Finite difference approximations to the Dirichlet problem for elliptic systems (Q1082052)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Finite difference approximations to the Dirichlet problem for elliptic systems
scientific article

    Statements

    Finite difference approximations to the Dirichlet problem for elliptic systems (English)
    0 references
    0 references
    0 references
    0 references
    1986
    0 references
    The authors consider systems of difference equations \[ (1)\quad \Delta^ hu+f(x,u)=0,\quad in\quad \Omega_ h\subset {\mathbb{R}}^ m,\quad u\in {\mathbb{R}}^ n,\quad u=0\quad on\quad \Gamma_ h \] where \(\Delta^ h\) is usual difference approximation of the Laplacian and \(\partial f_ i/\partial u_ j\geq 0\), \(i\neq j\), \(\partial f_ i/\partial u_ i\geq - L\). Based on inverse isotonity of \(I-\lambda \Delta^ h\), \(\lambda\geq 0\), monotone iterative schemes for solving (1) are constructed. These schemes are motivated by Crank-Nicholson and alternating direction implicit discretization of the corresponding parabolic problem. Unlike the stability problem the monotonity condition obtained in the article for the Crank-Nicholson scheme only distinguishes the purely implicit case, when no restriction on the time step (iteration parameter) arises. The necessity of this condition is supported by a numerical experiment. The authors use their schemes to obtain upper and lower solutions converging to minimal and maximal solutions, respectively, and to check uniqueness of the numerical solutions. Numerical results are given for a system from mathematical biology \((m=1\), \(n=2)\).
    0 references
    0 references
    Poisson equation
    0 references
    systems of difference equations
    0 references
    inverse isotonity
    0 references
    monotone iterative schemes
    0 references
    Crank-Nicholson
    0 references
    alternating direction implicit discretization
    0 references
    stability
    0 references
    upper and lower solutions
    0 references
    minimal and maximal solutions
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references