English

Jacobi method for symmetric $4\times4$ matrices converges for every cyclic pivot strategy

Numerical Analysis 2018-10-09 v2

Abstract

The paper studies the global convergence of the Jacobi method for symmetric matrices of size 44. We prove global convergence for all 720720 cyclic pivot strategies. Precisely, we show that inequality S(A[t+3])γS(A[t])S(A^{[t+3]})\leq\gamma S(A^{[t]}), t1t\geq1, holds with the constant γ<1\gamma<1 that depends neither on the matrix AA nor on the pivot strategy. Here A[t]A^{[t]} stands for the matrix obtained from AA after tt full cycles of the Jacobi method and S(A)S(A) is the off-diagonal norm of AA. We show why three consecutive cycles have to be considered. The result has a direct application on the JJ-Jacobi method.

Keywords

Cite

@article{arxiv.1701.02387,
  title  = {Jacobi method for symmetric $4\times4$ matrices converges for every cyclic pivot strategy},
  author = {Erna Begovic and Vjeran Hari},
  journal= {arXiv preprint arXiv:1701.02387},
  year   = {2018}
}

Comments

16 pages