleneck identification based on the optimized industrial chain knowledge graph, and finally generate decision-making suggestions.

[0046] In a third aspect, a computing device includes:

[0047] One or more processors;

[0048] A storage device for storing one or more programs, which when executed by the one or more processors cause the one or more processors to implement the method.

[0049] In a fourth aspect, a computer-readable storage medium stores a program, and when the program is executed by a processor, the method is implemented.

[0050] The above solution of the present invention has at least the following beneficial effects:

[0051] By collecting multi-source heterogeneous data such as demand information, resource status, process characteristics, and operation and maintenance data in the industrial chain, and performing preprocessing to form a standardized data set, it can comprehensively reflect the operation status and characteristics of the industrial chain. This helps to eliminate data islands and improve the availability and consistency of data. Perform semantic vectorization processing on the text data in the standardized data set, and perform node embedding on the multi-relationship graph to generate a low-dimensional representation of entities that integrates text and structural features. This method can make full use of the information in text data and structural data, capture the complex relationships between entities, and improve the accuracy and richness of entity representation.

[0052] Construct an industrial chain ontology based on the low-dimensional representation of entities, including structured data, semi-structured text, and unstructured building information models, and further construct an industrial chain knowledge graph. This helps to formally represent the entities, relationships, and attributes in the industrial chain, form a computable and inferable knowledge system, and provide strong support for the analysis and optimization of the industrial chain. Dynamically optimize the production line configuration and task assignment strategy through a two-stage deep Q-network, and update the optimization results to the node attributes and relationship weights of the knowledge graph to generate an optimized dynamic industrial chain knowledge graph. This method can respond to changes in the industrial chain in real time, improve production efficiency and resource utilization rate, reduce production costs, and enhance the competitiveness of the industrial chain.

[0053] Based on the optimized dynamic industrial chain knowledge graph, calculate the node importance weights through the graph attention network, and combine with the risk propagation model for industrial trend prediction, risk propagation path analysis, and resource bottleneck identification. This helps to discover potential problems and risks in the industrial chain in advance, provide support for decision-making, and avoid or reduce losses. The final decision-making suggestions are based on comprehensive data analysis and in-depth model reasoning, and are scientific and targeted. These suggestions can guide enterprises in the industrial chain to carry out reasonable production line configuration, task allocation, risk prevention and control, and resource scheduling, improving the overall operation efficiency and stability of the industrial chain. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 is a schematic flowchart of a method for modeling an industrial chain graph based on deep learning provided by an embodiment of the present invention.

[0055] Figure 2 is a schematic diagram of a system for modeling an industrial chain graph based on deep learning provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] Hereinafter, exemplary embodiments of the present disclosure will be described in more detail with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0057] As Figure 1 shown, an embodiment of the present invention proposes a method for modeling an industrial chain graph based on deep learning, and the method includes the following steps:

[0058] Step 1, collect multi-source heterogeneous data in the industrial chain, including demand information, resource status, process characteristics, and operation and maintenance data, and preprocess the multi-source heterogeneous data to form a standardized data set;

[0059] Step 2, perform semantic vectorization processing on the text data in the standardized data set, and perform node embedding on the multi-relationship graph to generate a low-dimensional representation of entities that integrates text and structural features;

[0060] Step 3, construct an industrial chain ontology based on the low-dimensional representation of entities, including structured data, semi-structured text, and unstructured building information models, and construct an industrial chain knowledge graph based on the industrial chain ontology;

[0061] Step 4, based on the industrial chain knowledge graph, dynamically optimize the production line configuration and task allocation strategies through a two-stage deep Q-network, and update the optimization results to the node attributes and relationship weights of the knowledge graph to generate an optimized dynamic industrial chain knowledge graph;

[0062] Step 5: According to the optimized dynamic industrial chain knowledge graph, calculate the node importance weights through the graph attention network, and combine with the risk propagation model to conduct industrial trend prediction, risk propagation path analysis, and resource bottleneck identification, and finally generate decision-making suggestions.

[0063] In the embodiment of the present invention, by collecting multi-source heterogeneous data such as demand information, resource status, process characteristics, and operation and maintenance data in the industrial chain, the key information of each link in the industrial chain can be comprehensively covered. Preprocessing the multi-source heterogeneous data to form a standardized data set helps to eliminate the format differences and noises between the data, and improve the data quality and consistency.

[0064] Step 2: Perform semantic vectorization processing on the text data in the standardized data set, which can convert text information into a vector form that can be understood by a computer and capture the semantic information in the text. Embed the nodes of the multi-relationship graph to generate a low-dimensional representation of the entity that fuses text and structural features, which can consider both the text description of the entity and its structural position in the graph, and improve the accuracy and comprehensiveness of the entity representation.

[0065] Step 3: Construct an industrial chain ontology based on the low-dimensional representation of the entity, including structured data, semi-structured text, and unstructured building information models, which can clearly define the entities, attributes, and relationships in the industrial chain, and provide a clear framework and semantic basis for the construction of the knowledge graph. Construct an industrial chain knowledge graph according to the industrial chain ontology, which can present the entities and relationships in the industrial chain in a graphical way, and intuitively display the structure and association relationships of the industrial chain.

[0066] Step 4: Dynamically optimize the production line configuration and task allocation strategy through a two-stage deep Q network, which can adjust the production line configuration and task allocation in real time according to the current production status and demand, and improve the production efficiency and resource utilization rate. Update the optimization results to the node attributes and relationship weights of the knowledge graph to generate an optimized dynamic industrial chain knowledge graph, which can reflect the latest status and changes of the industrial chain in real time.

[0067] Step 5: Calculate the node importance weights through the graph attention network, which can accurately evaluate the importance of each node in the industrial chain and identify key nodes and potential risk points. Combine with the risk propagation model to conduct industrial trend prediction, risk propagation path analysis, and resource bottleneck identification, which can discover potential risks and bottlenecks in the industrial chain in advance, provide scientific decision-making suggestions for decision-makers, and help enterprises formulate effective risk prevention and control and resource allocation strategies. The finally generated decision-making suggestions can provide specific action guidelines for enterprises, help enterprises optimize the industrial chain structure, improve production efficiency, reduce risks, and achieve sustainable development.

[0068] In a preferred embodiment of the present invention, in step 1 above, multi-source heterogeneous data in the industrial chain is collected, including demand information, resource status, process characteristics, and operation and maintenance data, and the multi-source heterogeneous data is preprocessed to form a standardized data set, which may include:

[0069] In an embodiment of the present invention, demand information of customers, suppliers, partners, etc. is collected through means such as questionnaires and interviews. Historical sales data is analyzed to predict future demand trends, and macro demand information is obtained by referring to industry research reports, market analysis articles, etc. Resource status data is collected in real time through devices such as sensors and RFID, and resource status information is extracted from the ERP system. For data that cannot be automatically collected, it is entered into the system manually. Operation and maintenance data is collected in real time through the equipment monitoring system, historical maintenance records are analyzed, key operation and maintenance information is extracted, and user feedback on the equipment usage situation is collected as a supplement to the operation and maintenance data.

[0070] Duplicate data is removed using methods such as hash algorithms and sorting comparisons, and missing values are processed using methods such as filling, deletion, or interpolation according to the data characteristics. Outliers are detected and processed through statistical methods, machine learning algorithms, etc. Data in different formats (such as text, pictures, videos, etc.) is converted into a unified format. Character encodings, date formats, etc. are converted into standard encodings and formats, and data in different units is unified into the same unit.

[0071] The data is scaled to the interval [0, 1] or [-1, 1]. The data is converted into a standard normal distribution with a mean of 0 and a variance of 1. The cleaned, transformed, normalized / standardized data is integrated into a unified data set, and information such as labels and annotations is added to the data set.

[0072] In a preferred embodiment of the present invention, in step 2 above, semantic vectorization processing is performed on the text data in the standardized data set, and node embedding is performed on the multi-relationship graph to generate a low-dimensional representation of entities that integrates text and structural features, which may include:

[0073] Step 210, word segmentation and part-of-speech tagging are performed on the text data in the standardized data set to extract the preliminary semantic features of the text, including word frequency vectors and co-occurrence matrices; entities are extracted from the standardized data set, including enterprises, equipment, processes, and association relationships, including supply-demand relationships and collaboration relationships, to construct the initial structure of the multi-relationship graph;

[0074] Step 211, the word frequency vector and the co-occurrence matrix are mapped to the corresponding nodes of the multi-relationship graph, and preliminary semantic representations are added to each node;

[0075] Step 212, calculate the structural features of each node in the multi-relational graph, including degree centrality and local clustering coefficient features, and splice and combine the structural features with the semantic features to form a composite feature vector containing text and structural information;

[0076] Step 213, perform a linear projection operation on the composite feature vector, and perform a non-linear transformation on the projected features to generate a low-dimensional representation vector of the entity.

[0077] In the embodiments of the present invention, a word segmentation tool (such as Jieba, NLTK, etc.) is used to segment the text data in the standardized dataset into individual words. Part-of-speech tagging is performed on the segmented words (such as nouns, verbs, adjectives, etc.) to identify the grammatical roles of the words in the sentence. The frequency of each word in the text is counted to form a word frequency vector, which reflects the importance of the word in the text, and the co-occurrence times of the words are calculated to form a co-occurrence matrix, which reflects the correlation between the words.

[0078] Entity extraction: Extract entities from the standardized dataset, such as enterprises, equipment, processes, etc.

[0079] Relationship extraction: Identify the association relationships between entities, such as supply-demand relationships, cooperation relationships, etc.

[0080] Construct an initial structure: Use entities as nodes and relationships as edges to construct the initial structure of the multi-relational graph.

[0081] Step 211, for each node in the multi-relational graph, map the corresponding word frequency vector and co-occurrence matrix information to the node according to the occurrence of the entity represented by the node in the text. Take the values in the word frequency vector and co-occurrence matrix as the attributes of the node, or integrate the information in multiple texts into a single node in some way (such as weighted average). After mapping, each node contains the text semantic information related to it, forming an initial semantic representation of the node.

[0082] Step 212, calculate the degree of each node in the multi-relational graph (i.e., the number of edges connected to the node), which reflects the importance of the node in the graph. Calculate the local clustering coefficient of each node, which reflects the closeness between the node and its neighbor nodes. Splice the structural features (degree centrality, local clustering coefficient) of the node with the semantic features (information after mapping of the word frequency vector and co-occurrence matrix) to form a composite feature vector containing text and structural information.

[0083] Step 213, perform a linear projection on the composite feature vector to map it from a high-dimensional space to a low-dimensional space, which can be achieved through matrix multiplication, where the projection matrix can be obtained through training. Perform a non-linear transformation (such as using an activation function) on the projected features to increase the expressive power of the model. After the linear projection and non-linear transformation, a low-dimensional representation vector of the entity is obtained, which integrates text and structural features.

[0084] Suppose there is a dataset on the automotive manufacturing industry chain, which includes text data (such as enterprise news, process descriptions, etc.) and structured data (such as supply and demand relationships, collaboration relationships, etc. between enterprises).

[0085] Perform word segmentation and part-of-speech tagging on the news text, and extract the word frequency vector and co-occurrence matrix. For example, for the news "A certain automotive enterprise reaches a cooperation agreement with a supplier", after word segmentation, it is "A certain automotive enterprise reaches a cooperation agreement with a supplier", and calculate the word frequency and co-occurrence relationship.

[0086] Extract entities: automotive enterprise, supplier.

[0087] Extract relationships: supply and demand relationship (automotive enterprise - supplier);

[0088] Construct an initial graph structure.

[0089] Map the word frequency and co-occurrence information in the news to the "automotive enterprise" and "supplier" nodes in the graph. Calculate the degree centrality and local clustering coefficient of the nodes, splice the structural features and semantic features to form a composite feature vector. Perform a linear projection and non-linear transformation on the composite feature vector to obtain the low-dimensional representation vectors of the "automotive enterprise" and "supplier".

[0090] By integrating text and structural features, the generated low-dimensional representation vector can more comprehensively reflect the characteristics and association relationships of entities, improve the expressive power of the model. The low-dimensional representation vector reduces the dimension of the data, reduces the risk of overfitting, and enhances the generalization ability of the model. The low-dimensional representation vector is convenient for subsequent machine learning tasks such as similarity calculation, clustering analysis, and classification, providing strong support for the analysis and decision-making of the industry chain. Through dimensionality reduction processing, the storage and calculation costs of data are reduced, and the analysis efficiency is improved. Integrating text and structural features helps to discover potential association relationships between entities, providing new ideas for the optimization and upgrading of the industry chain.

[0091] In a preferred embodiment of the present invention, in the above step 3, constructing an industry chain ontology based on the low-dimensional representation of entities, including structured data, semi-structured text, and unstructured building information models, and constructing an industry chain knowledge graph based on the industry chain ontology may include:

[0092] Step 310, define the ontology model of the industrial chain according to the entity low-dimensional representation, including structured data, semi-structured text, and unstructured building information models;

[0093] Step 311, map the entities, relationships, and attributes in the ontology model to the nodes, edges, and attribute fields of the graph database to generate an industrial chain knowledge graph containing nodes, edges, and relationship weights.

[0094] In the embodiment of the present invention, read and analyze the entity low-dimensional representation vectors generated in the previous steps. These vectors are obtained by performing semantic vectorization processing on text data and node embedding on the multi-relationship graph, and integrate text and structural features. According to the entity types and their associated relationships revealed by the entity low-dimensional representation, start constructing the ontology model framework of the industrial chain. This framework needs to cover structured data (such as enterprise information, transaction records, product specifications, etc.), semi-structured text (such as process descriptions, technical document summaries, market analysis reports, etc.), and unstructured building information models (such as 3D factory layouts, equipment models, process flowcharts, etc.).

[0095] Within the ontology model framework, clearly define various entities (such as enterprises, equipment, processes, raw materials, etc.), the relationships between them (such as supply-demand relationships, production cooperation relationships, technology dependence relationships, etc.), and the attribute fields of each entity and relationship (such as enterprise names, equipment models, process parameters, transaction amounts, etc.). By comparing the entity low-dimensional representation with the definitions of the ontology model, verify the accuracy and integrity of the model. If inconsistencies or omissions are found, the machine will automatically adjust the definitions of the ontology model to ensure that the model can accurately reflect the actual structure and characteristics of the industrial chain.

[0096] Step 311, select a suitable graph database (such as Neo4j, ArangoDB, JanusGraph, etc.) to store and query the industrial chain knowledge graph. The graph database can efficiently process the relationships between nodes and edges and is suitable for storing and querying complex knowledge graphs. Map each entity in the ontology model to a node in the graph database. Each node contains all the attribute fields of the entity for storing detailed information about the entity. Map each relationship in the ontology model to an edge in the graph database. Each edge connects two nodes, representing the relationship between them. At the same time, the edge also contains relationship type and weight information for describing the nature and tightness of the relationship. The relationship weight can be set according to the association strength or similarity information in the entity low-dimensional representation. After the mapping of entities to nodes and relationships to edges, the machine generates an industrial chain knowledge graph containing nodes, edges, and relationship weights, which graphically shows the entities and relationships in the industrial chain.

[0097] Suppose there is a knowledge graph construction task for the automotive manufacturing industrial chain.

[0098] Based on the low-dimensional representation of entities, an ontology model including entities such as automobile manufacturing enterprises, parts suppliers, production equipment, and technological processes is defined. At the same time, relationship types such as supply-demand relationships, production collaboration relationships, and technology dependence relationships are defined, as well as attribute fields such as enterprise names, equipment models, process parameters, and transaction amounts. Neo4j is selected as the graph database, and the entities in the ontology model are mapped to nodes and the relationships are mapped to edges. For example, "a certain automobile manufacturing enterprise" is mapped to a node, which includes attributes such as enterprise name, address, and contact information; "a certain parts supplier" is mapped to another node, which includes attributes such as supplier name, product specifications, and price; "the supply-demand relationship between a certain automobile manufacturing enterprise and a certain parts supplier" is mapped to an edge, which includes relationship type and weight information (such as transaction frequency, transaction amount, etc.). After mapping, an automobile manufacturing industry chain knowledge graph containing multiple nodes and edges is generated. This graph shows the entities and relationships in the automobile manufacturing industry chain in a graphical way.

[0099] By constructing the industry chain ontology and knowledge graph, it is possible to integrate and manage various types of data such as structured data, semi-structured text, and unstructured building information models, realizing the comprehensive integration and unified management of data. The industry chain knowledge graph shows the entities and relationships in the industry chain in a graphical way, making the data more intuitive and easy to understand. At the same time, the efficient query and traversal functions of the graph database can quickly locate and analyze the key entities and relationships in the industry chain, improving the efficiency of data analysis. The industry chain knowledge graph can reveal the potential association relationships and rules between entities, providing new ideas and directions for the optimization and upgrading of the industry chain. For example, by analyzing the supply-demand relationship and production collaboration relationship, the bottleneck links and potential risk points in the industry chain can be found. Based on the analysis results of the industry chain knowledge graph, scientific decision-making suggestions and support can be provided for decision-makers. At the same time, by mining and analyzing historical data, the future development trends and potential risks of the industry chain can also be predicted, providing strong support for the strategic planning and risk management of enterprises. The construction of the industry chain knowledge graph helps to promote the collaboration and cooperation between upstream and downstream enterprises in the industry chain, and promotes technological innovation and industrial upgrading. For example, by sharing the information and resources in the knowledge graph, enterprises can carry out R & D cooperation and market expansion activities more efficiently.

[0100] In another preferred embodiment of the present invention, in the industry chain knowledge graph, the nodes represent the entities in the industry chain, including enterprises, equipment, and processes, and the edges represent the relationships between entities, including supply-demand, collaboration, and dependence, and may include:

[0101] In the embodiments of the present invention, the nodes in the industrial chain knowledge graph represent entities in the industrial chain, including enterprises, equipment, processes, etc. Each node contains detailed attribute information, such as enterprise name, address, contact information, etc.; equipment model, specifications, manufacturer, etc.; process name, parameters, steps, etc.

[0102] Edges represent: Edges represent the relationships between entities, including supply-demand relationships (such as the relationship between an enterprise and its supplier), collaboration relationships (such as the cooperation or alliance relationship between enterprises), dependency relationships (such as the dependency of a process on equipment or raw materials), etc. Each edge contains relationship type and weight information, which is used to describe the nature and tightness of the relationship.

[0103] In a preferred embodiment of the present invention, in step 4 above, according to the industrial chain knowledge graph, through a two-stage deep Q-network, the production line configuration and task allocation strategy are dynamically optimized, and the optimization results are updated to the node attributes and relationship weights of the knowledge graph to generate an optimized dynamic industrial chain knowledge graph, which may include:

[0104] Step 410, according to the current state of the industrial chain knowledge graph, use the two-stage deep Q-network to simulate the combination of production line configuration adjustment and task allocation strategy;

[0105] Step 411, use the two-stage deep Q-network to dynamically adjust and optimize the production line configuration and task allocation strategy in the knowledge graph environment, calculate the target Q value and update the parameters during the iterative learning process to determine the final combination of production line configuration and task allocation strategy;

[0106] Step 412, according to the final combination of production line configuration and task allocation strategy, update the attributes of the equipment nodes in the knowledge graph;

[0107] Step 413, after the attributes of the equipment nodes are updated, re-evaluate and adjust the relationship weights between the nodes in the knowledge graph according to the new task allocation strategy;

[0108] Step 414, integrate all the updated node attributes and relationship weights to generate an optimized dynamic industrial chain knowledge graph.

[0109] In the embodiments of the present invention, the current state is read from the industrial chain knowledge graph, including the attributes of device nodes (such as device type, production capacity, maintenance status, etc.), the attributes of task nodes (such as task type, task volume, priority, etc.), and the relationship weights between nodes (such as supply-demand relationship, collaboration relationship, etc.). Initialize the model parameters of the Double Deep Q-Network (Double DQN), including the weights of the Q-network and the target Q-network. The Double Deep Q-Network reduces the overestimation problem of Q-values by introducing two Q-networks, improving the stability of learning. Using the Double Deep Q-Network, according to the current knowledge graph state, simulate different combinations of production line configuration adjustments and task allocation strategies. Each strategy combination corresponds to a Q-value, representing the expected return of the strategy combination.

[0110] Step 411, execute the current strategy combination in the knowledge graph environment, observe the environmental feedback (such as production line production efficiency, task completion time, etc.), and calculate the immediate reward according to the feedback. Select the next action (i.e., production line configuration adjustment or task allocation strategy) according to the ε-greedy strategy to balance exploration and exploitation.

[0111] Calculate the target Q-value of the next state using the target Q-network, and update the parameters of the Q-network by the gradient descent method according to the difference between the current Q-value and the target Q-value. In the iterative learning process, the machine continuously repeats the above steps, gradually optimizing the production line configuration and task allocation strategies. After multiple iterations, determine the final combination of production line configuration and task allocation strategies, which has the highest expected return.

[0112] Step 412, parse the final combination of production line configuration and task allocation strategies, and extract the configuration information related to device nodes (such as device enable / disable status, device production parameter adjustment, etc.). According to the parsing result, update the attributes of device nodes in the knowledge graph. For example, update the production status of a certain device from "idle" to "enabled", and adjust its production parameters to meet the new task requirements.

[0113] Step 413, analyze the new task allocation strategy, identify the affected node pairs (such as device and task nodes, device and device nodes, etc.). According to the new task allocation strategy, re-evaluate the relationship weights between the affected nodes. For example, if a certain device is assigned more tasks, the supply-demand relationship weight between it and the relevant task nodes may increase. According to the re-evaluation result, adjust the relationship weights between each node in the knowledge graph to reflect the impact of the new task allocation strategy on the industrial chain structure.

[0114] Step 414: Integrate all the updated device node attributes, task node attributes, and the relationship weights between nodes to form complete knowledge graph data. Using the integrated data, generate an optimized dynamic industrial chain knowledge graph. This graph reflects the latest production line configuration and task allocation strategies, as well as their impact on the industrial chain structure.

[0115] Suppose there is an optimization task for an intelligent manufacturing production line.

[0116] Read the current state from the industrial chain knowledge graph, including the production capacity and maintenance status of each device, as well as the current task allocation situation. Use the two-stage deep Q-network to simulate different combinations of production line configuration adjustments and task allocation strategies, and optimize the strategy combinations through iterative learning. For example, it may be found that enabling an idle device and assigning it to a high-priority task can improve the overall production efficiency. According to the optimized strategy combination, update the attributes of the device nodes in the knowledge graph. For example, change the status of a device from "idle" to "enabled" and adjust its production parameters. Analyze the new task allocation strategy and re-evaluate the relationship weights between the affected nodes. For example, if a device is assigned more tasks, the supply-demand relationship weight between it and the related task nodes increases. Integrate all the updated node attributes and relationship weights to generate an optimized dynamic industrial chain knowledge graph. This graph reflects the latest production line configuration and task allocation strategies, as well as their impact on the intelligent manufacturing production line structure.

[0117] Dynamically optimizing the production line configuration and task allocation strategies through the two-stage deep Q-network can significantly improve the production efficiency and task completion speed of the production line. The optimized industrial chain knowledge graph reflects the latest production line configuration and task allocation strategies, which helps to enhance the collaboration and flexibility between upstream and downstream enterprises in the industrial chain and improve the overall response speed. Based on the optimized knowledge graph, scientific decision-making suggestions and support can be provided for decision-makers. At the same time, by mining and analyzing historical data, the machine can also predict the future development trends and potential risks of the industrial chain. The construction of the dynamic industrial chain knowledge graph provides strong support for the continuous innovation and optimization of the industrial chain. Enterprises can continuously adjust and optimize the production line configuration and task allocation strategies according to the information and suggestions in the knowledge graph to adapt to market changes and customer needs. The optimized knowledge graph visually displays the entities and relationships in the industrial chain in a graphical way, making the data more intuitive and easy to understand. At the same time, the efficient query and traversal functions of the graph database can quickly locate and analyze the key entities and relationships in the industrial chain, improving the efficiency of data analysis.

[0118] In another preferred embodiment of the present invention, in step 411 above, a two-stage deep Q-network is used to dynamically adjust and optimize the production line configuration and task allocation strategy in the knowledge graph environment, calculate the target Q-value and update the parameters during the iterative learning process to determine the final combination of production line configuration and task allocation strategy, including:

[0119] Calculate the difference ratio of the efficiency between the new state and the old state, the cost difference between the old state and the new state, and the difference ratio of the resource utilization rate between the new state and the old state to obtain the immediate reward;

[0120] Predict the next state based on the current state, and select the action that maximizes the predicted Q-value of the current Q-network as the final action for the corresponding predicted state;

[0121] Input the future state and the final action into the target Q-network to obtain the Q-value prediction of the future state;

[0122] Fuse the immediate reward and the Q-value prediction of the future state to obtain the target Q-value.

[0123] In the embodiment of the present invention, initialize the parameters of the Q-network and the target Q-network and , is a delayed copy of. Set the discount factor = 0.9, representing the current value of future rewards, set the weight coefficient = 1.0, = 1.5, = 1.2, respectively representing the importance of efficiency, cost, and resource utilization rate in reward calculation. For each state transition, calculate the immediate reward ; where, and are the efficiencies in the new state and the old state respectively, and are the costs in the new state and the old state respectively, and are the resource utilization rates in the new state and the old state respectively. Use the current Q-network to predict the Q-value of the next state and select the action that maximizes the Q-value, that is, . Input the future state and the selected final action into the target Q-network to obtain the Q-value prediction of the future state, where, is the Q-value predicted by the current Q-network (whose parameters are and action ) given the state .

[0124] Fuse the immediate reward and the Q-value prediction of the future state to obtain the target Q-value, i.e., target Q-value = .

[0125] Use the error between the target Q-value and the predicted value of the current Q-network to update the parameters of the Q-network , using the gradient descent method. Every certain number of steps, copy the parameters of the Q-network to the parameters of the target Q-network to maintain the stability of the target Q-network. By calculating the immediate reward and future reward in real time, the two-stage deep Q-network can dynamically adjust the production line configuration and task allocation strategy to adapt to the changing production environment. By optimizing the reward calculation of efficiency, cost, and resource utilization, the system can tend to select actions that can improve production efficiency and resource utilization, thereby reducing production costs. The two-stage deep Q-network reduces the bias of Q-value estimation by introducing the target Q-network, improving the scientificity and accuracy of decision-making. At the same time, the use of the discount factor enables the system to balance the importance of immediate reward and future reward and make more long-term decisions. By continuously iterating and learning and updating parameters, it can gradually adapt to different production environments and task requirements, improving the adaptability and robustness of the system.

[0126] In a preferred embodiment of the present invention, in step 5 above, according to the optimized dynamic industrial chain knowledge graph, calculate the node importance weights through the graph attention network, and combine with the risk propagation model for industrial trend prediction, risk propagation path analysis, and resource bottleneck identification, and finally generate decision-making suggestions, which may include:

[0127] Step 510, extract node features, edge features, and topological structure information from the optimized dynamic industrial chain knowledge graph;

[0128] Step 511, use the graph attention network to process the node features, edge features, and topological structure information, automatically learn the association strength between nodes, and assign an importance weight to each node;

[0129] Step 512, combine the node importance weights with the predefined risk propagation model to evaluate the risk propagation path and potential impact in the industrial chain, and obtain the risk propagation analysis result;

[0130] Step 513, according to the risk propagation analysis result, predict the future development trend of the industrial chain, including identifying key nodes and potential risk points in the industrial chain, and obtain the industrial trend prediction result;

[0131] Step 514: Analyze the node importance weights, risk propagation model, and industrial trend prediction results to identify key risk propagation nodes and paths, as well as resource bottleneck nodes and potential resource shortage problems in the industrial chain, and generate decision-making suggestions, including adjustments to production line configuration, optimization of task allocation, formulation of risk prevention and control measures, and planning of resource allocation and scheduling.

[0132] In the embodiment of the present invention, each node in the knowledge graph is traversed, and attribute information related to the node is extracted, such as the type of the node (enterprise, product, technology, etc.), the scale of the node (such as the number of employees and output value of the enterprise), and the status of the node (such as the operation status of the enterprise and the market demand of the product). For each edge in the knowledge graph, information such as the type of the edge (such as supply relationship, cooperation relationship, etc.) and the weight of the edge (representing the strength or frequency of the relationship) is extracted, and the connection relationship between the nodes is recorded, and an adjacency matrix is constructed to represent the topological structure of the knowledge graph.

[0133] Step 511: Construct a Graph Attention Network (GAT) model with node features and edge features as inputs. Through the self-attention mechanism of the GAT model, automatically learn the association strength between nodes, that is, the importance contribution of each node to its neighbor nodes. According to the learned association strength, assign an importance weight to each node, indicating the relative importance of the node in the knowledge graph.

[0134] Step 512: Define a risk propagation model, such as a variant based on the SIR (Susceptible-Infected-Recovered) model or the SEIR (Susceptible-Exposed-Infected-Recovered) model, considering the impact of node importance weights on risk propagation. Input the node importance weights into the risk propagation model to simulate the risk propagation process in the industrial chain, and record the infection status and propagation path of each node. According to the simulation results, evaluate the potential impact of the risk on each node in the industrial chain, such as production interruption, demand decline, etc.

[0135] Step 513: According to the risk propagation analysis results, identify the key nodes that have the greatest impact on the overall stability of the industrial chain. Analyze the risk propagation paths and potential impacts to identify potential risk points in the industrial chain, such as supply chain interruption, market demand changes, etc. Combining the analysis of key nodes and potential risk points, predict the future development trend of the industrial chain, such as the market demand for certain products will increase, and certain technologies will be gradually phased out, etc.

[0136] Step 514: Analyze the risk propagation analysis results, identify the key risk propagation nodes and paths, and combine the node importance weights and industrial trend prediction results to identify the resource bottleneck nodes in the industrial chain, that is, those nodes that have the greatest impact on the overall operation efficiency of the industrial chain. Analyze the resource requirements and supply situations of the resource bottleneck nodes to identify potential resource shortage problems. Based on the analysis of risk propagation nodes and paths, resource bottleneck nodes, and potential resource shortage problems, generate decision-making suggestions, such as adjusting production line configurations, optimizing task assignments, formulating risk prevention and control measures, planning resource allocation and scheduling, etc.

[0137] Suppose there is a dynamic knowledge graph of the automotive industrial chain, which includes nodes such as automobile manufacturers, parts suppliers, and distributors, as well as edges such as supply relationships and cooperation relationships between them. Extract the characteristics of each node (such as enterprise scale, product type, etc.), the characteristics of the edges (such as the intensity of the supply relationship), and topological structure information (such as the adjacency matrix). Use the graph attention network to process this information and assign an importance weight to each node. For example, large automobile manufacturers may have a higher importance weight.

[0138] Combine the node importance weights with the risk propagation model to simulate the propagation paths and potential impacts of risks (such as supply chain disruptions) in the industrial chain. For example, if there is a problem with a certain key parts supplier, it may lead to production interruptions for multiple automobile manufacturers. According to the risk propagation analysis results, predict the future development trends of the automotive industrial chain. For example, predict that the market demand for some new energy vehicles will increase, while the market demand for some traditional fuel vehicles will decline. Analyze the node importance weights, risk propagation model, and industrial trend prediction results to identify the key risk propagation nodes and paths (such as some key parts suppliers), resource bottleneck nodes (such as some manufacturers with limited production capacity), and potential resource shortage problems (such as the insufficient supply of some key parts). Then generate decision-making suggestions, such as increasing the inventory of key parts, optimizing production line configurations to improve production capacity, etc.

[0139] By combining the graph attention network and the risk propagation model, it is possible to more accurately evaluate the importance of nodes and risk propagation paths, thereby generating more scientific decision-making suggestions. By identifying the key risk propagation nodes and paths, it is possible to formulate risk prevention and control measures in advance to reduce the impact of risks on the industrial chain. By identifying resource bottleneck nodes and potential resource shortage problems, it is possible to more reasonably plan resource allocation and scheduling and improve resource utilization efficiency. By predicting industrial trends and identifying key nodes, it is possible to guide enterprises in the industrial chain to strengthen cooperation, achieve coordinated development, and improve the competitiveness of the entire industrial chain.

[0140] As Figure 2 shown, an embodiment of the present invention also provides a deep learning-based industrial chain graph modeling system, including:

[0141] A data processing module, configured to collect multi-source heterogeneous data from the industrial chain, including demand information, resource status, process characteristics, and operation and maintenance data, and preprocess the multi-source heterogeneous data to form a standardized data set;

[0142] A knowledge fusion module, configured to perform semantic vectorization processing on the text data in the standardized data set, and perform node embedding on the multi-relationship graphs to generate a low-dimensional representation of entities that integrates text and structural features;

[0143] A graph construction module, configured to construct an industrial chain knowledge graph by using the low-dimensional representation of entities that integrates text and structural features, where nodes represent different entities in the industrial chain, and edges represent the relationships between entities;

[0144] A reinforcement learning module, configured to dynamically optimize the production line configuration and task allocation strategy through a two-stage deep Q network, and generate an optimized dynamic industrial chain knowledge graph;

[0145] A decision support module, configured to perform industrial trend prediction, risk propagation path analysis, and resource bottleneck identification based on the optimized industrial chain knowledge graph, and finally generate decision-making suggestions.

[0146] It should be noted that this system corresponds to the above method. All implementation manners in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0147] An embodiment of the present invention further provides a computing device, including: a processor and a memory storing a computer program. When the computer program is run by the processor, it executes the method as described above. All implementation manners in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0148] An embodiment of the present invention further provides a computer-readable storage medium storing instructions. When the instructions are run on a computer, the computer is caused to execute the method as described above. All implementation manners in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0149] The above is the preferred implementation manner of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A method for modeling an industrial chain map based on deep learning, characterized in that The method includes: Collecting multi-source heterogeneous data in the industrial chain, including demand information, resource status, process characteristics, and operation and maintenance data, and preprocessing the multi-source heterogeneous data to form a standardized data set; Performing semantic vectorization processing on the text-type data in the standardized data set, and performing node embedding on the multi-relationship graph to generate a low-dimensional representation of entities that integrates text and structural features; Constructing an industrial chain ontology based on the low-dimensional representation of entities, including structured data, semi-structured text, and unstructured building information models, and constructing an industrial chain knowledge graph based on the industrial chain ontology. In the industrial chain knowledge graph, nodes represent entities in the industrial chain, including enterprises, equipment, and processes, and edges represent the relationships between entities, including supply and demand, collaboration, and dependence; According to the industrial chain knowledge graph, through a two-stage deep Q-network, dynamically optimize the production line configuration and task allocation strategy, and update the optimization result to the node attributes and relationship weights of the knowledge graph to generate an optimized dynamic industrial chain knowledge graph, including: According to the current state of the industrial chain knowledge graph, use a two-stage deep Q-network to simulate the combination of production line configuration adjustment and task allocation strategy; Use a two-stage deep Q-network to dynamically adjust and optimize the production line configuration and task allocation strategy in the knowledge graph environment, calculate the target Q value and update the parameters during the iterative learning process to determine the final production line configuration and task allocation strategy combination; According to the final production line configuration and task allocation strategy combination, update the attributes of the equipment nodes in the knowledge graph; After the attributes of the equipment nodes are updated, re-evaluate and adjust the relationship weights between the nodes in the knowledge graph according to the new task allocation strategy; Integrate all updated node attributes and relationship weights to generate an optimized dynamic industrial chain knowledge graph; According to the optimized dynamic industrial chain knowledge graph, calculate the node importance weights through a graph attention network, and combine with a risk propagation model to conduct industrial trend prediction, risk propagation path analysis, and resource bottleneck identification, and finally generate decision-making suggestions.

2. The method for modeling an industrial chain map based on deep learning according to claim 1, wherein, Performing semantic vectorization processing on the text-type data in the standardized data set, and performing node embedding on the multi-relationship graph to generate a low-dimensional representation of entities that integrates text and structural features, including: Performing word segmentation and part-of-speech tagging on the text-type data in the standardized data set, and extracting the preliminary semantic features of the text, including word frequency vectors and co-occurrence matrices; Extracting entities from the standardized data set, including enterprises, equipment, processes, and association relationships, including supply and demand relationships and collaboration relationships, to construct the initial structure of the multi-relationship graph; Mapping the word frequency vectors and co-occurrence matrices to the corresponding nodes of the multi-relationship graph, and adding preliminary semantic representations to each node; Calculating the structural features of each node in the multi-relationship graph, including degree centrality and local clustering coefficient features, and splicing and combining the structural features and semantic features to form a composite feature vector containing text and structural information; Performing a linear projection operation on the composite feature vector, and performing a non-linear transformation on the projected features to generate a low-dimensional representation vector of the entity.

3. The method for modeling an industrial chain map based on deep learning according to claim 2, wherein, Construct an industrial chain ontology based on the low-dimensional representation of entities, including structured data, semi-structured text, and unstructured building information models, and construct an industrial chain knowledge graph based on the industrial chain ontology, including: Define the ontology model of the industrial chain according to the low-dimensional representation of entities, including structured data, semi-structured text, and unstructured building information models; Map the entities, relationships, and attributes in the ontology model to the nodes, edges, and attribute fields of the graph database Generate an industrial chain knowledge graph containing nodes, edges, and relationship weights.

4. The method for modeling an industrial chain map based on deep learning according to claim 3, wherein, Use the two-stage deep Q-network to dynamically adjust and optimize the production line configuration and task allocation strategy in the knowledge graph environment, calculate the target Q value and update the parameters during the iterative learning process to determine the final combination of production line configuration and task allocation strategy, including: Calculate the difference ratio of the efficiency between the new state and the old state, the cost difference between the old state and the new state, and the difference ratio of the resource utilization rate between the new state and the old state to obtain the immediate reward; Predict the next state based on the current state and select the action that maximizes the predicted Q value of the current Q-network as the final action under the corresponding predicted state; Input the future state and the final action into the target Q-network to obtain the Q value prediction of the future state; Fuse the immediate reward and the Q value prediction of the future state to obtain the target Q value.

5. The method for modeling an industrial chain map based on deep learning according to claim 4, wherein According to the optimized dynamic industrial chain knowledge graph, calculate the node importance weights through the graph attention network, and combine the risk propagation model for industrial trend prediction, risk propagation path analysis, and resource bottleneck identification, and finally generate decision-making suggestions, including: Extract node features, edge features, and topological structure information from the optimized dynamic industrial chain knowledge graph; Use the graph attention network to process the node features, edge features, and topological structure information, automatically learn the association strength between nodes, and assign an importance weight to each node; Combine the node importance weights with the predefined risk propagation model to evaluate the propagation path and potential impact of risks in the industrial chain to obtain the risk propagation analysis results; Predict the future development trend of the industrial chain based on the risk propagation analysis results, including identifying the key nodes and potential risk points in the industrial chain to obtain the industrial trend prediction results; Analyze the node importance weights, risk propagation model, and industrial trend prediction results to identify the key risk propagation nodes and paths, as well as the resource bottleneck nodes and potential resource shortage problems in the industrial chain, and generate decision-making suggestions, including the adjustment of production line configuration, the optimization of task allocation, the formulation of risk prevention and control measures, and the planning of resource allocation and scheduling.

6. A deep learning-based industrial chain map modeling system, which implements the method described in any one of claims 1 to 5, characterized in that, Including: A data processing module for collecting multi-source heterogeneous data from the industrial chain, including demand information, resource status, process characteristics, and operation and maintenance data, and preprocessing the multi-source heterogeneous data to form a standardized data set; A knowledge fusion module for performing semantic vectorization processing on the text data in the standardized data set and performing node embedding on the multi-relationship graph to generate a low-dimensional representation of entities that combines text and structural features; A spectrum construction module for constructing an industrial chain knowledge graph by using the low-dimensional representation of entities that fuses text and structural features, where nodes represent different entities in the industrial chain and edges represent the relationships between entities; A reinforcement learning module for dynamically optimizing the production line configuration and task assignment strategy through a two-stage deep Q-network to generate an optimized dynamic industrial chain knowledge graph; A decision support module for predicting industrial trends, analyzing risk propagation paths, and identifying resource bottlenecks based on the optimized industrial chain knowledge graph, and finally generating decision-making suggestions.

7. A computing device, characterized in that, Including: One or more processors; A storage device for storing one or more programs, which when executed by the one or more processors cause the one or more processors to implement the method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, A program is stored in the computer-readable storage medium, and when the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images