Orthogonal polynomials through the invariant theory of binary forms (Q323661): Difference between revisions

This paper connects the invariant theory of binary forms with generalized orthogonal polynomial systems in several variables. Classical orthogonal polynomials, like the Hermite polynomials, arise here as special cases. The starting point, the invariant theory of binary forms, is due to Kung and Rota: by denoting \(\mathbb{F}[\mathbf{x},\mathbf{y}]\) the polynomial ring with infinite sets of variables \(\mathbf{x}=\{x_i\mid i\in\mathbb{N}\}\) and \(\mathbf{y}=\{y_i\mid i\in\mathbb{N}\}\), one considers polynomials that are \textit{invariant} under the action of \(\mathrm{GL}l_2(\mathbb{F})\), when acting on pairs of variables \((x_i,y_i)\), and, on the other hand, \textit{covariants} of binary forms. The latter can be obtained by means of a linear, umbral operator, from which also the \textit{apolar covariant} evolves. By studying forms that are apolar to a given form, the authors are led to a determinantal formula for the joint-covariant. In Section~4, \textit{generalized orthogonal polynomial systems} are introduced, as families of polynomials \(\{p_{nm}(x_0)\mid 1\leq m\leq n\}\) with \(\deg p_{nm}=n\) such that \(\mathbf{E} x_0^k p_{mn}(x_0)=0\) for all \(0\leq k\leq n-m\) for some linear functional \(\mathbf{E}:\mathbb{F}[x_0]\to\mathbb{F}\). Proposition~5 then relates such orthogonal polynomial systems with binary forms; it allows one to apply the symbolic methods of invariant theory to orthogonal polynomials. Moreover, Theorems~6 and~7 state a determinantal formula and a Heine integral formula for generalized orthogonal polynomial systems. The following section shows a connection between the symbolic expression of a covariant and its presentation in terms of the roots, using Sylvester's theorem. Moreover, it introduces a general family of covariants that includes also transvectants. These covariants are shown to be the average value of the classical basis of symmetric polynomials over a set of roots of suitable orthogonal polynomials.