Workflow scheduling method based on depth enhancement learning

A technology that enhances learning and scheduling methods, applied in neural learning methods, biological neural network models, instruments, etc., can solve the problems of increasing algorithm generalization performance, long execution time, and poor generalization, so as to increase generalization performance and ensure The effect of time efficiency

Inactive Publication Date: 2016-12-14
UNIV OF ELECTRONIC SCI & TECH OF CHINA
View PDF3 Cites 68 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The present invention overcomes the deficiencies of the prior art, provides a workflow scheduling method based on deep enhanced learning, solves the defects of long execution time and poor generalization of the workflow scheduling method in the current distributed e

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
  • Workflow scheduling method based on depth enhancement learning
  • Workflow scheduling method based on depth enhancement learning
  • Workflow scheduling method based on depth enhancement learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0037] like Figure 1-3 As shown, the present invention includes a workflow scheduling method based on deep reinforcement learning, comprising the following steps:

[0038] Step A) collect task execution DAG workflow directed acyclic graph M pieces in the actual execution environment as a sample pool;

[0039] Step B) Carry out MDP Markov decision process modeling on each DAG workflow directed acyclic graph, and generate task state set S;

[0040] Step C) According to the neural network training method DQN, the task state set S and the corresponding known action set A generated by M pieces of DAG workflow directed acyclic graph are used as input, and substituted into the deep neural network formula Q(s, a; θ i ), to obtain the neural network parameter matrix θ when performing task i i , Q is the action value function, s is one of the task state set S, a is a scheduling scheme in the action set A;

[0041] Step D) Determine whether the task state set S generated by the DAG ...

Embodiment 2

[0084] This embodiment is preferably as follows on the basis of Embodiment 1: it also includes step F) after the sample pool has accumulated to a certain extent, repeat step C) to the deep neural network formula Q (s, a; θ i ) to recalculate to get a new θ i value and the new deep neural network Q(s,a; θ i ) is used for DAG workflow DAG scheduling calculation of subsequent input.

[0085] When the sample pool has accumulated to a certain extent, it means that the number of DAG workflow DAG samples accumulated in the sample pool exceeds 100, and step C) is started once for calculation, and the training samples will randomly sample 100 DAG work from the sample pool Stream DAGs for computation.

[0086] With the continuous increase of DAG workflow directed acyclic graph in the sample pool, the neural network training method DQN is used to continuously update θ i value, and then calculate the optimal scheduling scheme, so that the scheduling machine can learn the scheduling str...

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 present invention discloses a workflow scheduling method based on depth enhancement learning. The method comprises the following steps of A) collecting M task execution workflow directed acyclic graphs (DAG) in an actual execution environment as a sample pool; B) carrying out the Markov decision process (MDP) modeling on each workflow DAG to generate a task state set S; C) according to a training method DQN of a neural network, using the task state sets S generated by the M workflow DAG and the corresponding known action sets A as the input to substitute into a deep neural network formula, and obtaining the value of a neural network parameter matrix. According to the present invention, and by the above method, the defects that the execution time of the workflow scheduling method is long and the generalization performance is poor under a current distributed environment, are solved, the time efficiency of an algorithm is guaranteed acceleratedly, at the same time, the generalization performance of the algorithm itself is increased, and a scheduling machine can learn a scheduling strategy autonomously according to the actual scene characteristics.

Description

technical field [0001] The invention relates to the field of computer software, in particular to a workflow scheduling method based on deep reinforcement learning. Background technique [0002] In a distributed computing environment, workflow scheduling has always been one of the optimization problems in the computer field. The workflow scheduling problem is actually to give a scheduling scheme to schedule the tasks on the workflow to the appropriate execution nodes in a certain order to achieve the minimum execution cost. Its mathematical model is as follows: [0003] A specific computing application can be represented by a directed acyclic graph (DAG) G(T,E), where T is a set of n tasks {t 1 ,t 2 ,...,t n}, E is the set of dependencies between tasks. Each dependency e(i,j)∈E represents task t j required in task t i Execution cannot begin until execution is complete. Given a finite set of machines M, M contains n nodes {m 1 ,m 2 ,...,m n}. Let χ denote the set o...

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): G06Q10/06G06N3/08
CPCG06Q10/06313G06N3/08
Inventor 段翰聪闵革勇张建王瑾
Owner UNIV OF ELECTRONIC SCI & TECH OF CHINA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap