On a method for interpolating functions on chaotic nets (Q1277536): Difference between revisions

The many dimensional apparatus of an interpolation problem is considered. A function of the form \(\sum^N_{j=1}\lambda_jK(x-x_j)\), \(\lambda_j\in \mathbb{R}\); \(x,x_j\in \mathbb{R}^n\) where \(P(x)\) is a polynomial of degree \(\leq m/2\) depending on the set of variables \[ K(x)=\begin{cases} | x|^m,\quad &\text{for }m\text{ odd}\\ | x|^m\ln| x|,\quad &\text{for }m\text{ even}\end{cases} \] conciding with a given function \(f(x)\) at the points \(x_1,\dots,x_N\) is constructed. The approximational properties of this interpolation polynomial are investigated. The straight theorem has been proved: The approximation of a derivative function from the Sobolev space \(W_p^K\) by means of derivatives of this interpolation polynomial is estimated by a smooth modulus of the approximated function in \(L_q\) (theorem 2). The existence of the interpolation polynomial which is a unique solution of one variational problem has been established (theorem 1).