Business process guide method based on visualization algorithm

Abstract

The invention discloses a business process guide method based on a visualization algorithm, and the method comprises the following steps: collecting and sorting business process data, including each process step, task distribution, role responsibility and dependency; the data is converted into a standardized graphic model by using an advanced graphic algorithm, and two-dimensional or three-dimensional visual display is realized through an embedding algorithm and a data dimension reduction technology. The system displays the model in a business guide interface and provides an interaction function, and a user can zoom and drag nodes and customize a view. Real-time data updating is achieved through a simulation algorithm, the future state of the process is predicted, and a guide map is automatically adjusted according to actual data. A multi-level view switching function and an intelligent optimization algorithm are provided, a flow path and resource configuration are optimized by analyzing data association, finally an optimization result is stored, and a user is allowed to compare and backtrack historical flow versions. According to the method, the visual management and optimization efficiency of the business process is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of business process management and data visualization, and in particular to a business process navigation method based on visualization algorithm. Background Art

[0002] In modern enterprise management, efficient management and optimization of business processes have become one of the key factors in improving enterprise competitiveness. With the continuous expansion of business, the complexity of processes is increasing. How to effectively display, analyze and optimize these complex business processes has become an important challenge for enterprises. Although existing business process management tools can help enterprises model and manage business processes to a certain extent, they still have many limitations.

[0003] Traditional business process management tools mostly rely on static flowcharts and manual updates, which leads to rigid and unintuitive process display. Existing technologies often use two-dimensional static graphics to display complex business processes. This method is difficult to fully express the dynamic changes and complex dependencies in the process. Especially when facing large-scale and multi-level business processes, it is difficult for users to fully understand and grasp the operation status of the entire process through a static graphical interface. In addition, the manual update of flowcharts is not only inefficient, but also prone to errors, and it is difficult to timely reflect real-time changes in business processes.

[0004] Existing technologies lack interactivity and flexibility in the visualization of business processes. When operating flowcharts, users can usually only view predefined views and lack the ability to operate graphical models in real time, such as zooming, dragging, and node alignment. This lack of interactivity limits users' ability to deeply analyze and understand processes and cannot meet the personalized needs of managers at different levels. At the same time, since it is impossible to dynamically adjust views according to user needs, existing technologies are unable to cope with complex process analysis tasks.

[0005] The existing business process management system is not intelligent enough in terms of optimization and analysis. Traditional optimization methods mostly rely on manual experience and are difficult to handle large-scale data associations and complex dependencies. Even if some automated optimization algorithms are introduced, these algorithms are often limited to simple path optimization or resource allocation, and are unable to conduct in-depth analysis of potential problems in the process, and are also difficult to provide comprehensive and real-time optimization suggestions. Especially when faced with large-scale data sets, traditional methods show obvious deficiencies in identifying key patterns and removing process noise. Existing business process management tools also have major defects in historical version management and backtracking functions. Most tools can only simply record and display historical process versions, lack in-depth analysis and intelligent synthesis functions of differences between versions, and users often cannot obtain valuable decision support when comparing and backtracking process versions. In addition, existing tools lack intelligent prediction and decision support for backtracking operations in version management. The backtracking process is often just a simple version replacement, and cannot provide optimization suggestions based on the change trends between different versions, affecting the flexibility and effectiveness of process management.

[0006] Therefore, how to provide a business process navigation method based on visualization algorithm is a problem that technical personnel in this field urgently need to solve. Summary of the invention

[0007] One purpose of the present invention is to propose a business process navigation method based on visualization algorithm. The present invention makes full use of graph embedding and high-dimensional data visualization algorithm, quantum molecular dynamics simulation algorithm and random matrix theory optimization algorithm, and describes in detail how to achieve efficient management of complex business processes through graphical display, real-time update and intelligent optimization. The method has the advantages of intuitive display, flexible operation, intelligent optimization and efficient management.

[0008] A business process navigation method based on a visualization algorithm according to an embodiment of the present invention includes the following steps:

[0009] S1. Collect and organize business process data, including process steps, task allocation, role responsibilities and dependencies;

[0010] S2. Convert the business process data into a standardized graphical model using a graph embedding and high-dimensional data visualization algorithm, wherein the graph embedding algorithm is used to represent the flow chart as an embedding vector in a low-dimensional space to capture the deep relationship between nodes, and the high-dimensional data visualization algorithm reduces the dimension of high-dimensional data through UMAP technology to achieve a two-dimensional or three-dimensional intuitive display;

[0011] S3. Display a standardized graphical model on the business navigation interface. The interface provides interactive functions, including node zooming, dragging, positioning and customizing views. Users can view and adjust different perspectives of the business process in real time.

[0012] S4. Based on the standardized graphical model displayed, the real-time data update function is integrated using the quantum molecular dynamics simulation algorithm. By simulating the dynamic behavior of tasks and resources, the future state of the process is predicted, and the navigation map is automatically updated according to the actual data to reflect the latest process status;

[0013] S5. Based on the updated navigation map, a multi-level view is provided, and users can switch between the overall flowchart and task view to meet the needs of managers at different levels;

[0014] S6. Based on the updated process data, the random matrix theory optimization algorithm is used to intelligently optimize the business process. By analyzing large-scale data associations, key patterns and process noise are identified, paths and resource allocation are optimized, and optimization suggestions are provided in a visual interface.

[0015] S7. Record and save the optimized process version, provide the comparison and backtracking function of historical process versions, so that users can compare different versions of flowcharts through a visual interface and analyze the process improvement effect;

[0016] S8. Based on the saved process version, it supports multi-user collaboration, allowing multiple users to view and edit business process diagrams at the same time, and realize synchronous update and conflict detection through a visual interface.

[0017] Optionally, the S2 specifically includes:

[0018] S21. Structuring the business process data into a graph G(V,E), where V is a set of nodes, representing each step or task in the business process, and E is a set of edges, representing the dependency or order between nodes. i Corresponding to the specific steps and tasks in the business process, ij Represents task v i and v j The dependencies between them;

[0019] S22. Process the graph G(V,E) through the graph embedding algorithm and embed each node v i Mapped to an embedding vector h in a low-dimensional space i ∈R d , where d is the dimension of the embedding space. The embedding process captures the deep relationship between nodes by optimizing the objective function:

[0020]

[0021] in, is the sigmoid function, h u and h v are the embedding vectors of nodes u and v respectively;

[0022] S23, using UMAP technology to embed the vector set H = {h 1 ,h 2 ,…,h n} performs dimensionality reduction processing and embeds the high-dimensional vector h i Mapping to low-dimensional space z i ∈R k , where k = 2 or k = 3;

[0023] S24. Construct a neighbor graph in high-dimensional space and calculate the similarity between nodes. The weight of the neighbor graph is calculated using the Gaussian kernel function:

[0024]

[0025] Among them, σ i is the local scale parameter of node i;

[0026] S25. In low-dimensional space, construct an optimization objective function so that the distance between points in low-dimensional space reflects the proximity relationship in high-dimensional space:

[0027]

[0028] in, represents the total loss of point pairs in low-dimensional space, ||z i -z j || 2 represents the distance between nodes i and j in the low-dimensional space, w ij is the weight of edge (i,j), is the difference between the distances of point pairs in high-dimensional space;

[0029] S26, generate a two-dimensional or three-dimensional standardized graphical model, and use the node representation Z={z 1 ,z 2 ,…,z n}Shown in the visual interface, each node z i The edge (z i ,z j ) represents the dependency relationship between tasks.

[0030] Optionally, the S3 specifically includes:

[0031] S31, the standardized graphical model Z = {z 1 ,z 2 ,…,z n} is mapped to the visualization area of ​​the business navigation interface, so that each node z i Position and edge in the interface (z i ,zj )’s connection relationship maintains its relative position in the low-dimensional space;

[0032] S32. Implement node zoom function in the business navigation interface. Users can adjust node z by using mouse wheel and other zoom controls. i The display size of

[0033] S33, provide node dragging function, users can click and drag the node z i Move freely in the visualization area, and keep the z coordinates of adjacent nodes during the movement of nodes. j Real-time update of connection relationships;

[0034] S34, provide intelligent alignment function, users drag node z i When the system automatically detects the node and its adjacent nodes z j The nodes are aligned and the automatic adsorption function is used to align the nodes, keeping the connection relationship between the nodes neat and consistent:

[0035] z′ i =Align(z i ,z j )+Δz align ;

[0036] Among them, Align(z i ,z j ) is the node position after automatic alignment, Δz align is the increment for adjustment after alignment;

[0037] S35, provides positioning function, users can automatically position the current view center to the specified node z by clicking the control in the interface t The positioning process is performed by calculating the node z t The offset Δz from the center of the current view t And implement view translation;

[0038] S36, introduce adaptive layout function, users adjust node z i and view, the system automatically rearranges other nodes z j To optimize the overall layout, the adaptive layout process is based on the following optimization objective function:

[0039]

[0040] Among them, λ is the weight parameter, Edge_crossing(z i ,z j ) represents the number of intersections between edges. The system minimizes the objective function L layout Automatically optimize node positions;

[0041] S37. Provide custom view function, users can select different business process perspectives through interface controls, including zooming in on a specific process area or zooming in on the whole to see a global view. The view switching operation updates the view parameters, including the perspective, zoom ratio and visible area.

[0042] S38. When users view and adjust different perspectives of a business process in real time, the display status of all nodes and edges is dynamically updated according to user operations, and the graphical model reflects the latest perspective and layout of the business process.

[0043] Optionally, the S4 specifically includes:

[0044] S41, standardize the graphical model Z = {z 1 ,z 2 ,…,z n}Extract each node z i The initial state of each node z i Represents a task or resource in a business process. Each edge (z i ,z j ) represents the interdependencies between tasks or resources;

[0045] S42. Establish a quantum molecular dynamics model and define node z i The state vector r i (t) and momentum vector p i (t), respectively represent the location and change rate of the task or resource at time t, and the interaction between nodes is expressed by the interaction potential energy function V(r i ,r j )describe:

[0046]

[0047] Among them, k ij For node z i and z j The interaction strength between them, ||r i -r j || is the distance between two nodes;

[0048] S43, through quantum molecular dynamics simulation, use the following motion equation to predict node z i The state at the future time t+Δt:

[0049]

[0050]

[0051] Among them, m i For node z iThe quality of the corresponding task or resource, Δt is the time step;

[0052] S44, update the node status according to the simulated future status, and map these updated status to the graphical model Z′={z′ 1 ,z′ 2 ,…,z′ n}, where the updated state z′ of each node i By r i (t+Δt) determines;

[0053] S45. Adjust the parameter k of the quantum molecular dynamics model in real time through data feedback in the actual business process ij and m i , and recalculate the node status to generate new process data and reflect the latest business status;

[0054] S46, mapping the updated graphical model Z′ back to the business navigation interface to generate an updated navigation map, through which the user can view and adjust possible future process changes and resource allocations in real time.

[0055] Optionally, the S5 specifically includes:

[0056] S51, from the generated updated navigation map Z′={z′ 1 ,z′ 2 ,…,z′ n} to extract the overall business process view, which contains all nodes z′ i and their interconnected edges (z′ i ,z′ j );

[0057] S52, grouping related tasks and step nodes in the business process to form multiple task subsets C k , where each task subset C k ={z′ k1 ,z′ k2 ,…} represents a sub-process or task group, generating a local task view Z′ for the user k Provide a foundation;

[0058] S53. Keep the global view Z′ and the local task view Z′ in the system k The interface controls provide a view switching function, allowing users to switch between the overall view Z′ and the local view Z′ k Freely switch between views, and dynamically adjust view parameters during view switching;

[0059] S54. Added an automatic focus function. When the user switches from the global view to the local task view, it automatically identifies and focuses on the task or node that the user is currently interested in, automatically zooms in on the relevant details of these nodes when the view is switched, and automatically adjusts the viewing angle to best display the overall process when switching back to the global view;

[0060] S55. Introduce the multi-view collaboration function, so that all operations of users on different levels of views can be synchronized in real time in the updated navigation map Z′, and automatically updated to all related views to ensure coordination and consistency between multi-level views.

[0061] Optionally, the S6 specifically includes:

[0062] S61. Extract node z′ from updated process data i The state and task dependency (z′ i ,z′ j ) and other relevant business process parameters to provide input data for subsequent optimization analysis;

[0063] S62. Use random matrix theory to construct the correlation matrix R of process data, where the matrix elements R ij Represents the task node z′ i and z′ j By analyzing the eigenvalues ​​and eigenvectors of the correlation matrix R, we can identify the key patterns and potential noises in the business process.

[0064] S63. By calculating the spectral density ρ(λ) of the correlation matrix R, the eigenvalue λ related to the process noise is identified, and filtering technology is applied to remove the noise, retaining the characteristic information that is meaningful to the business process optimization:

[0065]

[0066] Where δ is the Diracdelta function and N is the total number of eigenvalues;

[0067] S64. Based on the processed correlation matrix, the path and resource configuration of the business process are optimized. The optimal task execution path is determined by maximizing the high correlation feature value on the path. The goal of optimizing the path is to maximize the path correlation and minimize the execution time:

[0068]

[0069] Among them, t ij is the task node z′ i and z′ j The execution time between

[0070] S65. These suggestions are presented to the user through a visual interface. The user can view the optimized path and resource allocation plan in the interface, and make corresponding business process adjustments to generate an optimized process version.

[0071] Optionally, the S7 specifically includes:

[0072] S71. Save the generated optimized path and resource configuration as an independent process version and record the version identifier V n , where V n Indicates the nth optimized process version;

[0073] S72. For each saved process version V n Recording Flowchart Z n The state of the node z′ i The location and task dependency (z′ i ,z′ j ), path optimization information and resource allocation status;

[0074] S73. Provide version management function, allowing users to view historical process versions in a visual interface 1 ,V 2 ,…,V n , the user can select a specific version to load and the corresponding flowchart Z n Map to the visual interface for display;

[0075] S74, using the dynamic process synthesis function, the user selects multiple historical versions V i and V j When comparing, a synthetic version V is automatically generated s , synthesize the optimal path and resource configuration, generate and display the synthetic version based on the predefined optimization goals;

[0076] S75, provide intelligent backtracking decision function, users can choose to backtrack to the historical version V m When based on the current version V n With V m Predict the retrospective impact of changes between the two, and generate a variety of retrospective strategies for users to choose from. Each strategy includes analysis of process efficiency, resource usage, and potential risks to help users make decisions;

[0077] S76. During the backtracking process, the system updates the navigation map in real time and rolls back the process status to version V m Flowchart after Z m Mapped to the visual interface, users can immediately view the status after backtracking and select the synthetic version or backtracking strategy for adjustment and optimization as needed.

[0078] The beneficial effects of the present invention are:

[0079] The present invention uses graph embedding and high-dimensional data visualization algorithms to transform complex business process data into standardized graphical models, and intuitively displays them in two or three dimensions. Users can flexibly operate the flowchart through functions such as zooming, dragging, and node alignment, making it easier to understand and grasp the overall picture and details of the business process. This flexible interactive method overcomes the rigidity of traditional static flowchart display and significantly improves user experience.

[0080] By integrating the quantum molecular dynamics simulation algorithm, the present invention can update the process data in real time and automatically adjust the navigation map when the business process changes to reflect the latest process status. At the same time, the algorithm can simulate the dynamic behavior of tasks and resources and predict the future state of the process, so that users can identify potential problems in advance and make more scientific decisions, thereby greatly improving the real-time and forward-looking nature of process management. Using the random matrix theory optimization algorithm, the present invention can intelligently analyze and optimize large-scale process data, optimize task paths and resource allocation by identifying key patterns and removing process noise, and provide optimization suggestions in a visual interface. Compared with traditional manual experience or simple automated optimization methods, the present invention shows higher efficiency and accuracy in processing complex processes and big data associations, thereby significantly improving the overall operating efficiency of business processes.

[0081] The present invention provides a multi-level view function, allowing users to switch freely between the overall flowchart and the task view to meet the needs of managers at different levels. At the same time, it supports multi-user collaborative operations, allowing multiple users to view and edit business process diagrams at the same time, and achieves synchronous updates and conflict detection through a visual interface, further enhancing the efficiency and accuracy of teamwork. Finally, through the recording and backtracking functions of historical versions, users can easily compare flowcharts of different versions, analyze process improvement effects, and perform intelligent backtracking and optimization when necessary. The comprehensive application of these functions enables the present invention to achieve effects in business process management, optimization and collaboration that are difficult to achieve with the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0083] Figure 1 A flowchart of a business process navigation method based on a visualization algorithm proposed by the present invention;

[0084] Figure 2 A schematic diagram of using a quantum molecular dynamics simulation algorithm to achieve real-time data update and future state prediction in the present invention;

[0085] Figure 3 A schematic diagram of the random matrix theory optimization algorithm used in the present invention for business process optimization;

[0086] Figure 4 It is a schematic diagram of the dynamic synthesis and intelligent backtracking decision-making functions of the historical process versions in the present invention. DETAILED DESCRIPTION

[0087] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0088] refer to Figure 1-4 , a business process navigation method based on visualization algorithm, comprising the following steps:

[0089] S1. Collect and organize business process data, including process steps, task allocation, role responsibilities and dependencies;

[0090] S2. Convert the business process data into a standardized graphical model using a graph embedding and high-dimensional data visualization algorithm, wherein the graph embedding algorithm is used to represent the flow chart as an embedding vector in a low-dimensional space to capture the deep relationship between nodes, and the high-dimensional data visualization algorithm reduces the dimension of high-dimensional data through UMAP technology to achieve a two-dimensional or three-dimensional intuitive display;

[0091] S3. Display a standardized graphical model on the business navigation interface. The interface provides interactive functions, including node zooming, dragging, positioning and customizing views. Users can view and adjust different perspectives of the business process in real time.

[0092] S4. Based on the standardized graphical model displayed, the real-time data update function is integrated using the quantum molecular dynamics simulation algorithm. By simulating the dynamic behavior of tasks and resources, the future state of the process is predicted, and the navigation map is automatically updated according to the actual data to reflect the latest process status;

[0093] S5. Based on the updated navigation map, a multi-level view is provided, and users can switch between the overall flowchart and task view to meet the needs of managers at different levels;

[0094] S6. Based on the updated process data, the random matrix theory optimization algorithm is used to intelligently optimize the business process. By analyzing large-scale data associations, key patterns and process noise are identified, paths and resource allocation are optimized, and optimization suggestions are provided in a visual interface.

[0095] S7. Record and save the optimized process version, provide the comparison and backtracking function of historical process versions, so that users can compare different versions of flowcharts through a visual interface and analyze the process improvement effect;

[0096] S8. Based on the saved process version, it supports multi-user collaboration, allowing multiple users to view and edit business process diagrams at the same time, and realize synchronous update and conflict detection through a visual interface.

[0097] In this implementation, S2 specifically includes:

[0098] S21. Structuring the business process data into a graph G(V,E), where V is a set of nodes, representing each step or task in the business process, and E is a set of edges, representing the dependency or order between nodes. i Corresponding to the specific steps and tasks in the business process, ij Represents task v i and v j The dependencies between them;

[0099] S22. Process the graph G(V,E) through the graph embedding algorithm and embed each node v i Mapped to an embedding vector h in a low-dimensional space i ∈R d , where d is the dimension of the embedding space. The embedding process captures the deep relationship between nodes by optimizing the objective function:

[0100]

[0101] in, is the sigmoid function, h u and h v are the embedding vectors of nodes u and v respectively;

[0102] S23, using UMAP technology to embed the vector set H = {h 1 ,h 2 ,…,h n} performs dimensionality reduction processing and embeds the high-dimensional vector h i Mapping to low-dimensional space z i ∈R k , where k = 2 or k = 3;

[0103] S24. Construct a neighbor graph in high-dimensional space and calculate the similarity between nodes. The weight of the neighbor graph is calculated using the Gaussian kernel function:

[0104]

[0105] Among them, σ i is the local scale parameter of node i;

[0106] S25. In low-dimensional space, construct an optimization objective function so that the distance between points in low-dimensional space reflects the proximity relationship in high-dimensional space:

[0107]

[0108] in, represents the total loss of point pairs in low-dimensional space, ||z i -z j || 2 represents the distance between nodes i and j in the low-dimensional space, w ij is the weight of edge (i,j), is the difference between the distances of point pairs in high-dimensional space;

[0109] S26, generate a two-dimensional or three-dimensional standardized graphical model, and use the node representation Z={z 1 ,z 2 ,…,z n}Shown in the visual interface, each node z i Shown by its position in low-dimensional space, edge z i ,z j ) represents the dependency relationship between tasks.

[0110] In this implementation, S3 specifically includes:

[0111] S31, the standardized graphical model Z = {z 1 ,z 2 ,…,z n} is mapped to the visualization area of ​​the business navigation interface, so that each node z i Position and edge in the interface (z i ,z j )’s connection relationship maintains its relative position in the low-dimensional space;

[0112] S32. Implement node zoom function in the business navigation interface. Users can adjust node z by using mouse wheel and other zoom controls. i The display size of

[0113] S33, provide node dragging function, users can click and drag the node z i Move freely in the visualization area, and keep the z coordinates of adjacent nodes during the movement of nodes. j Real-time update of connection relationships;

[0114] S34, provide intelligent alignment function, users drag node z i When the system automatically detects the node and its adjacent nodes z jThe nodes are aligned and the automatic adsorption function is used to align the nodes, keeping the connection relationship between the nodes neat and consistent:

[0115] z′ i =Align(z i ,z j )+Δz align ;

[0116] Among them, Align(z i ,z j ) is the node position after automatic alignment, Δz align is the increment for adjustment after alignment;

[0117] S35, provides positioning function, users can automatically position the current view center to the specified node z by clicking the control in the interface t The positioning process is performed by calculating the node z t The offset Δz from the center of the current view t And implement view translation;

[0118] S36, introduce adaptive layout function, users adjust node z i and view, the system automatically rearranges other nodes z j To optimize the overall layout, the adaptive layout process is based on the following optimization objective function:

[0119]

[0120] Among them, λ is the weight parameter, Edge_crossing(z i ,z j ) represents the number of intersections between edges. The system minimizes the objective function L layout Automatically optimize node positions;

[0121] S37. Provide custom view function, users can select different business process perspectives through interface controls, including zooming in on a specific process area or zooming in on the whole to see a global view. The view switching operation updates the view parameters, including the perspective, zoom ratio and visible area.

[0122] S38. When users view and adjust different perspectives of a business process in real time, the display status of all nodes and edges is dynamically updated according to user operations, and the graphical model reflects the latest perspective and layout of the business process.

[0123] In this implementation, S4 specifically includes:

[0124] S41, standardize the graphical model Z = {z 1 ,z 2 ,…,z n}Extract each node z i The initial state of each node z i Represents a task or resource in a business process. Each edge (z i ,z j ) represents the interdependencies between tasks or resources;

[0125] S42. Establish a quantum molecular dynamics model and define node z i The state vector r i (t) and momentum vector p i (t), respectively represent the location and change rate of the task or resource at time t, and the interaction between nodes is expressed by the interaction potential energy function V(r i ,r j )describe:

[0126]

[0127] Among them, k ij For node z i and z j The interaction strength between them, ||r i -r j || is the distance between two nodes;

[0128] S43, through quantum molecular dynamics simulation, use the following motion equation to predict node z i The state at the future time t+Δt:

[0129]

[0130]

[0131] Among them, m i For node z i The quality of the corresponding task or resource, Δt is the time step;

[0132] S44, update the node status according to the simulated future status, and map these updated status to the graphical model Z′={z′ 1 ,z′ 2 ,…,z′ n}, where the updated state z′ of each node i By r i (t+Δt) determines;

[0133] S45. Adjust the parameter k of the quantum molecular dynamics model in real time through data feedback in the actual business process ij and m i , and recalculate the node status to generate new process data and reflect the latest business status;

[0134] S46, mapping the updated graphical model Z′ back to the business navigation interface to generate an updated navigation map, through which the user can view and adjust possible future process changes and resource allocations in real time.

[0135] In this implementation manner, S5 specifically includes:

[0136] S51, from the generated updated navigation map Z′={z′ 1 ,z′ 2 ,…,z′ n} to extract the overall business process view, which contains all nodes z′ i and their interconnected edges (z′ i ,z′ j );

[0137] S52, grouping related tasks and step nodes in the business process to form multiple task subsets C k , where each task subset C k ={z′ k1 ,z′ k2 ,…} represents a sub-process or task group, generating a local task view Z′ for the user k Provide a foundation;

[0138] S53. Keep the global view Z′ and the local task view Z′ in the system k The interface controls provide a view switching function, allowing users to switch between the overall view Z′ and the local view Z′ k Freely switch between views, and dynamically adjust view parameters during view switching;

[0139] S54. Added an automatic focus function. When the user switches from the global view to the local task view, it automatically identifies and focuses on the task or node that the user is currently interested in, automatically zooms in on the relevant details of these nodes when the view is switched, and automatically adjusts the viewing angle to best display the overall process when switching back to the global view;

[0140] S55. Introduce the multi-view collaboration function, so that all operations of users on different levels of views can be synchronized in real time in the updated navigation map Z′, and automatically updated to all related views to ensure coordination and consistency between multi-level views.

[0141] In this implementation manner, S6 specifically includes:

[0142] S61. Extract node z′ from updated process data i The state and task dependency (z′ i ,z′ j) and other relevant business process parameters to provide input data for subsequent optimization analysis;

[0143] S62. Use random matrix theory to construct the correlation matrix R of process data, where the matrix elements R ij Represents the task node z′ i and z′ j By analyzing the eigenvalues ​​and eigenvectors of the correlation matrix R, we can identify the key patterns and potential noises in the business process.

[0144] S63. By calculating the spectral density ρ(λ) of the correlation matrix R, the eigenvalue λ related to the process noise is identified, and filtering technology is applied to remove the noise, retaining the characteristic information that is meaningful to the business process optimization:

[0145]

[0146] Where δ is the Diracdelta function and N is the total number of eigenvalues;

[0147] S64. Based on the processed correlation matrix, the path and resource configuration of the business process are optimized. The optimal task execution path is determined by maximizing the high correlation feature value on the path. The goal of optimizing the path is to maximize the path correlation and minimize the execution time:

[0148]

[0149] Among them, t ij is the task node z′ i and z′ j The execution time between

[0150] S65. These suggestions are presented to the user through a visual interface. The user can view the optimized path and resource allocation plan in the interface, and make corresponding business process adjustments to generate an optimized process version.

[0151] In this implementation manner, the S7 specifically includes:

[0152] S71. Save the generated optimized path and resource configuration as an independent process version and record the version identifier V n , where V n Indicates the nth optimized process version;

[0153] S72. For each saved process version V n Recording Flowchart Z n The state of the node z′ i The location and task dependency (z′ i ,z′ j), path optimization information and resource allocation status;

[0154] S73. Provide version management function, allowing users to view historical process versions in a visual interface 1 ,V 2 ,…,V n , the user can select a specific version to load and the corresponding flowchart Z n Map to the visual interface for display;

[0155] S74, using the dynamic process synthesis function, the user selects multiple historical versions V i and V j When comparing, a synthetic version V is automatically generated s , synthesize the optimal path and resource configuration, generate and display the synthetic version based on the predefined optimization goals;

[0156] S75, provide intelligent backtracking decision function, users can choose to backtrack to the historical version V m When based on the current version V n With V m Predict the retrospective impact of changes between the two, and generate a variety of retrospective strategies for users to choose from. Each strategy includes analysis of process efficiency, resource usage, and potential risks to help users make decisions;

[0157] S76. During the backtracking process, the system updates the navigation map in real time and rolls back the process status to version V m Flowchart after Z m Mapped to the visual interface, users can immediately view the status after backtracking and select the synthetic version or backtracking strategy for adjustment and optimization as needed.

[0158] Embodiment 1:

[0159] In order to verify the feasibility of the present invention in implementation, the present invention is applied to a large manufacturing enterprise with a complex production process involving the coordination of multiple departments and the allocation of numerous tasks. The problem faced by the enterprise is that with the increase in order volume and the diversification of product types, the business process has become extremely complex, and traditional process management methods are difficult to cope with, resulting in a decrease in production efficiency and serious waste of resources. The enterprise needs a business process management tool that can intuitively display the process, update data in real time, optimize resource allocation, and support multi-user collaboration to improve overall production efficiency and resource utilization.

[0160] In order to solve this problem, the company decided to introduce the business process navigation method based on visualization algorithm proposed in this invention. The specific application scenarios are as follows: During the order production process, the company needs to coordinate the task allocation and resource scheduling of multiple departments such as production workshops, procurement departments, and logistics departments. Traditional process management systems can only display processes in the form of static charts, and it is difficult to dynamically reflect changes in the production process, such as temporary adjustments to production lines, real-time inventory changes of raw materials, changes in transportation routes, etc. This static management mode leads to information asymmetry and cannot reflect the actual working status of each department in a timely manner, thus affecting the overall production efficiency.

[0161] In the application of the present invention, the enterprise first collects and organizes the business process data to structure the various production steps, task allocations, role responsibilities and dependencies involved into a graphical model. Then, the system uses graph embedding and high-dimensional data visualization algorithms to convert these complex business process data into standardized graphical models and display them in two-dimensional and three-dimensional ways. Through this intuitive graphical interface, enterprise managers can view and adjust each link in the process in real time. For example, when a production line is temporarily adjusted, the system will automatically update the flowchart and allow users to intuitively see the adjusted production flowchart through zooming and dragging functions. In addition, users can customize views according to their needs and focus on key tasks or bottleneck links, so as to more effectively monitor production progress.

[0162] Enterprises use quantum molecular dynamics simulation algorithms to update data in real time. When the purchasing department finds that a certain raw material is insufficient in stock, or the logistics department finds that transportation is delayed, the system can automatically update the business process based on real-time data and predict possible future impacts. For example, the system can predict how the delayed arrival of raw materials will affect the production plan, and dynamically display this change in the flow chart, so that all relevant departments can prepare for it in advance. In this way, enterprises can minimize the problem of low production efficiency caused by information lag.

[0163] In terms of resource allocation and task optimization, the system uses random matrix theory optimization algorithms to intelligently optimize the paths and resource allocation in business processes. The system analyzed the production data in the past six months and found that the resource allocation in the logistics link was unreasonable, which often caused the production line to stop working while waiting for transportation. By optimizing the resource allocation of logistics nodes, the system recommends reallocating part of the transportation resources to key logistics nodes. After the adjustment, the average waiting time for the production line to be transported was shortened from 120 minutes to 45 minutes, and the overall production efficiency was improved by 37%. The multi-level view function and multi-user collaboration function provided enable different management levels of the enterprise to view the flow chart according to their needs. For example, senior managers can see the key nodes of the entire production process through the global view, while workshop managers can focus on the specific operation links of the production line through the local view. At the same time, staff from multiple departments can view and edit business process diagrams in the system at the same time, and avoid data conflicts caused by multi-person operations through the system's conflict detection function. This collaborative function significantly improves the efficiency of cross-departmental communication and reduces misoperations caused by information asymmetry.

[0164] During the implementation process, the system also recorded each optimized process version and supported the comparison and backtracking of historical versions. By comparing different versions of flowcharts, the company can intuitively see the changes in production efficiency after each adjustment and select the optimal process version based on actual conditions. For example, by comparing the production process versions of the last three months, the company found that by adding the raw material inventory warning function of the procurement department, production efficiency increased by 20%. Based on this discovery, the company decided to apply this improvement to all future production processes.

[0165] Specific data prove the beneficial effects of the present invention. Since the implementation of the present invention, the production efficiency of enterprises has been significantly improved: the waiting time of the production line has been reduced by an average of 50%, and the resource utilization rate has increased by 30%. In addition, production delays caused by information lags have been reduced by 40%, and the efficiency of the logistics link has been improved by 37%. The multi-user collaboration function has increased the efficiency of cross-departmental communication by 25% and reduced the incidence of data conflicts by 80%. Through the backtracking and comparison of historical versions, enterprises can manage production processes more scientifically, production costs have been reduced by 15%, and overall operational efficiency has been improved by 20%.

[0166] Table 1 Comparison of key performance indicators before and after the implementation of the present invention

[0167]

[0168]

[0169] Table 2 Historical version retrospective analysis and production efficiency improvement table

[0170]

[0171] Through the summary of the table data, the present invention has achieved remarkable results in the application of large manufacturing enterprises. First, the average waiting time of the production line was reduced from 300 minutes before implementation to 105 minutes, a reduction of 65%. This significant reduction shows that the present invention has played an important role in optimizing production processes, reducing bottlenecks and improving production efficiency. At the same time, the resource utilization rate of the enterprise has increased from 70% before implementation to 91%, an increase of 30%. This result shows that through intelligent resource allocation and optimization, the present invention has effectively improved the utilization efficiency of resources and reduced resource waste. The number of production delays caused by information lag has been reduced from 25 times per month to 14 times, a reduction of 44%. This shows that the present invention has significantly reduced the production delay problem caused by information lag through real-time data update and dynamic adjustment of the process. In terms of cross-departmental communication, efficiency has also been significantly improved, reaching a 25% improvement. The incidence of data conflicts has been reduced from 20 times per month to 4 times, a reduction of 80%. This significant reduction shows that in multi-user collaborative work, the conflict detection and synchronization functions of the present invention greatly reduce errors caused by inconsistent operations and ensure the smooth operation of the process. After implementing the present invention, production costs dropped by 15%, while overall operational efficiency increased by 20%. This result fully demonstrates that the present invention is effective in helping enterprises optimize processes, improve resource utilization and reduce waste, bringing significant economic benefits and operational advantages to enterprises.

[0172] Through retrospective analysis of historical versions, the company also found that the production efficiency was significantly improved after each process optimization. Specifically, through the comparison and dynamic synthesis function of historical versions, the production efficiency was increased from 10% of version 1 to 22.2% of version 3. This continuous improvement effect shows that the present invention can not only solve the current problem, but also provide continuous improvement space for enterprises through continuous optimization and adjustment, further improving the competitiveness and market responsiveness of enterprises.

[0173] In summary, the application of the present invention in this large-scale manufacturing enterprise has significantly improved the management and optimization capabilities of business processes, effectively solved multiple pain points in traditional process management methods, greatly improved the enterprise's production efficiency and resource utilization, and showed excellent improvement effects in multiple key performance indicators. By improving production efficiency, optimizing resource utilization, reducing production delays and data conflicts, and reducing production costs, the overall operational efficiency and economic benefits of the enterprise have been greatly improved.

[0174] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A business process navigation method based on visualization algorithm, characterized in that: The steps include: S1. Collect and organize business process data, including process steps, task allocation, role responsibilities and dependencies; S2. Convert the business process data into a standardized graphical model using a graph embedding and high-dimensional data visualization algorithm, wherein the graph embedding algorithm is used to represent the flow chart as an embedding vector in a low-dimensional space to capture the deep relationship between nodes, and the high-dimensional data visualization algorithm reduces the dimension of high-dimensional data through UMAP technology to achieve a two-dimensional or three-dimensional intuitive display; S3. Display a standardized graphical model on the business navigation interface. The interface provides interactive functions, including node zooming, dragging, positioning and customizing views. Users can view and adjust different perspectives of the business process in real time. S4. Based on the standardized graphical model displayed, the real-time data update function is integrated using the quantum molecular dynamics simulation algorithm. By simulating the dynamic behavior of tasks and resources, the future state of the process is predicted, and the navigation map is automatically updated according to the actual data to reflect the latest process status; S5. Based on the updated navigation map, a multi-level view is provided, and users can switch between the overall flowchart and task view to meet the needs of managers at different levels; S6. Based on the updated process data, the random matrix theory optimization algorithm is used to intelligently optimize the business process. By analyzing large-scale data associations, key patterns and process noise are identified, paths and resource allocation are optimized, and optimization suggestions are provided in a visual interface. S7. Record and save the optimized process version, provide the comparison and backtracking function of historical process versions, so that users can compare different versions of flowcharts through a visual interface and analyze the process improvement effect; S8. Based on the saved process version, it supports multi-user collaboration, allowing multiple users to view and edit business process diagrams at the same time, and realize synchronous update and conflict detection through a visual interface.

2. A business process navigation method based on visualization algorithm according to claim 1, characterized in that: The S2 specifically includes: S21. Structuring the business process data into a graph G(V,E), where V is a set of nodes, representing each step or task in the business process, and E is a set of edges, representing the dependency or order between nodes. i Corresponding to the specific steps and tasks in the business process, ij Represents task v i and v j The dependencies between them; S22. Process the graph G(V,E) through the graph embedding algorithm and embed each node v i Mapped to an embedding vector h in a low-dimensional space i ∈R d , where d is the dimension of the embedding space. The embedding process captures the deep relationship between nodes by optimizing the objective function: in, is the sigmoid function, h u and h v are the embedding vectors of nodes u and v respectively; S23, using UMAP technology to embed the vector set H = {h1,h2,…,h n } performs dimensionality reduction processing and embeds the high-dimensional vector h i Mapping to low-dimensional space z i ∈R k , where k = 2 or k = 3; S24. Construct a neighbor graph in high-dimensional space and calculate the similarity between nodes. The weight of the neighbor graph is calculated using the Gaussian kernel function: Among them, σ i is the local scale parameter of node i; S25. In low-dimensional space, construct an optimization objective function so that the distance between points in low-dimensional space reflects the proximity relationship in high-dimensional space: in, Represents the total loss of point pairs in low-dimensional space, ∥z i -z j ∥2 represents the distance between nodes i and j in the low-dimensional space, w ij is the weight of edge (i,j), is the difference between the distances of point pairs in high-dimensional space; S26, generate a two-dimensional or three-dimensional standardized graphical model, and use the node representation Z={z1,z2,…,z n }Shown in the visual interface, each node z i Shown by its position in low-dimensional space, edge z i ,z j ) represents the dependency relationship between tasks.

3. The business process navigation method based on visualization algorithm according to claim 1, characterized in that: The S3 specifically includes: S31, standardize the graphical model Z = {z1, z2, ..., z n } is mapped to the visualization area of ​​the business navigation interface, so that each node z i Position and edge in the interface (z i ,z j )’s connection relationship maintains its relative position in the low-dimensional space; S32. Implement node zoom function in the business navigation interface. Users can adjust node z by using mouse wheel and other zoom controls. i The display size of S33, provide node dragging function, users can click and drag the node z i Move freely in the visualization area, and keep the z coordinates of adjacent nodes during the movement of nodes. j Real-time update of connection relationships; S34, provide intelligent alignment function, users drag node z i When the system automatically detects the node and its adjacent nodes z j The nodes are aligned and the automatic adsorption function is used to align the nodes, keeping the connection relationship between the nodes neat and consistent: z′ i =Align(z i ,z j )+Δz align ; Among them, Align(z i ,z j ) is the node position after automatic alignment, Δz align is the increment for adjustment after alignment; S35, provides positioning function, users can automatically position the current view center to the specified node z by clicking the control in the interface t The positioning process is performed by calculating the node z t The offset Δz from the center of the current view t And implement view translation; S36, introduce adaptive layout function, users adjust node z i and view, the system automatically rearranges other nodes z j To optimize the overall layout, the adaptive layout process is based on the following optimization objective function: Among them, λ is the weight parameter, Edge_crossing(z i ,z j ) represents the number of intersections between edges. The system minimizes the objective function L layout Automatically optimize node positions; S37. Provide custom view function, users can select different business process perspectives through interface controls, including zooming in on a specific process area or zooming in on the whole to see a global view. The view switching operation updates the view parameters, including the perspective, zoom ratio and visible area. S38. When users view and adjust different perspectives of a business process in real time, the display status of all nodes and edges is dynamically updated according to user operations, and the graphical model reflects the latest perspective and layout of the business process.

4. The business process navigation method based on visualization algorithm according to claim 1, characterized in that: The S4 specifically includes: S41, standardize the graphical model Z = {z1, z2, ..., z n }Extract each node z i The initial state of each node z i Represents a task or resource in a business process. Each edge (z i ,z j ) represents the interdependencies between tasks or resources; S42. Establish a quantum molecular dynamics model and define node z i The state vector r i (t) and momentum vector p i (t), respectively represent the location and change rate of the task or resource at time t, and the interaction between nodes is expressed by the interaction potential energy function V(r i ,r j )describe: Among them, k ij For node z i and z j The interaction strength between them, ∥r i -r j ∥ is the distance between two nodes; S43, through quantum molecular dynamics simulation, use the following motion equation to predict node z i The state at the future time t+Δt: Among them, m i For node z i The quality of the corresponding task or resource, Δt is the time step; S44, updating the states of the nodes according to the future states obtained by simulation, and mapping these updated states to the graphical model Z′={z′1,z′2,…,z′ n }, where the updated state z′ of each node i By r i (t+Δt) determines; S45. Adjust the parameter k of the quantum molecular dynamics model in real time through data feedback in the actual business process ij and m i , and recalculate the node status to generate new process data and reflect the latest business status; S46, mapping the updated graphical model Z′ back to the business navigation interface to generate an updated navigation map, through which the user can view and adjust possible future process changes and resource allocations in real time.

5. The business process navigation method based on visualization algorithm according to claim 1, characterized in that: The S5 specifically includes: S51, from the generated updated navigation map Z′={z′1,z′2,…,z′ n } to extract the overall business process view, which contains all nodes z′ i and their interconnected edges (z′ i ,z j ′); S52, grouping related tasks and step nodes in the business process to form multiple task subsets C k , where each task subset C k ={z′ k1 ,z′ k2 ,…} represents a sub-process or task group, generating a local task view Z′ for the user k Provide a foundation; S53. Keep the global view Z′ and the local task view Z′ in the system k The interface controls provide a view switching function, allowing users to switch between the overall view Z′ and the local view Z′ k Freely switch between views, and dynamically adjust view parameters during view switching; S54. Added an automatic focus function. When the user switches from the global view to the local task view, it automatically identifies and focuses on the task or node that the user is currently interested in, automatically zooms in on the relevant details of these nodes when the view is switched, and automatically adjusts the viewing angle to best display the overall process when switching back to the global view; S55. Introduce the multi-view collaboration function, so that all operations of users on different levels of views can be synchronized in real time in the updated navigation map Z′, and automatically updated to all related views to ensure coordination and consistency between multi-level views.

6. A business process navigation method based on visualization algorithm according to claim 1, characterized in that: The S6 specifically includes: S61. Extract node z′ from updated process data i The state and task dependency (z′ i ,z′ j ) and other relevant business process parameters to provide input data for subsequent optimization analysis; S62. Use random matrix theory to construct the correlation matrix R of process data, where the matrix elements R ij Represents the task node z′ i and z j By analyzing the eigenvalues ​​and eigenvectors of the correlation matrix R, the key patterns and potential noises in the business process are identified; S63. By calculating the spectral density ρ(λ) of the correlation matrix R, the eigenvalue λ related to the process noise is identified, and filtering technology is applied to remove the noise, retaining the feature information that is meaningful to the business process optimization: Where δ is the Diracdelta function and N is the total number of eigenvalues; S64. Based on the processed correlation matrix, the path and resource configuration of the business process are optimized. The optimal task execution path is determined by maximizing the high correlation feature value on the path. The goal of optimizing the path is to maximize the path correlation and minimize the execution time: Among them, t ij is the task node z′ i and z j 'Execution time between S65. These suggestions are presented to the user through a visual interface. The user can view the optimized path and resource allocation plan in the interface, and make corresponding business process adjustments to generate an optimized process version.

7. The business process navigation method based on visualization algorithm according to claim 1, characterized in that: The S7 specifically includes: S71. Save the generated optimized path and resource configuration as an independent process version and record the version identifier V n , where V n Indicates the nth optimized process version; S72. For each saved process version V n Recording Flowchart Z n The state of the node z′ i The location and task dependency (z′ i ,z j ′), path optimization information and resource allocation status; S73, provide version management function, allowing users to view historical process versions V1, V2, ..., V n , the user can select a specific version to load and the corresponding flowchart Z n Map to the visual interface for display; S74, using the dynamic process synthesis function, the user selects multiple historical versions V i and V j When comparing, a synthetic version V is automatically generated s , synthesize the optimal path and resource configuration, generate and display the synthetic version based on the predefined optimization goals; S75, provide intelligent backtracking decision function, users can choose to backtrack to the historical version V m When based on the current version V n With V m Predict the retrospective impact of changes between the two, and generate a variety of retrospective strategies for users to choose from. Each strategy includes analysis of process efficiency, resource usage, and potential risks to help users make decisions; S76. During the backtracking process, the system updates the navigation map in real time and rolls back the process status to version V m Flowchart after Z m Mapped to the visual interface, users can immediately view the status after backtracking and select the synthetic version or backtracking strategy for adjustment and optimization as needed.

Images