A dynamically assembled process knitting method and system

Abstract

The present invention discloses a dynamically assembled process weaving method and system, the method includes the following steps: S1, automatic process discovery; S2, process link disassembly; S3, process intelligent navigation; S4, process dynamic weaving; S5, process robot creation; S6, process simulation optimization; the system includes: an AI intelligent server and several groups of PC terminals set in various departments of the enterprise; several groups of PC terminals interact with the AI ​​intelligent server through a local area network, and are used to execute the dynamically assembled process weaving method according to the PC terminal instructions. The present invention can intelligently reorganize the dynamically released process weaving technical method and system, can realize the on-demand recommendation of process advancement roadmap through process intelligent navigation and intelligent assembly, realize the dynamic weaving of process models, avoid a large amount of repetitive process modeling work, and at the same time, through process robots, replace manual automatic execution of repetitive and cumbersome user operations, improve output, and reduce errors.

Description

Technical Field

[0001] The present invention relates to the technical field of workflow engines, and more specifically, to a dynamically assembled process weaving method; the present invention also relates to a dynamically assembled process weaving system. Background Art

[0002] Workflow technology is to decompose the business process into a series of standardized business process definition descriptions such as tasks and roles, and provide interpretation, execution and orchestration for the business process definition through a workflow engine, so that people and various application tools can work together. Usually a workflow model contains multiple links, and each link needs to circulate according to pre-set rules to ensure the accuracy of the workflow output results. In actual work, it is inevitable that there are cross-system and cross-business situations. The process needs to be implemented in stages or departments, and the basis for the later implementation may be the output of the previous implementation results. Therefore, when combing the business process, it is necessary to construct a separate workflow model for various possible situations. In addition, the same business process may be reused in different implementation stages, and it is impossible to update it in one place and everywhere. Therefore, based on the existing problems, the present invention provides a dynamically assembled process weaving method and system. Summary of the invention

[0003] The purpose of the present invention is to solve the shortcomings existing in the prior art, and to propose a process weaving method and system that can be dynamically assembled, which can intelligently reorganize the dynamically released process weaving technical method and system, and can realize on-demand recommendation of process advancement roadmaps through process intelligent navigation and intelligent assembly, thereby realizing dynamic weaving of process models and avoiding a large amount of repetitive process modeling work. At the same time, through process robots, it can replace manual automatic execution of repetitive and tedious user operations, thereby improving output and reducing errors.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A dynamically assembled process knitting method comprises the following steps:

[0006] S1. Automatic process discovery, using machine learning to automatically analyze and integrate multiple workflow records from different users;

[0007] S2. Decompose the process links, identify the key elements in the links, and decompose them into reusable material components of roles, tasks, and rules;

[0008] S3, process intelligent navigation, combines process links to generate a process navigation map, and generates multiple combination paths for selection as needed;

[0009] S4, dynamic weaving of processes, extracting the links and flow directions on the route according to the route selected on the navigation map for dynamic assembly and weaving;

[0010] S5. Create process robots to automatically perform repetitive and tedious user operations instead of humans, thereby increasing output and reducing errors.

[0011] S6. Process simulation optimization, generate process instances and simulate operation, collect and analyze flow log operation process information.

[0012] Preferably, in step S1, the specific implementation process of automatic process discovery is:

[0013] 1) Process capture, which records necessary data by capturing the interaction between the user and the application. The data recording adopts one or more of the following methods: mouse click, keyboard operation, handwriting board input, and voice input;

[0014] 2) Process analysis, using machine learning to automatically analyze process discovery records and analyze user operations;

[0015] 3) Process identification: identify the key links in the process and prioritize them to prepare for the subsequent process automation.

[0016] Preferably, in step S2, the process link disassembly specifically includes:

[0017] 1) Determine which people are needed for each link and break down the independent roles;

[0018] 2) Break down the tasks into independent ones by understanding what needs to be done in each link;

[0019] 3) Break down independent rules by corresponding the links between people and things, people and people, and things and things.

[0020] Preferably, in step S3, the specific process of process intelligent navigation is:

[0021] 1) Process link combination: combine all process links to generate a complete process navigation map. Click on the route or link on the navigation map to display detailed information of the route or link.

[0022] 2) Business processing route query: input business processing parameters, including processing items, previous items, and sponsors, and perform route query based on the input parameters;

[0023] 3) Calculate and generate the optimal route, and find multiple process advancement routes from the starting node to the ending node through the depth traversal and breadth traversal of the graph;

[0024] 4) Mark the optimal route on the navigation map. The currently selected route is highlighted, and the unselected ones are marked in light color. Click on different routes to switch.

[0025] Preferably, in step S4, the specific process of dynamic weaving is:

[0026] 1) Process information extraction: according to the route selected on the navigation map, from the starting node to the ending node, extract the links, roles, rules and other information in sequence;

[0027] 2) Dynamic process assembly: After a series of assembly guidance steps, the extracted links, roles, rules, etc. are dynamically assembled to generate a complete process model;

[0028] 3) Improve the circulation rules, analyze the generated process model, and supplement and improve some links where the business rules are unclear.

[0029] Preferably, in step S5, the specific steps of creating a process robot are:

[0030] 1) Analyze and filter out repetitive operations based on the information extracted from each link;

[0031] 2) Automatically create intelligent process robots through AI algorithms to simulate the human-computer interaction between users and applications;

[0032] Among them, the AI ​​algorithm workflow is:

[0033] Load the training model and dataset sample labels, obtain the tensor in the training model, and convert the input image into numpy array format; use the CNN convolution layer to extract features from the input image; use the region proposal network RPN to generate a proposal window; map the proposal window to the last convolution feature map of the CNN convolution layer, and use the target detection special layer ROIPooling to generate a fixed-size feature map for each region of interest ROI; perform classification regression and position regression; and perform position annotation and category prediction on the dynamic model in the original input image.

[0034] Preferably, the process simulation optimization described in step S6 has the following specific process:

[0035] 1) Generate a temporary process instance based on the published business process to simulate the effect of actual flow;

[0036] 2) Automatically receive and run the process from the start node to the end node, simulating the flow of each link;

[0037] 3) For some nodes that cannot be automatically transferred, manual intervention is used to ensure the normal flow of the process;

[0038] 4) Record and analyze the circulation log and other information, identify any problematic areas, and return to step S4 to further optimize the circulation rules.

[0039] A dynamically assembled process knitting system, comprising:

[0040] AI intelligent servers and several groups of PC terminals set up in various departments of the enterprise;

[0041] Several groups of PC terminals interact with the AI ​​intelligent server through a local area network, and are used to execute the dynamically assembled process knitting method according to PC terminal instructions;

[0042] As a preferred dynamically assembled process weaving system, several groups of PC terminals need to log in with a password, and the password adopts a pin+token mode. The pin is held by the AI ​​intelligent server operation and maintenance administrator account, and the token is held by several groups of PC terminal staff. When the PC terminal staff logs in to the PC terminal, the AI ​​intelligent server executes the dynamically assembled process weaving method according to the PC terminal instructions.

[0043] Technical effects and advantages of the present invention:

[0044] The present invention provides a dynamically assembled process weaving method and system. Compared with the traditional workflow engine technology, the present invention can intelligently reorganize the dynamically released process weaving technical method and system, and can realize on-demand recommendation of process advancement roadmaps through process intelligent navigation and intelligent assembly, thereby realizing dynamic weaving of process models and avoiding a large amount of repetitive process modeling work. At the same time, through process robots, repetitive and cumbersome user operations can be automatically performed instead of manual operations, thereby improving output and reducing errors.

[0045] On the basis of the above beneficial effects, the process weaving method that can be intelligently reorganized and dynamically published proposed in the present invention is to disassemble a series of pre-modeled workflow models into reusable material components, and then calculate the actual process advancement roadmap through rule algorithms according to the metadata information of the current business and the output information of the previous business, and then obtain the required components from the material warehouse for intelligent assembly to realize the dynamic weaving of the process model. Moreover, through the process weaving method that can be intelligently reorganized and dynamically published, the entire process of process assembly and publication does not require the writing of professional code, thus avoiding a large amount of repeated modeling workload, and also reducing the difficulty of workflow development. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 The present invention is a flow chart of a dynamically assembled process weaving method. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0048] The present invention provides Figure 1 A dynamically assembled process weaving method and system, wherein the dynamically assembled process weaving method comprises the following steps:

[0049] S1. Automatic process discovery, using machine learning to automatically analyze and integrate multiple workflow records from different users;

[0050] In step S1, the specific implementation process of automatic process discovery is as follows:

[0051] 1) Process capture, which records necessary data by capturing the interaction between the user and the application. Data recording uses one or more of the following methods: mouse click, keyboard operation, handwriting tablet input, and voice input;

[0052] 2) Process analysis, using machine learning to automatically analyze process discovery records and analyze user operations;

[0053] 3) Process identification: identifying the key links in the process and prioritizing them to prepare for the subsequent process automation;

[0054] S2. Decompose the process links, identify the key elements in the links, and decompose them into reusable material components of roles, tasks, and rules;

[0055] In step S2, the process steps are specifically broken down into:

[0056] 1) Determine which people are needed for each link and break down the independent roles;

[0057] 2) Break down the tasks into independent ones by understanding what needs to be done in each link;

[0058] 3) Use the links to correspond to the connections between people and things, people and people, and things and things, and disassemble independent rules;

[0059] S3, process intelligent navigation, combines process links to generate a process navigation map, and generates multiple combination paths for selection as needed;

[0060] In step S3, the specific process of process intelligent navigation is as follows:

[0061] 1) Process link combination: combine all process links to generate a complete process navigation map. Click on the route or link on the navigation map to display detailed information of the route or link.

[0062] 2) Business processing route query: input business processing parameters, including processing items, previous items, and sponsors, and perform route query based on the input parameters;

[0063] 3) Calculate and generate the optimal route, and find multiple process advancement routes from the starting node to the ending node through the depth traversal and breadth traversal of the graph;

[0064] 4) Mark the optimal route on the navigation map. The currently selected route is highlighted, and the unselected ones are marked in light colors. Click on different routes to switch;

[0065] S4, dynamic weaving of processes, extracting the links and flow directions on the route according to the route selected on the navigation map for dynamic assembly and weaving;

[0066] In step S4, the specific process of dynamic weaving is as follows:

[0067] 1) Process information extraction: according to the route selected on the navigation map, from the starting node to the ending node, extract the links, roles, rules and other information in sequence;

[0068] 2) Dynamic process assembly: After a series of assembly guidance steps, the extracted links, roles, rules, etc. are dynamically assembled to generate a complete process model;

[0069] 3) Improve the circulation rules, analyze the generated process model, and supplement and improve some links where the business rules are unclear;

[0070] S5. Create process robots to automatically perform repetitive and tedious user operations instead of humans, thereby increasing output and reducing errors.

[0071] In step S5, the specific steps for creating a process robot are:

[0072] 1) Analyze and filter out repetitive operations based on the information extracted from each link;

[0073] 2) Automatically create intelligent process robots through AI algorithms to simulate the human-computer interaction between users and applications;

[0074] Among them, the AI ​​algorithm workflow is:

[0075] Load the training model and dataset sample labels, obtain the tensor in the training model, and convert the input image into numpy array format;

[0076] Use CNN convolutional layer to extract features from input images;

[0077] Generate proposal windows using the region proposal network RPN;

[0078] Map the proposed window to the last convolutional feature map of the CNN convolutional layer, and generate a fixed-size feature map for each region of interest ROI through the target detection special layer ROIPooling;

[0079] Perform classification regression and position regression; perform position annotation and category prediction on the dynamic model in the original input image;

[0080] S6, process simulation optimization, generate process instances and simulate operation, collect and analyze the flow log operation process information;

[0081] The specific process of the process simulation optimization of step S6 is as follows:

[0082] 1) Generate a temporary process instance based on the published business process to simulate the effect of actual flow;

[0083] 2) Automatically receive and run the process from the start node to the end node, simulating the flow of each link;

[0084] 3) For some nodes that cannot be automatically transferred, manual intervention is used to ensure the normal flow of the process;

[0085] 4) Record and analyze the circulation log and other information, identify any problematic areas, and return to step S4 to further optimize the circulation rules.

[0086] This method breaks down a series of pre-modeled workflow models into reusable material components, calculates the actual process advancement roadmap through rule algorithms based on the metadata information of the current business and the output information of the previous business, and then obtains the required components from the material warehouse for intelligent assembly to achieve dynamic weaving of process models. In addition, through the process weaving method that can intelligently reorganize and dynamically publish, the entire process of process assembly and publication does not require the writing of professional code, avoiding a large amount of repeated modeling workload, and also reducing the difficulty of workflow development.

[0087] The main technical advantages of the above method are:

[0088] 1. Automatically discover the process model and disassemble it according to certain rules;

[0089] 2. Automatically generate multiple process assembly plans based on the algorithm and provide intelligent assembly guidance;

[0090] 3. Reorganize and release the process model according to the selected optimal route;

[0091] 4. Realize process automation through process robots.

[0092] This embodiment also provides a dynamically assembled process knitting system, including:

[0093] AI intelligent servers and several groups of PC terminals set up in various departments of the enterprise;

[0094] Several groups of PC terminals interact with the AI ​​intelligent server through a local area network to execute a dynamically assembled process knitting method according to PC terminal instructions;

[0095] Furthermore, several groups of PC terminals need to log in with a password. The password adopts a pin+token mode. The pin is held by the AI ​​intelligent server operation and maintenance administrator account, and the token is held by several groups of PC terminal staff. When the PC terminal staff logs in to the PC terminal, the AI ​​intelligent server executes the dynamically assembled process weaving method according to the PC terminal instructions.

[0096] In summary, it is possible to use intelligent process navigation to recommend processes on demand and to perform intelligent assembly, thereby dynamically weaving process models and avoiding a large amount of repetitive process modeling work. At the same time, process robots can be used to replace manual labor to automatically perform repetitive and tedious user operations, thereby improving output and reducing errors.

[0097] As an optional, dynamically assembled process braiding system, including:

[0098] Process definition module, business data module and process operation module;

[0099] The process definition module is used to define at least one workflow based on a digital signature algorithm;

[0100] The process running module is also used to: obtain data encryption rules according to business requests, and the data encryption rules include: no encryption, partial encryption, and full encryption;

[0101] Based on the data encryption rule, according to the key pair corresponding to at least one target node included in the target workflow, and in accordance with the flow order of at least one target node, execute at least one target node;

[0102] The business data module is used to receive a business request and determine a target workflow corresponding to the business request according to at least one workflow;

[0103] The process running module is used to run the target workflow according to the business request to build the target workflow instance corresponding to the business request;

[0104] Specifically, the process running module is further used for: if the data encryption rule is partial encryption or full encryption, in the process of executing the current target node, using the public key corresponding to the previous target node to verify the digital signature of the previous target node, and using the private key corresponding to the current target node to decrypt the encrypted data submitted by the previous target node;

[0105] If the verification passes and the decryption is successful, the decrypted data and the unencrypted data submitted by the previous target node are processed, the processed data is digitally signed using the private key and flow order corresponding to the current target node, and the processed data to be encrypted is encrypted using the public key corresponding to the next target node, and then the processed data is submitted to the next target node;

[0106] If the verification fails or the decryption fails, the failure to build the target workflow instance is recorded.

[0107] If the above system is adopted, the effect is:

[0108] Using digital signatures to control the permissions of each node in the workflow solves the technical problem that the identity information of the workflow nodes in the prior art needs to be obtained from the business system, and reduces the coupling between the workflow engine and the business system;

[0109] In addition, during the workflow running process, the process running module can serialize the data corresponding to the workflow, so as to realize the call between multiple workflows, and then realize the call of cross-system workflows.

[0110] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A dynamically assembled process knitting method, characterized in that: The steps include: S1. Automatic process discovery, using machine learning to automatically analyze and integrate multiple workflow records from different users; 1) Process capture, which records necessary data by capturing the interaction between the user and the application. The data recording adopts one or more of the following methods: mouse click, keyboard operation, handwriting board input, and voice input; 2) Process analysis, using machine learning to automatically analyze process discovery records and analyze user operations; 3) Process identification: identifying the key links in the process and prioritizing them to prepare for the subsequent process automation; S2. Decompose the process links, identify the key elements in the links, and decompose them into reusable material components of roles, tasks, and rules; 1) Determine which people are needed for each link and break down the independent roles; 2) Break down the tasks into independent ones by understanding what needs to be done in each link; 3) Use the links to correspond to the connections between people and things, people and people, and things and things, and disassemble independent rules; S3, process intelligent navigation, combines process links to generate a process navigation map, and generates multiple combination paths for selection as needed; 1) Process link combination: combine all process links to generate a complete process navigation map. Click on the route or link on the navigation map to display detailed information of the route or link. 2) Business processing route query: input business processing parameters, including processing items, previous items, and sponsors, and perform route query based on the input parameters; 3) Calculate and generate the optimal route, and find multiple process advancement routes from the starting node to the ending node through the depth traversal and breadth traversal of the graph; 4) Mark the optimal route on the navigation map. The currently selected route is highlighted, and the unselected ones are marked in light colors. Click on different routes to switch; S4, dynamic weaving of processes, extracting the links and flow directions on the route according to the route selected on the navigation map for dynamic assembly and weaving; 1) Process information extraction: according to the route selected on the navigation map, from the starting node to the ending node, extract the link, role, and rule information in sequence; 2) Dynamic process assembly: After a series of assembly guidance steps, the extracted links, roles, and rules are dynamically assembled to generate a complete process model; 3) Improve the circulation rules, analyze the generated process model, and supplement and improve some links where the business rules are unclear; S5. Create process robots to automatically perform repetitive and tedious user operations instead of humans, thereby increasing output and reducing errors. S6. Process simulation optimization, generate process instances and simulate operation, collect and analyze flow log operation process information.

2. A dynamically assembled process knitting method according to claim 1, characterized in that: In step S5, the specific steps for creating a process robot are: 1) Analyze and filter out repetitive operations based on the information extracted from each link; 2) Automatically create intelligent process robots through AI algorithms to simulate the human-computer interaction between users and applications; Among them, the AI ​​algorithm workflow is: Load the training model and dataset sample labels, obtain the tensor in the training model, and convert the input image into numpy array format; use the CNN convolution layer to extract features of the input image; use the region proposal network RPN to generate a proposal window; map the proposal window to the last convolution feature map of the CNN convolution layer, and use the target detection special layer ROIPooling to generate a fixed-size feature map for each region of interest ROI; perform classification regression and position regression; and perform position annotation and category prediction on the dynamic model in the original input image.

3. A dynamically assembled process knitting method according to claim 1, characterized in that: The specific process of the process simulation optimization described in step S6 is as follows: 1) Generate a temporary process instance based on the published business process to simulate the effect of actual flow; 2) Automatically receive and run the process from the start node to the end node, simulating the flow of each link; 3) For some nodes that cannot be automatically transferred, manual intervention is used to ensure the normal flow of the process; 4) Record and analyze the flow log information, identify any problematic areas, and return to step S4 to further optimize the flow rules.

4. A dynamically assembled process knitting system, characterized in that: include: AI intelligent servers and several groups of PC terminals set up in various departments of the enterprise; Several groups of PC terminals interact with the AI ​​intelligent server through a local area network, and are used to execute the method described in any one of claims 1 to 3 according to PC terminal instructions.

5. A dynamically assembled process knitting system according to claim 4, characterized in that: Several groups of PC terminals need to log in with a password. The password adopts a pin+token mode. The pin is held by the AI ​​smart server operation and maintenance administrator account, and the token is held by several groups of PC terminal staff.

Images