Resource allocation methods, devices, equipment, media and products
By acquiring task feature information of AI computing tasks and using deep reinforcement learning models for classification, combined with resource topology maps and health scores, a scheduling scheme is generated, which solves the problem of inaccurate resource allocation in existing technologies and improves resource utilization and task execution efficiency.
CN122309151APending Publication Date: 2026-06-30CRRC QINGDAO SIFANG CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CRRC QINGDAO SIFANG CO LTD
- Filing Date
- 2026-03-26
- Publication Date
- 2026-06-30
Smart Images
Figure CN122309151A_ABST
Abstract
This application relates to the fields of artificial intelligence and cloud computing technology, and provides a resource allocation method, apparatus, device, medium, and product. The method includes: acquiring task feature information of an AI computing task to be processed; classifying the AI computing task based on the task feature information using a preset deep reinforcement learning model to obtain a task classification result; determining a scheduling scheme template matching the task classification result, and generating an initial scheduling scheme based on the matching scheduling scheme template; and determining the allocated resources corresponding to the AI computing task based on the initial scheduling scheme, a pre-determined resource topology map, and the health scores of resource nodes in the resource topology map. Based on this application's solution, the accuracy of resource allocation for AI computing tasks can be improved.
Need to check novelty before this filing date? Find Prior Art