Effective recursive algorithm for judging the positive-definiteness of matrices of high dimension (Q3798148)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 4064365
Language Label Description Also known as
default for all languages
No label defined
    English
    Effective recursive algorithm for judging the positive-definiteness of matrices of high dimension
    scientific article; zbMATH DE number 4064365

      Statements

      Effective recursive algorithm for judging the positive-definiteness of matrices of high dimension (English)
      0 references
      0 references
      1988
      0 references
      Hermitian matrices
      0 references
      positive-definiteness
      0 references
      high-dimensional matrices
      0 references
      For an arbitrary Hermitian \(n\times n\) matrix P, the author presents a recursive formula constructing a sequence \(P^{(i)}\) \((i=0,1,2,...[n/k]- 1\), \(k>0)\) of Hermitian matrices, where \(P^{(0)}=P\) and \(P^{i)}\) is an (n-i)\(\times (n-i)\) matrix. It follows from a theorem of \textit{E. Kreindler} and \textit{A. Jameson} [IEEE Trans. Autom. Control 17, 147-148 (1972; Zbl 0262.15015)] that P is positive definite if and only if the upper-left corner \(k\times k\) submatrix of \(P^{(i)}\) is positive definite for all \(i=0,1,2,...,[n/k]-1\). The author especially analyzes the case \(k=1\), when such a submatrix is a scalar entry of \(P^{(i)}\). In this case the algorithm (having n-1 iterations) gives a criterion simpler than Sylvester's one; it needs only multiplication of vectors and subtraction of matrices, without calculation of determinants. Moreover, it saves computer memory, needing only access to one row of \(P^{(i)}\) at a time - the rest of the matrix may be stored on a disk. This gives the possibility to judge the positive-definiteness of very high- dimensional matrices.
      0 references

      Identifiers