The inverse of banded matrices (Q455833)

From MaRDI portal
Revision as of 20:08, 19 March 2024 by Openalex240319060354 (talk | contribs) (Set OpenAlex properties.)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
scientific article
Language Label Description Also known as
English
The inverse of banded matrices
scientific article

    Statements

    The inverse of banded matrices (English)
    0 references
    0 references
    0 references
    22 October 2012
    0 references
    A \(2r-1\) banded matrix \(B(n,r)\) of size \({n\times n}\) with \(b_{ij}=0\) for \(|i-j|\geq r\) has an LU decomposition with \(L\) and \(U\) having \(r\) nonzero diagonals. However \(U^{-1}\) is a full triangular matrix but its elements can be expressed in terms of determinants of matrices \(U(i,j)\) of size \({(j-i)\times(j-i)}\) which are upper Hessenberg submatrices consisting of rows \(i,\ldots,j-1\) and columns \(i+1,\ldots,j\) of \(U\), i.e., whose right top element is \(u_{ij}\). Similarly the elements of \(L^{-1}\) can be expressed in terms of determinants of lower Hessenberg matrices \(L(i,j)\). These can then be used to compute explicit expressions for the elements of the inverse \(B^{-1}\) by multiplying \(U^{-1}L^{-1}\).
    0 references
    triangular matrix
    0 references
    Hessenberg matrix
    0 references
    inverse
    0 references
    r-banded matrix
    0 references

    Identifiers