Cross-time-scale cloud edge collaborative resource management and task scheduling optimization method

CN122019140APending Publication Date: 2026-05-12CHINESE PEOPLES LIBERATION ARMY UNIT 61618
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202512028293.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-30
Publication Date
2026-05-12

AI Technical Summary

Technical Problem

Existing cloud-edge collaborative resource management and task scheduling methods suffer from insufficient cross-timescale collaboration, lack of budget constraint modeling, and poor adaptability to dynamic environments, resulting in a disconnect between resource allocation and actual costs, low resource utilization, and increased task latency.

Method used

A cross-timescale cloud-edge collaborative resource management and task scheduling optimization method is adopted. By dividing the system runtime into large-scale and small-scale periods, a service deployment and resource allocation strategy is generated using a deep deterministic strategy gradient algorithm and a near-end strategy optimization algorithm. The strategy is then dynamically adjusted through a two-way feedback mechanism to adapt to both slow-changing and fast-changing factors.

Benefits of technology

It achieves a balance between long-term planning and short-term load fluctuations, reduces the overhead of frequent reconfiguration and scheduling latency, improves resource utilization and task processing efficiency, enhances the system's adaptability to sudden tasks and server status changes, and ensures that resource allocation takes into account both performance and cost.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122019140A_ABST
    Figure CN122019140A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of edge computing and cloud computing, and relates to a cross-time-scale cloud edge collaborative resource management and task scheduling optimization method. Through hierarchical reinforcement learning and a bidirectional feedback mechanism, long-term optimization of service deployment and resource allocation under a large time scale and real-time response of task scheduling under a small time scale are realized, long-term planning and short-term load fluctuation are effectively balanced, and frequent reconfiguration overhead and scheduling delay are reduced; through explicit modeling budget and physical resource constraint, it is ensured that performance and cost are considered in resource configuration, and overload is avoided; the adaptability of the system to sudden tasks and server state changes is enhanced, the resource utilization rate and the task processing efficiency are improved, and efficient and stable support is provided for edge intelligent application.
Need to check novelty before this filing date? Find Prior Art