A resource coordination scheduling method combining ontology model and large model

CN122111691AActive Publication Date: 2026-05-29GUDOU TECHNOLOGY (CHENGDU) CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
GUDOU TECHNOLOGY (CHENGDU) CO LTD
Filing Date
2026-04-29
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

In existing resource collaborative scheduling technologies, the text segmentation method is crude, resulting in the loss of time binding relationships. Knowledge extraction lacks the identification of time-sensitive relationships, the relationship completion has a high misjudgment rate, and there is a lack of accurate perception of scheduling rules.

Method used

By assigning time anchors to candidate event statements, dividing event blocks by participating objects and action types, using a large language model for structured extraction, constructing entity alignment operations and resource constraint graph neural networks, generating a complete scheduling graph, and completing the relationships that conform to resource constraints.

Benefits of technology

Ensuring the integrity of time-bound relationships improves the temporal accuracy of knowledge extraction, reduces the misjudgment rate of relationships, and generates scheduling graphs that better fit actual scheduling scenarios, supporting the systematic and standardized accumulation of scheduling knowledge.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122111691A_ABST
    Figure CN122111691A_ABST
Patent Text Reader

Abstract

The application discloses a resource collaborative scheduling method based on ontology model and large model, mainly relates to the technical field of collaborative scheduling, and aims to solve the problem that the existing scheme has rough text segmentation method, and lacks precise perception of scheduling core elements in relation extraction and relation completion. It comprises the following steps: based on the constraint violation factor, the trained resource constraint graph neural network is used to calculate the relation score of the candidate relation; according to the relation score, the preset resource constraint table and the trained resource constraint graph neural network, the candidate relation between the node combination in the initial scheduling graph is completed, and the complete scheduling graph after completion is obtained; according to the complete scheduling graph, the corresponding concept class, object attribute and data attribute are generated; according to the complete scheduling graph and the preset resource constraint table, the corresponding concept axiom, relation axiom and scheduling rule with time delay window are generated, and are unified into a deployable scheduling domain ontology; the scheduling domain ontology is deployed into the resource collaborative scheduling system.
Need to check novelty before this filing date? Find Prior Art