A state-aware based dynamic execution link generation and adjustment method

By using a state-aware approach, similarity calculations between sub-intent feature vectors and resource tag vectors are employed to generate a hybrid string parallel execution link and monitor resource status in real time. This solves the problems of inaccurate resource matching, link rigidity, and high cost of anomaly recovery in distributed software architecture, thereby improving system efficiency and robustness.

CN122363780APending Publication Date: 2026-07-10SHENZHEN ZHISOFT TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
SHENZHEN ZHISOFT TECH CO LTD
Filing Date
2026-03-19
Publication Date
2026-07-10

AI Technical Summary

Technical Problem

Existing technologies suffer from low resource matching accuracy, rigid execution links, weak state awareness, and high anomaly recovery costs in distributed software architectures, resulting in low system efficiency and poor user experience.

Method used

By using a state-aware approach, similarity calculations are performed between sub-intent feature vectors and resource tag vectors to generate a hybrid string parallel execution link. Resource status is monitored in real time, and an improved greedy algorithm is used to optimize the link, enabling dynamic adjustment and local replanning. Combined with service degradation strategies, rapid fault tolerance is achieved.

Benefits of technology

It achieves dynamic optimization of the execution chain, improves the system's execution efficiency and robustness, and reduces the impact and cost of anomaly recovery.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122363780A_ABST
    Figure CN122363780A_ABST
Patent Text Reader

Abstract

This invention relates to the field of computer software technology and discloses a state-aware dynamic execution link generation and adjustment method. The method first receives a list of sub-intents decomposed from business intents, matches optimal execution resources for each sub-intent using cosine similarity calculation, and monitors the health status of resources in real time for hierarchical classification. Subsequently, based on sub-intent dependencies, matching resource status, and business rules, an improved greedy algorithm is used to generate an initial execution link in a hybrid serial-parallel manner. During execution, changes in resources, rules, and data are monitored in real time. Once a triggering condition is met, an impact range analysis algorithm is used to locate affected nodes, performing resource re-matching and link replanning only locally, and triggering a three-level service degradation strategy. This invention achieves dynamic link optimization and rapid fault tolerance, improving system execution efficiency and robustness.
Need to check novelty before this filing date? Find Prior Art