Method, device, equipment, medium and program product for handling an operational problem
By using two large language models for hierarchical verification and vectorization, the problems of modeling failure and poor system stability in complex operations research problems are solved, and efficient and accurate solutions to operations research problems are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA UNITED NETWORK COMM GRP CO LTD
- Filing Date
- 2026-03-23
- Publication Date
- 2026-06-26
AI Technical Summary
Existing technologies, when dealing with complex operations research problems, suffer from several drawbacks: lack of multi-round interactions leading to information loss and modeling failure; lack of independent verification causing logical errors; and static knowledge bases limiting knowledge transfer, resulting in low modeling success rates and poor system stability.
Two pre-built large language models are used for model generation and hierarchical verification, respectively. By obtaining the natural language problem description input by the user, the initial modeling information is determined. The second large model is used for hierarchical verification. After vectorization, relevant experience knowledge is retrieved from the pre-set vector database. The retrieval results are then fused with the complete modeling information to generate enhanced modeling information. Finally, the solution is determined based on the enhanced modeling information and the two models.
It improves the success rate of modeling and system stability, ensures the professionalism and accuracy of modeling, enhances the adaptability and decision-making basis of the system, reduces modeling errors, and improves the efficiency and accuracy of solving operations research problems.
Smart Images

Figure CN122285689A_ABST