On convergence of integrated radial basis function approximations on specific stencils with numerical applications

Fazlollah Soleymani, Shengfeng Zhu, Mahdiar Barfeie

Research output: Contribution to journalArticlepeer-review

Abstract

One strategy to enhance the performance of the localized meshless radial basis function scheme in finite difference mode (RBF-FD) is to incorporate a polynomial term into the RBF-FD framework, known as the integrated radial basis function (IRBF) approach. In this work, we propose a comprehensive framework for obtaining the approximation weights for the IRBF method. Then, we derive analytical expressions for the weighting coefficients in the one-dimensional scenario for specific stencils, enabling us to explore the theoretical high-order convergence properties at the nodal points utilizing the Gaussian RBF. Furthermore, these weights are extended to be used for solving high-dimensional partial differential problems, such as the Poisson equation and the four-dimensional time-dependent Black-Scholes equation.

Original languageEnglish
Pages (from-to)1137-1166
Number of pages30
JournalJournal of Applied Mathematics and Computing
Volume71
Issue numberSuppl 1
DOIs
StatePublished - Sep 2025

Keywords

  • Integrated radial basis function
  • Polynomial augmentation
  • Radial basis function
  • Specific stencils
  • Weights

Fingerprint

Dive into the research topics of 'On convergence of integrated radial basis function approximations on specific stencils with numerical applications'. Together they form a unique fingerprint.

Cite this