Fast algorithms for the solution of pseudodifferential equations
algorithmboundary element methodCauchy type singular integral equationcomputational complexityconvergencecorrected collocation methoderror estimatesnumerical resultsparametrix-Galerkin method
Boundary value problems for PDEs with pseudodifferential operators (35S15) Integral equations with kernels of Cauchy type (45E05) Stability and convergence of numerical methods for boundary value problems involving PDEs (65N12) Error bounds for boundary value problems involving PDEs (65N15) Boundary element methods for boundary value problems involving PDEs (65N38) Numerical methods for integral equations (65R20) Complexity and performance of numerical algorithms (65Y20)
In many cases the boundary element method produces an integral equation having a kernel singularity of the type \(x^{-1}\), a so-called Cauchy type singular integral equation given on the smooth closed curve \(\Gamma\) \[ a(t)u(t) +{b(t) \over \pi i} \int_\Gamma {u(\tau)d \tau \over \tau-t} +\int_\Gamma k(t,\tau) u(\tau)d \tau=f(t), \quad t\in \Gamma, \tag{1} \] where \(k(t,\tau)\) is smoother than \((\tau-t)^{-1}\). The present paper is mainly devoted to the solution of this kind of equations. We give an algorithm called parametrix-Galerkin method being based on the corrected collocation method. Its computational complexity is only \(O(n\log n)\) and it has a better convergence behaviour in ``negative Sobolev norms than the collocation method. Further, we describe some other fast algorithms for equation (1) given on smooth closed curves \(\Gamma\) or on the interval \([-1,1]\) and prove their convergence in several scales of function spaces, which we define in Section 2. In Section 3 we deal with special properties of equation (1). The ``classical approximation methods are described in Section 4, whereas in Section 5 the fast algorithms are proposed. All of the algorithms consist of two steps. In the first step, a parametrix for the pseudodifferential operator \(A\) is used in order to get an approximate solution on a fine grid. In a second step this solution is corrected in a certain manner by the solution on a coarse grid. Mostly, we only give an overview of the algorithms since the results are partly published yet. The errror estimates in Hölder-Zygmund norms (Subsection 5.3) are new. With some remarks on further investigations (Section 6) and numerical results (Section 7) we close our paper. For the entire collection see [Zbl 0877.00013].
- Fast solvers for partial differential equations. Abstracts from the workshop held May 18--24, 2008.
- Uniform convergence of a fast algorithm for Cauchy singular integral equations
- Coulomb interactions on planar structures: Inverting the square root of the Laplacian
- Efficient algorithms for solving partial differential equations with discontinuous solutions
- A fast trigonometric collocation method for some elliptic pseudodifferential equations
- Fast solvers for partial differential equations. Abstracts from the workshop held May 11--16, 2014.
- Fast solvers of integral and pseudodifferential equations on closed curves
- The fast solution of periodic pseudodifferential equations
- scientific article; zbMATH DE number 1129765 (Why is no real title available?)
- Fast and Oblivious Algorithms for Dissipative and Two-dimensional Wave Equations
- A fast algorithm for Prandtl's integro-differential equation
This page was built for publication: Fast algorithms for the solution of pseudodifferential equations
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4371001)
