A Markov-Nikolskii type inequality for absolutely monotone polynomials of order \(K\) (Q628375)

A function \(Q\) is called absolutely monotone of order \(k\) on an interval \(I\) if \(Q^{(l)}(x)\geq 0\) for all \(x\in I\) and \(l=0,\dots,k\). The main result of this paper is the following: Theorem. Let \(k,m,n\in\mathbb N\), \(1\leq k\leq n\), \(0\leq m\leq k/2\). Let \(0<q\leq p\leq\infty\). There exist absolute constants \(c_1>0\) and \(c_2>0\) and constants \(c_{p,q}>0\) and \(c_{p,q}'>0\) depending only on \(p\) and \(q\) such that \[ c_{p,q}\left(\left(\frac q{q+1}\right)^2\frac{c_1n^2}{k}\right)^{m+1/q-1/p}\leq\sup_Q\frac{\|Q^{(m)}\|_{L_p[-1,1]}}{\|Q\|_{L_q[-1,1]}}\leq c_{p,q}'\left(\frac{c_2n^2}{k}\right)^{m+1/q-1/p}, \] where the supremum is taken over all not identically zero, absolutely monotone polynomials \(Q\) of degree at most \(n\) and of order \(k\) on \([-1,1]\). By an observation due to Bernstein, the classical Markov inequality in the uniform norm for monotone polynomials is not essentially better than that for all polynomials (see [\textit{G. Szegö}, Orthogonal Polynomials. 4th ed. American Mathematical Society (AMS) (1975; Zbl 0305.42011)]). It follows from the author's estimates that a similar result holds in the case of higher derivatives of polynomials and Markov-Nikolskii type inequalities (in \(L_p\)-norm).