A Multi-Agent Data Spatial Interoperability Method for Multi-Source Heterogeneous Data

By collecting and standardizing the business logic and semantic associations of multi-source heterogeneous data, a multi-subject data space interoperability model is constructed, which solves the problem of insufficient data exchange capabilities in multi-source heterogeneous data sharing and interoperability, and improves the data sharing and interoperability capabilities across departments and organizations.

CN117931913BActive Publication Date: 2026-11-17NORTHWESTERN POLYTECHNICAL UNIV
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202410021916.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-01-08
Publication Date
2026-11-17
Estimated Expiration
2044-01-08

AI Technical Summary

Technical Problem

In the process of sharing and interoperating multi-source heterogeneous data, the lack of specific and feasible implementation plans for standardized data semantic publishing leads to insufficient data sharing capabilities across departments and organizations. Furthermore, the lack of unified professional terminology, unified data formats, and a unified metadata framework results in weak data exchange capabilities.

Method used

By employing real-time and batch data collection methods, we can mine the business logic and semantic relationships of multi-source heterogeneous data, standardize data semantic publishing, and construct a multi-subject data space interoperability model, including subject modules, global modules, data models, control centers, and information exchange centers. We can also use national/industry standards and business scenarios to build a unified professional terminology and metadata framework to ensure data consistency and security.

Benefits of technology

It enables cross-departmental and cross-organizational sharing of multi-source heterogeneous data, ensures the accuracy and security of data exchange, enhances the data interoperability capabilities among multiple entities, provides a unified platform and communication mechanism, and improves the efficiency of data sharing and exchange.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117931913B_ABST
    Figure CN117931913B_ABST
Patent Text Reader

Abstract

To address the shortcomings of existing methods for collecting, sharing, and interoperating multi-source heterogeneous data, such as insufficient sharing capabilities and weak data exchange capabilities during interoperation, this invention proposes a multi-entity data space interoperability method for multi-source heterogeneous data. This invention collects and mines the business logic and semantic relationships within multi-source heterogeneous data from different enterprises and customers, and achieves data sharing through standardized semantic publishing. Then, the shared data is extracted, transformed, and loaded into a unified target system, followed by cleaning, dimensionality reduction, and transformation. Next, based on data concept models, national / industry standards, specific business scenarios, and metadata sources, a unified framework of professional terminology, data formats, and metadata standards is constructed to form a multi-entity data space interoperability model. Finally, the data in this interoperability model is exported and queried. This invention enhances the data interoperability capabilities between entities, providing data support for business operation models.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to a multi-agent data space interoperability method for multi-source heterogeneous data. Background Technology

[0002] With the rapid development of new-generation information technology, emerging cloud computing, big data, the Internet of Things, artificial intelligence, and other technologies, data resources within and between manufacturing / supply chain network nodes are growing exponentially. The resulting massive amounts of business data are characterized by multi-source heterogeneity, low utilization, and difficulty in effective integration.

[0003] Data generated by internal and inter-enterprise business systems encompasses various fields such as sales, marketing, customer relationship management, supply chain, and human resources. It includes both structured data (e.g., relational databases) and unstructured data (e.g., documents, emails, logs), and commonly suffers from data quality issues (e.g., missing data, erroneous data, duplicate data). Because internal and inter-enterprise business systems operate independently, data is often stored in different formats, structures, and naming conventions across various databases, file systems, or applications. The lack of unified terminology, data formats, and metadata frameworks weakens data exchange capabilities and hinders data sharing among multiple entities within and across enterprises during data interoperability.

[0004] AsiaInfo Blue Ocean (Jiangsu) Data Technology Co., Ltd.'s invention patent "A Deep Learning-Based Method for Processing Unstructured Data" (201710044292.2) is based on the Hadoop big data framework. It solves the problems of storing, managing, and accessing unstructured data, and preprocesses unstructured data to achieve storage of unstructured data and enhance the efficiency and accuracy of unstructured data processing.

[0005] The invention patent "A Semantic Modeling Method for Relational Data" (201711428542.9) published by the Shenyang Institute of Automation, Chinese Academy of Sciences, involves a semantic modeling method for relational data. Based on multi-dimensional heterogeneous data sources, it extracts data information from multiple data sources, obtains database and data table information, and establishes a semantic data model based on multi-dimensional relational database data and rules. This provides underlying support for summarizing and analyzing data from multiple data sources and improves the interoperability of resources.

[0006] The invention patent "An Analysis Model of Drilling Speed-Up Influencing Factors Based on Big Data" (201910037016.2) published by Changqing Drilling Company of China National Petroleum Corporation Sichuan Drilling Engineering Co., Ltd. realizes data integration of distributed multi-source heterogeneous data sources based on Kettle; and builds retrieval services for each distributed heterogeneous data center based on SOA business intelligent cloud service platform, so as to enable a single functional department within the enterprise to obtain shared information provided by all departments in one stop, which is conducive to data sharing and thus promotes advanced experience.

[0007] The invention patent "An Industrial Internet Information Exchange Method" (202111258730.8) published by Hefei SIO Interconnect Technology Co., Ltd. proposes an industrial internet information exchange method. By developing heterogeneous network information interoperability standards and establishing a multi-level network information model system, it achieves cross-system interoperability, promotes the integration of information technology networks and production control networks, and uses semantic description methods and mandatory grammatical data structure requirements to form standardized norms. This solves the problems of differences between multi-source heterogeneous systems at the edge layer, enterprise layer, and industrial chain layer, as well as the difficulty in integrating data sharing services, which cause information to be blocked, not smooth, and not accurate.

[0008] The invention patent "Unified SQL Query Method for Heterogeneous Data Sources" (202311065187.9) published by Nanjing University of Posts and Telecommunications discloses a unified SQL query method for heterogeneous data sources. The framework includes modules such as access adaptation, parsing transformation, optimization verification, and SQL execution, which ensures high-accuracy interoperability between different data sources and effectively improves the scalability and application scope of the unified query system for heterogeneous sources.

[0009] The aforementioned research and inventions have played a certain role in advancing the collection, sharing, and interoperability of multi-source heterogeneous data, but the following problems still exist:

[0010] 1. When sharing multi-source heterogeneous data, there is a lack of exploration into the business logic and semantic relationships of the data, a lack of specific and feasible implementation plans for standardized semantic release of data, and an inability to guide enterprises to standardize the semantic release of data in specific business scenarios and application environments. This can easily lead to misunderstandings of the contextual meaning of data due to differences in semantic understanding, resulting in insufficient data sharing capabilities across departments and organizations.

[0011] 2. When constructing a multi-entity data space interoperability model, there is a lack of inputs such as data conceptual models, metadata sources, national / industry standards, and specific business scenarios. There is no unified professional terminology, unified data format, or unified metadata framework, which makes it easy for ambiguities to be understood between enterprises. Different systems and applications have difficulty interacting and integrating data. There is a lack of consistent descriptions and definitions of data elements, attributes, and relationships, which makes the data exchange capability between multiple entities weak when conducting data interoperability.

[0012] 3. The lack of modeling for data exchange and information interaction within the entity makes it impossible to clearly define the flow and transmission methods of data within the enterprise. The lack of a clear description of the logical structure of the data makes it difficult for different systems within the enterprise to understand and interpret the meaning of the data. Furthermore, the lack of a unified platform and communication mechanism during information interaction results in insufficient information sharing capabilities among entities. Summary of the Invention

[0013] To address the technical problems of insufficient sharing capabilities and weak data exchange capabilities during interoperation in existing methods for collecting, sharing, and interoperating multi-source heterogeneous data, this invention proposes a multi-subject data space interoperability method for multi-source heterogeneous data.

[0014] The technical solution of this invention is:

[0015] A method for multi-agent data space interoperability for multi-source heterogeneous data, characterized by the following steps:

[0016] Step 1: Sharing of multi-source heterogeneous data

[0017] For multi-source heterogeneous data of manufacturing / supply chain network node enterprises, data is collected using real-time and batch data collection methods. The business logic and semantic relationships of the multi-source heterogeneous data are mined, and data is standardized and semantically published to achieve the sharing of multi-source heterogeneous data.

[0018] Step 2: Extract, transform, and load the shared data from Step 1 into the target system;

[0019] Step 3: Preprocess the data in the target system after loading the data, including cleaning, dimensionality reduction and normalization.

[0020] Step 4: Construct a multi-agent data space interoperability model;

[0021] For each manufacturing / supply chain network node enterprise, a unified glossary of professional terms is constructed based on national and / or industry standards and business scenarios. The format of the data after preprocessing in step 3 is unified according to the data concept model. A unified metadata standard framework is constructed based on the metadata source, thereby forming a multi-entity data space interoperability model. The metadata standard framework includes technical metadata, business metadata, and data quality metadata. The multi-entity data space interoperability model includes a main module, a global module, a data model, a control center, sub-modules, and an information exchange center. There are multiple main modules, each corresponding to a different enterprise. The global module integrates data from different sub-modules into a unified view, making data easier to access, query, and analyze. The data model defines the data organization structure, including entities and attributes. The control center enables data to be transmitted between different sub-modules, between different sub-modules and the main module, and between different main modules according to set rules and processes. There are multiple sub-modules, each corresponding to a department within each main module. The information exchange center provides a unified platform and communication mechanism to promote data exchange and interoperability between different sub-modules and the global module.

[0022] Step 5: Export and query the data from the multi-agent data space interoperability model constructed in Step 4.

[0023] Furthermore, the real-time data and batch data acquisition methods mentioned in step 1 are Storm, Sqoop, Kettle, or DataX data acquisition technologies.

[0024] Furthermore, in step 1, after collecting tabular data, document data, and XML data using real-time data and batch data collection methods, a data table will be formed. When sharing, the data table needs to be converted into RDF format data first, and then the RDF format data is shared with the collected database data.

[0025] Furthermore, step 2 specifically involves:

[0026] Step 2.1: Extract business-related data from the multi-source heterogeneous data shared in Step 1;

[0027] Step 2.2: Perform at least one of the following operations on the business-related data extracted in Step 2.1: data format conversion, data normalization, field calculation, and data merging, to ensure that the data meets the requirements of data quality, consistency, and availability;

[0028] Step 2.3: Load the transformed data from Step 2.2 into the target system, which is a data warehouse, data mart, or other analysis platform.

[0029] Furthermore, in step 3: principal component analysis is used to reduce the dimensionality of the data; and appropriate methods are selected from sum standardization, maximum standardization and range standardization to normalize the data according to actual needs.

[0030] Furthermore, the specific method for constructing a unified vocabulary of professional terms in step 4 is as follows:

[0031] Step 4.1.1 Based on the national and / or industry standards that will be input into the multi-subject data space interoperability model and the specific business scenarios, determine which fields of professional terms the vocabulary to be constructed should include;

[0032] Step 4.1.2 Collect terminology information, including the term itself, term definition, term examples, and term field / category;

[0033] Step 4.1.3 Standardize the expression of professional terminology and formulate standards or guidelines to ensure the consistency and correct use of professional terminology;

[0034] Step 4.1.4 Develop a glossary of standardized technical terms to standardize and record the technical terms used within the entity, so as to ensure the accuracy and consistency of terminology during data interoperability.

[0035] Furthermore, the method for unifying the format of the preprocessed data in step 3 according to the data conceptual model in step 4 is as follows:

[0036] Step 4.2.1 Describe the relationships between data based on the data concept model that will be input into the multi-agent data space interoperability model, and clarify the concepts and structure of data;

[0037] Step 4.2.2 Based on business requirements, convert the data from its original format to a unified target format;

[0038] Step 4.2.3 Adjust the data structure according to the requirements of the unified target format, including merging fields, splitting fields, and renaming fields, to ensure that the data has a consistent structure;

[0039] Step 4.2.4 unifies data types, including converting string types to numeric types and date types to a unified date format, for subsequent calculations and analysis;

[0040] Step 4.2.5 Standardize the data to ensure that the data has consistent units, range and accuracy.

[0041] Furthermore, in step 4, the technical metadata includes the data table structure, data file format, data model, data dictionary, data transmission protocol and interface; the business metadata includes business rules, business lineage, business processes and data flow, data ownership and metadata permissions; and the data quality metadata includes data accuracy, data integrity, data consistency, data timeliness and data uniqueness.

[0042] Furthermore, the specific method for constructing the multi-agent data space interoperability model in step 4 is as follows:

[0043] Step 1) Define requirements:

[0044] Clarify the requirements and objectives of the multi-entity interoperability model, namely, determine the functions of the main modules and their required data integration methods, the data access, query and analysis requirements of the global module, the entity, attribute and relationship definitions of the data model, the data flow management rules and error handling mechanisms of the control center, the departments and functions corresponding to the sub-modules, and the security and privacy protection requirements of the information exchange center.

[0045] Step 2) Design the architecture:

[0046] Based on the requirements identified in step 1), design the overall architecture of the multi-entity interoperability model, namely, determining the relationships and interaction methods between the main module, global module, data model, control center, sub-modules, and information exchange center; considering the needs of data integration, access, and analysis, determine the functions and interfaces that the global module should have; design the data model, defining entities, attributes, and the relationships between them; determine the data flow rules and error handling mechanisms of the control center; map the sub-modules to the various departments of the main module, and define their functions and interfaces; design a unified platform and communication mechanism for the information exchange center, and consider the security and privacy protection measures for data transmission and interoperability.

[0047] Step 3) Implement the module:

[0048] Based on the architecture design in step 2), implement each module step by step; develop multiple main modules, each corresponding to a different enterprise, ensuring they can operate independently and meet their respective needs; develop a global module to implement data integration, access, query, and analysis functions; implement a data model, defining entities, attributes, and relationships, and providing corresponding interfaces and operation methods; develop a control center to implement data flow and transmission management functions, including tracking data sources, targets, and paths, error logging, error handling, and alarm notifications; develop corresponding functional modules for each submodule and interface with the various departments of the main modules; implement an information exchange center, providing a unified platform and communication mechanism, and incorporating security measures to protect the security and privacy of data transmission and interoperability;

[0049] Step 4) Testing and Optimization:

[0050] After completing the development of each module, testing and performance optimization are carried out. The tests check whether each module functions normally, whether data integration is accurate and effective, whether data access and query are fast and reliable, whether data transmission and flow follow rules and processes, and whether information exchange is secure and reliable. Based on the test results, optimizations are made to fix existing problems and improve the system's performance and stability.

[0051] Step 5) Deployment and Operation:

[0052] The process of deploying the system to actual modules is as follows:

[0053] Identify the main modules: Divide the manufacturing / supply chain network node enterprises into several main modules based on upstream and downstream relationships and business connections, with each main module corresponding to one enterprise;

[0054] Design global modules: Determine the functions of global modules and the data content to be integrated, including designing a unified view and methods for data access, querying, and analysis;

[0055] Define the data model: Based on the enterprise's needs, define the data organization structure, including the relationships between entities and attributes, as well as the logical structure of the data; wherein, the relationships between entities and attributes are defined based on the vocabulary of the unified professional terminology, and the logical structure of the data is defined based on the metadata standard framework.

[0056] Design Control Center: Define the functions and responsibilities of the control center, including managing data flow and transmission, tracking data sources, destinations, and paths, and detecting and handling potential anomalies and errors; wherein, data flow and transmission are managed according to the unified terminology glossary and unified data format, data sources, destinations, and paths are tracked according to the metadata standard framework, and potential anomalies and errors are detected and handled.

[0057] Divide into sub-modules: Based on the requirements of the main module, divide it into different sub-modules, each sub-module corresponding to an internal department or function;

[0058] Design an information exchange center: Determine the functions and communication mechanisms of the information exchange center to promote data exchange and interoperability between sub-modules and global modules, and ensure data security and privacy;

[0059] Implementation and Testing: Build the corresponding software system based on the designed modules, and conduct testing and verification to ensure the stability and reliability of the system;

[0060] Operation and optimization: Perform operation and maintenance management on the built modules, and optimize and improve them according to actual usage to improve the system's performance and efficiency.

[0061] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program is used to implement the above-described method for multi-subject data space interoperability for multi-source heterogeneous data when run.

[0062] The present invention also provides an electronic device, including a processor and a storage medium, wherein a computer program is stored on the storage medium, and the computer program is characterized in that: when the computer program is run, it is used to implement the above-described method for multi-subject data space interoperability for multi-source heterogeneous data.

[0063] The beneficial effects of this invention are:

[0064] This invention collects multi-source heterogeneous data from manufacturing / supply chain network node enterprises through various data acquisition methods, mines the business logic and semantic relationships between multi-source heterogeneous data from different enterprises and customers, and performs standardized semantic publishing of data, avoiding misunderstanding of the contextual meaning of data due to differences in semantic understanding, and realizing the sharing of multi-source heterogeneous data across departments and organizations.

[0065] By extracting, transforming, and loading multi-source heterogeneous data, data from multiple heterogeneous data sources is integrated into a unified target system, achieving data integration and consistency.

[0066] Preprocessing operations such as cleaning, dimensionality reduction, and transformation are performed on the data to ensure its accuracy, integrity, and consistency, while retaining the main features that are meaningful to the data.

[0067] Based on this, a unified professional terminology, unified data format, and unified metadata standard framework are constructed for manufacturing / supply chain network node enterprises according to data concept models, national / industry standards, specific business scenarios, and metadata sources. This forms a multi-entity data space interoperability model to ensure the security and feasibility of data exchange and information sharing among multiple entities.

[0068] Finally, the data in the data space is exported and queried. A data export module, a data index module, a query method module, a query interface module, and a query security module are designed. Through operations such as adding, deleting, modifying, and querying indexed data, the data interoperability between entities is further enhanced, providing data support for the business operation model. Attached Figure Description

[0069] Figure 1 This is a flowchart of the method of the present invention.

[0070] Figure 2 This is a data acquisition architecture diagram of DataX.

[0071] Figure 3It is a flowchart for data sharing among multi-source heterogeneous data.

[0072] Figure 4 This is an example diagram of a unified professional terminology vocabulary.

[0073] Figure 5 It is a diagram of a unified standard metadata framework. Detailed Implementation

[0074] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0075] like Figure 1 As shown, the multi-agent data space interoperability method for multi-source heterogeneous data provided by the present invention includes the following steps:

[0076] Step 1: Sharing of multi-source heterogeneous data

[0077] First, data collection is performed on multi-source heterogeneous data from manufacturing / supply chain network node enterprises using both real-time and batch data collection methods to achieve offline batch transmission and synchronization of large-scale data.

[0078] Then, data sharing is performed on the collected multi-source heterogeneous data. This multi-source heterogeneous data includes tabular data, document data, XML data, and database data. After collecting tabular, document, and XML data using real-time and batch data acquisition methods, data tables are formed. Before sharing, these data tables need to be converted to RDF format for easier data sharing, and then the RDF data is shared with the collected database data. Real-time and batch data acquisition methods include Storm, Sqoop, Kettle, or DataX data acquisition technologies. Taking DataX data acquisition technology as an example... Figure 2 As shown, its architecture includes the following main components:

[0079] Reader: Responsible for reading data from the data source. DataX provides a variety of built-in reader plugins that support various data source types, such as relational databases, HDFS, and Hive.

[0080] Converter: An optional component used to transform, clean, and process the read data. The converter can perform operations such as format conversion, field mapping, and data filtering.

[0081] Writer: Responsible for writing data to the target storage system or data warehouse. DataX also provides a variety of built-in writer plugins to adapt to various data storage systems, such as relational databases, HDFS, Hive, etc.

[0082] Channels: Serving as buffers between readers, converters, and writers, channels are used to handle data transmission and exchange. Different implementation methods can be chosen depending on the specific needs, such as memory queues or file queues.

[0083] The specific method for achieving multi-source heterogeneous data sharing in this step is as follows:

[0084] Step 1.1: For a specific business scenario (such as procurement, manufacturing, or warehousing), conduct a correlation analysis of business logic and semantics, uncover the hidden relationships in the data, and add them as tags to the original collected data. Specifically, the Apriori algorithm or the FP-Growth algorithm can be used. The following is the FP-Growth algorithm process: Traverse the dataset, calculate the support of each item, and construct an FP-tree; for each frequent item, construct the corresponding conditional pattern base by traversing the FP-tree; for each conditional FP-tree, continue the recursive operation to construct new conditional pattern bases and conditional FP-trees until no more can be constructed; by traversing the leaf nodes of the conditional FP-tree layer by layer, connect the parent node paths of each node to form a frequent itemset; for each frequent item, divide it into two non-empty subsets and calculate the confidence score; based on the set confidence threshold, filter out the association rules that meet the requirements; based on the set support and confidence thresholds, sort and filter the frequent itemsets and association rules to obtain the final frequent itemsets and association rules, and add the final frequent itemsets and association rules as labels to the corresponding data tables and databases collected using real-time data and batch data acquisition methods.

[0085] Step 1.2: Configure data table mapping for the data tables obtained after the association analysis in Step 1.1. The open API provided by the RDF (Resource Description Framework) platform can be used to automatically match terms that meet the data table mapping configuration conditions. This includes matching the corresponding categories in the data table and ontology based on the data table and field description information, and matching the fields with the attributes in the ontology. In this way, the ontology mapping configuration of data tables and fields can be completed quickly and accurately.

[0086] Step 1.3: Perform ontology mapping and appending. This involves matching and mapping the data table, fields, and ontology to establish relationships between them. For content that doesn't match, new terms can be automatically added to the existing ontology. This process includes term extraction, matching, and ontology updating to improve the ontology's content. After completing ontology mapping and appending, use a D2RQ conversion tool to convert the data table to RDF format.

[0087] Step 1.4: Store the database data and associated RDF format data in a graph database or triplet storage system, such as GraphDB or Virtuoso. Then, by providing a SPARQL query interface or RESTful API, standardize the semantic publishing of the data, enabling other users to access and query this business-related database data and RDF format data, ultimately achieving multi-source heterogeneous data sharing within manufacturing / supply chain network node enterprises and between different enterprises.

[0088] Step 2: Extract, transform, and load (ETL) the shared data between manufacturing / supply chain network node enterprises and users to enable data to be migrated from the source system to the target system in a reliable and efficient manner, which helps to ensure data quality, data consistency, and data availability.

[0089] Step 2.1: Data Extraction

[0090] Extract various business-related data (such as order data, production data, procurement data, etc.) from the database and RDF format data obtained in step 1. By establishing connections with the corresponding database and triplet storage system, select or query relevant data tables or databases to extract field data related to manufacturing or supply chain processes, including order number, customer information, product information, production information, procurement forms, etc.

[0091] Step 2.2: Data Transformation

[0092] After data extraction, the data is transformed to ensure its quality, consistency, and availability, thereby meeting the needs of a data warehouse. The transformation steps include at least one of the following operations: data format conversion, data normalization, field calculation, and data merging. Specifically:

[0093] Data format conversion: Unify the format of extracted business-related data fields, such as converting different date formats into a unified standard format.

[0094] Field calculations: Calculate additional metrics based on order data, production data, procurement data, etc., such as total sales and average order amount.

[0095] Data merging: Merging data from multiple tables or data sources to create a complete dataset.

[0096] Data normalization: unify fields with the same meaning, such as using "order" to represent both "indent" and "order".

[0097] Step 2.3: Data Loading

[0098] The transformed data is loaded into the target system, which can be a data warehouse, data mart, or other analytics platform. Loading can be a full load (loading all data into the target system) or an incremental load (loading only the changed data). During the loading process, the transformed data from step 2.2 is inserted, updated, or replaced in the corresponding positions of the relevant business tables in the target system. Specifically, the INSERT statement or the loading function provided by ETL tools can be used to complete the data loading task.

[0099] Step 3: Preprocess the data in the target system after loading the data to eliminate potential problems in the original data and provide a data foundation for the subsequent construction of a multi-agent data space interoperability model.

[0100] Step 3.1: Data Cleaning

[0101] Data cleaning involves inspecting, correcting, and transforming raw data to eliminate missing, abnormal, duplicate, and non-standard parts, ensuring the accuracy, completeness, and consistency of the data. Specifically:

[0102] Detecting and Handling Missing Values: Missing values ​​are empty values ​​or unrecorded data. They can be detected using statistical methods (such as counting, percentages), visualization methods (such as heatmaps, missing value plots), or programming methods. For detected missing values, appropriate methods can be chosen to fill or delete them. Common methods include: 1) deleting rows or columns containing missing values; 2) filling missing values ​​with descriptive statistics such as mean, median, and mode; 3) estimating missing values ​​using interpolation methods (such as linear interpolation and polynomial interpolation); and 4) predicting missing values ​​based on the values ​​of other relevant variables.

[0103] Outlier detection and handling: Outliers are extreme values ​​that are significantly different from other observations. They can be detected using statistical methods (such as box plots, outlier detection algorithms), visualization methods (such as scatter plots, histograms), or existing specialized anomaly detection algorithms. For detected outliers, they are replaced with reasonable values. Replacement can be based on statistical indicators (such as the mean, median, or mode) or using interpolation techniques (such as linear interpolation, polynomial interpolation, or spline interpolation).

[0104] Handling duplicate values: Duplicate values ​​refer to records in the dataset that are completely or partially repeated. The deduplication operation can be used to remove duplicate values ​​to ensure that each record is unique.

[0105] Addressing Inconsistencies: Inconsistencies may arise between different fields in the data, or between different values ​​for the same field. To resolve these inconsistencies, formatting or standardization operations are necessary. Examples include standardizing date formats, standardizing text, and merging similar categories.

[0106] Step 3.2: Data Dimensionality Reduction

[0107] Dimensionality reduction refers to reducing the number of dimensions or features of data by removing redundant information, noise, and redundant features, while retaining the main features that are meaningful to the data. This simplifies data representation, reduces computational complexity, and improves model performance. Principal component analysis (PCA) can be used for dimensionality reduction. By extracting the largest individual differences revealed by the principal components, the number of data indicators is reduced, while retaining as much information as possible. The specific steps for dimensionality reduction using PCA are as follows:

[0108] Step 3.2.1: Normalization of linear functions

[0109] By using a normalization formula to perform a linear transformation on the original data, the result is mapped to the range [0,1], achieving proportional scaling of the original data. The normalization formula is as follows:

[0110]

[0111] Among them, X norm Let X represent a data value after linear transformation, and let X represent a data point in the data sequence before linear transformation. min Let X represent the minimum value in the data sequence before the linear transformation. max This represents the maximum value in the data sequence before the linear transformation.

[0112] Step 3.2.2: Calculate the covariance matrix

[0113] The covariance matrix reflects the correlation between different features; its elements represent the covariance between two features. The formula for calculating the covariance matrix is ​​as follows:

[0114]

[0115] Where v represents the calculated covariance matrix, n is the number of samples, and k is the feature dimension. The matrix obtained by subtracting the mean (zero centering) from all normalized samples.

[0116] Step 3.2.3: Eigenvalue Decomposition

[0117] The covariance matrix calculated in step 3.2.2 is subjected to eigenvalue decomposition to obtain eigenvalues ​​and corresponding eigenvectors. Eigenvalues ​​represent the variance of the covariance matrix in each eigenvector direction, while eigenvectors represent the principal directions of the covariance matrix. This problem can be represented as a maximization problem, as follows:

[0118]

[0119] Where ω represents the covariance matrix, ω T This represents the transpose of the covariance matrix.

[0120] Step 3.2.4: Select principal components

[0121] Based on the magnitude of the eigenvalues ​​obtained from the decomposition in step 3.2.3, the most important eigenvectors are selected as principal components. The eigenvectors corresponding to the top d eigenvalues ​​are then used to map the n-dimensional samples to d dimensions using the following mapping.

[0122]

[0123] New x i The d-th dimension is x i The d-th principal component ω d Projecting along a direction involves selecting the eigenvectors corresponding to the d largest eigenvalues ​​and discarding features (noise) with smaller variances, ensuring that each n-dimensional column vector x... i Mapped to a d-dimensional column vector x i '.

[0124] Step 3.2.5: Adjust data along principal component axes

[0125] The normalized data is projected onto the principal components selected in step 3.2.4 to obtain the dimensionality-reduced dataset. This can be achieved by multiplying the original data with the principal component matrix.

[0126] Step 3.3: Data Transformation

[0127] Data transformation refers to a series of operations and conversions performed on data to change its form, representation, or distribution, revealing hidden patterns and relationships, thereby making the data more suitable for specific analysis, modeling, or application needs.

[0128] To eliminate the influence of dimensions between data features, the features retained after dimensionality reduction in step 3.2 are normalized to ensure comparability between different indicators. Data normalization can be viewed as a specific data transformation method specifically designed to scale data to a uniform scale. Data normalization includes the following three methods:

[0129] Method 1: Sum-based Standardization

[0130] The features retained after dimensionality reduction in step 3.2 are summed and standardized according to the following formula. The summed and standardized data are between (0,1) and their sum is 1. The summed and standardized formula is as follows:

[0131]

[0132] Where, xij Let x' represent the value in the i-th row and j-th column of the data matrix. ij This represents the value of the data matrix after summation and standardization.

[0133] Method 2: Maximum Standardization

[0134] The features retained after dimensionality reduction in step 3.2 are standardized for maximum values ​​using the following formula. In the new data after maximum standardization, the maximum value of each feature is 1, and all other values ​​are less than 1. The maximum standardization formula is as follows:

[0135]

[0136] Where, x ij Let x represent the value in the i-th row and j-th column of the data matrix. ” ij This represents the value of the data matrix after the maximum value standardization process.

[0137] Method 3: Range Standardization

[0138] The features retained after dimensionality reduction in step 3.2 are standardized by range using the following formula. After range standardization, the maximum value of each feature is 1, the minimum value is 0, and all other values ​​are between 0 and 1. The range standardization formula is as follows:

[0139]

[0140] Where, x ij Let x represent the value in the i-th row and j-th column of the data matrix, min{x} ij} and max{x ij} refers to x ij Minimum and maximum values ​​in the same column, x i ' j "" indicates the value of the data matrix after range standardization.

[0141] The above three different data normalization methods can be selected according to different needs. Among them, summation normalization is often used for probability distributions, proportions and percentages, maximum normalization is suitable for situations where data needs to be mapped to a specific range, and range normalization is used for situations where data needs to be mapped to any interval.

[0142] Step 4: Construct a multi-agent data space interoperability model based on the data preprocessed in Step 3.

[0143] Step 4.1: Construct a unified glossary of professional terms

[0144] Step 4.1.1 Based on the national and / or industry standards that will be input into the multi-subject data space interoperability model and the specific business scenarios, clarify the goals and scope of creating the terminology glossary, that is, determine which fields of professional terms need to be included.

[0145] Step 4.1.2 Collect terminology information, including the term itself, term definition (provide a clear and concise definition to ensure that everyone can understand the meaning of the term), term examples (to help further understand its context and usage), and term field / category (categorize the term into the corresponding field or category for better organization and retrieval).

[0146] Step 4.1.3 Standardize the expression of professional terminology. In the glossary of professional terms, some standards or guidelines should be formulated to ensure the consistency and correct use of professional terminology. For example, the preferred spelling, abbreviation or correct spelling rules of specific professional terms can be specified.

[0147] Step 4.1.4 Develop a unified glossary of professional terms (example shown). Figure 4 This is to standardize and record the technical terms used within the entity, so as to ensure the accuracy and consistency of terminology during data interoperability.

[0148] Step 4.2: Standardize the data format of the data after preprocessing in Step 3.

[0149] Step 4.2.1 uses the data concept model, which will be input into the multi-agent data space interoperability model later, to describe the relationships between data and clarify the concepts and structure of the data. A data concept model is a tool or method used in database design and data modeling to describe concepts, entities, attributes, and relationships in the data domain, as well as the connections between these elements.

[0150] Step 4.2.2 Based on business requirements, convert the data from its original format (such as CSV, Excel, JSON, RDF, etc.) into a unified target format.

[0151] Step 4.2.3 Adjust the data structure according to the requirements of the unified target format, including operations such as merging fields, splitting fields, and renaming fields, to ensure that the data has a consistent structure.

[0152] Step 4.2.4 unifies data types, including converting string types to numeric types and date types to a unified date format, to facilitate subsequent calculations and analysis.

[0153] Step 4.2.5 Perform data standardization. For certain specific fields, standardization is required to ensure that the data has consistent units, ranges, and precision.

[0154] Step 4.3: Construct a unified metadata standard framework

[0155] Based on the metadata sources that will be input into the multi-entity data space interoperability model, the metadata is divided into technical metadata, business metadata, and data quality metadata, resulting in a unified metadata standard framework, such as... Figure 5 As shown.

[0156] Technical metadata describes and records data technology aspects, and includes the following elements:

[0157] Table structure: This refers to information such as the table name, field names, data types, data length, constraints (e.g., primary keys, foreign keys, uniqueness), and indexes. This metadata describes the data structure stored in the database.

[0158] Data file format: refers to the format and organization of data files, such as CSV, XML, JSON, etc. This includes information such as the data file delimiter, encoding, and row and column structure.

[0159] Data model: A model used to represent data structures, such as the entity-relationship model in relational databases or the object model in object-oriented databases. A data model describes the relationships, attributes, and constraints between data.

[0160] A data dictionary refers to the definition, meaning, business rules, and relationships between data fields and other fields. A data dictionary can provide information such as field descriptions, value ranges, units, and data lineage, which helps in understanding and using the data.

[0161] Data transmission protocols and interfaces: These refer to the protocols and interfaces used for data transmission, such as HTTP, FTP, SOAP, REST, etc. Data transmission protocols and interfaces can describe information such as data transmission methods, port numbers, authentication and authorization.

[0162] Business metadata describes and records data assets and related business information. It provides information about the business meaning, purpose, rules, and association with business processes of the data, and includes the following elements:

[0163] Business rules: These describe the business rules and constraints governing the data. These rules define the requirements for data validity, compliance, and consistency. Examples include the value range of a field, calculation formulas, and data validation rules.

[0164] Business lineage: Describes the source, transmission path, and impact of data. Business lineage tracing can track the flow of data from the source system to the target system and record the dependencies between data, enabling business users to trace the source and history of data changes.

[0165] Business processes and data flows: This connects data assets to business processes, describing the role and flow of data within those processes. This helps business users understand the role and value of data in their business activities.

[0166] Data ownership: Designating the owner and person in charge of the data. This helps clarify responsibilities for data management and maintenance, ensuring data accuracy and compliance.

[0167] Metadata permissions: Define access permissions and control policies for business metadata. This ensures that only authorized users can view and modify metadata, protecting sensitive business information.

[0168] Data quality metadata describes and records information about data quality. It is used to assess, monitor, and improve data quality levels and includes the following elements:

[0169] Data accuracy: Describes whether the data is accurate and consistent with real-world entities and facts.

[0170] Data integrity: Describes whether the data is complete and whether it contains all the required fields and records.

[0171] Data consistency: describes whether data is consistent across different systems or data sources, and whether there are conflicts or contradictions.

[0172] Data timeliness: describes the frequency and timeliness of data updates, and whether it can reflect the latest situation in a timely manner.

[0173] Data uniqueness: describes whether there are duplicate records in the data, and whether there is a unique identifier to ensure the uniqueness of each record.

[0174] Step 4.4: Based on the unified professional terminology glossary constructed in Step 4.1, the data after the unified data format in Step 4.2, and the unified metadata standard framework constructed in Step 4.3, construct a multi-subject data space interoperability model.

[0175] The constructed multi-agent data space interoperability model includes a principal module, a global module, a data model, a control center, sub-modules, and an information exchange center. The definition of the multi-agent data space interoperability model is as follows:

[0176]

[0177] Where i represents the entity's number, such as the number of company A; j represents the number of the module belonging to the entity after dividing the entity with number i into multiple modules; NM i This represents the main module numbered i (e.g., company A); GM i This indicates a global module, which supports combining different submodules into a single module (SM).ij Data is integrated into a unified view, making it easier to access, query, and analyze; DM i A data model represents the organizational structure of data, including entities and attributes. By defining the relationships between entities and attributes, it describes the logical structure of the data, enabling different enterprise management systems (such as ERP, SCM, CRM, etc.) to understand and interpret the meaning of the data; CC i The control center is responsible for managing the flow and transmission of data. By tracking the source, destination, and path of data, it ensures that data is transmitted according to predetermined rules and processes between different sub-modules, between different sub-modules and the main module, and between different main modules. It can also detect and handle anomalies and errors that may occur during data interoperability, resolving issues promptly by recording error logs, sending alarm notifications, and providing corresponding error handling mechanisms. ij This represents a submodule, corresponding to different departments within a main entity (such as the sales department, purchasing department, etc.); CH ij It represents the information exchange center, providing a unified platform and communication mechanism to facilitate communication between different sub-modules (SM). ij With the global module GM i Data exchange and interoperability between them, and through the adoption of various security measures such as encryption, authentication and access control, ensure the security and privacy of data during transmission and interoperability.

[0178] The principle of establishing a multi-agent interoperability model:

[0179] This approach integrates data from entities within a specific domain (such as a company or customer) using a wrapper / mediator model, establishing a globally unified information exchange and control center within each entity. Different entities are managed in a federated manner, using a unified data exchange protocol to construct a federated model of their global schemas. Unified data interoperability and data storage technologies are employed within each entity; cross-entity data interoperability may require the conversion and mapping of technical terminology, data formats, and metadata frameworks.

[0180] Methods for establishing multi-agent interoperability models:

[0181] 1) Define requirements:

[0182] Before construction begins, it is necessary to clarify the requirements and objectives of the multi-entity interoperability model. This includes determining the functions of the main modules and their required data integration methods, the data access, query, and analysis requirements of the global modules, the entity, attribute, and relationship definitions of the data model, the data flow management rules and error handling mechanisms of the control center, the departments and functions corresponding to the sub-modules, and the security and privacy protection requirements of the information exchange center.

[0183] 2) Design Architecture:

[0184] Based on the requirements identified in step 1), design the overall architecture of the multi-entity interoperability model, namely, determining the relationships and interaction methods between the main module, global module, data model, control center, sub-modules, and information exchange center. Considering the needs for data integration, access, and analysis, determine the functions and interfaces that the global module should possess. Design the data model, defining entities, attributes, and the relationships between them. Determine the data flow rules and error handling mechanisms for the control center. Map the sub-modules to the various departments of the main module and define their functions and interfaces. Design a unified platform and communication mechanism for the information exchange center, considering security and privacy protection measures for data transmission and interoperability.

[0185] 3) Implementation module:

[0186] Based on the architecture design in step 2), implement each module step by step. Develop multiple main modules, each corresponding to a different enterprise, ensuring they can operate independently and meet their respective needs. Develop a global module to implement data integration, access, query, and analysis functions. Implement the data model, defining entities, attributes, and relationships, and providing corresponding interfaces and operation methods. Develop a control center to manage data flow and transmission, including tracking data sources, destinations, and paths, error logging, error handling, and alarm notifications. Develop corresponding functional modules for each submodule and interface with the various departments of the main modules. Implement an information exchange center, providing a unified platform and communication mechanism, and incorporating security measures to protect the security and privacy of data transmission and interoperability.

[0187] 4) Testing and optimization:

[0188] After completing the development of each module, testing and performance optimization are performed. Testing ensures each module functions correctly, data integration is accurate and effective, data access and querying are fast and reliable, data transmission and flow follow rules and procedures, and information exchange is secure and reliable. Based on the test results, optimizations are made to fix existing problems and improve the system's performance and stability.

[0189] 5) Deployment and Operation:

[0190] After testing and optimization are completed, the process of deploying the system to the actual module is as follows:

[0191] Identify the main modules: Divide the manufacturing / supply chain network node enterprises into several main modules based on upstream and downstream relationships and business connections, with each main module corresponding to a specific enterprise.

[0192] Design global modules: Determine the functions of global modules and the data content to be integrated, including designing a unified view and methods for data access, querying, and analysis.

[0193] Define the data model: Based on the enterprise's needs, define the data organization structure, including the relationships between entities and attributes, as well as the logical structure of the data. Tools such as Entity-Relationship Diagrams (ER Diagrams) can be used for data model design. Use standardized terminology and data formats to ensure the consistency and accuracy of entities, attributes, and relationships. Simultaneously, use a standardized metadata framework as a reference to design the logical structure and organization of the data.

[0194] Design the control center: Define the functions and responsibilities of the control center, including managing data flow and transmission, tracking data sources, destinations, and paths. Establish rules and processes for managing data flow and transmission based on a unified glossary of technical terms and data formats. Track data sources, destinations, and paths according to the established metadata standard framework, and perform error handling and anomaly detection.

[0195] Divide into sub-modules: Based on the requirements of the main module, divide it into different sub-modules, each sub-module corresponding to an internal department or function.

[0196] Design the information exchange center: Determine the functions and communication mechanisms of the information exchange center to facilitate data exchange and interoperability between sub-modules and global modules, and ensure data security and privacy.

[0197] Implementation and Testing: Build the corresponding software system based on the designed modules, and conduct testing and verification to ensure the stability and reliability of the system.

[0198] Operation and optimization: Perform operation and maintenance management on the built modules, and optimize and improve them according to actual usage to improve the system's performance and efficiency.

[0199] Step 5: Export and query the data in the multi-subject data space interoperability model constructed in Step 4 to promote data exchange and sharing, and provide data support for the business operation model.

[0200] To export and query data based on the multi-agent data space interoperability model constructed in step 4, it is necessary to build a data export module, a data index module, a query method module, and a query interface module. Each module is used to implement the following functions:

[0201] The data export module is used to export data from the multi-agent data space interoperability model. Structured data is loaded into an XML file, and unstructured data is attached to the XML file. The resulting XML data file contains the specific data and its attributes.

[0202] The data indexing module is used to index exported XML data, including operations such as adding, deleting, modifying and querying indexed data. By creating an index structure, the access speed of data can be accelerated. The index can organize data according to the values ​​of one or more fields and provide fast data search, filtering and sorting functions, reducing the time and resource consumption of full table scans.

[0203] The query method module is used to analyze the data query process and meet the basic search requirements through basic queries and compound queries. By analyzing the structure and conditions of the query statement, it determines how to execute the query optimally, including index selection, query rewriting, query plan generation and execution plan optimization, in order to reduce query response time and system resource consumption.

[0204] The query interface module constructs extended query interfaces and designs the interface specifications and security to meet the needs of enterprises and customers for structured and unstructured data queries.

[0205] The query security module implements security measures and access control by building a unified security protocol, restricting users' access to data queries, and ensuring that only authorized users can perform specific query operations by defining user roles, permissions, and access policies, thus protecting the security of sensitive data.

Claims

1. A multi-agent data space interoperability method for multi-source heterogeneous data, characterized in that, Includes the following steps: Step 1: Sharing of multi-source heterogeneous data For multi-source heterogeneous data of manufacturing / supply chain network node enterprises, data is collected using real-time and batch data collection methods. The business logic and semantic relationships of the multi-source heterogeneous data are mined, and data is standardized and semantically published to achieve the sharing of multi-source heterogeneous data. Step 2: Extract, transform, and load the shared data from Step 1 into the target system; Step 3: Preprocess the data in the target system after loading the data, including cleaning, dimensionality reduction and normalization. Step 4: Construct a multi-agent data space interoperability model; For each manufacturing / supply chain network node enterprise, a unified glossary of professional terms is constructed based on national and / or industry standards and business scenarios. The format of the data after preprocessing in step 3 is unified according to the data concept model. A unified metadata standard framework is constructed based on the metadata source, thereby forming a multi-entity data space interoperability model. The metadata standard framework includes technical metadata, business metadata, and data quality metadata. The multi-entity data space interoperability model includes a main module, a global module, a data model, a control center, sub-modules, and an information exchange center. There are multiple main modules, each corresponding to a different enterprise. The global module integrates data from different sub-modules into a unified view, making the data easier to access, query, and analyze. The data model defines the data organization structure, including entities and attributes. The control center enables data to be transmitted between different sub-modules, between different sub-modules and the main module, and between different main modules according to set rules and processes. There are multiple sub-modules, each corresponding to a department within each main module. The information exchange center provides a unified platform and communication mechanism to promote data exchange and interoperability between different sub-modules and the global module. The method for constructing a unified vocabulary of professional terms is as follows: Step 4.1.1 Based on the national and / or industry standards that will be input into the multi-subject data space interoperability model and the specific business scenarios, determine which fields of professional terms the vocabulary to be constructed should include; Step 4.1.2 Collect terminology information, including the term itself, term definition, term examples, and term field / category; Step 4.1.3 Standardize the expression of professional terminology and formulate standards or guidelines to ensure the consistency and correct use of professional terminology; Step 4.1.4 Develop a glossary of standardized technical terms to standardize and record the technical terms used within the entity, so as to ensure the accuracy and consistency of terminology during data interoperability. The technical metadata includes data table structure, data file format, data model, data dictionary, data transmission protocol and interface; the business metadata includes business rules, business lineage, business processes and data flow, data ownership and metadata permissions; the data quality metadata includes data accuracy, data integrity, data consistency, data timeliness and data uniqueness. Step 5: Export and query the data from the multi-agent data space interoperability model constructed in Step 4.

2. The multi-source heterogeneous data oriented multi-agent data space interoperability method according to claim 1, characterized in that: The real-time and batch data acquisition methods mentioned in step 1 are Storm, Sqoop, Kettle, or DataX data acquisition technologies; In step 1, after collecting tabular data, document data, and XML data using real-time and batch data acquisition methods, a data table will be formed. When sharing, the data table needs to be converted into RDF format data first, and then the RDF format data is shared with the collected database data.

3. The multi-source heterogeneous data oriented multi-agent data space interoperability method according to claim 1, characterized in that: Step 2 is as follows: Step 2.1: Extract business-related data from the multi-source heterogeneous data shared in Step 1; Step 2.2: Perform at least one of the following operations on the business-related data extracted in Step 2.1: data format conversion, data normalization, field calculation, and data merging, to ensure that the data meets the requirements of data quality, consistency, and availability; Step 2.3: Load the transformed data from Step 2.2 into the target system, which is a data warehouse, data mart, or other analysis platform.

4. The multi-source heterogeneous data oriented multi-agent data space interoperability method according to claim 1, characterized in that: In step 3: Principal component analysis is used to reduce the dimensionality of the data; based on actual needs, appropriate methods are selected from summation standardization, maximum standardization, and range standardization to normalize the data.

5. The multi-source heterogeneous data oriented multi-agent data space interoperability method according to claim 1, characterized in that: The method for unifying the format of the preprocessed data in step 3 according to the data conceptual model in step 4 is as follows: Step 4.2.1 Describe the relationships between data based on the data concept model that will be input into the multi-agent data space interoperability model, and clarify the concepts and structure of data; Step 4.2.2 Based on business requirements, convert the data from its original format to a unified target format; Step 4.2.3 Adjust the data structure according to the requirements of the unified target format, including merging fields, splitting fields, and renaming fields, to ensure that the data has a consistent structure; Step 4.2.4 unifies data types, including converting string types to numeric types and date types to a unified date format, for subsequent calculations and analysis; Step 4.2.5 Standardize the data to ensure that the data has consistent units, range and accuracy.

6. The method for multi-agent data spatial interoperability for multi-source heterogeneous data according to claim 1, characterized in that: The specific method for constructing the multi-agent data space interoperability model in step 4 is as follows: Step 1) Define requirements: Clarify the requirements and objectives of the multi-entity interoperability model, namely, determine the functions of the main modules and their required data integration methods, the data access, query and analysis requirements of the global module, the entity, attribute and relationship definitions of the data model, the data flow management rules and error handling mechanisms of the control center, the departments and functions corresponding to the sub-modules, and the security and privacy protection requirements of the information exchange center. Step 2) Design the architecture: Based on the requirements identified in step 1), design the overall architecture of the multi-entity interoperability model, namely, determining the relationships and interaction methods between the main module, global module, data model, control center, sub-modules, and information exchange center; considering the needs of data integration, access, and analysis, determine the functions and interfaces that the global module should have; design the data model, defining entities, attributes, and the relationships between them; determine the data flow rules and error handling mechanisms of the control center; map the sub-modules to the various departments of the main module, and define their functions and interfaces; design a unified platform and communication mechanism for the information exchange center, and consider the security and privacy protection measures for data transmission and interoperability. Step 3) Implement the module: Based on the architecture design in step 2), implement each module step by step; develop multiple main modules, each corresponding to a different enterprise, to ensure that they can run independently and meet their respective needs; develop a global module to implement data integration, access, query and analysis functions; implement the data model, define entities, attributes and relationships, and provide corresponding interfaces and operation methods; Develop a control center to manage data flow and transmission, including tracking data sources, destinations and paths, error logging, error handling, and alarm notifications; Develop corresponding functional modules for each submodule and connect them with the various departments of the main module; To establish an information exchange center, providing a unified platform and communication mechanism, and incorporating security measures to protect the security and privacy of data transmission and interoperability; Step 4) Testing and Optimization: After the development of each module is completed, testing and performance optimization are carried out; the functions of each module are tested to see if they are running normally, whether the data integration is accurate and effective, whether the data access and query are fast and reliable, whether the data transmission and flow are carried out according to rules and processes, and whether the information exchange is secure and reliable. Optimize based on test results, fix existing problems, and improve system performance and stability; Step 5) Deployment and Operation: The process of deploying the system to actual modules is as follows: Identify the main modules: Divide the manufacturing / supply chain network node enterprises into several main modules based on upstream and downstream relationships and business connections, with each main module corresponding to one enterprise; Design global modules: Determine the functions of global modules and the data content to be integrated, including designing a unified view and methods for data access, querying, and analysis; Define the data model: Based on the enterprise's needs, define the data organization structure, including the relationships between entities and attributes, as well as the logical structure of the data; wherein, the relationships between entities and attributes are defined based on the vocabulary of the unified professional terminology, and the logical structure of the data is defined based on the metadata standard framework. Design Control Center: Define the functions and responsibilities of the control center, including managing data flow and transmission, tracking data sources, destinations, and paths, and detecting and handling potential anomalies and errors; wherein, data flow and transmission are managed according to the unified terminology glossary and unified data format, data sources, destinations, and paths are tracked according to the metadata standard framework, and potential anomalies and errors are detected and handled. Divide into sub-modules: Based on the requirements of the main module, divide it into different sub-modules, each sub-module corresponding to an internal department or function; Design the information exchange center: Determine the functions and communication mechanisms of the information exchange center to promote data exchange and interoperability between sub-modules and global modules, and ensure data security and privacy; Implementation and Testing: Build the corresponding software system based on the designed modules, and conduct testing and verification to ensure the stability and reliability of the system; Operation and optimization: Perform operation and maintenance management on the built modules, and optimize and improve them according to actual usage to improve the system's performance and efficiency.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is executed to implement the multi-agent data space interoperability method for multi-source heterogeneous data as described in any one of claims 1-6.

8. An electronic device, comprising a processor and a storage medium, wherein a computer program is stored on the storage medium, characterized in that: The computer program is executed to implement the multi-agent data space interoperability method for multi-source heterogeneous data as described in any one of claims 1-6.

Citation Information

Patent Citations

  • Method for processing unstructured data on basis of deep learning

    CN106844636A

  • A Semantic Modeling Method for Relational Data

    CN110019555B

  • Drilling acceleration influence factor analysis model based on big data

    CN111435344A

  • An Industrial Internet Information Exchange Method

    CN113992769B

  • Unified SQL query method oriented to heterogeneous data sources

    CN117093599A