Can one hear a matrix? Recovering a real symmetric matrix from its spectral data (Q2135180): Difference between revisions

It is known that the spectrum of a symmetric real \(N \times N\) matrix determines the matrix up to unitary equivalence. So, more spectral data is needed together with some sign indicators in order to remove this unitary ambiguities. In this paper, the spectral information about the matrix comes from the spectra of its \(N\) nested main minors. Such a spectral information determines the matrix up to a finite number of possibilities. The reconstruction can be then made unique by providing additional discrete data encoded in the signs of certain expressions involving entries of the desired matrix. The first part of this manuscript deals with the inverse problem of full, real symmetric matrices of size \(N \times N\), where the number of unknown entries is \(\dfrac{1}{2}N(N+1)\). The spectral data to be used is the union of the spectra of the first \(N\) nested main minors of the matrix of size \(1,2,\ldots,N\), and \(\dfrac{1}{2}N(N-1)\) sign indicators needed for the complete reconstruction. The technique, called telescopic construction, is inductive: given a matrix \(A\) and a minor \(A^{(n)}\) of dimension \(n\), its next neighbor \(A^{(n+1)}\) is obtained by computing the \((n+1)\)-th column from the given spectra and sign indicators. The uniqueness of the resulting solution is proved. \par The values \(s^{(n)}_j\), that we are going to define, play the role of the sign indicators in this paper. Let \(A\) be a real symmetric matrix of size \(N \times N\). \(A^{(n)}\), \(1 \leq n \leq N\), denotes its \(n \times n\) upper main diagonal minor, so \(A^{(1)}=A_{11}\) and \(A^{(N)}=A\). \(|x^{(n)}\rangle\) denotes the column vector in dimension \(n\) with entries \(x^{(n)}_j\), \(j=1,2,\ldots,n\), and the scalar product is denoted by \(\langle x^{(n)} | y^{(n)}\rangle\). The authors also introduce the unit vectors \(|e^{(n)}(j)\rangle\), \(1 \leq j \leq n\), whose entries all vanish but for the \(j\)'th which is unity. For every minor \(A^{(n)}\), \(1 \leq n \leq (N-1)\), define the upper off-diagonal part of the next column \[ |a^{(n)}\rangle:=\left( \begin{array}{c} A_{1,n+1}^{(n+1)} \\ \vdots \\ A_{n,n+1}^{(n+1)} \end{array} \right) = \sum_{j=1}^n A_{j,n+1}^{(n+1)} | e^{(n+1)}(j)\rangle. \] The spectra of \(A^{(n)}\), \(1 \leq n \leq N\), is denoted by \(\sigma^{(n)}= \{ \lambda^{(n)}_j\}_{j=1}^n\), with \(\lambda^{(n)}_k \geq \lambda^{(n)}_j\), \(\forall k \geq j\), and the corresponding normalized eigenvectors are \(\{ |v^{(n)}(j)\rangle \}_{j=1}^n\). The overlaps of \(|a^{(n)}\rangle\) with the eigenvectors of \(A^{(n)}\) is denoted by \(s^{(n)}_j \xi^{(n)}_j\) where \[ \left \{ s^{(n)}_j:= \mathrm{Sign} \left(\langle a^{(n)} | v^{(n)}(j)\rangle \right) \right\}_{j=1}^n, \] with \[ \mathrm{Sign}(x)=\left\{ \begin{array} {ccc} +1,& & \mbox{for} \ x\geq 0, \\ -1,& & \mbox{for} \ x < 0. \end{array} \right. \] \par In the second part of the paper, the authors use the telescopic method restricted to banded matrices with band width \(D=2d+1\) much smaller than \(N\). The only sign information required consists of the signs of the off-diagonal matrix elements in the diagonal which is \(d\) steps away from the main diagonal. The proposed method is proved to provide a unique solution only for generic \(D\)-diagonal matrices. \par Finally, in the last part, it is shown that the large redundancy which exists in the telescopic approach can be reduced by studying a different construction: one considers only the \(d+1\) dimensional principal minors \(M^{(n)}\) which are distinguished by the position of their upper corner of \(M^{(n)}\) along the diagonal.