Generic symmetric matrix pencils with bounded rank (Q2662229)

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 7332248
Language Label Description Also known as
default for all languages
No label defined
    English
    Generic symmetric matrix pencils with bounded rank
    scientific article; zbMATH DE number 7332248

      Statements

      Generic symmetric matrix pencils with bounded rank (English)
      0 references
      0 references
      0 references
      0 references
      9 April 2021
      0 references
      The rank of a pencil \(\lambda A+B\) of complex matrices is the rank of \( \lambda A+B\) over the polynomial ring \(\mathbb{C}[\lambda ]\). The pencil is symmetric if both \(A\) and \(B\) are (square) symmetric matrices. The present paper describes the set \textrm{PENCIL}\(_{m\times n}^{s}(r)\) of symmetric pencils with rank at most \(r\) in terms of the closure in the Euclidean metric of some simpler subsets. Define matrices of dimensions \(d\times (d+1)\) and \(\ell \times \ell \), respectively, by \[ \mathcal{L}_{d}:= \left\{ \begin{array}{c} \lambda \text{ if }i=j, \\ 1\text{ if }i+1=j, \\ 0\text{ otherwise}, \end{array} \right. \] \[ \mathcal{J}_{\ell }(\mu ):= \left\{ \begin{array}{c} 1\text{ if }i=\ell -j, \\ \lambda -\mu \text{ if }i=\ell +1-j ,\\ 0\text{ otherwise} ,\end{array} \right. \] where \(\mu \in \mathbb{C}\) and \[ \mathcal{M}_{d}:=\left[ \begin{array}{cc} 0 & \mathcal{L}_{d}^{\top } \\ \mathcal{L}_{d} & \end{array} \right] , \] \[ \mathcal{J}_{\ell }(\infty ):=\left\{ \begin{array}{c} \lambda \text{ if }i=\ell -j ,\\ 1\text{ if }i=\ell +1-j, \\ 0\text{ otherwise}. \end{array} \right. \] It is known that every symmetric pencil is congruent to a block diagonal pencil whose diagonal blocks are of the forms \(\mathcal{M}_{d},\mathcal{J} _{\ell }(\mu )\) and \(\mathcal{J}_{\ell }(\infty )\) for various values of the parameters, and this canonical form is unique up to the order of the blocks (see [\textit{R. C. Thompson}, Linear Algebra Appl. 147, 323--371 (1991; Zbl 0726.15007)]). For any symmetric pencil \(\lambda A+B\), the congruence bundle \(\mathbf{B}^{c}(\lambda A+B)\) is defined to be the set of all symmetric pencils which have canonical forms which differ from the canonical form for \(\lambda A+B\) only in the choice of the parameters \(\mu \) and \(\infty \) (the eigenvalues); briefly, these are the pencils with the same generic eigenstructure as \(\lambda A+B\). The main theorem describes the generic eigenstructure of a symmetric matrix pencil as follows. Let \(n\) and \(r\) be integers with \(1\leq r<n\). For each integer \(a\) with \(0\leq a\leq r/2\) write \(a=(n-r)q+s\) where \( q:=\left\lfloor a/(n-r)\right\rfloor\). Then for any distinct complex numbers \(\mu _{1},...,\mu _{r-2a}\) we define \(\mathcal{K}_{a}(\lambda )\) be the block diagonal matrix with \(s\) copies of \(\mathcal{M}_{q+1}\), \(n-r-s\) copies of \(\mathcal{M}_{q}\) and single copies of \(\mathcal{J}_{1}(\mu _{1}),...,\mathcal{J}_{1}(\mu _{r-2q})\). (Theorem 3): \textrm{PENCIL}\(_{n\times n}^{s}(r)\) is equal to the closure of \(\bigcup_{a}\mathcal{K}_{a}(\lambda )\).
      0 references
      0 references
      matrix pencil
      0 references
      symmetric pencil
      0 references
      strict equivalence
      0 references
      congruence
      0 references
      orbit
      0 references
      bundle
      0 references
      spectral information
      0 references
      complete eigenstructure
      0 references
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references