Abstract <p> For functions from the Sobolev space <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\mathring{W}^n_\infty[0,1]\)</EquationSource> </InlineEquation> and an arbitrary point <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(a\in(0,1)\)</EquationSource> </InlineEquation>, we study sharp estimating functions <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(A_{n,k,\infty}(a) \)</EquationSource> </InlineEquation> in the inequality <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(|f^{(k)}(a)|\le A_{n,k,\infty}(a) \cdot \|f^{(n)}\|_{L_\infty [0,1]}\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(0 \le k &lt; n\)</EquationSource> </InlineEquation>. We establish a link between the functions <InlineEquation ID="IEq6"> <EquationSource Format="TEX">\(A_{n,k,\infty}\)</EquationSource> </InlineEquation> and best approximations of special splines by polynomials in <InlineEquation ID="IEq7"> <EquationSource Format="TEX">\(L_1[0,1]\)</EquationSource> </InlineEquation>. We introduce the Markov set <InlineEquation ID="IEq8"> <EquationSource Format="TEX">\(\mathcal{A}_{n,k}\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq9"> <EquationSource Format="TEX">\(0 \le k &lt; n\)</EquationSource> </InlineEquation>, of values of the parameter <InlineEquation ID="IEq10"> <EquationSource Format="TEX">\(a\)</EquationSource> </InlineEquation>, on which we obtain the representation <InlineEquation ID="IEq11"> <EquationSource Format="TEX">\(A_{n, k, \infty}(a)= 2^{-(n-k)}|V_n^{(k)}(2a-1)|\)</EquationSource> </InlineEquation> of the function <InlineEquation ID="IEq12"> <EquationSource Format="TEX">\(A_{n,k,\infty}\)</EquationSource> </InlineEquation> in terms of the absolute value of the <InlineEquation ID="IEq13"> <EquationSource Format="TEX">\(k\)</EquationSource> </InlineEquation>th derivative of the Peano kernel <InlineEquation ID="IEq14"> <EquationSource Format="TEX">\(V_n\)</EquationSource> </InlineEquation> of order <InlineEquation ID="IEq15"> <EquationSource Format="TEX">\(n\)</EquationSource> </InlineEquation>. For arbitrary <InlineEquation ID="IEq16"> <EquationSource Format="TEX">\(n\)</EquationSource> </InlineEquation> and <InlineEquation ID="IEq17"> <EquationSource Format="TEX">\(k\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq18"> <EquationSource Format="TEX">\(0 \le k \le n-2\)</EquationSource> </InlineEquation>, we show that the embedding constant <InlineEquation ID="IEq19"> <EquationSource Format="TEX">\(\Lambda_{n,k,\infty}\)</EquationSource> </InlineEquation> of the Sobolev spaces <InlineEquation ID="IEq20"> <EquationSource Format="TEX">\(\mathring{W}^n_\infty[0,1] \hookrightarrow \mathring{W}^k_\infty[0,1]\)</EquationSource> </InlineEquation> is the maximum value of the function <InlineEquation ID="IEq21"> <EquationSource Format="TEX">\(2^{-(n-k)}|V_n^{(k)}(2a-1)| \)</EquationSource> </InlineEquation> on the interval <InlineEquation ID="IEq22"> <EquationSource Format="TEX">\([0,1]\)</EquationSource> </InlineEquation>. For odd <InlineEquation ID="IEq23"> <EquationSource Format="TEX">\(n\)</EquationSource> </InlineEquation> and even <InlineEquation ID="IEq24"> <EquationSource Format="TEX">\(k\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq25"> <EquationSource Format="TEX">\(0 \le k &lt; n\)</EquationSource> </InlineEquation>, the embedding constant <InlineEquation ID="IEq26"> <EquationSource Format="TEX">\(\Lambda_{n,k,\infty}\)</EquationSource> </InlineEquation> is refined as <InlineEquation ID="IEq27"> <EquationSource Format="TEX">\(\Lambda_{n,k,\infty}=2^{-(n-k)}|V_n^{(k)}(0)|\)</EquationSource> </InlineEquation>. We also express the embedding constants <InlineEquation ID="IEq28"> <EquationSource Format="TEX">\(\Lambda_{n,k,\infty}\)</EquationSource> </InlineEquation> for odd <InlineEquation ID="IEq29"> <EquationSource Format="TEX">\(n\)</EquationSource> </InlineEquation> and even <InlineEquation ID="IEq30"> <EquationSource Format="TEX">\(k\)</EquationSource> </InlineEquation> in terms of hypergeometric functions and study the asymptotic behavior of the embedding constants as <InlineEquation ID="IEq31"> <EquationSource Format="TEX">\(n=2m+1 \to \infty\)</EquationSource> </InlineEquation> with fixed <InlineEquation ID="IEq32"> <EquationSource Format="TEX">\(k\)</EquationSource> </InlineEquation> or <InlineEquation ID="IEq33"> <EquationSource Format="TEX">\(n-k\)</EquationSource> </InlineEquation>. </p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

On Sharp Uniform Estimates of Intermediate Even-Order Derivatives in Sobolev Spaces

  • D. D. Kazimirov,
  • I. A. Sheipak

摘要

Abstract

For functions from the Sobolev space \(\mathring{W}^n_\infty[0,1]\) and an arbitrary point \(a\in(0,1)\) , we study sharp estimating functions \(A_{n,k,\infty}(a) \) in the inequality \(|f^{(k)}(a)|\le A_{n,k,\infty}(a) \cdot \|f^{(n)}\|_{L_\infty [0,1]}\) , \(0 \le k < n\) . We establish a link between the functions \(A_{n,k,\infty}\) and best approximations of special splines by polynomials in \(L_1[0,1]\) . We introduce the Markov set \(\mathcal{A}_{n,k}\) , \(0 \le k < n\) , of values of the parameter \(a\) , on which we obtain the representation \(A_{n, k, \infty}(a)= 2^{-(n-k)}|V_n^{(k)}(2a-1)|\) of the function \(A_{n,k,\infty}\) in terms of the absolute value of the \(k\) th derivative of the Peano kernel \(V_n\) of order \(n\) . For arbitrary \(n\) and \(k\) , \(0 \le k \le n-2\) , we show that the embedding constant \(\Lambda_{n,k,\infty}\) of the Sobolev spaces \(\mathring{W}^n_\infty[0,1] \hookrightarrow \mathring{W}^k_\infty[0,1]\) is the maximum value of the function \(2^{-(n-k)}|V_n^{(k)}(2a-1)| \) on the interval \([0,1]\) . For odd \(n\) and even \(k\) , \(0 \le k < n\) , the embedding constant \(\Lambda_{n,k,\infty}\) is refined as \(\Lambda_{n,k,\infty}=2^{-(n-k)}|V_n^{(k)}(0)|\) . We also express the embedding constants \(\Lambda_{n,k,\infty}\) for odd \(n\) and even \(k\) in terms of hypergeometric functions and study the asymptotic behavior of the embedding constants as \(n=2m+1 \to \infty\) with fixed \(k\) or \(n-k\) .