Performance of Richardson extrapolation on some numerical methods for a singularly perturbed turning point problem whose solution has boundary layers (Q2877775)

From MaRDI portal





scientific article; zbMATH DE number 6334122
Language Label Description Also known as
English
Performance of Richardson extrapolation on some numerical methods for a singularly perturbed turning point problem whose solution has boundary layers
scientific article; zbMATH DE number 6334122

    Statements

    0 references
    0 references
    25 August 2014
    0 references
    singularly perturbed turning point problems
    0 references
    finite difference scheme
    0 references
    exponentially fitted difference scheme
    0 references
    Shishkin meshes
    0 references
    error estimate
    0 references
    Richardson extrapolation
    0 references
    boundary layers
    0 references
    interior layers
    0 references
    numerical result
    0 references
    Performance of Richardson extrapolation on some numerical methods for a singularly perturbed turning point problem whose solution has boundary layers (English)
    0 references
    The authors apply the Richardson extrapolation technique to some numerical schemes for solving singularly perturbed turning point problems. Generally, the solutions of singularly perturbed turning point problems exhibit either exponential boundary layers or interior layers. In this paper, the authors consider the first case, to solve these problems numerically; exponentially fitted difference schemes (also known as fitted operator methods) and finite difference schemes on piecewise-uniform Shishkin meshes are available in the literature. Also, the Richardson extrapolation technique is well-known, the post-processing technique exists in the literature to improve the order of accuracy. Here, the authors combine these two techniques to improve the accuracy of the numerical solution of singularly perturbed turning point problems. But there is no improvement in the final results. After extrapolation also, the convergence rate still remain first-order only. The authors do not provide any mathematical reasons for not achieving the second-order convergence rate. The numerical results remain almost the same both before and after the Richardson extrapolation, which requires calculating the solution twice on two different grids (computationally expensive).
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references