A numerical study of the higher-dimensional Gelfand-Bratu model

Sehar Iqbal*, Paul Andries Zegeling

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

In this article, a higher dimensional nonlinear boundary-value problem, viz., Gelfand-Bratu (GB) problem, is solved numerically. For the three-dimensional case, we present an accurate and efficient nonlinear multigrid (MG) approach and investigate multiplicities depending on the bifurcation parameter λ. We adopt a nonlinear MG approach Full approximation scheme (FAS) extended with a Krylov method as a smoother to handle the computational difficulties for obtaining the upper branches of the solutions. Further, we examine the numerical bifurcation behaviour of the GB problem in 3D and identify the existence of two new bifurcation points. Experiments illustrate the convergence of the numerical solutions and demonstrate the effectiveness of the proposed numerical strategy for all parameter values λ∈(0,λc]. For higher dimensions, we transform the GB problem, using n-dimensional spherical coordinates, to a nonlinear ordinary differential equation (ODE). The numerical solutions of this nonlinear ODE are computed by a shooting method for a range of values of the dimension parameter n. Numerical experiments show the existence of several types of solutions for different values of n and λ. These results confirm the bifurcation behaviour of the higher dimensional GB problem as predicted from theoretical results in literature.

Original languageEnglish
Pages (from-to)1619-1633
Number of pages14
JournalComputers and Mathematics with Applications
Volume79
Issue number6
DOIs
Publication statusPublished - 15 Mar 2020

Funding

Sehar Iqbal acknowledges the financial support by the Schlumberger Foundation (Faculty for the Future award).

Keywords

  • Bifurcation behaviour
  • Boundary-value problems
  • Bratu problem
  • Finite differences
  • Nonlinear multigrid

Fingerprint

Dive into the research topics of 'A numerical study of the higher-dimensional Gelfand-Bratu model'. Together they form a unique fingerprint.

Cite this