Exact constants in Jackson-type inequalities for approximation of analytic functions from \(H_p^r\) with \(1\leq p\leq 2\) (Q2455181)

The Jackson inequalities are relations in which the error of appron of an individual function \(f\) is estimated in terms of the modulus of continuity of the function to be approximated or of some of its derivatives. Concretely, let \(X\), \(Y\) be normed spaces of analytic functions \(f(z)\) in a bounded domain in the complex plane and let \(X^r\) be the set of functions \(f(z) \in X\) whose \(r\)th derivative \(f^{(r)}(z)\) exists and belongs to \(X\), and such that the corresponding boundary value \(F^{(r)}(\theta)\) is continuous and constant nowhere on the boundary. IF \(X^r\subset Y\) and \(\mathcal{R}_n\) is a \(n\)-dimensional subspace in \(Y\) then the Jackson inequalities are relations of the form \[ E(f, \mathcal{R}_n)_Y:= \inf_{\xi \in \mathcal{R}_n} | | f- \xi | | _Y \leq M_r \alpha_{nr}^{-1} \omega_m (F^{(r)}, \gamma )_X \quad (f \in X^r), \] where \(\alpha_{nr}=n!/(n-r)!\), \(\omega_m(F^{(r)}, \gamma )_X\) is the \(m\)th-order modulus of continuity of the function \(F^{(r)}\) in \(X\), \(M_r\) is a constant that depends on \(m\) and \(r\), \(\gamma\) are fixed numbers. The problem of finding the minimum constant in the inequality is equivalent to calculate the sharp bound for the Jackson value \(\displaystyle\sup_{f\in X^r, \;F^{(r)} \neq \text{const.}} \frac{\alpha_{nr} E(f, \mathcal{R}_n)_Y}{\omega_m (F^{(r)}, \gamma)_X}\). In this paper the authors derive the Jackson inequality for the best approximation of functions on the Hardy space \(H^p\) (\(1\leq p \leq 2\)). They consider not only the derivative \(f^{(r)}\) but also the rth-derivative \(f_a^{(r)} \) with respect to the argument, and also specify the exact Jackson values at fixed values of \(\gamma\).