CN-121235435-B - Dynamic workflow automatic optimization management method and system based on artificial intelligence

CN121235435BCN 121235435 BCN121235435 BCN 121235435BCN-121235435-B

Abstract

The invention belongs to the technical field of computer technology and industrial automation, and relates to an automatic optimization management method and system of dynamic workflow based on artificial intelligence, comprising the following steps of acquiring multi-source heterogeneous real-time data and calibrating task charge attributes for a task to be executed according to a preset conversion rule; the method comprises the steps of constructing a real-time task potential field diagram based on task charge attributes of each task through a preset acting force calculation model, identifying and extracting task conflict points and cooperation opportunity points from potential energy gradient information in the real-time task potential field diagram to generate a task situation list, generating disturbance intervention instruction streams according to potential energy gradient characteristic differences in the task situation list, executing the disturbance intervention instruction streams, collecting result data after instruction execution, and feeding back and updating task charge attributes. The invention solves the problems that the tasks are generally regarded as isolated individuals in the traditional mode, and complex coordination and competition relations between the tasks are difficult to quantify and process.

Inventors

LI XIAOQIANG
WEI CHAODONG
Qiu Zhuyao

Assignees

福建龙创未来软件科技有限公司
厦门龙创未来人工智能有限公司

Dates

Publication Date: 20260512
Application Date: 20251202

Claims (8)

1. The automatic dynamic workflow optimization management method based on artificial intelligence is characterized by comprising the following steps of: S1, acquiring multi-source heterogeneous real-time data, and calibrating task charge attributes comprising a quality value, an electrical value and a priority potential energy value for a task to be executed according to a preset conversion rule, wherein the step of calibrating the task charge attributes comprises the steps of calculating the quality value according to the resource consumption and processing complexity of the task, defining the task with a cooperative relationship as opposite charges attracted to each other according to the interaction between the tasks, defining the task with a resource competition relationship as same charges which are mutually exclusive to determine the electrical value, calculating the priority potential energy value according to the delivery date of the task, periodically increasing the priority potential energy value of the task if the waiting time of the task exceeds a preset time threshold, and increasing a risk added value for the quality value of the task if the multi-source heterogeneous real-time data indicates that resources related to the task have fault early warning; s2, constructing a real-time task potential field diagram through a preset acting force calculation model based on task charge attributes of each task; s3, identifying and extracting task conflict points and collaboration opportunity points from potential energy gradient information in a real-time task potential field diagram, and generating a task situation list; S4, generating a disturbance intervention instruction stream according to potential energy gradient characteristic differences in the task situation list; s5, executing disturbance intervention instruction flow, collecting result data after instruction execution, and feeding back to S1 for updating task charge attribute.
2. The automatic optimization management method for dynamic workflow based on artificial intelligence according to claim 1, wherein the construction of the real-time task potential field diagram comprises the following steps: instantiating the task into particles with quality values, electrical values and priority potential energy values, and placing the particles in a multidimensional virtual coordinate system representing a task resource space according to the resource requirements of the particles; Calculating interaction force generated between task particles based on task charge attributes by using an acting force calculation model; and constructing a real-time task potential field diagram based on the calculated interaction force and the priority potential energy value of each task.
3. The automatic optimization management method for dynamic workflow based on artificial intelligence according to claim 1, wherein the task situation list is generated, comprising the steps of: identifying an area with potential energy higher than a preset conflict threshold value in a real-time task potential field diagram as a task conflict point, and extracting corresponding high potential energy gradient characteristics; identifying a region or an attractive force track with potential energy lower than a preset opportunity threshold in a real-time task potential field diagram as a cooperative opportunity point, and extracting a corresponding low potential energy gradient characteristic; and summarizing the identified task conflict points, the identified coordination opportunity points and the corresponding potential energy gradient characteristics thereof to form a task situation list.
4. A dynamic workflow automatic optimization management method based on artificial intelligence according to claim 3, wherein generating disturbance intervention instruction stream comprises the steps of: Aiming at task conflict points and high potential energy gradient characteristics thereof in a task situation list, generating a time offset instruction for suggesting a specific task to perform time adjustment; And generating resource link instructions for suggesting binding or distributing the mutually attracted tasks to the same resource according to the collaborative opportunity points and the low potential energy gradient characteristics thereof in the task situation list.
5. The automatic optimization management method for dynamic workflow based on artificial intelligence according to claim 3, wherein the step of generating disturbance intervention instruction stream further comprises the steps of: if the task is endowed with a risk added value due to the occurrence of fault early warning of the associated resource, and thus a high potential energy gradient characteristic is formed, generating a repulsive force field instruction for increasing repulsive force around the fault early warning resource region when constructing a real-time task potential field diagram.
6. The automatic optimization management method for dynamic workflow based on artificial intelligence according to claim 1, further comprising the steps of, after the step of generating the task situation list: displaying a real-time task situation map and a task situation list through a graphical interaction interface; receiving a calibration instruction which is input through a graphical interaction interface and is used for adjusting acting force of a specific area or temporarily improving a priority potential energy value of a specific task; The calibration instructions are converted into adjustment parameters and input into S1 to update task charge attributes.
7. The automatic optimization management method for dynamic workflow based on artificial intelligence according to claim 1, wherein the disturbance intervention instruction stream is executed and fed back, comprising the steps of: Sending a disturbance intervention instruction stream to a task execution site; Collecting execution start time, completion time and resource consumption changes of tasks associated with the disturbance intervention instruction stream to form execution result data; Based on the execution result data, the task charge attribute of the affected task is updated to trigger a new round of real-time task potential field diagram construction.
8. The automatic dynamic workflow optimization management system based on artificial intelligence is characterized by comprising the following modules: The task attribute calibration module is used for acquiring multi-source heterogeneous real-time data and calibrating task charge attributes comprising a quality value, an electrical value and a priority potential energy value for a task to be executed according to a preset conversion rule, wherein the step of calibrating the task charge attributes comprises the steps of calculating the quality value according to the resource consumption and the processing complexity of the task, defining the task with a cooperative relationship as opposite charges attracted to each other according to the interaction between the tasks, defining the task with a resource competition relationship as the same charges which repel each other so as to determine the electrical value, calculating the priority potential energy value according to the delivery date of the task, periodically increasing the priority potential energy value of the task if the waiting time of the task exceeds a preset time threshold, and increasing the risk added value for the quality value of the task if the multi-source heterogeneous real-time data indicates that resources related to the task have fault early warning; The real-time potential field construction module is used for constructing a real-time task potential field diagram through a preset acting force calculation model based on task charge attributes of each task; The task situation identification module is used for identifying and extracting task conflict points and collaboration opportunity points from potential energy gradient information in the real-time task situation map, and generating a task situation list; the intervention instruction generation module generates a disturbance intervention instruction stream according to potential energy gradient characteristic differences in the task situation list; the instruction execution and feedback module is used for executing disturbance intervention instruction flow, collecting result data after instruction execution, and feeding back the result data to the task attribute calibration module for updating task charge attributes.

Description

Dynamic workflow automatic optimization management method and system based on artificial intelligence Technical Field The invention belongs to the technical field of computer technology and industrial automation, and relates to an automatic optimization management method and system for dynamic workflow based on artificial intelligence. Background In modern manufacturing industry and complex business process management, a production site or business environment is a complex system with dynamic changes, and is full of multi-source real-time information, such as equipment states, material inventory, personnel availability, emergency order insertion and the like, and the factors act together, so that the execution relationship among tasks presents competitiveness and synergy, the traditional scheduling method is difficult to comprehensively capture and respond to the dynamic complexity in real time, and the problems of resource waste, efficiency bottleneck, delivery delay and the like are caused by the disconnection between a production plan and actual execution. The solutions currently in common use in the industry rely primarily on embedded scheduling modules in a manufacturing execution system or enterprise resource planning system, which typically assign and order tasks based on preset static rules, limited throughput scheduling algorithms, or simple priority queues. For example, a dispatcher may manually or semi-automatically drag schedule on the Gantt chart based on the delivery date and process route of the work order, or the system may automatically dispatch tasks based on fixed rules such as first in first out, shortest job time first in, etc. These methods can play a role in coping with stable, small-lot production modes, and are currently the mainstream production scheduling infrastructure. However, the conventional approach generally treats the tasks as isolated individuals, and it is difficult to quantify and process complex synergic and competing relationships between the tasks. Disclosure of Invention In a first aspect, the present invention provides an artificial intelligence based automatic optimization management method for dynamic workflow, which adopts the following technical scheme: an artificial intelligence-based automatic optimization management method for dynamic workflow comprises the following steps: s1, acquiring multi-source heterogeneous real-time data, and calibrating task charge attributes comprising a quality value, an electrical value and a priority potential energy value for a task to be executed according to a preset conversion rule; s2, constructing a real-time task potential field diagram through a preset acting force calculation model based on task charge attributes of each task; s3, identifying and extracting task conflict points and collaboration opportunity points from potential energy gradient information in a real-time task potential field diagram, and generating a task situation list; S4, generating a disturbance intervention instruction stream according to potential energy gradient characteristic differences in the task situation list; s5, executing disturbance intervention instruction flow, collecting result data after instruction execution, and feeding back to S1 for updating task charge attribute. The further scheme of the invention, the task charge attribute calibration method comprises the following steps: Calculating a quality value according to the resource consumption and the processing complexity of the task; defining tasks with cooperative relationship as attractive opposite charges according to the interaction between the tasks, and defining tasks with resource competition relationship as repulsive same charges so as to determine an electrical value; and calculating the priority potential energy value according to the delivery date of the task. According to a further scheme, the method comprises the steps of calibrating task charge attributes, and further comprises the following steps of: if the waiting time of the task exceeds a preset time threshold, periodically increasing the priority potential energy value of the task; if the multi-source heterogeneous real-time data indicate that the resource related to the task has fault early warning, the risk added value is increased for the quality value of the task. The invention further provides a scheme for constructing a real-time task potential field diagram, which comprises the following steps: instantiating the task into particles with quality values, electrical values and priority potential energy values, and placing the particles in a multidimensional virtual coordinate system representing a task resource space according to the resource requirements of the particles; Calculating interaction force generated between task particles based on task charge attributes by using an acting force calculation model; and constructing a real-time task potential field diagram based on the calculated interaction force a