CN117291494A - Spare part inventory matching control method based on knowledge graph - Google Patents
Spare part inventory matching control method based on knowledge graph Download PDFInfo
- Publication number
- CN117291494A CN117291494A CN202210666257.5A CN202210666257A CN117291494A CN 117291494 A CN117291494 A CN 117291494A CN 202210666257 A CN202210666257 A CN 202210666257A CN 117291494 A CN117291494 A CN 117291494A
- Authority
- CN
- China
- Prior art keywords
- spare part
- graph
- knowledge
- clustering
- control method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 83
- 238000004458 analytical method Methods 0.000 claims abstract description 18
- 238000012545 processing Methods 0.000 claims abstract description 13
- 238000007726 management method Methods 0.000 claims description 51
- 238000005065 mining Methods 0.000 claims description 37
- 238000013461 design Methods 0.000 claims description 35
- 239000013598 vector Substances 0.000 claims description 15
- 238000012937 correction Methods 0.000 claims description 9
- 238000007621 cluster analysis Methods 0.000 claims description 8
- 230000003993 interaction Effects 0.000 claims description 8
- 238000004519 manufacturing process Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000012423 maintenance Methods 0.000 abstract description 13
- 239000000463 material Substances 0.000 description 16
- 238000010586 diagram Methods 0.000 description 12
- 238000005457 optimization Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 7
- 230000008439 repair process Effects 0.000 description 7
- 230000000007 visual effect Effects 0.000 description 7
- 230000004044 response Effects 0.000 description 6
- 238000010276 construction Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 229910000831 Steel Inorganic materials 0.000 description 3
- 230000002776 aggregation Effects 0.000 description 3
- 238000004220 aggregation Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 239000003795 chemical substances by application Substances 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 238000007429 general method Methods 0.000 description 3
- 239000010959 steel Substances 0.000 description 3
- 238000012790 confirmation Methods 0.000 description 2
- 238000007667 floating Methods 0.000 description 2
- 238000001556 precipitation Methods 0.000 description 2
- 238000005728 strengthening Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 230000007306 turnover Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/087—Inventory or stock management, e.g. order filling, procurement or balancing against orders
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/906—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Computational Linguistics (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- General Business, Economics & Management (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Artificial Intelligence (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Finance (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Accounting & Taxation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The spare part inventory matching control method based on the knowledge graph is used for establishing cooperative control between a device management system and a purchasing management system based on the spare part inventory, and setting an intelligent spare part retrieval system based on the knowledge graph; the spare part intelligent retrieval system is arranged independently of the equipment management system and the purchase management system in a mode of being capable of communicating with the equipment management system and the purchase management system; and setting a clustering analysis model based on a knowledge graph in the intelligent spare part retrieval system, and completing real-time dynamic clustering processing of the spare parts through the clustering analysis model so as to provide retrieval service with a clustering request for the purchase management system. According to the knowledge-graph-based spare part inventory matching control method, corresponding application of real-time dynamic clustering operation and setting is completed through the clustering analysis model established based on the knowledge graph, personalized dependence of general spare part search expansion on field experts is effectively reduced, the utilization rate of spare part inventory is improved, and maintenance cost investment of enterprises is effectively reduced.
Description
Technical Field
The invention belongs to the field of spare part inventory control in the steel manufacturing industry, and particularly relates to a spare part inventory matching control method based on a knowledge graph.
Background
The efficient management of spare parts is a work for long-term promotion of Bao-steel stock, is important content for reducing operation cost and improving turnover efficiency of two companies, and is also important content for the change of multi-base equipment management of one company. The development of the spare part inventory similarity retrieval system can realize the one-key inquiry function of new material inventory, repair inventory and machine side inventory information, provides a replaceable inventory information reference for new product purchase approval, is a powerful means for strengthening source management and control, is beneficial to improving inventory utilization efficiency and improves spare part plan management and inventory management level.
The management of the Bao-steel stock materials is based on the material codes and the 12-bit codes of the equipment, and the life management of the materials is realized through the code tree and the equipment tree in the information system. The material code (C code for short) is used as a key word for new product purchase, spare part repair, stock (including machine side) and machine up and down management, and is an important reference in material management. The 12-bit code of the equipment focuses on equipment information and records information such as the installed position, the installed quantity and the like of the spare parts. The C code and the 12-bit code rule of the equipment are in a tree structure, and a corresponding relation is established through comparison of the two codes, so that effective coordination of equipment management and purchase management is realized.
For each stock spare, a series of material description parameters, which are based on material codes, and other relevant stock information are stored in the data warehouse. The existing code of PSCS system is 98 ten thousand, with 67 ten thousand valid states and increasing at a rate of over ten thousand per year. The model specification information of two codes is mainly maintained in the system by spot inspection on site. In actual production, stock backlog and stock resources cannot be effectively utilized due to the current situations of incomplete material type rule information, incomplete technical parameters and one-object multi-code. Meanwhile, the material class and shape rule standards are very complex, some adopt international standards, some adopt factory specifications, on-site users lack comprehensive professional knowledge of material replacement, effective inventory resources are difficult to find under emergency conditions, and a passive situation of difficult material emergency response is caused.
Search engines are a typical leading technology in the internet era, whose implementation mechanism is quite different from information queries based on traditional relational database retrieval. Typical search engines can quickly find relevant documents based on text similarity search algorithms based on keywords entered by users, and return the results after sorting. The search engine technology reduces the requirement on keyword input, improves the quality of a search result set and the response speed of a search process, is widely applied in the Internet era, and is also a basic platform for realizing the similarity retrieval of spare parts.
In view of practical operation difficulty, key technical parameters of spare part materials are not completely covered in the spare part management system design, but the delivery model of the spare part materials is a necessary basic parameter in the system design, and the realization of alternative spare part retrieval through spare part model clustering is a feasible approach. A general search engine based on text analysis can realize the judgment of similarity between a motor and a motor, but cannot realize the judgment of similarity between motors of two types of Y180L-8 and 1LG6186-8, and an independent cluster analysis system is required to give a cluster rule base.
The traditional spare part clustering rule base management is usually highly dependent on manual work in the whole operation and maintenance work of equipment, has high dependency on experience and capability of people, and has the possibility of generating misjudgment and missed judgment. Knowledge points related to spare part clustering rule base management are complex, multidimensional field experience is needed, so that corresponding expert posts have high requirements on experience and capability, staffs are difficult to cultivate, and core expert knowledge and experience cannot be effectively accumulated and deposited.
The application number is: the invention application of CN201811085023.1 discloses a multi-agent-based maintenance resource bidirectional joint scheduling policy decision method, which can realize the joint scheduling of maintenance resources of multiple stages and multiple guarantee points and support the comprehensive scheduling decision of maintenance workers and spare parts after different equipment faults. The method comprises the following steps: 1. and (5) simulation modeling. Equipment, maintenance resources, etc. are packaged as individual agents and define attributes within each agent. 2. And (5) simulating configuration. The maintenance resource scheduling strategy, the fixed matching scheduling strategy, the shortest distance scheduling strategy and the maximum stock scheduling strategy are defined, and in addition, the simulation times are defined. 3. And generating a simulation evaluation result. And respectively counting the single simulation cost and the average cost after N times of simulation under three different dispatching strategies. 4. And (5) sorting maintenance resource scheduling strategies. And sequencing maintenance resource scheduling strategies according to the average cost.
The application number is: the invention application of CN201910193000.0 discloses a spare part inventory optimization method and device, and the method is applied to the technical fields of inventory control and logistics management. The method comprises the following steps: setting an availability index of an equipment spare part system; based on the availability index, establishing a spare part inventory optimization model by taking the availability of spare parts as a constraint condition and taking the minimum total guarantee cost of the spare parts as an optimization target; and solving the spare part inventory optimization model by adopting a marginal effect method, and determining an optimal spare part inventory scheme.
The application number is: the invention application of CN202011420720.5 discloses a method for configuring the joint inventory of maintenance and guarantee resources based on a hyper heuristic algorithm, which comprises the following steps: analyzing and maintaining a resource inventory system; constructing an objective function; constructing constraint conditions; generating an initial configuration scheme; optimizing the initial inventory configuration scheme and outputting a configuration result.
The application number is: the invention application of CN202210412990.4 discloses an intelligent matching method and device based on combination of a material model and a knowledge graph, wherein the method comprises the following steps: according to the types of different repair services, a plurality of sets of repair time standard-reaching rules are formulated; analyzing the received repair mail, and matching the equipment model and fault type information of the repair equipment obtained through analysis with a bill of materials recorded with spare part model information to obtain replaceable spare part product information; if the local warehouse does not have the replaceable spare part products, the replaceable spare part products of other warehouses are allocated to be delivered according to the nearby matching principle.
Disclosure of Invention
In order to solve the problems, the invention provides a spare part inventory matching control method based on a knowledge graph, which has the following technical scheme:
the utility model provides a spare part inventory matching control method based on a knowledge graph, which is used for establishing cooperative control between a device management system and a purchasing management system based on spare part inventory, and is characterized in that:
setting an intelligent searching system of spare parts based on a knowledge graph;
the spare part intelligent retrieval system is arranged independently of the equipment management system and the purchase management system in a mode of being capable of communicating with the equipment management system and the purchase management system;
and setting a clustering analysis model based on a knowledge graph in the intelligent spare part retrieval system, and completing real-time dynamic clustering processing of the spare parts through the clustering analysis model so as to provide retrieval service with a clustering request for a purchase management system.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
based on the practice of the actual scene of field production and fault rush-repair and the technical parameters of two dimensions of a manufacturer product sample manual, a triplet data file is established by combining entity design, entity attribute design, relationship design and relationship attribute design in the knowledge graph, and a general spare part knowledge graph is generated and a data source for operation of a cluster analysis model is formed.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
and the cluster analysis model completes cluster mining in a mode of establishing an undirected graph and solving the maximum group of the undirected graph according to the generated data source represented by the general spare part knowledge graph and the set expert rules.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
the undirected graph takes each independent spare part in the universal spare part knowledge graph as a node, and takes specific judging conditions in expert rules and the consistency judging result of the conditions applied between the two spare parts as edges to finish establishment.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
solving the maximum clique of the undirected graph is performed in a manner of processing deterministic rules in expert rules first and then processing the remaining rules.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
after the first solving of the maximum group of the undirected graph is completed, judging the result based on the set rule, and outputting the result as a final result if the result meets the set rule; otherwise, correcting the expert rules which do not accord with the set rules, and carrying out the solving of the maximum group again after finishing the correction, and repeating the steps until the set rules accord with.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
and (3) finishing the adjacent matrix representation of the undirected graph based on the network X of Python, and finishing the solving operation of the maximum clique based on the solving maximum clique function of the network X.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
triggering operation by the real-time dynamic clustering process according to the set triggering condition;
the triggering conditions include:
when the spare part stock entity enters, leaves and corrects;
when the stock types of spare parts are newly increased and corrected;
and when the spare part clustering rule changes.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
triggering a clustering request at the same time when a purchase management system sends a search request to a spare part intelligent search system;
and after receiving the clustering request, the spare part intelligent retrieval system performs corresponding retrieval in the clustering result completed by the clustering analysis model.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
after receiving the search request, the spare part intelligent search system calculates word frequency of any one of the received search information text and the index library after word segmentation is established, and forms search result recommendation according to calculation results.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
and establishing a search instruction word frequency direction U based on the search information, establishing a spare part characteristic word frequency vector V based on an index library, and forming search result recommendation by calculating the cosine similarity of inner product space included angles between vectors U, V.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
the spare part characteristic word frequency vector V comprises a spare part self characteristic label and a spare part clustering label.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
and giving corresponding weights to the search results according to the distance from the physical distance of the search position, and generating the ordered recommendation of the priority order according to the weights.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
the mining result forms two output forms of a webpage list and an Excel form.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
and forming the presentation of the mining result through the set man-machine interaction interface.
The invention relates to a spare part inventory matching control method based on a knowledge graph, which is characterized by comprising the following steps of:
the presentation of the mining results includes: global result display, inter-group comparison result display and intra-group comparison result display;
the comparison result display among the groups focuses on displaying the difference between every two clustering groups, and the comparison result display among the groups focuses on displaying the commonality inside the same clustering group.
According to the spare part inventory matching control method based on the knowledge graph, on the basis of referencing the existing spare part management system management logic and the data structure, the knowledge graph mode design of the universal spare part is formed according to field practice experience. And (3) utilizing a special knowledge graph platform supporting tool to compare the entity design, entity attribute design, relationship design and relationship attribute design in the general spare part knowledge graph mode with the general spare part core technical parameters mainly derived from the product sample manual of the manufacturer to generate a general spare part knowledge graph. The system support field expert builds and manages alternative general spare part clustering rules (expert rules) to infer one or more alternative general spare part clustering groups on the existing general spare part map. The system provides a visual interaction interface for spare part clustering rule mining model construction and iterative optimization, wherein the visual interaction interface comprises three parts, namely global result display, inter-group comparison result display and intra-group comparison display. The global result display page can select a plurality of groups at most, and select any two rules in the mining rule group to display the scatter diagram, so that inter-group and intra-group aggregation effects are intuitively displayed. The inter-group comparison result display page displays left and right comparison graphs of various indexes of the two motor groups under the set rule, and focuses on displaying the difference between the two clustering groups. The intra-group comparison result display page displays the left-right comparison graphs of the indexes of the two bodies in the same group under the set rule, focuses on displaying the commonality of the inside of the same cluster group, and improves the efficiency of iterative optimization of domain experts. In order to overcome the defect of slow search response time in a big data environment in a general method, the method is specially designed when the spare part clustering rule-based mining result is integrated in a spare part retrieval system. And adding a clustering label in the index library of the retrieval system, mining a result according to the spare part clustering rule and operating the retrieval system practically, and dynamically realizing real-time clustering of the index library of the retrieval system in the background. The real-time clustering trigger event includes: the stock entity of spare parts is subjected to warehouse entry, warehouse exit, correction and other changes; the stock types of spare parts are changed such as new addition, correction and the like; the spare part clustering rule changes; the knowledge graph of the general spare parts changes, etc. When the search system receives a search request from a user, a clustering request is added according to the spare part clustering rule mining result on the basis of the initialization request, and the search request is integrated into the same search set according to different search weights after the search is completed and fed back to the user.
In summary, the knowledge-graph-based spare part inventory matching control method provided by the invention completes real-time dynamic clustering operation and corresponding established application through the clustering analysis model established based on the knowledge graph, effectively reduces personalized dependence of general spare part search expansion on domain experts, is beneficial to effective precipitation and inheritance of internal knowledge of enterprises, improves the utilization rate of spare part inventory, and effectively reduces investment of maintenance cost of the enterprises.
Drawings
FIG. 1 is a schematic diagram of the structural principle of the present invention;
FIG. 2 is a schematic diagram of a knowledge graph of a general motor in an embodiment of the invention;
FIG. 3 is a diagram of a knowledge graph logic architecture in an embodiment of the present invention;
FIG. 4 is a diagram of a business person and business specialist executable use case in an embodiment of the present invention;
FIG. 5 is a schematic diagram of a business expert in an embodiment of the present invention running a constructed rule;
FIG. 6 is a schematic flow chart of an excavation algorithm in an embodiment of the present invention;
FIG. 7 is a diagram showing global results in an embodiment of the present invention;
FIG. 8 is a schematic diagram showing the comparison results between groups in the embodiment of the present invention;
fig. 9 is a schematic diagram showing the results of intra-group comparison in the embodiment of the present invention.
Detailed Description
The invention further provides a spare part inventory matching control method based on a knowledge graph according to the specification, the drawings and the specific embodiments.
The spare part inventory matching control method based on the knowledge graph is used for establishing cooperative control between a device management system and a purchasing management system based on the spare part inventory, and setting an intelligent spare part retrieval system based on the knowledge graph;
the spare part intelligent retrieval system is arranged independently of the equipment management system and the purchase management system in a mode of being capable of communicating with the equipment management system and the purchase management system;
and setting a clustering analysis model based on a knowledge graph in the intelligent spare part retrieval system, and completing real-time dynamic clustering processing of the spare parts through the clustering analysis model so as to provide retrieval service with a clustering request for a purchase management system.
Wherein,
based on the practice of the actual scene of field production and fault rush-repair and the technical parameters of two dimensions of a manufacturer product sample manual, a triplet data file is established by combining entity design, entity attribute design, relationship design and relationship attribute design in the knowledge graph, and a general spare part knowledge graph is generated and a data source for operation of a cluster analysis model is formed.
Wherein,
and the cluster analysis model completes cluster mining in a mode of establishing an undirected graph and solving the maximum group of the undirected graph according to the generated data source represented by the general spare part knowledge graph and the set expert rules.
Wherein,
the undirected graph takes each independent spare part in the universal spare part knowledge graph as a node, and takes specific judging conditions in expert rules and the consistency judging result of the conditions applied between the two spare parts as edges to finish establishment.
Wherein,
solving the maximum clique of the undirected graph is performed in a manner of processing deterministic rules in expert rules first and then processing the remaining rules.
Wherein,
after the first solving of the maximum group of the undirected graph is completed, judging the result based on the set rule, and outputting the result as a final result if the result meets the set rule; otherwise, correcting the expert rules which do not accord with the set rules, and carrying out the solving of the maximum group again after finishing the correction, and repeating the steps until the set rules accord with.
Wherein,
and (3) finishing the adjacent matrix representation of the undirected graph based on the network X of Python, and finishing the solving operation of the maximum clique based on the solving maximum clique function of the network X.
Wherein,
triggering operation by the real-time dynamic clustering process according to the set triggering condition;
the triggering conditions include:
when the spare part stock entity enters, leaves and corrects;
when the stock types of spare parts are newly increased and corrected;
and when the spare part clustering rule changes.
Wherein,
triggering a clustering request at the same time when a purchase management system sends a search request to a spare part intelligent search system;
and after receiving the clustering request, the spare part intelligent retrieval system performs corresponding retrieval in the clustering result completed by the clustering analysis model.
Wherein,
after receiving the search request, the spare part intelligent search system calculates word frequency of any one of the received search information text and the index library after word segmentation is established, and forms search result recommendation according to calculation results.
Wherein,
and establishing a search instruction word frequency direction U based on the search information, establishing a spare part characteristic word frequency vector V based on an index library, and forming search result recommendation by calculating the cosine similarity of inner product space included angles between vectors U, V.
Wherein,
the spare part characteristic word frequency vector V comprises a spare part self characteristic label and a spare part clustering label.
Wherein,
and giving corresponding weights to the search results according to the distance from the physical distance of the search position, and generating the ordered recommendation of the priority order according to the weights.
Wherein,
the mining result forms two output forms of a webpage list and an Excel form.
Wherein,
and forming the presentation of the mining result through the set man-machine interaction interface.
Wherein,
the presentation of the mining results includes: global result display, inter-group comparison result display and intra-group comparison result display;
the comparison result display among the groups focuses on displaying the difference between every two clustering groups, and the comparison result display among the groups focuses on displaying the commonality inside the same clustering group.
Working principles, procedures and examples
For the convenience of visual understanding, the following description will be made with the motor as the body. And is to be understood in general in conjunction with the drawings.
The specific implementation process is as follows:
5.1 general Motor knowledge graph Pattern design
Based on reference of management logic and data structure of ready-to-use spare part management system, core technical parameters of the spare part are selected according to practical experience of actual scene of on-site production and fault first-aid repair, and a knowledge graph special technology is combined to complete a universal motor knowledge graph mode as shown in figure 2.
5.2 general Motor spare part knowledge graph Generation
The knowledge graph platform support tool comprises a whole set of knowledge graph full-flow design construction work, and performs platform and tool function integration and optimization from original data source (including structured data and unstructured data) and data set management to design and construction of a graph and graph management and application, and comprises the following steps: the system comprises a map home page, map management, map application, timing diagram application, machine learning modeling, map mode design, entity and relationship labeling, data source management, data set management, knowledge acquisition, system management, similar recommendation and other functional modules.
And (3) utilizing a knowledge graph platform supporting tool to compare entity design, entity attribute design, relationship design and relationship attribute design in a general spare part knowledge graph mode with general spare part core technical parameters mainly derived from a manufacturer product sample manual to form a technical parameter triplet data file corresponding to the general motor knowledge graph mode, and further importing the technical parameter triplet data file into a knowledge graph platform to complete general spare part knowledge graph generation (see figure 3).
5.3 general motor spare part clustering rule mining model
The system supports a user to construct and manage alternative general spare part clustering rules (expert rules), sets entity attribute value constraints associated with general spare parts in an ontology, constructs a set of alternative general spare part group judging rules, and can infer one or more alternative general spare part clustering groups on an existing general spare part map by applying the rules on the general spare part map.
General spare part clustering rules (expert rules) define:
1) The expert rule base comprises n expert rules;
2) Each expert rule contains n general spare part attribute parameters, and defines different parameter categories and parameter judgment standards;
3) Each expert rule may generate a plurality of general spare part groupings, different general spare part groupings including alternative general spare part models within different numerical ranges under the same class of parameter categories and parameter judgment criteria.
The system user may perform as shown, for example, in fig. 4.
The business specialist can run the built rules, run the mining general motor spare part group,
1) When a new rule is to be added,
2) When an existing rule is to be modified,
3) After the entity is newly entered, the corresponding flow is shown in fig. 5.
The constraint condition of the attribute condition designed according to expert rules in the method comprises the following steps:
attribute data type | Judgment condition |
Enumerating type attributes | Identical, non-limiting |
Numerical attributes | Equal, unequal, floating by x%, not limited |
The general spare part map only stores various characteristics of the spare part itself, but does not contain a comparison relation between the various characteristics of different spare parts. For a given set of expert rules, it is difficult to directly mine alternative universal spare part clusters on a universal spare part map. In order to solve the convenience, each independent spare part in the universal spare part map is taken as a node, and a specific judging condition in expert rules and a coincidence judging result of the condition applied between the two spare parts are taken as edges to form a new undirected graph. Thereby converting the original problem into a maximum clique solving problem in the undirected graph. The maximum group problem is an NP complete problem, and in a specific solution, the efficiency of solving the maximum group of the problem can be effectively improved by firstly processing a deterministic rule, such as 'motor installation size aa is equal', dividing the whole graph into sub-graphs with smaller scales and then processing a flexible rule, such as 'motor rated power difference < = 1%'.
NetworkX is a widely used graph algorithm Python toolkit, in which the biggest group solution tool is based on the implementation of an improved algorithm proposed in 1973 by Bron and Kerbosch (1973), 2006 Tomita, tanaka, takahashi, etc. The knowledge graph mining flow is shown in fig. 6.
5.4 visual interaction interface design for spare part clustering rule mining model construction and iterative optimization
The visual interface design comprises two parts of rule setting, execution and mining result viewing.
The rule setting function is divided into three parts of setting basic information, setting rules, and confirmation information. The set basic information page contains the names of the rules, descriptions of the rules and mining targets. The setting rule page can select the pattern mode design to set the rules of the set entity and the attribute on the entity, and can set different rules according to different data types (integer values, floating point values can be set equal, > =, <=, relative values are between absolute values, enumeration strings can be set containing relations, text strings can be set equal). The confirmation information page can confirm the information set in the first step and the second step and create a rule.
The mining result display function comprises three parts, namely global result display, inter-group comparison result display and intra-group comparison display. The global result display page can select a plurality of groups at most, and any two rules in the mining rule group are selected to display the scatter diagram, so that inter-group and intra-group aggregation effects are intuitively displayed (see fig. 7).
The inter-group comparison result display page displays a left-right comparison graph of the respective indexes of the two motor groups under the set rule, focusing on displaying the difference between the two cluster groups (see fig. 8).
The intra-group comparison result display page displays a left-right comparison chart of each index of two motors in the same group under the set rule, focusing on displaying the commonality inside the same cluster group (see fig. 9).
There are two forms of final output of the mining result: checking mining results in two modes of a webpage list and an Excel table; the output result in the form of an Exclel table can be directly integrated into an online spare part retrieval system as a configuration file.
5.5 spare part retrieval system strengthening method based on spare part clustering rule mining result
In order to overcome the defect of slow search response time in a big data environment in a general method, unlike a general processing method, the method is specially designed when the spare part clustering rule-based mining result is integrated in a spare part retrieval system:
conventional processing patterns to be applied in user searches based on spare part clustering rule mining results may result in a spare part retrieval system that is too long in response time. The method dynamically realizes the real-time clustering (4 dynamic clustering) of the index library of the retrieval system in the background by adding a cluster label in the index library of the system in advance. The dynamic clustering trigger event includes: the stock entity of spare parts is subjected to warehouse entry, warehouse exit, correction and other changes; the stock types of spare parts are changed such as new addition, correction and the like; the spare part clustering rule changes; the general spare part knowledge graph changes, etc. (triggering comes from 3 index library updating). The criteria for dynamic clustering are derived from the rule-based knowledge mining output described in section 5.3 (1 cluster analysis model). The specific implementation form of dynamic clustering is to assign a cluster label to each independent spare part in a spare part inventory index library to a feature mark overall of a group of spare parts which can be replaced with each other, for example, a group of motor model codes of motors which can be replaced with each other.
When the search system receives a search request from a user, a clustering request is added according to the spare part clustering rule mining result on the basis of the initialization request, and the search request is integrated into the same search set according to different search weights after the search is completed and fed back to the user. The specific implementation is as follows:
in order to improve the effectiveness of a search system, the method is specially designed for front-end user access, specifically, in the spare part searching process, operators are used to copy a section of equipment description text to directly serve as a search key word, a large amount of invalid information such as 'model specification', 'random spare parts', 'maintenance supplies' and the like are often contained, and meanwhile, the low-value information is also commonly stored in a spare part inventory index library of a background. The background search essentially comprises the steps of carrying out word frequency calculation on a group of given search keyword texts and any example existing in an index library after word segmentation to form a search instruction word frequency vector u and a spare part characteristic word frequency vector v, and recommending a search result on the basis of calculating the cosine similarity of an inner product space included angle between the vectors u and v. The search keywords submitted by the user are simply adopted, so that the ineffective interference exists in the corresponding word frequency vector. The technology performs targeted search instruction preprocessing based on the statistics of the existing spare part inventory information, eliminates the similar invalid information, highlights the essential intention characteristics of search, and effectively improves the search quality (6 semantic analysis).
For spot inspectors, if spare parts of the same manufacturer type and specification and spare parts of different manufacturer types exist at the same time but can be replaced by each other, the spare parts of the same manufacturer type and specification are required to be arranged in the front in search. In order to achieve the purpose, when a search instruction is generated, the method improves the effectiveness of a search instruction word frequency vector u through (6 semantic analysis), and the spare part characteristic word frequency vector v simultaneously comprises a spare part self characteristic label v1 and a spare part clustering label v2, so that the spare parts with the same manufacturer model specification are arranged in the front (7 search instruction model) in a search result.
According to the spare part inventory matching control method based on the knowledge graph, on the basis of referencing the existing spare part management system management logic and the data structure, the knowledge graph mode design of the universal spare part is formed according to field practice experience. And (3) utilizing a special knowledge graph platform supporting tool to compare the entity design, entity attribute design, relationship design and relationship attribute design in the general spare part knowledge graph mode with the general spare part core technical parameters mainly derived from the product sample manual of the manufacturer to generate a general spare part knowledge graph. The system support field expert builds and manages alternative general spare part clustering rules (expert rules) to infer one or more alternative general spare part clustering groups on the existing general spare part map. The system provides a visual interaction interface for spare part clustering rule mining model construction and iterative optimization, wherein the visual interaction interface comprises three parts, namely global result display, inter-group comparison result display and intra-group comparison display. The global result display page can select a plurality of groups at most, and select any two rules in the mining rule group to display the scatter diagram, so that inter-group and intra-group aggregation effects are intuitively displayed. The inter-group comparison result display page displays left and right comparison graphs of various indexes of the two motor groups under the set rule, and focuses on displaying the difference between the two clustering groups. The intra-group comparison result display page displays the left-right comparison graphs of the indexes of the two bodies in the same group under the set rule, focuses on displaying the commonality of the inside of the same cluster group, and improves the efficiency of iterative optimization of domain experts. In order to overcome the defect of slow search response time in a big data environment in a general method, the method is specially designed when the spare part clustering rule-based mining result is integrated in a spare part retrieval system. And adding a clustering label in the index library of the retrieval system, mining a result according to the spare part clustering rule and operating the retrieval system practically, and dynamically realizing real-time clustering of the index library of the retrieval system in the background. The real-time clustering trigger event includes: the stock entity of spare parts is subjected to warehouse entry, warehouse exit, correction and other changes; the stock types of spare parts are changed such as new addition, correction and the like; the spare part clustering rule changes; the knowledge graph of the general spare parts changes, etc. When the search system receives a search request from a user, a clustering request is added according to the spare part clustering rule mining result on the basis of the initialization request, and the search request is integrated into the same search set according to different search weights after the search is completed and fed back to the user.
In summary, the knowledge-graph-based spare part inventory matching control method provided by the invention completes real-time dynamic clustering operation and corresponding established application through the clustering analysis model established based on the knowledge graph, effectively reduces personalized dependence of general spare part search expansion on domain experts, is beneficial to effective precipitation and inheritance of internal knowledge of enterprises, improves the utilization rate of spare part inventory, and effectively reduces investment of maintenance cost of the enterprises.
Claims (16)
1. The utility model provides a spare part inventory matching control method based on a knowledge graph, which is used for establishing cooperative control between a device management system and a purchasing management system based on spare part inventory, and is characterized in that:
setting an intelligent searching system of spare parts based on a knowledge graph;
the spare part intelligent retrieval system is arranged independently of the equipment management system and the purchase management system in a mode of being capable of communicating with the equipment management system and the purchase management system;
and setting a clustering analysis model based on a knowledge graph in the intelligent spare part retrieval system, and completing real-time dynamic clustering processing of the spare parts through the clustering analysis model so as to provide retrieval service with a clustering request for a purchase management system.
2. The knowledge-graph-based spare part inventory matching control method as claimed in claim 1, wherein the method comprises the following steps:
based on the practice of the actual scene of field production and fault rush-repair and the technical parameters of two dimensions of a manufacturer product sample manual, a triplet data file is established by combining entity design, entity attribute design, relationship design and relationship attribute design in the knowledge graph, and a general spare part knowledge graph is generated and a data source for operation of a cluster analysis model is formed.
3. The knowledge-graph-based spare part inventory matching control method as claimed in claim 1, wherein the method comprises the following steps:
and the cluster analysis model completes cluster mining in a mode of establishing an undirected graph and solving the maximum group of the undirected graph according to the generated data source represented by the general spare part knowledge graph and the set expert rules.
4. The knowledge-graph-based spare part inventory matching control method according to claim 3, wherein the method comprises the following steps of:
the undirected graph takes each independent spare part in the universal spare part knowledge graph as a node, and takes specific judging conditions in expert rules and the consistency judging result of the conditions applied between the two spare parts as edges to finish establishment.
5. The knowledge-graph-based spare part inventory matching control method according to claim 3, wherein the method comprises the following steps of:
solving the maximum clique of the undirected graph is performed in a manner of processing deterministic rules in expert rules first and then processing the remaining rules.
6. The knowledge-graph-based spare part inventory matching control method according to claim 3, wherein the method comprises the following steps of:
after the first solving of the maximum group of the undirected graph is completed, judging the result based on the set rule, and outputting the result as a final result if the result meets the set rule; otherwise, correcting the expert rules which do not accord with the set rules, and carrying out the solving of the maximum group again after finishing the correction, and repeating the steps until the set rules accord with.
7. The knowledge-graph-based spare part inventory matching control method according to claim 3, wherein the method comprises the following steps of:
and (3) finishing the adjacent matrix representation of the undirected graph based on the network X of Python, and finishing the solving operation of the maximum clique based on the solving maximum clique function of the network X.
8. The knowledge-graph-based spare part inventory matching control method as claimed in claim 1, wherein the method comprises the following steps:
triggering operation by the real-time dynamic clustering process according to the set triggering condition;
the triggering conditions include:
when the spare part stock entity enters, leaves and corrects;
when the stock types of spare parts are newly increased and corrected;
and when the spare part clustering rule changes.
9. The knowledge-graph-based spare part inventory matching control method as claimed in claim 1, wherein the method comprises the following steps:
triggering a clustering request at the same time when a purchase management system sends a search request to a spare part intelligent search system;
and after receiving the clustering request, the spare part intelligent retrieval system performs corresponding retrieval in the clustering result completed by the clustering analysis model.
10. The knowledge-graph-based spare part inventory matching control method as claimed in claim 9, wherein the method comprises the following steps:
after receiving the search request, the spare part intelligent search system calculates word frequency of any one of the received search information text and the index library after word segmentation is established, and forms search result recommendation according to calculation results.
11. The knowledge-graph-based spare part inventory matching control method as claimed in claim 10, wherein the method comprises the following steps:
and establishing a search instruction word frequency direction U based on the search information, establishing a spare part characteristic word frequency vector V based on an index library, and forming search result recommendation by calculating the cosine similarity of inner product space included angles between vectors U, V.
12. The knowledge-graph-based spare part inventory matching control method as claimed in claim 11, wherein the method comprises the following steps:
the spare part characteristic word frequency vector V comprises a spare part self characteristic label and a spare part clustering label.
13. The knowledge-graph-based spare part inventory matching control method as claimed in claim 10, wherein the method comprises the following steps:
and giving corresponding weights to the search results according to the distance from the physical distance of the search position, and generating the ordered recommendation of the priority order according to the weights.
14. The knowledge-graph-based spare part inventory matching control method according to claim 3, wherein the method comprises the following steps of:
the mining result forms two output forms of a webpage list and an Excel form.
15. The knowledge-graph-based spare part inventory matching control method according to claim 3, wherein the method comprises the following steps of:
and forming the presentation of the mining result through the set man-machine interaction interface.
16. The knowledge-graph-based spare part inventory matching control method as claimed in claim 15, wherein the method comprises the following steps:
the presentation of the mining results includes: global result display, inter-group comparison result display and intra-group comparison result display;
the comparison result display among the groups focuses on displaying the difference between every two clustering groups, and the comparison result display among the groups focuses on displaying the commonality inside the same clustering group.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210666257.5A CN117291494A (en) | 2022-06-14 | 2022-06-14 | Spare part inventory matching control method based on knowledge graph |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210666257.5A CN117291494A (en) | 2022-06-14 | 2022-06-14 | Spare part inventory matching control method based on knowledge graph |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117291494A true CN117291494A (en) | 2023-12-26 |
Family
ID=89255813
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210666257.5A Pending CN117291494A (en) | 2022-06-14 | 2022-06-14 | Spare part inventory matching control method based on knowledge graph |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117291494A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117933883A (en) * | 2024-03-21 | 2024-04-26 | 释普信息科技(上海)有限公司 | Intelligent classification management method and device based on inventory intelligent lock |
-
2022
- 2022-06-14 CN CN202210666257.5A patent/CN117291494A/en active Pending
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117933883A (en) * | 2024-03-21 | 2024-04-26 | 释普信息科技(上海)有限公司 | Intelligent classification management method and device based on inventory intelligent lock |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Peng et al. | An integrated multi-criteria decision-making framework for sustainable supplier selection under picture fuzzy environment | |
Xiao et al. | Green supplier selection in steel industry with intuitionistic fuzzy Taxonomy method | |
Su | A hybrid fuzzy approach to fuzzy multi-attribute group decision-making | |
CN104573106A (en) | Intelligent urban construction examining and approving method based on case-based reasoning technology | |
Falatoonitoosi et al. | Modeling for green supply chain evaluation | |
Wu et al. | An approach of product usability evaluation based on Web mining in feature fatigue analysis | |
CN104077407A (en) | System and method for intelligent data searching | |
Yu et al. | An integrated sustainable supplier selection approach using compensatory and non-compensatory decision methods | |
Agrawal et al. | Benchmarking fuzzy logic and ANFIS approaches for leanness evaluation in an Indian SME: a case study | |
CN105138588A (en) | Database overlap mode abstract generating method based on multi-label propagation | |
Kenaria et al. | Selection of factors affecting the supply chain and green suppliers by the TODIM method in the dairy industry | |
Ren et al. | Multifactor correlation analysis and modeling for product low-carbon design | |
CN105740434A (en) | Network information scoring method and device | |
CN117291494A (en) | Spare part inventory matching control method based on knowledge graph | |
Zhang et al. | The Dominance Degree‐Based Heterogeneous Linguistic Decision‐Making Technique for Sustainable 3PRLP Selection | |
Hwang et al. | Web-based multi-attribute analysis model for make-or-buy decisions | |
Zaefarian et al. | Variance decomposition analysis: What is it and how to perform it–A complete guide for B2B researchers | |
Huang | [Retracted] Green Supply Chain Management and Its Impact on Economic‐Environmental Performance: Evidence from Asian Countries | |
Zhong et al. | Sustainable supply chain partner selection and order allocation: A hybrid fuzzy PL-TODIM based MCGDM approach | |
Teniwut et al. | Spatial intelligent decision support system for increasing productivity on natural rubber agroindustry by green productivity approach | |
Ali et al. | Identification of critical factors for the implementation of reverse logistics in the manufacturing industry of Pakistan | |
CN110310012A (en) | Data analysing method, device, equipment and computer readable storage medium | |
CN115438995B (en) | Business processing method and equipment for clothing customization enterprise based on knowledge graph | |
CN101334793B (en) | Method for automatic recognition for dependency relationship of demand | |
Chen et al. | Shifting bottleneck-driven TOCh for solving product mix problems |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |