Unlock instant, AI-driven research and patent intelligence for your innovation.

Multi-tenant task scheduling method based on reinforcement learning

A task scheduling and reinforcement learning technology, applied in neural learning methods, multi-programming devices, climate sustainability, etc., can solve problems such as poor scheduling accuracy, poor online scheduling effect, and inability to meet cluster scheduling goals, and improve accuracy. performance, optimize completion time, and optimize cluster resource utilization

Pending Publication Date: 2022-07-05
HAINAN UNIVERSITY
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] In order to solve the problems of poor scheduling accuracy, poor online scheduling effect and failure to meet the cluster scheduling goals in the existing technology, a multi-tenant task scheduling method based on reinforcement learning is proposed

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Multi-tenant task scheduling method based on reinforcement learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0033] like figure 1 As shown, this embodiment provides a multi-tenant task scheduling method based on reinforcement learning, including the following steps:

[0034] Establishing a task scheduling agent based on deep reinforcement learning includes the following steps:

[0035] Use experience playback technology to obtain historical task scheduling sample sets and randomly select several sample data;

[0036] The task scheduling sample includes the system status of historical task scheduling, the corresponding scheduling action and the corresponding reward value. The system status includes the cluster status and the task status. The cluster status includes the resource usage of the machines in the cluster. The task status includes the resource requirements of the task, runtime requirements and tenant information;

[0037] The formula for the reward value is:

[0038] r(t)=αUtil(t)-βRes(t)

[0039] In the formula, r(t) is the reward value of the t decision time step; Util(...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention belongs to the technical field of task scheduling, and discloses a reinforcement learning-based multi-tenant task scheduling method, which comprises the following steps of: establishing a task scheduling agent based on deep reinforcement learning; acquiring resource use conditions of machines in the cluster in real time; obtaining tasks uploaded by a plurality of tenants and updating corresponding task queues; generating a corresponding scheduling decision by using a task scheduling agent according to the task queue of each user and the resource use condition of the machines in the real-time cluster; and executing task scheduling according to the scheduling decision. The problems that in the prior art, the scheduling accuracy is poor, the online scheduling effect is poor, and the cluster scheduling target cannot be met are solved.

Description

technical field [0001] The invention belongs to the technical field of task scheduling, in particular to a multi-tenant task scheduling method based on reinforcement learning. Background technique [0002] With the rapid development of computer technology and the Internet economy, cloud computing has become one of the most promising and valuable research directions after big data and artificial intelligence in recent years. Cloud computing is a distributed computing system that provides software, CPU, memory, storage, and other computing resources. It provides pay-as-you-go services over the Internet, and cloud computing is used to build and run a cloud computing environment with virtualization technology. In cloud computing, cloud resource providers will create multiple virtual machines (VMs) based on physical resources to process computing tasks submitted by users. Allocate computing tasks to a virtual machine through a scheduling algorithm. Therefore, whether the sched...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06F9/48G06F9/50G06N3/08
CPCG06F9/4881G06F9/5038G06F9/5016G06N3/08Y02D10/00
Inventor 吉才敏杨瀚林叶春杨周辉
Owner HAINAN UNIVERSITY