Network relationship analysis method and system based on intelligent devices in smart factory manufacturing

Abstract

The application relates to the technical field of Internet of Things, in particular to a network relationship analysis method and system based on intelligent equipment in intelligent factory manufacturing, which comprises the following steps: collecting interaction, communication and cooperation data of equipment, and performing cleaning and standardization processing; abstracting the equipment into network nodes and constructing a network graph model; calculating the centrality index of the nodes, identifying key nodes, identifying node clusters through a spectral clustering algorithm, and constructing the group structure of the key nodes; analyzing the connection relationship between the nodes, calculating the optimal connection path of the key nodes and other nodes by using a weighted shortest path algorithm, and adjusting the node connection in the network according to the calculation result. The method can effectively analyze the network relationship of the equipment in the intelligent factory, optimize the connection between the equipment, and improve the production efficiency and system robustness.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things, and in particular to a network relationship analysis method and system based on smart devices in smart factory manufacturing. Background Art

[0002] Smart devices involve numerous complex interactions and communication relationships when working together, making network relationship analysis increasingly important. By analyzing the network relationships of smart devices, we can effectively uncover collaborative patterns between devices, optimize production processes, and improve overall operational efficiency.

[0003] Existing methods for analyzing smart device network relationships are often simplistic and lack comprehensiveness, making it difficult to fully reveal the complex interaction patterns between devices. Many methods are limited to analyzing a single data source (such as interaction logs or communication records), while ignoring the combined effects of multiple data sources. Furthermore, existing methods have limitations in identifying key nodes, constructing node clusters, and optimizing connection paths, failing to fully consider the balance between the global and local characteristics of the network structure.

[0004] Regarding the network relationships of smart devices, it is particularly necessary to propose a comprehensive analysis method based on smart device interaction logs, communication records and collaborative data to more accurately reveal the correlation between devices, identify key nodes and their cluster structure, and further improve the overall efficiency and stability of the smart factory manufacturing system by optimizing the connection relationship between nodes.

[0005] In view of this, the present invention proposes a network relationship analysis method and system based on smart devices in smart factory manufacturing. Summary of the Invention

[0006] To achieve the above objectives, the present invention provides a network relationship analysis method and system based on smart devices in smart factory manufacturing. The specific technical solutions are as follows:

[0007] The network relationship analysis method based on smart devices in smart factory manufacturing includes:

[0008] Collect interaction logs, communication records, and production collaboration data from smart devices, and perform data cleaning and standardization;

[0009] Abstract smart devices as nodes in the network, use the interactions and communication behaviors between devices as edges, and build a network graph model that reflects the node relationships;

[0010] Calculate the degree centrality, betweenness centrality and closeness centrality indicators of the nodes, set thresholds, and identify key nodes in the network based on the thresholds;

[0011] Based on the identified key nodes, the spectral clustering algorithm is used to identify node clusters and construct the group structure of key nodes in the network;

[0012] Based on the identified key nodes and the group structure of key nodes, the connection relationship between nodes is analyzed, and the weighted shortest path algorithm is used to calculate the optimal connection path between key nodes and other nodes. According to the calculation results, the connection between key nodes and nodes inside and outside the group is adjusted in the network.

[0013] Preferably, the interaction logs, communication records and production collaboration data of smart devices are collected and data cleaning and standardization is performed, including: collecting the interaction logs of smart devices and recording them as data sets ,in Indicates the Interaction log data, is the total number of interaction logs;

[0014] Collect communication records between devices and record them as data sets ,in Indicates the Communication records, is the total number of communication records;

[0015] Collect collaborative data of equipment in the production process, recorded as data set ,in Indicates the Collaborative data, is the total number of collaborative data;

[0016] Preprocess the collected data, including deduplication, format unification, missing value filling, data transcoding, and numerical data normalization;

[0017] The preprocessed dataset 、 and Denoted as 、 and .

[0018] Preferably, the smart devices are abstracted as nodes in the network, and the interaction and communication behaviors between devices are used as edges, including: pre-processing the smart device interaction log , communication records and production collaboration data Extract the smart device ID from the smart device ID to form a smart device ID set ,in Indicates the Smart devices, is the total number of smart devices;

[0019] Collected by smart device ID As a set of nodes in a network diagram, smart devices Abstract into nodes , each node Represents a smart device;

[0020] According to the interaction log , extract the interaction relationship between smart devices and construct directed edge sets ; For each record in the interaction log ,in is the source smart device ID, is the target smart device ID, is the interaction time, For interaction type, in the node and Directed edges between , the edge weight is ;

[0021] According to the communication records , extract the communication relationship between smart devices and construct directed edge sets ; For each record in the communication record ,in is the source smart device ID, is the target smart device ID, is the communication time, For the communication protocol, at the node and Directed edges between , the edge weight is ;

[0022] According to collaborative data , extract the collaborative relationship between smart devices and construct an undirected edge set ; For each record in the collaborative data ,in The ID of the smart device participating in the collaboration, For collaborative time, For collaborative tasks, here Undirected edges are established between smart device nodes to form undirected edges, and the edge weight is .

[0023] Preferably, constructing a network graph model reflecting node relationships includes: constructing a node network graph based on the interaction relationship between smart devices, the communication relationship between smart devices, and the collaborative relationship between smart devices. ,in is a node set, is an edge set, is an edge weight set;

[0024] For each edge in the edge set , the weight of the edge is calculated by weighted summation.

[0025] Preferably, the degree centrality, betweenness centrality and closeness centrality of the nodes are calculated, and a threshold is set to determine the key nodes in the network, including: for each node in the network graph , the degree centrality of the node is calculated. ;

[0026] The degree centrality includes in-degree centrality and out-degree centrality ;

[0027] For each node in the network graph , the betweenness centrality of the node is calculated. ;

[0028] For each node in the network graph , the closeness centrality of the node is calculated. ;

[0029] The degree centrality threshold , the betweenness centrality threshold and the closeness centrality threshold are set; if the degree centrality , the betweenness centrality and the closeness centrality of a node satisfy:

[0030]

[0031] , the node is identified as a key node in the network, and a key node network subgraph is constructed based on all key nodes.

[0032] Preferably, according to the identified key nodes, node clusters are identified by a spectral clustering algorithm to construct a group structure of key nodes in the network, including: for the key node network subgraph composed of key nodes, the key nodes are clustered using a spectral clustering algorithm to construct a group structure of key nodes in the network.

[0033] The spectral clustering algorithm clusters based on the eigenvalues and eigenvectors of the Laplacian matrix of the graph; and the subgraph is constructed. The adjacency matrix ;

[0034] Computational subgraph The degree matrix ;

[0035] Computational subgraph The Laplace matrix of ;

[0036] Laplacian matrix Perform eigenvalue decomposition to obtain eigenvalues and the corresponding eigenvector ;

[0037] Before selection The eigenvectors corresponding to the smallest non-zero eigenvalues ​​constitute the characteristic matrix ,in is the number of target categories for clustering, is a sparse matrix;

[0038] For the feature matrix Normalize each row of to get the normalized feature matrix ;

[0039] Normalize the feature matrix Each row is considered A point in space;

[0040] use -means algorithm clusters the points and obtains Clustering results .

[0041] Preferably, based on the identified key nodes, the node clusters are identified by using the spectral clustering algorithm to construct the group structure of the key nodes in the network, and further comprising: for each cluster , analyze the attribute characteristics of internal nodes and calculate the edge density within the cluster , Indicates the degree of connection between nodes within the cluster;

[0042] Calculate edge density between clusters , Represents clustering and The tightness of the connection between them.

[0043] Preferably, based on the identified key nodes and the key node group structure, the connection relationship between nodes is analyzed, including: key node network subgraph Each key node in , analyze key nodes The connection relationship between nodes in the network;

[0044] Defining key nodes With node The connection strength between , is the sum of the edge weights of all paths between nodes;

[0045] Calculate key nodes and its cluster The average connection strength between internal nodes ;

[0046] Calculate key nodes and clustering The average connection strength between internal nodes .

[0047] Preferably, the weighted shortest path algorithm is used to calculate the optimal connection path between the key node and other nodes, and according to the calculation result, the connection between the key node and the nodes inside and outside the group is adjusted in the network, including: for the key node With node The connection between nodes is calculated using the weighted shortest path algorithm;

[0048] Defining key nodes arrive The optimal connection path The path with the minimum sum of edge weights is:

[0049] For key nodes , calculate the optimal connection path between the internal nodes of the cluster and the internal nodes of the non-cluster, and obtain the optimal connection path set ;

[0050] Adjust the connection relationship between key nodes and other nodes in the network based on the calculated connection strength and optimal connection path;

[0051] For key nodes , if the key node The average connection strength between the nodes and other nodes in the cluster Below the preset threshold , then add new edges within the cluster or adjust the weights of existing edges;

[0052] For key nodes , if the key node The average connection strength between nodes outside the cluster Below the preset threshold , then add new edges between clusters or adjust the weights of existing edges;

[0053] For the optimal connection path set Each path in , if the weight of the edge on the path is lower than the preset threshold , then the value below the preset threshold The weights of the edges are adjusted.

[0054] A network relationship analysis system based on smart devices in smart factory manufacturing, which is used to implement the network relationship analysis method based on smart devices in smart factory manufacturing, including: a data acquisition module, a graph model construction module, a key node identification module, a node group construction module and a node connection adjustment module;

[0055] The data acquisition module is used to collect interaction logs, communication records and production collaboration data of smart devices, and perform data cleaning and standardization;

[0056] The graph model construction module is used to abstract smart devices into nodes in the network, use the interactions and communication behaviors between devices as edges, and construct a network graph model that reflects the node relationships;

[0057] The key node identification module is used to calculate the degree centrality, betweenness centrality and closeness centrality indicators of the nodes, and set thresholds to identify the key nodes in the network according to the thresholds;

[0058] The node group construction module identifies node clusters based on the identified key nodes through a spectral clustering algorithm and constructs a group structure of key nodes in the network;

[0059] The node connection adjustment module analyzes the connection relationship between nodes based on the identified key nodes and the group structure of key nodes, uses a weighted shortest path algorithm to calculate the optimal connection path between the key nodes and other nodes, and adjusts the connection between the key nodes and nodes inside and outside the group in the network based on the calculation results.

[0060] Beneficial effects of the present invention: The present invention systematically collects and cleans the interaction logs, communication records and production collaboration data of smart devices to ensure the accuracy and consistency of the data, provide a high-quality data foundation for subsequent network analysis, and reduce errors and noise.

[0061] The present invention converts smart devices into network nodes and uses their interaction and communication behaviors as edges to build an intuitive network graph model, which helps visualize the relationships and dependencies between devices, facilitates the identification of key connection points and optimizes space.

[0062] The present invention calculates degree centrality, betweenness centrality, and closeness centrality indicators and sets thresholds to accurately identify key nodes in the network, which helps to optimize resource allocation, enhance the reliability of important equipment, and reduce the overall risk of the system.

[0063] The present invention uses a spectral clustering algorithm to divide node clusters according to key nodes, build a group structure, identify groups of devices with similar functions or working together, optimize resource allocation and task coordination, and improve production efficiency and collaboration effects.

[0064] Based on key nodes and group structures, the present invention uses a weighted shortest path algorithm to calculate the optimal connection path and adjust the connection between key nodes and nodes inside and outside the group, thereby enhancing network connectivity and information flow efficiency, and improving system robustness and response speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 This is a flow chart of the network relationship analysis method based on smart devices in smart factory manufacturing provided by the present invention;

[0066] Figure 2 This is a structural diagram of the network relationship analysis system based on intelligent devices in intelligent factory manufacturing provided by the present invention. DETAILED DESCRIPTION

[0067] For a better understanding of the present invention, various aspects of the present invention will be described in more detail with reference to the accompanying drawings. It should be understood that these detailed descriptions are merely descriptions of exemplary embodiments of the present invention and are not intended to limit the scope of the present invention in any way. Throughout this specification, like reference numerals refer to like elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0068] In the accompanying drawings, the size, dimensions, and shapes of the elements have been slightly adjusted for ease of illustration. The accompanying drawings are for illustration purposes only and are not drawn strictly to scale. As used herein, the terms "substantially," "approximately," and similar terms are used to indicate approximate values, not degrees, and are intended to illustrate inherent deviations in measurements or calculations that would be recognized by a person of ordinary skill in the art. In addition, in the present invention, the order in which the steps are described does not necessarily represent the order in which these steps occur in actual operation, unless otherwise specified or inferred from the context.

[0069] It should also be understood that expressions such as "comprises," "including," "having," "includes," and / or "comprising" are open rather than closed expressions in this specification, indicating the presence of the stated features, elements, and / or components, but do not exclude the presence of one or more other features, elements, components, and / or combinations thereof. In addition, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features rather than just the individual elements in the list. In addition, when describing embodiments of the present invention, "may" is used to mean "one or more embodiments of the present invention." And, the term "exemplary" is intended to refer to an example or illustration.

[0070] Unless otherwise defined, all terms used herein (including engineering and scientific terms) have the same meaning as commonly understood by those skilled in the art to which this invention pertains. It should also be understood that, unless otherwise expressly stated herein, words defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and should not be interpreted in an idealized or overly formal sense.

[0071] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0072] Example 1

[0073] Reference Figure 1 , which is the first embodiment of the present invention, provides a network relationship analysis method based on smart devices in smart factory manufacturing.

[0074] Step 1: Collect interaction logs, communication records, and production collaboration data from smart devices, and perform data cleaning and standardization. This includes:

[0075] Collect interaction logs of smart devices, including device ID, interaction time, interaction type (such as data transmission, control instructions, etc.) and interaction object information, and record them as data sets ,in Indicates the Interaction log data, The total number of interaction logs.

[0076] Collect communication records between devices, including source device ID, target device ID, communication time, communication protocol and communication content, The total number of communication records.

[0077] Collect collaborative data of equipment in the production process, including task allocation, goods flow and product processing collaborative events, recorded as data sets ,in Indicates the Collaborative data, is the total number of collaborative data.

[0078] Preprocess the collected data, including deduplication, format unification, missing value filling, data transcoding, and numerical data normalization.

[0079] For example, the numerical field (such as communication duration) is normalized and mapped to The specific formula is:

[0080]

[0081] in, is the original value, and are the minimum and maximum values ​​of the field, respectively. is the normalized value.

[0082] For example, the categorical fields (such as interaction type, communication protocol, etc.) are encoded and converted into numerical variables; for example, one-hot encoding is used to map each category into a binary vector, where the position corresponding to the category in the vector is 1 and the remaining positions are 0; after the above data preprocessing steps, standardized smart device interaction logs, communication records and production collaboration data are obtained, laying a data foundation for subsequent network relationship analysis.

[0083] The preprocessed dataset 、 and Denoted as 、 and .

[0084] Step 2: Abstract smart devices as nodes in the network, use the interactions and communication behaviors between devices as edges, and build a network graph model that reflects the node relationships. This includes:

[0085] The pre-processed smart device interaction log , communication records and production collaboration data Extract the smart device ID from the smart device ID to form a smart device ID set ,in Indicates the Smart devices, The total number of smart devices.

[0086] Collected by smart device ID As a set of nodes in a network diagram, smart devices Abstract into nodes , each node Represents a smart device.

[0087] According to the interaction log , extract the interaction relationship between smart devices and construct directed edge sets ; For each record in the interaction log ,in is the source smart device ID, is the target smart device ID, is the interaction time, For interaction type, in the node and Directed edges between , the edge weight is ; Edge weight Calculated based on indicators such as interaction frequency or interaction duration.

[0088] According to the communication records , extract the communication relationship between smart devices and construct directed edge sets ; For each record in the communication record ,in is the source smart device ID, is the target smart device ID, is the communication time, For the communication protocol, at the node and Directed edges between , the edge weight is ; Edge weight Calculated based on indicators such as communication frequency or communication data volume.

[0089] According to collaborative data , extract the collaborative relationship between smart devices and construct an undirected edge set ; For each record in the collaborative data ,in The ID of the smart device participating in the collaboration, For collaborative time, For collaborative tasks, here Undirected edges are established between each pair of smart device nodes to form undirected edges, and the edge weight is ; Edge weight Calculated based on indicators such as collaboration frequency or collaboration duration.

[0090] Construct a node network diagram based on the interaction relationship, communication relationship and collaborative relationship between smart devices ,in is a node set, is an edge set, is the edge weight set.

[0091] For edge sets For each edge in Calculated by weighted summation, the formula is: ;in, 、 、 is the weight factor.

[0092] In step 2, smart devices are abstracted as nodes in the network graph, and the interactions, communications, and collaborative relationships between devices are abstracted as edges. A network graph model that reflects device relationships is constructed, providing a basis for subsequent network relationship analysis.

[0093] Step 3: Calculate the degree centrality, betweenness centrality, and closeness centrality of the nodes, set thresholds, and identify key nodes in the network based on the thresholds. Specifically, it includes:

[0094] For network diagrams Each node in , computing nodes Degree centrality .

[0095] The degree centrality includes indegree centrality and outdegree centrality , respectively reflecting the node's receiving and sending connection capabilities; in-degree centrality The calculation formula is: ;in is the adjacency matrix The elements in represent the slave nodes arrive The weight of the edge; out-degree centrality The calculation formula is: ;in, is the adjacency matrix The elements in represent the slave nodes arrive The weight of the edge; degree centrality The calculation formula is: .

[0096] For network diagrams Each node in , computing nodes Betweenness centrality , betweenness centrality indicates the degree to which a node plays an "intermediary" role in the network, that is, the frequency of being on the shortest path between other nodes; betweenness centrality The calculation formula is: ;in, Representation node and The number of shortest paths between Representation node and Passing through nodes The number of shortest paths; nodes with higher betweenness centrality play important information transmission and control functions in the network.

[0097] For network diagrams Each node in , computing nodes Closeness Centrality ,Closeness centrality represents the average distance between a node and other ,nodes in the network, reflecting the reachability of the node in the ,network.

[0098] Closeness Centrality The calculation formula is ;in Representation node and The shortest path length between nodes; nodes with higher proximity centrality have better accessibility in the network and can quickly interact and communicate with other nodes.

[0099] Setting degree centrality threshold , betweenness centrality threshold and proximity centrality threshold ; If the degree centrality of the node , betweenness centrality and closeness centrality satisfy:

[0100]

[0101] Then the node Identify key nodes in the network and build a key node network subgraph based on all key nodes .

[0102] Step 3 comprehensively considers degree centrality, betweenness centrality, and closeness centrality to rank the importance of nodes in the network and identify the most influential and critical smart devices in the network. Key devices need to be monitored and managed to ensure their stable operation and optimize their connection with other devices to improve the overall efficiency and reliability of the network. Through the above steps, using indicators such as degree centrality, betweenness centrality, and closeness centrality, key nodes and important devices in the smart device network are identified, providing decision support for subsequent network optimization and management.

[0103] Step 4: According to the identified key nodes, identify node clusters through the spectral clustering algorithm, and construct the group structure of the key nodes in the network.

[0104] For the key node network subgraph composed of key nodes , use the spectral clustering algorithm to cluster the key nodes and construct the group structure of the key nodes in the network.

[0105] The spectral clustering algorithm is based on the eigenvalues and eigenvectors of the Laplacian matrix of the graph, and the specific method steps are as follows:

[0106] Construct the adjacency matrix of the subgraph , ; wherein represents the weight of the edge between key nodes and .

[0107] Calculate the degree matrix of the subgraph , ; wherein , represents the degree of the key node , is the number of key nodes, represents a diagonal matrix;

[0108] Calculate the Laplacian matrix of the subgraph , wherein ;

[0109] Perform eigenvalue decomposition on the Laplacian matrix to obtain the eigenvalues and the corresponding eigenvectors ;

[0110] Select the eigenvectors corresponding to the first smallest non-zero eigenvalues to form the feature matrix , wherein is the target number of classes for clustering, is a sparse matrix;

[0111] Normalize each row of the feature matrix to obtain the normalized feature matrix ;

[0112] Each row of the normalized feature matrix is regarded as a point in space;

[0113] Use -means algorithm clusters the points and obtains Clustering results .

[0114] For each cluster , analyze the attribute characteristics of internal nodes, including node type, function, subsystem, etc., identify the common characteristics of clusters; calculate the edge density within the cluster , which indicates the degree of connection between nodes within the cluster: ;in, Represents clustering The number of internal nodes, Represents clustering The number of edges between internal nodes.

[0115] Calculate edge density between clusters , indicating clustering and The tightness of the connection between: ;in, Represents clustering and The number of edges between .

[0116] Step 4: Based on the clustering results, identify clusters of key equipment with similar functions or working together to provide decision support for production scheduling, task allocation, etc.; for clusters with high edge density within the cluster, it means that the interactions between the equipment within the cluster are frequent, and it is necessary to focus on optimizing their communication and collaboration mechanisms to improve the working efficiency of the cluster; for cluster pairs with high edge density between clusters, it means that the collaborative relationship between these clusters is close, and it is necessary to optimize the cross-cluster data exchange and task collaboration process to improve the collaboration efficiency between clusters; the clustering results are fed back to the production management system and equipment control system to dynamically adjust the collaborative relationship and communication strategy between equipment to achieve intelligent production scheduling and optimization.

[0117] Step 5: Based on the identified key nodes and the group structure of key nodes, analyze the connection relationship between nodes, use the weighted shortest path algorithm to calculate the optimal connection path between the key nodes and other nodes, and adjust the connection between the key nodes and nodes inside and outside the group in the network based on the calculation results. Specifically include:

[0118] Key node network subgraph Each key node in , analyze key nodes The connection relationship between nodes in the network.

[0119] Defining key nodes With node The connection strength between , is the sum of the edge weights of all paths between nodes: ;in, Indicates key nodes arrive The set of all paths of Indicates the path, represents the edges on the path, Represents an edge The weight of .

[0120] Calculate key nodes and its cluster The average connection strength between internal nodes : .

[0121] Calculate key nodes and clustering The average connection strength between internal nodes : .

[0122] For key nodes With node The connection between nodes is calculated using the weighted shortest path algorithm.

[0123] Defining key nodes arrive The optimal connection path The path with the minimum sum of edge weights is:

[0124] For key nodes , calculate the optimal connection path between the internal nodes of the cluster and the internal nodes of the non-cluster, and obtain the optimal connection path set .

[0125] According to the calculated connection strength and optimal connection path, the connection relationship between the key nodes and other nodes in the network is adjusted.

[0126] For key nodes , if the key node The average connection strength between the nodes and other nodes in the cluster Below the preset threshold , then new edges are added within the cluster or the weights of existing edges are adjusted.

[0127] For key nodes , if the key node The average connection strength between nodes outside the cluster Below the preset threshold , new edges are added between clusters or the weights of existing edges are adjusted.

[0128] For the optimal connection path set Each path in , if the weight of the edge on the path is lower than the preset threshold , then the value below the preset threshold The weights of the edges are adjusted.

[0129] Step 5: Based on the identified key nodes and the group structure of key nodes, the connection relationship between nodes is analyzed, and the optimal connection path between the key nodes and other nodes is calculated using the weighted shortest path algorithm. According to the calculation results, the connection between the key nodes and other nodes inside and outside the group is adjusted in the network to optimize the connection relationship of the network and improve the efficiency and robustness of the network.

[0130] Example 2

[0131] Reference Figure 2 , which is the second embodiment of the present invention, provides a network relationship analysis system based on smart devices in smart factory manufacturing.

[0132] The system includes: a data acquisition module, a graph model construction module, a key node identification module, a node group construction module and a node connection adjustment module.

[0133] The data acquisition module is used to collect interaction logs, communication records and production collaboration data of smart devices, and perform data cleaning and standardization processing.

[0134] The graph model construction module is used to abstract smart devices into nodes in the network, use the interaction and communication behaviors between devices as edges, and construct a network graph model that reflects the node relationship.

[0135] The key node identification module is used to calculate the degree centrality, betweenness centrality and closeness centrality indicators of the nodes, and set thresholds to identify the key nodes in the network according to the thresholds.

[0136] The node group construction module identifies node clusters based on the identified key nodes through a spectral clustering algorithm, and constructs a group structure of key nodes in the network.

[0137] The node connection adjustment module analyzes the connection relationship between nodes based on the identified key nodes and the group structure of key nodes, uses a weighted shortest path algorithm to calculate the optimal connection path between the key nodes and other nodes, and adjusts the connection between the key nodes and nodes inside and outside the group in the network based on the calculation results.

[0138] The above order for steps of the method is only for illustration, and the steps of the method of the present application are not limited to the order specifically described above, unless otherwise specifically stated.

[0139] Furthermore, in some embodiments, the present application can also be implemented as a program recorded in a recording medium, which includes machine-readable instructions for implementing the method according to the present application. Thus, the present application also covers the recording medium storing the program for executing the method according to the present application.

[0140] In addition, the parts of the above technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail, so as not to be too redundant.

[0141] The specific embodiments described above are further explained in connection with the purposes, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application, and is not used to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A network relationship analysis method based on smart devices in smart factory manufacturing, characterized by: include: Collect interaction logs, communication records, and production collaboration data from smart devices, and perform data cleaning and standardization; Abstract smart devices as nodes in the network, use the interactions and communication behaviors between devices as edges, and build a network graph model that reflects the node relationships; Calculate the degree centrality, betweenness centrality and closeness centrality of the nodes, set the threshold, and identify the key nodes in the network according to the threshold, including: Each node in , computing nodes Degree centrality ; The degree centrality includes indegree centrality and outdegree centrality In-degree centrality The calculation formula is: ;in is the adjacency matrix The elements in represent the slave nodes arrive The weight of the edge; out-degree centrality The calculation formula is: ;in, is the adjacency matrix The elements in represent the slave nodes arrive The weight of the edge; degree centrality The calculation formula is: ; For network diagrams Each node in , computing nodes Betweenness centrality Betweenness centrality The calculation formula is: ;in, Representation node and The number of shortest paths between Representation node and Passing through nodes The number of shortest paths; For network diagrams Each node in , computing nodes Closeness Centrality , closeness centrality The calculation formula is ;in Representation node and The shortest path length between Setting degree centrality threshold , betweenness centrality threshold and proximity centrality threshold ; If the degree centrality of the node , betweenness centrality and closeness centrality satisfy: ； Then the node Identify key nodes in the network and build a key node network subgraph based on all key nodes ; According to the identified key nodes, the node clusters are identified through the spectral clustering algorithm, and the group structure of key nodes in the network is constructed, including: for the key node network subgraph composed of key nodes ,Use spectral clustering algorithm to cluster key nodes and construct the group structure of key nodes in the network; The spectral clustering algorithm clusters based on the eigenvalues ​​and eigenvectors of the graph's Laplacian matrix; constructs subgraphs The adjacency matrix ; Computational subgraph The degree matrix ; Computational subgraph The Laplace matrix of ; For the Laplace matrix Perform eigenvalue decomposition to obtain eigenvalues and the corresponding eigenvector ; Before selection The eigenvectors corresponding to the smallest non-zero eigenvalues ​​constitute the characteristic matrix ,in is the number of target categories for clustering, is a sparse matrix; for the characteristic matrix Normalize each row of to get the normalized feature matrix ; Normalize the feature matrix Each row is considered A point in space; use -means algorithm clusters the points and obtains Clustering results ; Based on the identified key nodes and the group structure of key nodes, the connection relationship between nodes is analyzed, and the node clusters are identified through the spectral clustering algorithm to construct the group structure of key nodes in the network; for each cluster , analyze the attribute characteristics of internal nodes and calculate the edge density within the cluster , ,in, Represents clustering The number of internal nodes, Represents clustering the number of edges between internal nodes, Indicates the degree of connection between nodes within a cluster; calculates the edge density between clusters , indicating clustering and The degree of connection between: ;in, Represents clustering and The number of edges between ;Key node network subgraph Each key node in , analyze key nodes The connection relationship between nodes in the network; Defining key nodes With node The connection strength between , is the sum of the edge weights of all paths between nodes: , ;in, Indicates key nodes arrive The set of all paths of Indicates the path, represents the edges on the path, Represents an edge The weight of the key nodes and its cluster The average connection strength between internal nodes : ; Calculate key nodes and clustering The average connection strength between internal nodes : ; Use the weighted shortest path algorithm to calculate the optimal connection path between the key node and other nodes, and adjust the connection between the key node and the nodes inside and outside the group in the network according to the calculation results. With node The connection between nodes is calculated using the weighted shortest path algorithm; Defining key nodes arrive The optimal connection path The path with the smallest sum of edge weights: For key nodes , calculate the optimal connection path between the internal nodes of the cluster and the internal nodes of the non-cluster, and obtain the optimal connection path set ; According to the calculated connection strength and optimal connection path, the connection relationship between the key nodes and other nodes in the network is adjusted; For key nodes , if the key node The average connection strength between the nodes and other nodes in the cluster Below the preset threshold , then add new edges within the cluster or adjust the weights of existing edges; for key nodes , if the key node The average connection strength between nodes outside the cluster Below the preset threshold , then add new edges between clusters or adjust the weights of existing edges; For the optimal connection path set Each path in , if the weight of the edge on the path is lower than the preset threshold , then the value below the preset threshold The weights of the edges are adjusted.

2. The network relationship analysis method based on intelligent devices in intelligent factory manufacturing according to claim 1 is characterized in that: Collect interaction logs, communication records, and production collaboration data of smart devices, and perform data cleaning and standardization, including: Collect interaction logs of smart devices and record them as data sets ,in Indicates the Interaction log data, is the total number of interaction logs; Collect communication records between devices and record them as data sets ,in Indicates the Communication records, is the total number of communication records; Collect collaborative data of equipment in the production process, recorded as data set ,in Indicates the Collaborative data, is the total number of collaborative data; Preprocess the collected data, including deduplication, format unification, missing value filling, data transcoding, and numerical data normalization; The preprocessed dataset 、 and Denoted as 、 and .

3. The network relationship analysis method based on intelligent devices in intelligent factory manufacturing according to claim 2 is characterized in that: Abstract smart devices as nodes in the network, with the interaction and communication behaviors between devices as edges, including: pre-processing smart device interaction logs , communication records and production collaboration data Extract the smart device ID from the smart device ID to form a smart device ID set ,in Indicates the Smart devices, is the total number of smart devices; Collected by smart device ID As a set of nodes in a network diagram, smart devices Abstract into nodes , each node Represents a smart device; According to the interaction log , extract the interaction relationship between smart devices and construct directed edge sets ; For each record in the interaction log ,in is the source smart device ID, is the target smart device ID, is the interaction time, For interaction type, in the node and Directed edges between , the edge weight is ; According to the communication records , extract the communication relationship between smart devices and construct directed edge sets ; For each record in the communication record ,in is the communication time, For the communication protocol, at the node and Directed edges between , the edge weight is ; According to collaborative data , extract the collaborative relationship between smart devices and construct an undirected edge set ; For each record in the collaborative data ,in The ID of the smart device participating in the collaboration, For collaborative time, For collaborative tasks, here Undirected edges are established between smart device nodes to form undirected edges, and the edge weight is .

4. The network relationship analysis method based on intelligent devices in intelligent factory manufacturing according to claim 3 is characterized in that: Construct a network graph model that reflects node relationships, including: constructing a node network graph based on the interaction relationship between smart devices, the communication relationship between smart devices, and the collaborative relationship between smart devices ,in is a node set, is an edge set, is the edge weight set; For edge sets For each edge in Calculated by weighted summation.

5. A network relationship analysis system based on smart devices in smart factory manufacturing, which is used to implement the network relationship analysis method based on smart devices in smart factory manufacturing according to any one of claims 1 to 4, characterized in that: include: Data acquisition module, graph model construction module, key node identification module, node group construction module and node connection adjustment module; The data acquisition module is used to collect interaction logs, communication records and production collaboration data of smart devices, and perform data cleaning and standardization; The graph model construction module is used to abstract smart devices into nodes in the network, use the interactions and communication behaviors between devices as edges, and construct a network graph model that reflects the node relationships; The key node identification module is used to calculate the degree centrality, betweenness centrality and closeness centrality indicators of the nodes, and set thresholds to identify the key nodes in the network according to the thresholds; The node group construction module identifies node clusters based on the identified key nodes through a spectral clustering algorithm and constructs a group structure of key nodes in the network; The node connection adjustment module analyzes the connection relationship between nodes based on the identified key nodes and the group structure of key nodes, uses a weighted shortest path algorithm to calculate the optimal connection path between the key nodes and other nodes, and adjusts the connection between the key nodes and nodes inside and outside the group in the network based on the calculation results.

Images