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

Finite-time convergence new neural network design method

A limited time, neural network technology, applied in the field of neural network, can solve the problem of extra workload, achieve the effect of wide application field, strong practicability, avoid extra workload and cumbersome process

Inactive Publication Date: 2017-05-17
JISHOU UNIVERSITY
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to overcome the deficiencies of prior art and method, provide a kind of design method of the new neural network model real-time solution engineering / mathematical problem of finite time convergence, overcome the additional workload and the real-time processing of recursive neural network in the past cumbersome process

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
  • Finite-time convergence new neural network design method
  • Finite-time convergence new neural network design method
  • Finite-time convergence new neural network design method

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment 1

[0024] We consider the time-varying matrix inversion problem that arises frequently in engineering and science, mathematically defining the matrix inverse A -1 (t)∈R n×n The equation for A(t)X(t)=I, where I∈R n×n is the identity matrix, X(t)∈R n×n is the unknown matrix to be inverted. figure 2 It shows the error convergence of the previous recurrent neural network to solve the time-varying matrix inversion problem, and the convergence time is 6 seconds. image 3 It shows the state solution convergence of the previous recurrent neural network to solve the matrix time-varying inversion. and Figure 4 It shows the error convergence of solving the time-varying matrix inversion problem when the new evolution formula is used. The convergence time is 2.1 seconds, which is nearly 3 times faster, and the convergence performance is greatly improved. Figure 5 Shows the state solution convergence of the present invention to solve the time-varying matrix inversion problem when using...

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 provides a finite-time convergence new neural network design method, which comprises the following steps: 1) determining problems to be solved and describing the problems through a mathematical equation in an unified manner; 2) defining an indefinite unbounded error function through the mathematical equation in the step 1); 3) designing an ingenious evolution formula through the error function in the step 2); and carrying out development on the evolution formula in the step 3) to obtain a finite-time convergence new neural network model. The method is characterized in that, by designing the special evolution formula, a new neural network realizes finite-time convergence.

Description

technical field [0001] The present invention relates to the aspect of neural network in the field of artificial intelligence, more specifically, relates to a novel neural network design method with limited time convergence. Background technique [0002] As a new emerging technology, recurrent neural network has its own unique advantages, such as parallel processing ability, distributed storage ability, strong fault tolerance and strong self-adaptive ability. Therefore, it has been widely used in signal processing, pattern recognition, optimization combination, knowledge engineering, expert system, robot control and so on. However, the engineering / mathematical problem solving of recurrent neural networks in the past can only converge to the expected solution of the required problem when the time tends to infinity. The ideal situation is only exponential convergence, which cannot make the recurrent neural network converge to our desired solution in a limited time range. In s...

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): G05B13/04
CPCG05B13/04
Inventor 肖林
Owner JISHOU UNIVERSITY