Borg-type theorems for high-order equations with matrix coefficients (Q2477455)

The classical Borg Theorem is a theorem in inverse spectral problem for a Sturm-Liouville equation \(- y'' + q(x) y = \lambda y\). In this article, the Borg theorem is generalized to the \(n\)-th order equation \[ Ly := y^{(n)} + \sum_{k=1}^{n-1} Q_k (x) y^{(n-1-k)} = \lambda y, \] where \(y\) is the column vector \((y_1, y_2, \dots, y_p)\), \(Q_k(\cdot): [0, 1] \rightarrow C^{p\times p}\) is an analytic \(p \times p\) matrix, \(k \in {1, 2, \dots, n-1}\). Let \(I_p\) be the \(p \times p\) identity matrix, \(Y_i (x, \lambda)\) denote a \(p \times p\) matrix solution that satisfies the initial conditions \(Y_i^{(j)} (0, \lambda) = \delta_{i (j+1)} \cdot I_p\), with \(\delta_{i (j+1)}\) being the Dirac delta function, \(i \in {1, 2, \dots, n}, j \in {0, 1, \dots, n-1}\). Let \(W_L(\lambda)\) denote the (canonical) monodromy matrix with the ith column vector \[ (Y_i(1, \lambda), Y_i'(1, \lambda), \dots, Y_i^{(n-1)} (1, \lambda)), \quad i \in {1, 2, \dots, n}. \] The author shows that the \(n\)-th order equation is uniquely determined by any \(i\)th column of the monodromy matrix or by its first row \((Y_1 (1, \lambda), Y_2 (1, \lambda), \dots, Y_n (1, \lambda))\).