On the relation between Darboux transformations and polynomial mappings (Q390520)

The monic Hermite polynomials can be expressed in terms of the monic Laguerre polynomials as \(H_{2n}(\lambda )=L_n^{(-1/2)}(\lambda ^2)\) and \(H_{2n+1}(\lambda )=\lambda L_n^{(1/2)}(\lambda ^2)\). More generally, if \(P_j\) are monic orthogonal polynomials with respect to a measure \(d\mu \) on \([0,\infty )\) then the polynomials defined by \(\Phi _{2j}(\lambda )=P_j(\lambda ^2)\) and \(\Phi _{2j+1}(\lambda )=\lambda \tilde P_j(\lambda ^2)\), where \[ \tilde P_j(\lambda )=\frac{1}{\lambda }\left[ P_{j+1}(\lambda )-\frac{P_{j+1}(0)}{P_j(0)}P_j(\lambda )\right], \] are orthogonal with respect to the measure \(\frac{\text{sign}\, t}{2}d\mu (t^2)\) on the real line \(\mathbb{R}\). The unwrapping transformation \(P_j\mapsto \Phi _j\) can be expressed in terms of the Cauchy-Stieltjes transform as \(S(\lambda )\mapsto \lambda \, S(\lambda ^2)\). In this paper the author gives a matrix interpretation of this construction and shows that this transformation is essentially a Darboux transformation of the underlying Jacobi matrix. He also shows that the generalized Jacobi matrices appear in the theory of polynomial mappings and presents an introduction to the theory of generalized Jacobi matrices from a classical point of view.