Convergence properties of numerical discretizations and regridding methods (Q2366434)

From MaRDI portal
Revision as of 01:12, 15 February 2024 by RedirectionBot (talk | contribs) (‎Removed claims)
scientific article
Language Label Description Also known as
English
Convergence properties of numerical discretizations and regridding methods
scientific article

    Statements

    Convergence properties of numerical discretizations and regridding methods (English)
    0 references
    29 June 1993
    0 references
    The paper is devoted to the study of difference schemes for the following two problems: (P1) \(-u_ x+s(x)=0\), \(u(0)=u_ 0\) \((0<x<1)\) and (P2) \(u_{xx}+s(x)=0\), \(u(0)=u_ 0\), \(u(1)=u_ 1\) \((0<x<1)\). As is well known, both problems occur when studying the steady state solution of the transport or heat equation. After suitable difference discretization of the derivatives it is proved that for an arbitrary grid \(0=x_ 0<x_ 1<\dots<x_ N=1\) with \(h_ i=x_ i-x_{i-1}\), \(h=\max\{h_ i\}\) both the truncation errors \(T_ i\) and the global errors \(E_ i\) are of order \(O(h)\) in problem (P1) while in problem (P2) one has \(T_ i=O(h)\) and \(E_ i=O(h^ 2)\) (so there is a superconvergence phenomenon). Let us now put \(x_ i=g(\xi_ i)\), where \(\xi_ i\) is the uniform grid. Sufficient conditions for the function \(g:(0,1)\to (0,1)\) are given to obtain \(T_ i=O(h^ p)\), \(E_ i=O(h^ p)\) \((p=2,4)\) in both problems (P1), (P2).
    0 references
    convergence
    0 references
    nonuniform grids
    0 references
    regridding methods
    0 references
    difference schemes
    0 references
    truncation errors
    0 references
    global errors
    0 references
    superconvergence
    0 references

    Identifiers

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