Digital research and development demand automatic disassembling method based on knowledge reasoning

By using improved GraphRAG retrieval and Reflexion iterative reasoning, combined with constraint set verification and shadow scheduling, standardized requirements and resource plans are generated, solving the problem of unification and traceability of requirement decomposition under multi-source input, and realizing efficient and reliable R&D requirement decomposition and plan generation.

CN121390756AActive Publication Date: 2026-01-23DAISY (SHANGHAI) SOFTWARE CO LTD

Patent Information

Application Number
CN202511561264.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-10-29
Publication Date
2026-01-23
Estimated Expiration
2045-10-29

Smart Images

  • Figure CN121390756A_ABST
    Figure CN121390756A_ABST
Patent Text Reader

Abstract

The invention discloses a digital research and development demand automatic disassembling method based on knowledge reasoning. The method comprises the following steps: gathering multi-source demands, uniformly de-duplicating, and establishing a version baseline; a semantic graph is generated according to the standardized requirements, a support graph and a constraint set are constructed, and a resource portrait, historical measurement and a template library are formed; performing retrieval and sub-graph aggregation by using an improved GraphRAG model, completing alignment reasoning and outputting a fact set, a conflict list and a deletion list; the method comprises the following steps: generating a hierarchical task, dependency and priority by using an improved Reflexion algorithm, and performing shadow scheduling and acyclic verification to obtain an execution sequence table; generating a design / interface / test / verification / acceptance workpiece, and solving resources, milestones and key paths according to the design / interface / test / verification / acceptance workpiece; and a tracing link and an early warning mechanism are established, and rearrangement and resource redistribution are carried out when the time exceeds the limit. According to the method, the schedule forming time is shortened, the reworking and delay risks are reduced, and the workpiece coverage rate, the critical path controllability and the delivery traceability are improved.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Industrial knowledge graph completion and adaptive retrieval method based on large language model

    CN119336921A

  • Time sequence knowledge graph retrieval enhancement generation method and device and medium

    CN120069060A

  • Iterative retrieval reasoning enhanced knowledge graph RAG method based on introspection mechanism

    CN120216732A

  • Code automatic generation and optimization system based on multiple modes

    CN120276718A

  • Multi-task printing scheduling optimization method based on AI intelligent strategy generation

    CN120525098A

Cited By

  • Software demand disassembling method and device based on multi-source demand, equipment and medium

    CN122018858A

  • Software requirement decomposition method, device and equipment based on multi-source requirements and medium

    CN122018858B

  • Protective shoe AI generation method based on physical process constraint and safety standard verification

    CN122113138A