Error estimates for harmonic and biharmonic interpolation splines with annular geometry

Abstract: The main result in this paper is an error estimate for interpolation biharmonic polysplines in an annulus $A\left( r_{1},r_{N}\right) $, with respect to a partition by concentric annular domains $A\left( r_{1} ,r_{2}\right) ,$ ...., $A\left( r_{N-1},r_{N}\right) ,$ for radii $0<r_{1}<....<r_{N}.$ The biharmonic polysplines interpolate a smooth function on the spheres $\left\vert x\right\vert =r_{j}$ for $j=1,...,N$ and satisfy natural boundary conditions for $\left\vert x\right\vert =r_{1}$ and $\left\vert x\right\vert =r_{N}.$ By analogy with a technique in one-dimensional spline theory established by C. de Boor, we base our proof on error estimates for harmonic interpolation splines with respect to the partition by the annuli $A\left( r_{j-1},r_{j}\right) $. For these estimates it is important to determine the smallest constant $c\left( \Omega\right) ,$ where $\Omega=A\left( r_{j-1},r_{j}\right) ,$ among all constants $c$ satisfying [ \sup_{x\in\Omega}\left\vert f\left( x\right) \right\vert \leq c\sup _{x\in\Omega}\left\vert \Delta f\left( x\right) \right\vert ] for all $f\in C^{2}\left( \Omega\right) \cap C\left( \overline{\Omega }\right) $ vanishing on the boundary of the bounded domain $\Omega$ . In this paper we describe $c\left( \Omega\right) $ for an annulus $\Omega=A\left( r,R\right) $ and we will give the estimate [ \min{\frac{1}{2d},\frac{1}{8}}\left( R-r\right) ^{2}\leq c\left( A\left( r,R\right) \right) \leq\max{\frac{1}{2d},\frac{1}{8}}\left( R-r\right) ^{2}% ] where $d$ is the dimension of the underlying space.