Markov-Bernstein type inequalities on compact subsets of \({\mathbb R}\) (Q5932652)

Given a compact set \(K\subset{\mathbb{R}}\), the \(n\)-th Markov factor of \(K\) at \(0\) is defined by \[ M_n(K)=\sup\left\{|p_n'(0)|:p_n\in\Pi_n, \sup_{x\in K}|p_n(x)|\leq 1 \right\}, \] where \(\Pi_n\) denotes the set of algebraic polynomials of degree at most \(n\). There are known inequalities for this factor when the set \(K\) has a certain geometric structure (e.g., Cantor type sets, finitely many intervals, etc.). In this paper another possible way of generalizations is considered. The authors introduce two density functions to \(K\) at the point \(0\) and show that the Markov factor can be estimated from above in terms of these density functions. They also show that the estimates obtained are sharp for sets whose elements form a sequence tending to \(0\). Sharp estimates are obtained for further classes of compact sets as well. As application, Markov type and Bernstein type inequalities are discussed.