Automatic polygon and quadrilateral mesh subdivision method based on deep reinforcement learning

The method for automatic polygon quadrilateral mesh generation using deep reinforcement learning solves the problems of reliance on human experience and poor generalization ability in existing technologies. It achieves fully automatic and efficient polygon mesh generation, and the generated quadrilateral meshes perform well in terms of orthogonality, smoothness and aspect ratio.

CN122066802AActive Publication Date: 2026-05-19CALCULATION AERODYNAMICS INST CHINA AERODYNAMICS RES & DEV CENT
5 Cites 2 Cited by

Patent Information

Application Number
CN202610535036.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-22
Publication Date
2026-05-19
Estimated Expiration
2046-04-22

AI Technical Summary

Technical Problem

Existing quadrilateral mesh generation methods rely on human experience, have poor generalization ability, lack a global optimization perspective, are difficult to adapt to complex polygonal regions, and are difficult to process variable-length inputs, resulting in inconsistent mesh quality and low automation.

Method used

An automatic polygon quadrilateral mesh generation method based on deep reinforcement learning is adopted. By constructing a reinforcement learning environment and using a deep Q-network framework, the polygon generation process is modeled as a Markov decision process. A composite reward function and standardized steps are designed to achieve fully automatic generation and high-quality mesh generation.

Benefits of technology

It enables the processing of polygons with any number of vertices without human intervention, generating high-quality quadrilateral meshes, improving automation and generalization capabilities, significantly reducing preprocessing costs, and improving mesh quality and meshing efficiency.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a polygonal and quadrilateral mesh automatic subdivision method based on deep reinforcement learning, relates to the field of computational geometry and mesh generation, and constructs an automatic subdivision framework fusing geometric priori knowledge and a data driving strategy. According to the method, nine basic topology filling templates covering triangles to hexagons are predefined as discrete action spaces, and continuous geometric segmentation is converted into a sequence decision problem; learning a mapping relation between the polygon state characteristics and the optimal template selection strategy by using a deep Q network, and balancing exploration and utilization through a dynamic epsilon-greedy mechanism; and in combination with a vertex number priority scheduling strategy and a multi-dimensional quality award function, guiding an intelligent agent to adaptively generate a high-quality quadrilateral grid. According to the method, full-automatic and high-robustness subdivision of the complex polygon area is achieved, the grid orthogonality and the length-width ratio quality are remarkably improved, manual intervention is avoided, and the method is suitable for engineering scenes such as finite element analysis.
Need to check novelty before this filing date? Find Prior Art