Laurent series based RBF-FD method to avoid ill-conditioning

We propose a new approach to avoid the inherent ill-condition in the computation of RBF-FD weights, which is due to the fact that the RBF interpolation matrix is nearly singular. The new approach is based on the semi-analytical computation of the Laurent series of the inverse of the RBF interpolatio...

Full description

Saved in:
Bibliographic Details
Published in:Engineering analysis with boundary elements 2015-03, Vol.52, p.24-31
Main Authors: Gonzalez-Rodriguez, Pedro, Bayona, Victor, Moscoso, Miguel, Kindelan, Manuel
Format: Article
Language:English
Subjects:
Approximation
Computation
Finite differences
Interpolation
Inverse
Mathematical analysis
Mathematical models
Operators
Radial basis functions
Truncation errors
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We propose a new approach to avoid the inherent ill-condition in the computation of RBF-FD weights, which is due to the fact that the RBF interpolation matrix is nearly singular. The new approach is based on the semi-analytical computation of the Laurent series of the inverse of the RBF interpolation matrix. Once the Laurent series is obtained, it can be used to compute the RBF-FD weights of any differential operator exactly without extra cost. The proposed method also provides analytical formulas for the RBF-FD weights in terms of the parameters involved in the problem. These formulas can be used to derive the exact dependence of the truncation error in the approximation of any differential operator of a given function. Furthermore, from the analysis presented here one can derive the values of the parameters involved in the problem for which the RBF interpolation matrix becomes ill-conditioned and, hence, for which the weights cannot be obtained numerically.
ISSN:0955-7997
1873-197X
DOI:10.1016/j.enganabound.2014.10.018