Task unloading processing method, model training method and system

By using task offload processing methods in the edge network, a task offload strategy is generated and the task nodes are offloaded to the target device for processing, the problem of insufficient local computing resources is solved, and the efficiency of task processing and reasonable resource scheduling is achieved.

CN119629242AActive Publication Date: 2025-03-14TONGJI UNIV +1
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411537604.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-30
Publication Date
2025-03-14
Estimated Expiration
2044-10-30

AI Technical Summary

Technical Problem

In the case of insufficient local computing resources of the equipment, how to achieve task processing, especially in scenarios such as smart agricultural machinery and vehicles, it is difficult to update the on-board computing hardware, making it difficult to fully support the operation of software application services.

Method used

By adopting the task offload processing method in the edge network, the processing requirements information of the task nodes in the pending task process and the available resource information of the equipment are obtained, feature encoding and model input are performed, task offloading the task nodes to the target device for processing, and finally integrating the processing results.

Benefits of technology

In the case where the local resources of the device are insufficient, the coordinated processing of the edge network is achieved to improve the efficiency of task processing, reduce the consumption of local processing resources of the device, and enable the device to obtain the overall processing results of the pending task flow.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119629242A_ABST
    Figure CN119629242A_ABST
Patent Text Reader

Abstract

The embodiment of the invention discloses a task unloading processing method, a model training method and a model training system, which are used for realizing task processing under the condition that local computing power resources of equipment are insufficient. The scheme comprises the following steps: acquiring processing demand information respectively corresponding to a plurality of task nodes in a to-be-processed task process and available resource information of each device in an edge network; executing feature coding based on feature dimensions of a preset state space to obtain task state features and environment features, and inputting the task state features and the environment features into a pre-trained task unloading processing model to obtain a task unloading strategy corresponding to the to-be-processed task process; unloading the to-be-unloaded task node to corresponding target unloading equipment to execute processing based on the task unloading strategy; and generating a task processing result corresponding to the to-be-processed task flow based on a task node processing result fed back by the target unloading equipment.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Multi-level computing power network task scheduling method and device

    CN116708443A

  • Cloud edge computing power network system computing unloading method and system

    CN117749796A

  • Calculation unloading and resource allocation method and device based on Rainbow algorithm in Internet of Vehicles

    CN117880891A

  • Air-ground integrated network resource optimization method based on reinforcement learning and storage medium

    CN117979364A

  • Multi-user multi-task computing offloading method and system in mobile edge computing environment

    WO2023116460A1