基于动态图的多主体关系与意图协同闭环优化方法及装置
By using a dynamic graph-based collaborative closed-loop optimization method, the problem of jointly modeling collaborative relationships and behavioral intentions in multi-agent collaborative systems was solved, achieving higher accuracy and stable prediction results, and improving the system's adaptability in complex scenarios.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NAT UNIV OF DEFENSE TECH
- Filing Date
- 2026-03-30
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies struggle to effectively model collaborative relationships and behavioral intentions within a unified framework in multi-agent collaborative systems, leading to unstable inference results, high misjudgment rates, and a lack of responsiveness to interactive noise and rapid changes in complex scenarios.
A closed-loop optimization method based on dynamic graphs for multi-agent relationships and intentions is adopted. Through data preprocessing, graph representation encoding, collaborative relationship estimation, behavioral intention inference, and closed-loop feedback correction, a unified framework is constructed for iterative optimization to achieve synchronous output and mutual constraint of collaborative relationships and behavioral intentions.
The F1 score for predicting collaborative relationships was improved from 0.78 to 0.86, and the F1 score for inferring intent was improved from 0.74 to 0.83. The false positive rate and stability error were reduced, and the system's adaptability and overall stability in complex interaction scenarios were enhanced.
Smart Images
Figure CN121960577B_ABST