Markov-type inequalities and extreme zeros of orthogonal polynomials (Q2238043): Difference between revisions

The aim of this paper is to investigate extremal problems of Markov-type involving Hilbert spaces and linear operators. More precisely, the authors consider a Hilbert space \(H\), an orthogonal system \(\{p_m\}_{m\geq 0}\subset H\), a linear operator \(\mathcal{L}\) defined in \(H\) and the following Markov-type inequality \[ \|\mathcal{L} f\| \leq c_{n}(\mathcal{L})\|f\|, \quad f \in \pi_{n}, \tag{1}\] where \( \pi_{n}:= \operatorname{span}\left\{p_{0}, p_{1}, \ldots, p_{n}\right\}\), and \[ c_{n}(\mathcal{L})=\sup \left\{\|\mathcal{L} f\|: f \in \pi_{n},\|f\|=1\right\}.\tag{2} \] They show that if \(\mathcal{L}\) satisfies the following conditions: \begin{itemize} \item[(i)] \(\mathcal{L}\) vanishes on \( \pi_{k-1}\) for some non-negative integer \(k\) (this condition is void when \(k=0\)). \item[(ii)] There exist two sequences \(\left\{\alpha_{i}\right\}_{i=0}^{\infty}, \, \left\{\beta_{i}\right\}_{i=k}^{\infty}\) of non-zero real numbers such that \[ \mathcal{L} p_{m}=\beta_{m} \sum_{i=0}^{m-k} \alpha_{j} p_{j}, \quad m \geq k,\] \end{itemize} then, for each \(n\geq k\), the reciprocal of the squared sharp constant (2), \(\frac{1}{ c_{n}^{2}(\mathcal{L})}\) coincides with the smallest eigenvalue of either of the \((n-k+1) \times(n-k+1)\) positive definite Jacobi matrices \(\mathbf{J}_{n-k+1}\) and \(\widetilde{\mathbf{J}}_{n-k+1}\), (cf. Theorem 1.1). In addition, the authors provide the explicit form of such Jacobi matrices and apply their result to the estimation of the sharp constants in the Markov \(L_2\) inequalities with Laguerre weight function (see Theorem 1.3) and with Gegenbauer weight (see Theorem 1.4), respectively.