<p>A new generalization of shifted thin plate splines <Equation ID="Equ14"> <MediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_Equ14.gif" Format="GIF" Height="23" Rendition="HTML" Resolution="72" Type="Linedraw" Width="463" /> </MediaObject> <EquationSource Format="TEX">\(\begin{aligned} \phi (x)=(c^{2d}+||x||^{2d})\log \left( c^{2d}+||x||^{2d}\right) ,\qquad x\in \mathbb {R}^n, d\in \mathbb {N}, c&gt;0, \end{aligned}\)</EquationSource> <EquationSource Format="MATHML"><math display="block"> <mrow> <mtable> <mtr> <mtd columnalign="right"> <mrow> <mrow> <mi>ϕ</mi> <mrow> <mo stretchy="false">(</mo> <mi>x</mi> <mo stretchy="false">)</mo> </mrow> <mo>=</mo> <mo stretchy="false">(</mo> </mrow> <msup> <mi>c</mi> <mrow> <mn>2</mn> <mi>d</mi> </mrow> </msup> <mo>+</mo> <msup> <mrow> <mo stretchy="false">|</mo> <mo stretchy="false">|</mo> <mi>x</mi> <mo stretchy="false">|</mo> <mo stretchy="false">|</mo> </mrow> <mrow> <mn>2</mn> <mi>d</mi> </mrow> </msup> <mrow> <mo stretchy="false">)</mo> <mo>log</mo> </mrow> <mfenced close=")" open="("> <msup> <mi>c</mi> <mrow> <mn>2</mn> <mi>d</mi> </mrow> </msup> <mo>+</mo> <msup> <mrow> <mo stretchy="false">|</mo> <mo stretchy="false">|</mo> <mi>x</mi> <mo stretchy="false">|</mo> <mo stretchy="false">|</mo> </mrow> <mrow> <mn>2</mn> <mi>d</mi> </mrow> </msup> </mfenced> <mo>,</mo> <mspace width="2em" /> <mi>x</mi> <mo>∈</mo> <msup> <mrow> <mi mathvariant="double-struck">R</mi> </mrow> <mi>n</mi> </msup> <mo>,</mo> <mi>d</mi> <mo>∈</mo> <mi mathvariant="double-struck">N</mi> <mo>,</mo> <mi>c</mi> <mo>&gt;</mo> <mn>0</mn> <mo>,</mo> </mrow> </mtd> </mtr> </mtable> </mrow> </math></EquationSource> </Equation>is presented to further increase the accuracy of quasi-interpolation. When restricted to Euclidean spaces of even dimensionality <i>n</i>, the generalization enables the construction of a quasi-Lagrange operator that reproduces all polynomials of degree <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq1.gif" Format="GIF" Height="15" Rendition="HTML" Resolution="72" Type="Linedraw" Width="77" /> </InlineMediaObject> <EquationSource Format="TEX">\(n+2d-1\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>n</mi> <mo>+</mo> <mn>2</mn> <mi>d</mi> <mo>-</mo> <mn>1</mn> </mrow> </math></EquationSource> </InlineEquation>. The case complements the newly proposed generalized multiquadric <InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq2.gif" Format="GIF" Height="22" Rendition="HTML" Resolution="72" Type="Linedraw" Width="321" /> </InlineMediaObject> <EquationSource Format="TEX">\(\phi (x)=\sqrt{c^{2d}+||x||^{2d}},\quad x\in \mathbb {R}^n, d\in \mathbb {N}, c&gt;0\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>ϕ</mi> <mrow> <mo stretchy="false">(</mo> <mi>x</mi> <mo stretchy="false">)</mo> </mrow> <mo>=</mo> <msqrt> <mrow> <msup> <mi>c</mi> <mrow> <mn>2</mn> <mi>d</mi> </mrow> </msup> <mo>+</mo> <msup> <mrow> <mo stretchy="false">|</mo> <mo stretchy="false">|</mo> <mi>x</mi> <mo stretchy="false">|</mo> <mo stretchy="false">|</mo> </mrow> <mrow> <mn>2</mn> <mi>d</mi> </mrow> </msup> </mrow> </msqrt> <mo>,</mo> <mspace width="1em" /> <mi>x</mi> <mo>∈</mo> <msup> <mrow> <mi mathvariant="double-struck">R</mi> </mrow> <mi>n</mi> </msup> <mo>,</mo> <mi>d</mi> <mo>∈</mo> <mi mathvariant="double-struck">N</mi> <mo>,</mo> <mi>c</mi> <mo>&gt;</mo> <mn>0</mn> </mrow> </math></EquationSource> </InlineEquation>, which is restricted to odd dimensions, as shown in a previous work. This generalization improves the approximation order by a factor of <InlineEquation ID="IEq3"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq3.gif" Format="GIF" Height="23" Rendition="HTML" Resolution="72" Type="Linedraw" Width="81" /> </InlineMediaObject> <EquationSource Format="TEX">\(\mathcal {O}\left( h^{2(d-1)}\right) \)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi mathvariant="script">O</mi> <mfenced close=")" open="("> <msup> <mi>h</mi> <mrow> <mn>2</mn> <mo stretchy="false">(</mo> <mi>d</mi> <mo>-</mo> <mn>1</mn> <mo stretchy="false">)</mo> </mrow> </msup> </mfenced> </mrow> </math></EquationSource> </InlineEquation>, where <InlineEquation ID="IEq4"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq4.gif" Format="GIF" Height="14" Rendition="HTML" Resolution="72" Type="Linedraw" Width="42" /> </InlineMediaObject> <EquationSource Format="TEX">\(h&gt;0\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>h</mi> <mo>&gt;</mo> <mn>0</mn> </mrow> </math></EquationSource> </InlineEquation> denotes the fill distance and <InlineEquation ID="IEq5"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq5.gif" Format="GIF" Height="14" Rendition="HTML" Resolution="72" Type="Linedraw" Width="40" /> </InlineMediaObject> <EquationSource Format="TEX">\(d=1\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>d</mi> <mo>=</mo> <mn>1</mn> </mrow> </math></EquationSource> </InlineEquation> represents the classical thin plate spline. The results are then compared with the theoretical optimal approximation from the shift-invariant space generated by this function. Moreover, we introduce a new class of inverse multiquadrics <Equation ID="Equ15"> <MediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_Equ15.gif" Format="GIF" Height="27" Rendition="HTML" Resolution="72" Type="Linedraw" Width="411" /> </MediaObject> <EquationSource Format="TEX">\(\begin{aligned} \phi (x)=\left( c^\lambda +||x||^\lambda \right) ^\beta ,\qquad x\in \mathbb {R}^n, \lambda \in \mathbb {R},\beta \in \mathbb {R}\backslash \mathbb {N}, c&gt;0. \end{aligned}\)</EquationSource> <EquationSource Format="MATHML"><math display="block"> <mrow> <mtable> <mtr> <mtd columnalign="right"> <mrow> <mi>ϕ</mi> <mrow> <mo stretchy="false">(</mo> <mi>x</mi> <mo stretchy="false">)</mo> </mrow> <mo>=</mo> <msup> <mfenced close=")" open="("> <msup> <mi>c</mi> <mi>λ</mi> </msup> <mo>+</mo> <msup> <mrow> <mo stretchy="false">|</mo> <mo stretchy="false">|</mo> <mi>x</mi> <mo stretchy="false">|</mo> <mo stretchy="false">|</mo> </mrow> <mi>λ</mi> </msup> </mfenced> <mi>β</mi> </msup> <mo>,</mo> <mspace width="2em" /> <mi>x</mi> <mo>∈</mo> <msup> <mrow> <mi mathvariant="double-struck">R</mi> </mrow> <mi>n</mi> </msup> <mrow> <mo>,</mo> <mi>λ</mi> <mo>∈</mo> <mi mathvariant="double-struck">R</mi> <mo>,</mo> <mi>β</mi> <mo>∈</mo> <mi mathvariant="double-struck">R</mi> <mo stretchy="true">\</mo> <mi mathvariant="double-struck">N</mi> <mo>,</mo> <mi>c</mi> <mo>&gt;</mo> <mn>0</mn> <mo>.</mo> </mrow> </mrow> </mtd> </mtr> </mtable> </mrow> </math></EquationSource> </Equation>We provide an explicit representation of the generalized Fourier transform and discuss its asymptotic behaviour near the origin. Particular emphasis is placed on the case where <InlineEquation ID="IEq6"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq6.gif" Format="GIF" Height="14" Rendition="HTML" Resolution="72" Type="Linedraw" Width="15" /> </InlineMediaObject> <EquationSource Format="TEX">\(\lambda \)</EquationSource> <EquationSource Format="MATHML"><math> <mi>λ</mi> </math></EquationSource> </InlineEquation> and <InlineEquation ID="IEq7"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq7.gif" Format="GIF" Height="17" Rendition="HTML" Resolution="72" Type="Linedraw" Width="15" /> </InlineMediaObject> <EquationSource Format="TEX">\(\beta \)</EquationSource> <EquationSource Format="MATHML"><math> <mi>β</mi> </math></EquationSource> </InlineEquation> are both negative. It is demonstrated that, in dimensions <InlineEquation ID="IEq8"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq8.gif" Format="GIF" Height="15" Rendition="HTML" Resolution="72" Type="Linedraw" Width="43" /> </InlineMediaObject> <EquationSource Format="TEX">\(n\ge 3\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>n</mi> <mo>≥</mo> <mn>3</mn> </mrow> </math></EquationSource> </InlineEquation>, it is possible to build a quasi-Lagrange operator that reproduces all polynomials of degree <InlineEquation ID="IEq9"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq9.gif" Format="GIF" Height="14" Rendition="HTML" Resolution="72" Type="Linedraw" Width="41" /> </InlineMediaObject> <EquationSource Format="TEX">\(n-3\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>n</mi> <mo>-</mo> <mn>3</mn> </mrow> </math></EquationSource> </InlineEquation> when <i>n</i> is even and of degree <InlineEquation ID="IEq10"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq10.gif" Format="GIF" Height="22" Rendition="HTML" Resolution="72" Type="Linedraw" Width="24" /> </InlineMediaObject> <EquationSource Format="TEX">\(\frac{n-1}{2}\)</EquationSource> <EquationSource Format="MATHML"><math> <mfrac> <mrow> <mi>n</mi> <mo>-</mo> <mn>1</mn> </mrow> <mn>2</mn> </mfrac> </math></EquationSource> </InlineEquation> when <i>n</i> is odd. Furthermore, the uniform approximation error is given by <InlineEquation ID="IEq11"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq11.gif" Format="GIF" Height="23" Rendition="HTML" Resolution="72" Type="Linedraw" Width="128" /> </InlineMediaObject> <EquationSource Format="TEX">\(\mathcal {O}\left( h^{n-2}\log (1/h)\right) \)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi mathvariant="script">O</mi> <mfenced close=")" open="("> <msup> <mi>h</mi> <mrow> <mi>n</mi> <mo>-</mo> <mn>2</mn> </mrow> </msup> <mo>log</mo> <mrow> <mo stretchy="false">(</mo> <mn>1</mn> <mo stretchy="false">/</mo> <mi>h</mi> <mo stretchy="false">)</mo> </mrow> </mfenced> </mrow> </math></EquationSource> </InlineEquation> for <i>n</i> even and <InlineEquation ID="IEq12"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="41_2025_10169_Article_IEq12.gif" Format="GIF" Height="33" Rendition="HTML" Resolution="72" Type="Linedraw" Width="71" /> </InlineMediaObject> <EquationSource Format="TEX">\(\mathcal {O}\left( h^{\frac{n-3}{2}}\right) \)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi mathvariant="script">O</mi> <mfenced close=")" open="("> <msup> <mi>h</mi> <mfrac> <mrow> <mi>n</mi> <mo>-</mo> <mn>3</mn> </mrow> <mn>2</mn> </mfrac> </msup> </mfenced> </mrow> </math></EquationSource> </InlineEquation> for <i>n</i> odd.</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 Quasi-Interpolation and Their Associated Shift-Invariant Space Using a New Class of Generalized Thin Plate Splines and Inverse Multiquadrics

  • Mathis Ortmann,
  • Martin Buhmann

摘要

A new generalization of shifted thin plate splines \(\begin{aligned} \phi (x)=(c^{2d}+||x||^{2d})\log \left( c^{2d}+||x||^{2d}\right) ,\qquad x\in \mathbb {R}^n, d\in \mathbb {N}, c>0, \end{aligned}\) ϕ ( x ) = ( c 2 d + | | x | | 2 d ) log c 2 d + | | x | | 2 d , x R n , d N , c > 0 , is presented to further increase the accuracy of quasi-interpolation. When restricted to Euclidean spaces of even dimensionality n, the generalization enables the construction of a quasi-Lagrange operator that reproduces all polynomials of degree \(n+2d-1\) n + 2 d - 1 . The case complements the newly proposed generalized multiquadric \(\phi (x)=\sqrt{c^{2d}+||x||^{2d}},\quad x\in \mathbb {R}^n, d\in \mathbb {N}, c>0\) ϕ ( x ) = c 2 d + | | x | | 2 d , x R n , d N , c > 0 , which is restricted to odd dimensions, as shown in a previous work. This generalization improves the approximation order by a factor of \(\mathcal {O}\left( h^{2(d-1)}\right) \) O h 2 ( d - 1 ) , where \(h>0\) h > 0 denotes the fill distance and \(d=1\) d = 1 represents the classical thin plate spline. The results are then compared with the theoretical optimal approximation from the shift-invariant space generated by this function. Moreover, we introduce a new class of inverse multiquadrics \(\begin{aligned} \phi (x)=\left( c^\lambda +||x||^\lambda \right) ^\beta ,\qquad x\in \mathbb {R}^n, \lambda \in \mathbb {R},\beta \in \mathbb {R}\backslash \mathbb {N}, c>0. \end{aligned}\) ϕ ( x ) = c λ + | | x | | λ β , x R n , λ R , β R \ N , c > 0 . We provide an explicit representation of the generalized Fourier transform and discuss its asymptotic behaviour near the origin. Particular emphasis is placed on the case where \(\lambda \) λ and \(\beta \) β are both negative. It is demonstrated that, in dimensions \(n\ge 3\) n 3 , it is possible to build a quasi-Lagrange operator that reproduces all polynomials of degree \(n-3\) n - 3 when n is even and of degree \(\frac{n-1}{2}\) n - 1 2 when n is odd. Furthermore, the uniform approximation error is given by \(\mathcal {O}\left( h^{n-2}\log (1/h)\right) \) O h n - 2 log ( 1 / h ) for n even and \(\mathcal {O}\left( h^{\frac{n-3}{2}}\right) \) O h n - 3 2 for n odd.