A task flow generation system
Through the collaborative architecture of multi-system integration modules, historical task database, task model generation module, vector process generation module, and task decomposition module, the problems of decomposing new product categories without historical data coverage and information flow conflicts between multiple systems are solved. This enables accurate generation and adaptation of task processes, ensuring the stable execution of production and R&D projects.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU GETSOFT CO LTD
- Filing Date
- 2026-04-01
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, task flow generation relies on fixed historical task models, which cannot accurately break down new categories of tasks without historical data coverage, and cannot handle information flow differences and conflicts between multiple business systems, resulting in insufficient adaptability of the generated flow to actual business scenarios.
It adopts a collaborative architecture consisting of a multi-system interface module, a historical task library, a task model generation module, a vector process generation module, and a task decomposition module. It decomposes the target task into multiple independent topological modal components through topological modal decomposition, generates a task process vector topology graph, and achieves accurate decomposition and adaptation of the task process through techniques such as coupling degree calculation, modal basis function training, and dynamic correction.
It enables precise process generation for all categories of tasks, improves the adaptability of task processes to actual business scenarios, and ensures the stable execution of production and R&D projects.
Smart Images
Figure CN122414775A_ABST