A method for explicitating implicit knowledge for tobacco enterprise production scheduling
By constructing a lexicon of explicit and implicit knowledge and using ontology fusion technology, the problem of making implicit knowledge explicit in the production scheduling of tobacco enterprises was solved, achieving efficient and economical market adaptability and data integration in production planning, thereby improving production efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA TOBACCO ZHEJIANG IND CO LTD
- Filing Date
- 2023-10-26
- Publication Date
- 2026-05-15
AI Technical Summary
In the production planning and scheduling of tobacco companies, it is difficult to classify explicit and tacit knowledge, and tacit knowledge is difficult to standardize, resulting in data isolation and affecting the economic efficiency and market adaptability of production plans.
By employing natural language processing techniques and ontology mapping methods, a classification lexicon of explicit and implicit knowledge is constructed. A scheduling ontology is built using the n-gram matching algorithm and OWL language. Combined with the implicit knowledge IS meta-model, the implicit knowledge is made explicit and the ontology is integrated, thus consolidating explicit and implicit knowledge.
This has improved the efficiency and effectiveness of tobacco companies' production planning and scheduling, enhanced their ability to adapt to market changes, and ensured the economical and efficient execution of production plans.
Smart Images
Figure CN117453926B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of tobacco enterprise production management technology, and in particular relates to a method for making tacit knowledge explicit in tobacco enterprise production scheduling. Background Technology
[0002] With the continuous development and progress of society, more and more people have become aware of the harm smoking poses to their health. Coupled with the impact of various policies, cigarette sales reached a turning point in 2015, the market gradually slumped, and competition among tobacco companies intensified. To better meet the diverse and personalized needs of consumers and achieve the required profit, tax, and total industrial output, tobacco companies need to adjust their cigarette product categories in a timely manner and produce products economically and efficiently. This undoubtedly places higher demands on their production planning and scheduling.
[0003] Due to the influence of various factors such as market environment, technological processes, production scheduling experience, production equipment, material supply, and management level, production planning and scheduling involves a large amount of heterogeneous knowledge with different syntax, types, and formats both inside and outside the enterprise. This knowledge is divided into explicit knowledge and tacit knowledge. Explicit knowledge has good structural characteristics that facilitate extraction and application, while tacit knowledge usually exists in various forms such as web pages, images, audio, video, thoughts, and experience, and is characterized by high personalization and ambiguity. Therefore, how to effectively make this tacit knowledge explicit will have a significant impact on the production scheduling of tobacco enterprises. Summary of the Invention
[0004] To address the technical problems existing in the prior art, such as the relative isolation of data between different information systems (IS) within tobacco enterprises, the difficulty in classifying explicit and tacit knowledge, and the difficulty in standardizing tacit knowledge, this invention provides a method for making tacit knowledge explicit in tobacco enterprise production scheduling, which can help solve the problem of making tacit knowledge explicit in production scheduling.
[0005] The technical solution adopted in this invention is:
[0006] A method for making tacit knowledge explicit in tobacco enterprise production scheduling, characterized by the following steps:
[0007] Step S1: Extract and preprocess scheduling knowledge;
[0008] Data and knowledge related to production scheduling are extracted from internal and external information systems of tobacco companies. After the data extraction of tobacco companies’ production scheduling-related information systems is completed, the FoolNLTK tool is used to preprocess the data.
[0009] Step S2: Construct a thesaurus of explicit and implicit scheduling knowledge;
[0010] Based on expert knowledge of tobacco companies’ production scheduling and relevant internal experience, a lexicon of explicit and implicit knowledge classification is constructed using natural language processing technology.
[0011] The extracted data and knowledge are classified using a lexicon of explicit and implicit knowledge, which distinguishes between explicit and implicit knowledge.
[0012] Step S3: Extract keywords from implicit and explicit knowledge to make implicit scheduling knowledge explicit;
[0013] Keyword extraction is performed on tacit and explicit knowledge. The similarity between the two keywords is compared using an n-gram matching algorithm. If the similarity m is greater than or equal to the set threshold n, a scheduling knowledge framework is constructed based on the explicit knowledge. The scheduling knowledge framework is then used to integrate the set of tacit knowledge keywords to make the tacit knowledge explicit.
[0014] Step S4: Integrate explicit and implicit scheduling knowledge;
[0015] Based on explicit knowledge, the scheduling ontology is built from top to bottom using the OWL language and the Protege ontology development tool. OWL can describe the relationships between concepts, properties, and individuals to ensure powerful semantic expression capabilities and enable reasoning functions.
[0016] By using ontology mapping and ontology fusion methods, semantic integration between the scheduling IS semantic model and the scheduling ontology is achieved, thus integrating explicit and implicit scheduling knowledge.
[0017] Furthermore, in step S3, for implicit knowledge and explicit knowledge whose keyword similarity m is less than the set threshold n or for implicit knowledge that cannot be extracted by keywords, it is necessary to construct an implicit scheduling knowledge information system concept UML meta-model, namely the implicit scheduling knowledge IS meta-model.
[0018] Furthermore, the implicit scheduling knowledge IS metamodel does not contain any explicit knowledge, that is, it does not contain any explicit semantics. Its construction process includes two main steps: ① extraction of the database ER model; ② mapping of the ER model to the UML conceptual model.
[0019] Furthermore, the steps for constructing the scheduling IS semantic model specifically include the following:
[0020] Extract meaningful features from data related to tobacco companies' production scheduling so that the model can understand and learn;
[0021] The explicit knowledge of tobacco enterprise production scheduling and the extracted meaningful features, combined with expert knowledge and experience in production scheduling, are used as the model training set to train the nonlinear support vector machine model, i.e., the SVM model.
[0022] After the SVM model is trained, it is evaluated and supplemented by experts in the field of tobacco enterprise production scheduling to ensure its accuracy and completeness, and to keep it constantly updated, ultimately resulting in a semantic model of tobacco enterprise production scheduling IS that makes tacit knowledge explicit.
[0023] Furthermore, in step S3, in order to make the implicit semantics in the implicit scheduling knowledge IS metamodel explicit, a fact-oriented modeling method is adopted to construct the scheduling IS semantic model. Implicit fact terms are used to query the information semantics of the business domain, which can convert attributes into entities and relations to obtain knowledge units related to scheduling.
[0024] Ontologies, as conceptual modeling tools that describe information systems at the semantic and knowledge levels, have been widely applied in many fields since their introduction into the field of artificial intelligence in the early 1990s, including knowledge engineering, digital libraries, information retrieval, processing of heterogeneous information on the Web, and the Semantic Web. The tobacco company's production scheduling ontology encompasses technical data on the production of all cigarette brands and other cigarettes within the company, integrating all production-related business data. Utilizing ontology technology can reduce misunderstandings caused by different terminology or perspectives, and solve communication and exchange problems in the use of production scheduling knowledge with a unified architecture, thereby improving the efficiency and effectiveness of tobacco production planning and scheduling.
[0025] Compared with the prior art, the beneficial effects of the present invention are reflected in:
[0026] This invention extracts data from the internal production scheduling database of tobacco companies, constructs a scheduling ontology using explicit knowledge, and further improves the scheduling ontology by combining implicit knowledge explicitation and ontology mapping and fusion techniques. This makes the scheduling ontology more complete and comprehensive, so as to better meet the expected requirements. Attached Figure Description
[0027] Figure 1 The flowchart illustrates a method for making implicit knowledge explicit in tobacco enterprise production scheduling, as provided by this invention.
[0028] Figure 2 This is a schematic diagram illustrating the construction process of the scheduling implicit knowledge IS meta-model of the present invention.
[0029] Figure 3 This is a schematic diagram of the scheduling IS semantic model construction process of the present invention.
[0030] Figure 4 This is a schematic diagram of the scheduling ontology mapping framework of the present invention.
[0031] Figure 5 This is the mapping and transformation structure framework of the present invention.
[0032] Figure 6 This invention uses Protege to create a tobacco production scheduling ontology. Detailed Implementation
[0033] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.
[0034] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.
[0035] The present invention will now be described in detail with reference to the accompanying drawings and exemplary embodiments.
[0036] refer to Figures 1 to 6 The present invention provides a method for making tacit knowledge explicit in tobacco production scheduling, comprising the following steps:
[0037] Step S1: Extract and preprocess scheduling knowledge;
[0038] This invention extracts data and knowledge related to production scheduling from internal and external information systems of tobacco enterprises. After extracting the data from these systems, due to the significant differences between these multi-source heterogeneous data and the richness of the Chinese natural language, preprocessing such as data cleaning and impurity removal, word segmentation, part-of-speech tagging, entity recognition, and sentence structure analysis is required. This invention uses Natural Language Processing (NLP) technology for processing and employs the FoolNLTK tool to perform preprocessing.
[0039] Step S2: Construct a thesaurus of explicit and implicit scheduling knowledge;
[0040] Based on expert knowledge of tobacco companies’ production scheduling and relevant internal experience, a lexicon of explicit and implicit knowledge classification is constructed using natural language processing technology.
[0041] It is necessary to integrate the expert experience in production planning and scheduling of tobacco companies and construct a thesaurus of explicit and implicit knowledge related to production scheduling. This thesaurus contains a large number of keywords associated with explicit knowledge and can be continuously updated and supplemented during use.
[0042] After natural language processing, keywords corresponding to these data and knowledge can be obtained. By comparing these keywords with a classification lexicon of explicit and implicit knowledge using keyword matching technology, the classification of explicit and implicit knowledge can be achieved.
[0043] Step S3: Extract keywords from implicit and explicit knowledge to make implicit scheduling knowledge explicit;
[0044] Keyword extraction is performed on tacit and explicit knowledge. The similarity of keywords between the two is compared using an n-gram matching algorithm. If the similarity is greater than a certain threshold, a scheduling knowledge framework is constructed based on the explicit knowledge. This knowledge framework is then used to integrate the set of tacit knowledge keywords, thus making the tacit knowledge explicit. Knowledge integration is a dynamic process that allows for the reorganization of knowledge, discarding useless knowledge and organically integrating useful knowledge to make it more flexible, organized, and systematic. If necessary, the original knowledge system can also be reconstructed to form a new core knowledge system for the enterprise.
[0045] For the remaining tacit knowledge, an implicit scheduling knowledge information system conceptual UML metamodel is constructed, namely the implicit scheduling knowledge IS metamodel. This model does not contain any explicit knowledge, that is, it does not contain any explicit semantics. Its construction process includes two main steps: ① extraction of the database ER model; ② mapping the ER model to the UML conceptual model.
[0046] refer to Figure 2 The implicit scheduling knowledge information system concept UML metamodel construction method used in this invention can also be called the implicit scheduling knowledge IS metamodel. At this point, the model is a preliminary scheduling concept model and does not contain any clear explicit semantics.
[0047] To make the implicit semantics in the implicit scheduling knowledge IS metamodel explicit, a fact-oriented modeling approach is adopted to construct the scheduling IS semantic model. This method uses implicit factual terms to query the semantic information of the business domain, which can convert attributes into entities and relations to obtain scheduling-related knowledge units;
[0048] Step S4: Integrate explicit and implicit scheduling knowledge;
[0049] Based on explicit knowledge, the scheduling ontology is built from top to bottom using the OWL language and the Protege ontology development tool. OWL can describe the relationships between concepts, properties, and individuals to ensure powerful semantic expression capabilities and enable reasoning functions.
[0050] By employing methods such as ontology mapping and ontology fusion, semantic integration between the scheduling IS semantic model and the scheduling ontology is achieved, thus integrating explicit and implicit scheduling knowledge.
[0051] Compared to using only explicit knowledge from the enterprise information system database to build a scheduling ontology, combining implicit knowledge can make the constructed scheduling ontology more complete and better effective.
[0052] When making tacit knowledge explicit, the first step is to extract keywords from the processed tacit and explicit knowledge, and then compare the keyword similarity between the two using an n-gram matching algorithm. If the similarity m is greater than or equal to a set threshold n, a scheduling knowledge framework is constructed based on the explicit knowledge. This knowledge framework is then used to integrate the tacit knowledge keyword set, thus completing the explicitization of the tacit knowledge.
[0053] For implicit knowledge where the similarity m between two entities is less than a set threshold n, or where keyword extraction is not possible, a semantic model needs to be constructed to make it explicit. In this case, an implicit knowledge IS meta-model needs to be constructed, and then a series of model transformations are used to build the implicit knowledge IS meta-model into a semantic model.
[0054] The process of constructing the tacit knowledge IS metamodel includes two main steps: ① extraction of the database ER model; ② mapping the ER model to the UML conceptual model.
[0055] The steps for constructing the production scheduling IS semantic model include: extracting meaningful features from data related to tobacco enterprise production scheduling so that the model can understand and learn. These features include keywords, sentence structure, production scheduling basis, influencing factors, and other meaningful features; using the extracted explicit knowledge of tobacco enterprise production scheduling and the extracted meaningful features, combined with expert knowledge and experience in production scheduling as the model training set, to train a nonlinear support vector machine model, i.e., an SVM model; after the model is trained, it is evaluated and supplemented by experts in the field of tobacco enterprise production scheduling to ensure its accuracy and completeness, and to keep it continuously updated, ultimately obtaining a tobacco enterprise production scheduling IS semantic model that makes tacit knowledge explicit.
[0056] By combining these two methods, all tacit knowledge can be made explicit.
[0057] In constructing the scheduling ontology, this application adopts a top-down design approach, selecting the OWL language and utilizing the Protege ontology development tool. Ontology modeling is the process of establishing domain concepts and their relationships based on actual application requirements. In this process, the categories of objects of interest to users are considered first, followed by the concepts that describe these objects. If concept A depends on concept B, it indicates that concept A has an attribute whose value domain is an instance of concept B; therefore, to define concept A, concept B must be defined first. Based on these analyses, a lifecycle-oriented ontology construction method is chosen to complete the construction of the scheduling ontology.
[0058] An ontology, as a definition of concepts and a description of conceptual relationships, can be represented as a set: ONT = {set of concept classes}, {set of associations between concept classes}. Scheduled ontology modeling is the process of establishing domain concepts and their relationships based on actual application requirements. In this process, the categories of objects that users are interested in are considered first, and then the concepts that describe these objects are considered one by one. If concept A depends on concept B, it means that concept A has an attribute whose value domain is an instance of concept B; therefore, to complete the definition of concept A, concept B needs to be defined first. (Reference) Figure 6 The tobacco production scheduling ontology built using Protege includes six conceptual classes: product knowledge, manufacturing resources, manufacturing methods and tasks, organizational structure, and production enterprises. Each class has different subclasses and relationships.
[0059] The knowledge used in constructing the scheduling ontology is explicit knowledge. Therefore, the content of the scheduling ontology at this stage is not comprehensive enough. This application utilizes an existing method: ontology mapping, to complete the content integration between the scheduling IS semantic model and the scheduling ontology based on explicit knowledge.
[0060] refer to Figure 5 This document presents a mapping and transformation framework. Ontology mapping is a typical logic-based method that associates and maps concepts, entities, and relationships between different ontologies. Its main goal is to achieve semantic interoperability between different ontologies, and between ontologies and different databases or other semantic models, enabling them to understand and interact with each other. These mappings are typically constructed manually by domain experts or through various techniques, such as comparisons based on ontology structure or language semantics. Addressing the different knowledge structures, knowledge objects, and heterogeneous systems within the scheduling IS semantic model, this document employs different mapping methods to solve the semantic mapping problem, resolving the mapping between the scheduling ontology based on explicit knowledge and the scheduling IS semantic model.
[0061] To address the mapping issue between the scheduling ontology and the IS semantic model, this application first calculates the semantic similarity between the two through mapping. Building upon existing similarity calculation methods, this application considers ontology features such as concepts, attributes, relationships, and instances, integrating schema-level, instance-level, and structure-level methods to calculate concept similarity and complete the mapping between the scheduling IS semantic model and the scheduling ontology.
[0062] For mapping problems that semantic similarity calculation cannot solve, a reference method is adopted. Figure 6This application employs reasoning-based ontology fusion technology. It utilizes category theory to semi-automatically discover the semantic structure and core concepts of the model, thereby helping to identify related semantic concepts between the scheduling ontology and the scheduling IS semantic model. The application uses the Semantic Web Rule Language (SWRL) to describe the mapping rules of the ontology. Based on the defined SWRL mapping rules, the Pellet inference engine automatically calculates and infers similar or identical ontology concepts and reclassifies them.
[0063] Through this process, based on ontology mapping and ontology fusion technologies, the integration between the scheduling ontology and the scheduling IS semantic model is achieved, ultimately completing the integration of explicit and implicit scheduling knowledge and realizing the process of making implicit knowledge explicit.
[0064] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.
Claims
1. A method for making tacit knowledge explicit in tobacco enterprise production scheduling, characterized in that, Includes the following steps: Step S1: Extract and preprocess scheduling knowledge; Data and knowledge related to production scheduling are extracted from internal and external information systems of tobacco companies. After the data extraction of tobacco companies’ production scheduling-related information systems is completed, the FoolNLTK tool is used to preprocess the data. Step S2: Construct a thesaurus of explicit and implicit scheduling knowledge; Based on expert knowledge of tobacco companies’ production scheduling and relevant internal experience, a lexicon of explicit and implicit knowledge classification is constructed using natural language processing technology. The extracted data and knowledge are classified using a lexicon of explicit and implicit knowledge, which distinguishes between explicit and implicit knowledge. Step S3: Extract keywords from implicit and explicit knowledge to make implicit scheduling knowledge explicit; Keyword extraction is performed on tacit and explicit knowledge. The similarity between the two keywords is compared using an n-gram matching algorithm. If the similarity m is greater than or equal to the set threshold n, a scheduling knowledge framework is constructed based on the explicit knowledge. The scheduling knowledge framework is then used to integrate the set of tacit knowledge keywords to make the tacit knowledge explicit. Step S4: Integrate explicit and implicit scheduling knowledge; Based on explicit knowledge, the scheduling ontology is built from top to bottom using the OWL language and the Protege ontology development tool. OWL can describe the relationships between concepts, properties, and individuals to ensure powerful semantic expression capabilities and enable reasoning functions. By using ontology mapping and ontology fusion methods, semantic integration between the scheduling IS semantic model and the scheduling ontology is achieved, thus integrating explicit and implicit scheduling knowledge.
2. The method for making tacit knowledge explicit in tobacco enterprise production scheduling as described in claim 1, characterized in that, In step S3, for implicit knowledge and explicit knowledge whose keyword similarity m is less than the set threshold n or for implicit knowledge that cannot be extracted by keywords, it is necessary to construct a UML meta-model of implicit scheduling knowledge information system concept, namely the implicit scheduling knowledge IS meta-model.
3. The method for making tacit knowledge explicit in tobacco enterprise production scheduling as described in claim 2, characterized in that, The implicit scheduling knowledge IS metamodel does not contain any explicit knowledge, that is, it does not contain any explicit semantics. Its construction process includes two main steps: ① extraction of the database ER model; ② mapping of the ER model to the UML conceptual model.
4. The method for making tacit knowledge explicit in tobacco enterprise production scheduling as described in claim 3, characterized in that, The specific steps for constructing the scheduling IS semantic model include the following: Extract meaningful features from data related to tobacco companies' production scheduling so that the model can understand and learn; The explicit knowledge of tobacco enterprise production scheduling and the extracted meaningful features, combined with expert knowledge and experience in production scheduling, are used as the model training set to train the nonlinear support vector machine model, i.e., the SVM model. After the SVM model is trained, it is evaluated and supplemented by experts in the field of tobacco enterprise production scheduling to ensure its accuracy and completeness, and to keep it constantly updated, ultimately resulting in a semantic model of tobacco enterprise production scheduling IS that makes tacit knowledge explicit.
5. The method for making tacit knowledge explicit in tobacco enterprise production scheduling as described in claim 1, characterized in that, In step S3, in order to make the implicit semantics in the implicit scheduling knowledge IS metamodel explicit, a fact-oriented modeling method is adopted to construct the scheduling IS semantic model. Implicit fact terms are used to query the information semantics of the business domain, which can convert attributes into entities and relations to obtain knowledge units related to scheduling.