Two-sided bounds on the inverses of diagonally dominant tridiagonal matrices (Q1301308): Difference between revisions

Tridiagonal matrices arise in many applications, e.g. the numerical solution of partial differential equations and cubic spline interpolation. Therefore this class of matrices is studied very well and a lot of theory is known about tridiagonal matrices and their inverses. One of the most beautiful results is due to \textit{F. R. Gantmacher} and \textit{M. G. Krein} [Oscillation matrices, oscillation kernels, and small oscillations of mechanical systems (1941; Zbl 0060.03303)] who proved that the inverse of an irreducible symmetric tridiagonal matrix is a so-called Green's matrix. A similar result is also proved by \textit{W. W. Barrett} [Linear Algebra Appl. 27, 211-217 (1979; Zbl 0414.15004)]. This result is generalized in several directions. In the paper under review the author establishes upper and lower bounds for the entries of the inverses of diagonally dominant tridiagonal matrices. These bounds improve the bounds recently given by \textit{P. N. Shivakumar} and \textit{C. Ji} [ibid. 247, 297-316 (1996; Zbl 0862.65015)]. Moreover, he shows how to improve the bounds iteratively. For an \(n{\times}n\) \(M\)-matrix this iterative refinement yields the exact inverse after \(n-1\) steps.