Inverse problems for Hankel and Toeplitz matrices

From MaRDI portal





For \(u\in C^{n+1}\), \(g\in C^{m+1}\) any \((m+1)\times(n+1)\) Hankel matrix \(H\) satisfying \(Hu=g\) is of the form \(H=H_{mn}((g(\lambda^{- 1})\lambda^{-1}+p(\lambda))/u(\lambda))\) with \(p(\lambda)\) any polynomial of degree less than \(n\), and \(H_{mn}(f)=[s_{i+j}]^{mn}_{00}\), \(f(\lambda)=\sum s_ k\lambda^{-k-1}\), the Laurent expansion of \(f\) at infinity. For pairs of equations \(H_{\ell n}u=g\), \(H_{km}v=h\) with given vectors and Hankel matrices to be determined, one obtains solvability criteria, a uniqueness criterion and a solution formula. Special cases, in particular inverse spectral problems (i.e., \(u\) and \(v\) eigenvectors corresponding to given eigenvalues) and the inverse problem for homogeneous equations, are discussed. The paper continues with the inverse problem for a single Toeplitz equation, presenting an approach that seems better fitted to the Toeplitz structure than the obvious transformation into a Hankel problem, and including the Hermitian case. Finally, solvability and uniqueness criteria and a solution formula for pairs of Toeplitz equations and Hermitian Toeplitz equations are given.











This page was built for publication: Inverse problems for Hankel and Toeplitz matrices

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1184468)