Stable computations with flat radial basis functions using vector-valued rational approximations

One commonly finds in applications of smooth radial basis functions (RBFs) that scaling the kernels so they are 'flat' leads to smaller discretization errors. However, the direct numerical approach for computing with flat RBFs (RBF-Direct) is severely ill-conditioned. We present an algorithm for bypassing this ill-conditioning that is based on a new method for rational approximation (RA) of vector-valued analytic functions with the property that all components of the vector share the same singularities. This new algorithm (RBF-RA) is more accurate, robust, and easier to implement than the Contour-Pade method, which is similarly based on vector-valued rational approximation. In contrast to the stable RBF-QR and RBF-GA algorithms, which are based on finding a better conditioned base in the same RBF-space, the new algorithm can be used with any type of smooth radial kernel, and it is also applicable to a wider range of tasks (including calculating Hermite type implicit RBF-FD stencils). We present a series of numerical experiments demonstrating the effectiveness of this new method for computing RBF interpolants in the flat regime. We also demonstrate the flexibility of the method by using it to compute implicit RBF-FD formulas in the flat regime and then using these for solving Poisson's equation in a 3-D spherical shell.

[1]  Bengt Fornberg,et al.  A Stable Algorithm for Flat Radial Basis Functions on a Sphere , 2007, SIAM J. Sci. Comput..

[2]  Ana C. Matos,et al.  Algebraic properties of robust Padé approximants , 2015, J. Approx. Theory.

[3]  Graham F. Carey,et al.  A high-order compact formulation for the 3D Poisson equation , 1996 .

[4]  Gregory E. Fasshauer,et al.  Meshfree Approximation Methods with Matlab , 2007, Interdisciplinary Mathematical Sciences.

[5]  Guirong Liu,et al.  A point interpolation meshless method based on radial basis functions , 2002 .

[6]  Bengt Fornberg,et al.  On choosing a radial basis function and a shape parameter when solving a convective PDE on a sphere , 2008, J. Comput. Phys..

[7]  Edward J. Fuselier,et al.  Sobolev-type approximation rates for divergence-free and curl-free RBF interpolants , 2008, Math. Comput..

[8]  Holger Wendland,et al.  Fast evaluation of radial basis functions : methods based on partition of unity , 2002 .

[9]  Bengt Fornberg,et al.  Stabilization of RBF-generated finite difference methods for convective PDEs , 2011, J. Comput. Phys..

[10]  Roberto Cavoretto,et al.  Partition of unity interpolation on multivariate convex domains , 2014, Int. J. Model. Simul. Sci. Comput..

[11]  C. Shu,et al.  Local radial basis function-based differential quadrature method and its application to solve two-dimensional incompressible Navier–Stokes equations , 2003 .

[12]  A. I. Tolstykh,et al.  On using radial basis functions in a “finite difference mode” with applications to elasticity problems , 2003 .

[13]  Elisabeth Larsson,et al.  Radial basis function partition of unity methods for pricing vanilla basket options , 2016, Comput. Math. Appl..

[14]  Y. V. S. S. Sanyasiraju,et al.  Local radial basis function based gridfree scheme for unsteady incompressible viscous flows , 2008, J. Comput. Phys..

[15]  Robert Schaback,et al.  Error estimates and condition numbers for radial basis function interpolation , 1995, Adv. Comput. Math..

[16]  Bengt Fornberg,et al.  On the role of polynomials in RBF-FD approximations: I. Interpolation and accuracy , 2016, J. Comput. Phys..

[17]  Zongmin Wu,et al.  Hermite-Birkhoff interpolation of scattered data by radial basis functions , 1992, Approximation Theory and its Applications.

[18]  Robert Michael Kirby,et al.  A Radial Basis Function (RBF)-Finite Difference (FD) Method for Diffusion and Reaction–Diffusion Equations on Surfaces , 2014, Journal of Scientific Computing.

[19]  Y. Hon,et al.  Domain decomposition for radial basis meshless methods , 2004 .

[20]  B. Fornberg,et al.  Radial Basis Function-Generated Finite Differences: A Mesh-Free Method for Computational Geosciences , 2015 .

[21]  Bengt Fornberg,et al.  Stable calculation of Gaussian-based RBF-FD stencils , 2013, Comput. Math. Appl..

[22]  Elisabeth Larsson,et al.  A Radial Basis Function Partition of Unity Collocation Method for Convection–Diffusion Equations Arising in Financial Applications , 2015, J. Sci. Comput..

[23]  B. Fornberg,et al.  Some observations regarding interpolants in the limit of flat radial basis functions , 2003 .

[24]  M KirbyRobert,et al.  A Radial Basis Function (RBF)-Finite Difference (FD) Method for Diffusion and Reaction---Diffusion Equations on Surfaces , 2015 .

[25]  Elisabeth Larsson,et al.  Stable Computations with Gaussian Radial Basis Functions , 2011, SIAM J. Sci. Comput..

[26]  Erik Lehto,et al.  A guide to RBF-generated finite differences for nonlinear transport: Shallow water simulations on a sphere , 2012, J. Comput. Phys..

[27]  Y. Hon,et al.  Overlapping domain decomposition method by radial basis functions , 2003 .

[28]  John P. Boyd,et al.  Error saturation in Gaussian radial basis functions on a finite interval , 2010, J. Comput. Appl. Math..

[29]  Bengt Fornberg,et al.  On the role of polynomials in RBF-FD approximations: II. Numerical solution of elliptic PDEs , 2017, J. Comput. Phys..

[30]  B. Fornberg,et al.  A numerical study of some radial basis function based solution methods for elliptic PDEs , 2003 .

[31]  Bengt Fornberg,et al.  Scattered node compact finite difference-type formulas generated from radial basis functions , 2006, J. Comput. Phys..

[32]  T. Driscoll,et al.  Interpolation in the limit of increasingly flat radial basis functions , 2002 .

[33]  V. Maz'ya,et al.  On approximate approximations using Gaussian kernels , 1996 .

[34]  David Stevens,et al.  The use of PDE centres in the local RBF Hermitian method for 3D convective-diffusion problems , 2009, J. Comput. Phys..

[35]  A. U.S.,et al.  Stable Computation of Multiquadric Interpolants for All Values of the Shape Parameter , 2003 .

[36]  Bengt Fornberg,et al.  The Runge phenomenon and spatially variable shape parameters in RBF interpolation , 2007, Comput. Math. Appl..

[37]  M. Froissart,et al.  Approximation de Pade Application à la physique des particules élémentaires , 1969 .

[38]  Bengt Fornberg,et al.  A primer on radial basis functions with applications to the geosciences , 2015, CBMS-NSF regional conference series in applied mathematics.

[39]  L. Trefethen,et al.  Robust rational interpolation and least-squares , 2011 .

[40]  B. Fornberg,et al.  Theoretical and computational aspects of multivariate interpolation with increasingly flat radial basis functions , 2003 .

[41]  Bengt Fornberg,et al.  Fast calculation of Laurent expansions for matrix inverses , 2016, J. Comput. Phys..

[42]  Edward B. Saff,et al.  Low Complexity Methods For Discretizing Manifolds Via Riesz Energy Minimization , 2013, Found. Comput. Math..

[43]  Michael J. McCourt,et al.  Stable Evaluation of Gaussian Radial Basis Function Interpolants , 2012, SIAM J. Sci. Comput..

[44]  R. Schaback.  Multivariate Interpolation by Polynomials and Radial Basis Functions , 2005 .

[45]  Bengt Fornberg,et al.  Solving PDEs with radial basis functions * , 2015, Acta Numerica.