Material transfer optimization scheduling method and system based on large language model
By constructing prompt words and a large language model to generate offspring individuals, and combining validity detection and path fine-tuning mechanisms, the temperature and population structure are dynamically adjusted. This solves the problems of the generated sequence not satisfying validity and the path structure being difficult to improve in the single AGV material transfer optimization scheduling, thus improving convergence efficiency and solution quality.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CENT SOUTH UNIV
- Filing Date
- 2026-02-09
- Publication Date
- 2026-06-12
AI Technical Summary
In existing technologies, the single AGV material transfer optimization scheduling problem suffers from several drawbacks, including the generation sequence not meeting the validity requirements, the path structure being difficult to continuously improve, the recurrence of inferior offspring, and insufficient temperature population adaptive adjustment, which limits the convergence efficiency and solution quality.
By constructing prompt words, using a large language model to generate offspring individuals, and combining effectiveness detection and repair, path fine-tuning mechanism, dynamic sub-path obstacle avoidance and adaptive guidance of failed samples, temperature parameters and population structure are adjusted in a coordinated manner to optimize the material transfer scheduling of a single AGV.
It improves the convergence efficiency and solution quality of single AGV material transfer optimization scheduling, the generated path structure is more in line with the constraints, reduces the occurrence of inferior offspring, and improves the effectiveness of the search and the speed of path optimization.
Smart Images
Figure CN121660415B_ABST