Numerical solution of a certain hypersingular integral equation of the first kind (Q639954)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Numerical solution of a certain hypersingular integral equation of the first kind
scientific article

    Statements

    Numerical solution of a certain hypersingular integral equation of the first kind (English)
    0 references
    0 references
    0 references
    0 references
    0 references
    11 October 2011
    0 references
    The paper is devoted to the collocation method for the solution of the hypersingular integral equation \[ I(s,u) \equiv \frac{1}{4\pi} \int_0^{2\pi} \frac{u(x)}{\sin^2 \frac{x-s}{2}}dx = f(s), \quad s \in (0, 2\pi) \] with the compatibility condition \(\int_0^{2\pi} f(x)dx = 0.\) At first, the authors investigate the quadrature rule \(I(s,u) = L_{on}(s,u) + R_{on}(s,u),\) where \[ L_{on}(s,u) = \int_0^{2\pi}\left(\sum^n_{i=1} u(\bar x_i) \varphi_i(x)\right) \frac{dx}{\sin^2 \frac{x-s}{2}}, \tag{1} \] \(0 = x_0 < x_1 < \cdot < x_{n-1} < x_n = 2\pi,\) \(\bar x_i = (x_{i-1} + x_i)/2,\) \[ \varphi_i(x) = \begin{cases} 1, &x \in [x_{i-1}, x_i], \\ 0, &x \in [0, 2\pi] \backslash [x_{i-1}, x_i]. \end{cases} \] For a uniform mesh they have \(I(s,u) - L_{on}(s,u) = -4\pi u'(s) \tan \frac{\tau \pi}{2} + \epsilon(s),\) where \[ |\epsilon(s)| \leq \begin{cases} C\rho(s) \gamma^{-1}(s,h) h^{1+\alpha}, \, &u(x) \in C^{2+\alpha}[0, 2\pi],\\ C\rho(s) \gamma^{-1}(h,s) h^{2} |\ln h|, \, &u(x) \in C^{3}[0, 2\pi],\\ C\rho(s) \gamma^{-1}(h,s) h^{2}, \, &u(x) \in C^{3+\alpha}[0, 2\pi];\\ \end{cases} \] \(\tau=x_{m-1}+(\tau+1)h/2\) with \(\tau \in (-1,1)\); \(\rho(s), \gamma(h,s)\) are some functions, \(0 < \alpha < 1\). Using quadrature formula (1) they construct the collocation numerical scheme with collocation points \(\bar x_i,\) \(i=1,2,\dots,n.\) The error of this scheme is equal to \(C h^2\) for \(u \in C^{3+\alpha} [0, 2\pi].\)
    0 references
    0 references
    hypersingular integral equation
    0 references
    collocation method
    0 references
    error bounds
    0 references
    midpoint rule
    0 references
    superconvergence
    0 references
    quadrature rule
    0 references
    0 references
    0 references
    0 references
    0 references