Zhang Neural Networks for fast and accurate computations of the field of values

DOI10.1080/03081087.2019.1648375MaRDI QIDQ5125852zbMATH OpenOpenAlexWikidataFDO

Authors Frank Uhlig

Publication date 2 October 2020

Published in Linear and Multilinear Algebra (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1904.10568

zbMATH Keywords

eigenvalue computationfield of valuesnumerical matrix algorithmnumerical rangeparameter-varying matrix problemZhang neural network (ZNN)

Mathematics Subject Classification ID

Norms of matrices, numerical range, applications of functional analysis to matrix theory (15A60) Direct numerical methods for linear systems and matrix inversion (65F05) Numerical computation of eigenvalues and eigenvectors of matrices (65F15)

Abstract: In this paper a new and different neural network, called Zhang Neural Network (ZNN) is appropriated from discrete time-varying matrix problems and applied to the angle parameter-varying matrix field of values (FoV) problem. This problem acts as a test bed for newly discovered convergent 1-step ahead finite difference formulas of high truncation orders. The ZNN method that uses a look-ahead finite difference scheme of error order 6 gives us 15+ accurate digits of the FoV boundary in record time when applied to hermitean matrix flows $A(t)$.

Recommendations

• From Zhang neural network to scaled hyperpower iterations
• Varying-parameter Zhang neural network for approximating some expressions involving outer inverses
• The construction of high order convergent look-ahead finite difference formulas for Zhang neural networks
• From Zhang Neural Network to Newton Iteration for Matrix Inversion
• Zhang neural networks for a set of linear matrix inequalities with time-varying coefficient matrix
• A new design formula exploited for accelerating Zhang neural network and its application to time-varying matrix inversion
• Complex-valued neural networks
• Computing Time-Varying ML-Weighted Pseudoinverse by the Zhang Neural Networks
• Approximation by spherical neural networks with zonal functions

Cites work

• A 5-instant finite difference formula to find discrete time-varying generalized matrix inverses, matrix inverses, and scalar reciprocals
• Computation of multiple eigenvalues and generalized eigenvectors for matrices dependent on parameters
• Computer generated numerical ranges and some resulting theorems
• Faster and more accurate computation of the field of values boundary for \textit{n} by \textit{n} matrices
• Geometry of the numerical range of matrices
• scientific article; zbMATH DE number 925119 (Why is no real title available?)
• Neural networks based approach for computing eigenvectors and eigenvalues of symmetric matrix
• Neural networks for computing eigenvalues and eigenvectors
• Numerical Determination of the Field of Values of a General Complex Matrix
• On Smooth Decompositions of Matrices
• On the numerical range of a matrix
• Path-Following Method to Determine the Field of Values of a Matrix with High Accuracy
• Recurrent neural networks for nonlinear output regulation
• The construction of high order convergent look-ahead finite difference formulas for Zhang neural networks
• Time-varying matrix eigenanalyses via Zhang neural networks and look-ahead finite difference equations
• Zhang functions and various models

Cited in

(5)

• Time-varying matrix eigenanalyses via Zhang neural networks and look-ahead finite difference equations
• Coalescing Eigenvalues and Crossing Eigencurves of 1-Parameter Matrix Flows
• The construction of high order convergent look-ahead finite difference formulas for Zhang neural networks
• scientific article; zbMATH DE number 7692349 (Why is no real title available?)
• Zeroing Neural Networks : an Introduction to Predictive Computations for Time-varying Matrix Problems via ZNN

This page was built for publication: Zhang Neural Networks for fast and accurate computations of the field of values

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5125852)