Orthogonal polynomials in several variables for measures with mass points (Q1960251): Difference between revisions

From MaRDI portal
Set OpenAlex properties.
Importer (talk | contribs)
Changed an Item
Property / arXiv ID
 
Property / arXiv ID: 0911.2818 / rank
 
Normal rank

Revision as of 23:11, 18 April 2024

scientific article
Language Label Description Also known as
English
Orthogonal polynomials in several variables for measures with mass points
scientific article

    Statements

    Orthogonal polynomials in several variables for measures with mass points (English)
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    13 October 2010
    0 references
    Let \(d\mu\) be a positive finite measure on \(\mathbb R^d\) with finite moments and let \(\langle . \;, \;. \rangle_{\mu}\) denote the inner product defined by \[ \langle p , q \rangle_{\mu}:= \int_{\mathbb R^d} p(x) q(x) d\mu(x), \quad p,q \in \Pi^d \] where \(\Pi^d\) denotes the space of real polynomials in \(d\)-variables. Then orthogonal polynomials of several variables with respect to \(\langle . \;, \;. \rangle_{\mu}\) exist. Let \(N \in\mathbb N_0\) and let \(\xi_1,\xi_2,\dots, \xi_N\) be distinct points in \(\mathbb R^d\). Moreover, assume \(\Lambda\) to be a symmetric positive semi-definite matrix of size \(N \times N\). Next a new inner product \(\langle . \;, \;. \rangle_{\nu}\) is defined by \(\langle p , q \rangle_{\nu} = \langle p, q \rangle_{\mu} + \mathbf{p}(\xi)^{\text{tr}} \;\Lambda \;\mathbf{q}(\xi)\) where \(\mathbf{p}(\xi) = \left( p(\xi_1),p(\xi_2),\dots,p(\xi_N) \right)^{\text{tr}}\) and the superscript tr indicates the transpose. This new inner product can be defined via an integral with measure \(d\nu\) that is obtained by adding \(N\) mass points to \(d\mu\). In this paper orthogonal polynomials with respect to the new inner product \(\langle . \;, \;. \rangle_{\nu}\) are studied. It is found that these orthogonal polynomials and their associated reproducing kernels can be explicitly expressed in terms of those associated with \(d\mu\). As an example, the case of the Jacobi weight function \[ W_k(x) = x_1^{\kappa_1 - 1/2} \dots x_d^{\kappa_d-1/2} (1-x_1-\dots -x_d)^{\kappa_{d+1}-1/2}, \quad \kappa_i>0 \] on the simplex \(T^d:=\{ x=(x_1,\dots , x_d) \in\mathbb R^d : x_i \geq 0 , 1-x_1-\dots -x_d \geq 0 \}\) in \(\mathbb R^d\) is considered with mass points added at the vertices of \(T^d\) and \(\Lambda\) being a diagonal matrix, for which the obtained formulas can be further specified and expressed in terms of the classic Jacobi polynomials. The result is then used to study the asymptotic expansion of the Christoffel functions with respect to \(\langle.\;,\;.\rangle_{\nu}\).
    0 references
    multivariate orthogonal polynomials
    0 references
    Dirac mass
    0 references
    Jacobi polynomials
    0 references
    Christoffel functions
    0 references

    Identifiers