Remez-Type Inequality for Smooth Functions

Abstract

The classical Remez inequality bounds the maximum of the absolute value of a polynomial $P(x)$ of degree $d$ on $[-1,1]$ through the maximum of its absolute value on any subset $Z$ of positive measure in $[-1,1]$. Similarly, in several variables the maximum of the absolute value of a polynomial $P(x)$ of degree $d$ on the unit ball $B^n \subset {\mathbb R}^n$ can be bounded through the maximum of its absolute value on any subset $Z\subset Q^n_1$ of positive $n$-measure $m_n(Z)$. In \cite{Yom} a stronger version of Remez inequality was obtained: the Lebesgue $n$-measure $m_n$ was replaced by a certain geometric quantity $\omega_{n,d}(Z)$ satisfying $\omega_{n,d}(Z)\geq m_n(Z)$ for any measurable $Z$. The quantity $\omega_{n,d}(Z)$ can be effectively estimated in terms of the metric entropy of $Z$ and it may be nonzero for discrete and even finite sets $Z$. In the present paper we extend Remez inequality to functions of finite smoothness. This is done by combining the result of \cite{Yom} with the Taylor polynomial approximation of smooth functions. As a consequence we obtain explicit lower bounds in some examples in the Whitney problem of a $C^k$-smooth extrapolation from a given set $Z$, in terms of the geometry of $Z$.