Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Data center resource offline scheduling method based on deep reinforcement learning

A reinforcement learning, data center technology, applied in the computer field to achieve good applicability

Pending Publication Date: 2019-07-02
田文洪 +3
View PDF2 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

First, the decisions made by these systems are often highly repetitive, generating large amounts of training data for deep reinforcement learning

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
  • Data center resource offline scheduling method based on deep reinforcement learning
  • Data center resource offline scheduling method based on deep reinforcement learning
  • Data center resource offline scheduling method based on deep reinforcement learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0014] The specific implementation manners of the present invention will be described in further detail below according to the drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0015] figure 1 It is a schematic diagram of the deep reinforcement learning framework of the embodiment of the present invention.

[0016] Such as figure 1 As shown, the agent is interacting with the environment. At each time step t, the agent observes some state s_t and chooses an action a_t. After the action, the environment state transitions to s_(t+1), and the agent receives reward r_t. State transitions and rewards are stochastic and assumed to have Markov properties.

[0017] Further, the agent can only control its own behavior and has no prior knowledge of which state the environment will transition to or what the reward might be. During training, the agent can observe these quanti...

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 relates to the technical field of computers, in particular to a data center resource offline scheduling method based on deep reinforcement learning. The deep reinforcement learning can provide a feasible alternative scheme for a human heuristic method of resource scheduling management. Through continuous learning, the deep reinforcement learning method can optimize the specific working loads (such as periodic loads or random loads) and maintain high-quality optimization scheduling results under various conditions. And by taking the minimum average operation slowdown (system slowing down time) as an optimization target, and calculating a reward value of each scheduling in offline scheduling, the deep network is guided to be optimized towards the target and finally to be trained towards the optimal target. Results show that in a large number of tests in the embodiment of the invention, the slowdown of the offline scheduling method using the deep reinforcement learning is far lower than the SJF (shortest job priority algorithm) and other traditional optimized job scheduling methods, and the advantages of the deep reinforcement learning method in the field are embodied.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to an off-line scheduling method for data center resources based on deep reinforcement learning. Background technique [0002] Resource management is a fundamental problem in computer networks and operating systems. Resource allocation is usually a combinatorial problem that can be mapped to different NP-hard problems. Although each resource allocation scheme is specific, the general approach is to design efficient heuristic algorithms with performance guarantees under certain conditions. Recent studies have shown that machine learning can provide a viable alternative to human-induced heuristics for resource management, especially deep reinforcement learning, which has become an active area of ​​machine learning research. [0003] In fact, deep reinforcement learning methods are especially suitable for resource management systems. First, the decisions made by these systems are ...

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): G06F9/50
CPCG06F9/5061G06F9/5011G06F9/5027
Inventor 不公告发明人
Owner 田文洪
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com