A spectral transform for the matrix Hill's equation (Q1768060)

The paper aims at describing the range and mapping properties of the Floquet transform that associates the Floquet matrix \(\Psi_1(\lambda,Q)\) to a \(K\times K\) potential \(Q(x)\) in a matrix Hill equation. It has been shown recently that the Floquet matrix contains enough data for uniquely solvable inverse spectral problems. The most interesting mapping properties of the Floquet transform arise when the differences \(\Psi_1(\lambda,Q)-\Psi_1(\lambda,Q_0)\) are studied, where \(Q_0=\int_0^1 Q(x)dx\). In particular, the author examines the sampled differences \(\Psi_1(n^2\pi^2,Q)-\Psi_1(n^2\pi^2,Q_0)\), \(n=1,2,3,\dots\), which leads to some analytic map to a matrix-valued sequences space. It is shown that these samples come from functions of one of the Paley-Wiener Hilbert spaces of entire functions. It is also shown that only a small set of matrix Hill equations can have the spectral data of a scalar equation.