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

Real-time Scheduling Method for Weight Sharing Deep Networks Based on Dimensional Optimal Transformation

A deep network, dimension technology, applied in biological neural network models, data processing applications, instruments, etc.

Active Publication Date: 2021-06-08
ZHEJIANG UNIV OF TECH
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

At present, it seems that the traditional scheduling algorithm is difficult to meet the requirements of the industry in terms of processing massive scheduling data and quickly responding to scheduling problems.

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
  • Real-time Scheduling Method for Weight Sharing Deep Networks Based on Dimensional Optimal Transformation
  • Real-time Scheduling Method for Weight Sharing Deep Networks Based on Dimensional Optimal Transformation
  • Real-time Scheduling Method for Weight Sharing Deep Networks Based on Dimensional Optimal Transformation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0057] Attached below Figure 1-5 The technical scheme of the present invention is further described.

[0058] figure 1 A flow diagram of the method of the invention is shown.

[0059] Example overview

[0060] Set up the workshop production scheduling problem, n machines, p workpieces, and the workpieces have q processing processes. The purpose is to arrange processing machines for each process of each workpiece. Changing the values ​​of n, p, and q can change the scale of the problem, and set n=3, q=3, and q=3 for small problems. Medium-sized problem n=30, q=30, q=30. Large problems n=300, q=300, q=300. The specific data format is shown in Table 1 below:

[0061] Table 1

[0062] Workflow time consuming (input) sequence number (output) Artifact 1 Process 1 5 5 Artifact 1 Process 2 12 6 Artifact 1 Process 3 3 2 … … … Artifact 2 Process 1 4 1 Artifact 2 Process 1 8 4 Artifact 2 Process 1 6 8 … … … ...

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 real-time scheduling method of the weight sharing deep network based on the optimal transformation of dimensions includes: step 1. obtaining real-time data and scheduling data collected from actual scheduling occasions as training data; step 2. processing the real-time data obtained in step 1, Process into a multi-layer two-dimensional matrix form that satisfies the input of the deep network; step 3. use the multi-layer two-dimensional matrix in step 2 and the scheduling data obtained in step 1 as the input and output of the deep network respectively, and train the deep network; step 4. Use the convolutional neural network trained in step 3 in the actual scheduling environment; perform actual network scheduling.

Description

[0001] The invention relates to a real-time scheduling method of a weight sharing deep network. Background technique [0002] For the production scheduling problem, the current mainstream method uses a mathematical model combined with an optimization heuristic algorithm to solve it, which can obtain high solution accuracy. However, in the context of big data, the explosive growth of production parameters in the production environment and strict scheduling time indicators put forward further requirements for scheduling methods. At present, it seems that the traditional scheduling algorithm is difficult to meet the industry requirements in terms of processing massive scheduling data and quickly responding to scheduling problems. Contents of the invention [0003] The present invention overcomes the above-mentioned shortcomings of the prior art, and proposes a real-time big data scheduling technology method based on a deep convolutional network. [0004] The present invention ...

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
Patent Type & Authority Patents(China)
IPC IPC(8): G06Q10/06G06N3/04
CPCG06Q10/06312G06N3/045
Inventor 王万良臧泽林李伟琨王宇乐赵燕伟
Owner ZHEJIANG UNIV OF TECH