Dissertation > Excellent graduate degree dissertation topics show

The Improved Block Jacobi-Davidson Method for Solving Large Symmetric Eigenvalue Problems

Author: KangYanYan
Tutor: DaiHua
School: Nanjing University of Aeronautics and Astronautics
Course: Computational Mathematics
Keywords: Symmetric matrix Eigenvalue problem Block Jacobi-Davidson method Shrink Technology Methods to reconcile
CLC: O241.6
Type: Master's thesis
Year: 2010
Downloads: 45
Quote: 0
Read: Download Dissertation

Abstract


Block Jacobi-Davidson method is to calculate heavy or dense symmetric matrix eigenvalues ??effective methods. The method may also calculate the number of characteristics on an extreme , but generated in the iterative process has converged , the Ritz number of iterations will still participate in subsequent operation , which reduces the overall convergence rate of the method , and secondly , the block Jacobi-Davidson method calculating eigenvalues ??within a large amount of calculation . In order to improve the block Jacobi-Davidson method overall convergence rate , this paper dynamic contraction technique proposed dynamic contraction block Jacobi-Davidson method ; large symmetric matrix in order to calculate the internal characteristics of the paper will reconcile the Rayleigh-Ritz method and block Jacobi- Davidson method is proposed by combining reconcile block Jacobi-Davidson method and technology used to reconcile dynamic contraction block Jacobi-Davidson method , gives the dynamic contraction reconcile block Jacobi-Davidson method . Correction equations is the block Jacobi-Davidson type approach is the key , this is equivalent to using the correction equation augmented correction equation into the form of the classic saddle point problems , and their use of appropriate pretreatment technology . Numerical results show that dynamic contraction block Jacobi-Davidson method is superior to the block Jacobi-Davidson method , dynamic contraction reconcile block Jacobi-Davidson method can efficiently compute large symmetric matrix of internal heavy or dense eigenvalues ??.

Related Dissertations

  1. Solving large-scale non- symmetric matrix eigenvalue problem Weighted Arnoldi method,O241.6
  2. Possible Numbers of a in (a,b) Matrices with a Given Rank,O151.21
  3. The Solutions of a Few Constrained Matrix Equations and Their Optimal Approximations,O241.6
  4. Several special tridiagonal matrix with matrix equation X ~ TAX = B matrix inverse problem,O241.6
  5. Research on the Real Symmetric Matrix Inverse Eigenvalue Problems,O241
  6. Some Doubly Strucured Matrix Eigenvalue Problems,O241.6
  7. Generalized unitary matrices and generalized nature of the promotion of Hermite Matrices,O151.21
  8. An Improvement and Analysis of the Lanczos Algorithm for Symmetric Matrices Eigenproblem,O241.6
  9. Superconvergent Numerical Algorithms of Integral Equations and Their Compact Operators,O175.5
  10. Studies on Inverse Eigenvalue Problem of Generalized Centre-symmetric Matrix,O241.6
  11. On the group inverse of the symmetric matrix to maintain,O152
  12. A Hierarchy of Evolution Equations and Its Integrable System,O175.29
  13. The Technique of Acceleration and Preconditioning for Solving Large Symmetric Eigenvalue Problems,O241.6
  14. Solving Large Scale Matrix Eigenvalue Problem with Optimization Methods,O241.6
  15. Existence of Solutions for a Class Quasilinear Elliptic Eigenvalue Problems,O175
  16. Neutron transport eigenvalue problems and γ -ray detection efficiency of the Monte Carlo method,O571.5
  17. Achieved by vector form implicitly restarted Arnoldi method blocks,O241.6
  18. The Solutions of the Matrix Equation A~TXA=B on the Linear Manifolds,O151.21
  19. Second-order differential equation eigenvalue problems and transform complex discrete Hamiltonian systems,O175.8
  20. A Hermite multigrid method and elliptic eigenvalue problem Nonconforming Element,O241.5

CLC: > Mathematical sciences and chemical > Mathematics > Computational Mathematics > Numerical Analysis > Linear algebra method of calculating
© 2012 www.DissertationTopic.Net  Mobile