New methods for quasi-interpolation approximations: resolution of odd-degree singularities

Martin Buhmann,Janin Jäger,Joaquín Jódar,Miguel L. Rodríguez
2023-09-06
Abstract:In this paper, we study functional approximations where we choose the so-called radial basis function method and more specifically, quasi-interpolation. From the various available approaches to the latter, we form new quasi-Lagrange functions when the orders of the singularities of the radial function's Fourier transforms at zero do not match the parity of the dimension of the space, and therefore new expansions and coefficients are needed to overcome this problem. We develop explicit constructions of infinite Fourier expansions that provide these coefficients and make an extensive comparison of the approximation qualities and - with a particular focus - polynomial precision and uniform approximation order of the various formulae. One of the interesting observations concerns the link between algebraic conditions of expansion coefficients and analytic properties of localness and convergence.
Numerical Analysis,Functional Analysis
What problem does this paper attempt to address?
### What problem does this paper attempt to solve? This paper primarily investigates the issue of singularity handling in quasi-interpolation approximation methods, particularly in the case of odd-order singularities. Specifically: 1. **Research Background**: - The paper employs the Radial Basis Function (RBF) method, with a particular focus on quasi-interpolation techniques. - When the order of singularity of the Fourier transform of the radial basis function at the origin does not match the parity of the spatial dimension, new expansions and coefficients are needed to overcome this issue. 2. **Core Problem**: - When the order of singularity of the Fourier transform of the radial basis function at the origin is odd, traditional quasi-Lagrange functions cannot be directly applied. Therefore, new methods need to be developed to handle this situation. - Specifically, the paper focuses on two cases: Multiquadrics and Thin-Plate Splines: - Multiquadrics are suitable for odd-dimensional spaces to ensure the order of singularity is even. - Thin-Plate Splines are suitable for even-dimensional spaces, also ensuring the order of singularity is even. 3. **Solution Methods**: - For Thin-Plate Splines in 1-dimensional space, where the order of singularity is odd, the paper proposes using an infinite Fourier series to construct new quasi-Lagrange functions. - For Multiquadrics in 1-dimensional space, by selecting appropriate Fourier coefficients, a finite-term quasi-Lagrange function is constructed to achieve polynomial accuracy. 4. **Main Contributions**: - A series of new methods for handling odd-order singularities are proposed, including infinite Fourier series and finite-term Fourier coefficients. - The paper compares the approximation quality, polynomial accuracy, and uniform approximation order of different methods. - It discovers the relationship between polynomial accuracy and the algebraic conditions of Fourier coefficients, and explores the analytical properties of locality and convergence. In summary, this paper aims to address the approximation problem in radial basis function methods under odd-order singularity conditions and proposes a series of new methods and techniques to tackle this challenge.