Interpolation with lagrange polynomials a simple proof of Markov inequality and some of its generalizations |
| |
Authors: | A Shadrin |
| |
Institution: | 1. Computing Center Siberian Branch, Academy of Sciences of the USSR, prospekt akademika Lavrentieva 6, 630090, Novosibirsk, Russia
|
| |
Abstract: | The following theorem is provedTheorem 1. Let q be a polynomial of degree n (q∈Pn) with n distinct zeroes lying in the interval ?1,1] and $$|p^{(k)} (x)| \leqslant \max \{ |q^{(k)} (x)|,|\frac{1}{k}(x^2 - 1)q^{(k + 1)} (x) + xq^{(k)} (x)|\} .$$ If polynomial p∈Pn satisfies the inequality $$|p(\tau _l )| \leqslant |q(\tau _l )|,{\text{ }}\tau _l \in \Delta '_{q'} $$ then for each \(k = \overline {1,n} \) and any x∈?1,1] its k-th derivative satisfies the inequality $$||p_n^{(k)} || \leqslant ||T_n^{(k)} ||$$ This estimate leads to the Markov inequality for the higher order derivatives of polynomials if we set q=Tn, where Tn is Chebyshev polynomial least deviated from zero. Some other results are established which gives evidence to the conjecture that under the conditions of Theorem 1 the inequality ‖p(k)‖≤‖q(k)‖ holds. |
| |
Keywords: | |
本文献已被 CNKI SpringerLink 等数据库收录! |
|