scholarly journals A sharp error estimate of piecewise polynomial collocation for nonlocal problems with weakly singular kernels

Author(s):  
Minghua Chen ◽  
Wenya Qi ◽  
Jiankang Shi ◽  
Jiming Wu

Abstract As is well known, piecewise linear polynomial collocation (PLC) and piecewise quadratic polynomial collocation (PQC) are used to approximate the weakly singular integrals $$\begin{equation*}I(a,b,x) =\int^b_a \frac{u(y)}{|x-y|^\gamma}\textrm{d}y, \quad x \in (a,b),\quad 0< \gamma <1,\end{equation*}$$which have local truncation errors $\mathcal{O} (h^2 )$ and $\mathcal{O} (h^{4-\gamma } )$, respectively. Moreover, for Fredholm weakly singular integral equations of the second kind, i.e., $\lambda u(x)- I(a,b,x) =f(x)$, $\lambda \neq 0$, the global convergence rates are also $\mathcal{O} (h^2 )$ and $\mathcal{O} (h^{4-\gamma } )$ by PLC and PQC in Atkinson (2009, The Numerical Solution of Integral Equations of the Second Kind, Cambridge University Press). In this work we study the following nonlocal problems, which are similar to the above Fredholm integral equations: $$\begin{equation*}\int^b_a \frac{u(x)-u(y)}{|x-y|^\gamma}\textrm{d}y =f(x), \quad x \in (a,b),\quad 0< \gamma <1. \end{equation*}$$In the first part of this paper we prove that the weakly singular integrals $I(a,b,x)$ have optimal local truncation error $\mathcal{O}(h^4\eta _i^{-\gamma } )$ by PQC, where $\eta _i=\min \left \{x_i-a,b-x_i\right \}$ and $x_i$ coincides with an element junction point. Then the sharp global convergence orders $\mathcal{O}\left (h\right )$ and $\mathcal{O} (h^3)$ by PLC and PQC, respectively, are established for nonlocal problems. Finally, numerical experiments are shown to illustrate the effectiveness of the presented methods.

2018 ◽  
Vol 23 (4) ◽  
pp. 639-664 ◽  
Author(s):  
Sumaira Rehman ◽  
Arvet Pedas ◽  
Gennadi Vainikko

We discuss the bounds of fast solving weakly singular Fredholm integral equations of the second kind with a possible diagonal singularity of the kernel and certain boundary singularities of the derivatives of the free term when the information about the smooth coefficient functions in the kernel and about the free term is restricted to a given number of sample values. In this situation, a fast/quasifast solver is constructed. Thus the complexity of weakly singular integral equations occurs to be close to that of equations with smooth data without singularities. Our construction of fast/quasifast solvers is based on the periodization of the problem.


2000 ◽  
Vol 5 (1) ◽  
pp. 76-85
Author(s):  
K. Hakk

For the solution of weakly singular integral equations by the piecewise polynomial collocation method it is necessary to solve large linear systems. In the present paper a two‐grid iteration method for solving such systems is constructed and the convergence of this method is investigated.


1998 ◽  
Vol 3 (1) ◽  
pp. 104-113
Author(s):  
Kristiina Hakk ◽  
Arvet Pedas

The piecewise polynomial collocation method is discussed to solve second kind Fredholm integral equations with weakly singular kernels K (t, s) which may be discontinuous at s = d, d = const. The main result is given in Theorem 4.1. Using special collocation points, error estimates at the collocation points are derived showing a more rapid convergence than the global uniform convergence in the interval of integration available by piecewise polynomials.


2008 ◽  
Vol 13 (1) ◽  
pp. 29-36 ◽  
Author(s):  
R. Kangro ◽  
I. Kangro

Piecewise polynomial collocation methods on special nonuniform grids are efficient methods for solving weakly singular Fredholm and Volterra integral equations but there is a widespread belief that those methods are numerically unstable in the case of large values of the nonuniformity parameter r. We show that this method by itself is stable and discuss some implementation problems that may lead to unstable behavior of numerical results.


2011 ◽  
Vol 2011 ◽  
pp. 1-16
Author(s):  
Sudeshna Banerjea ◽  
Barnali Dutta ◽  
A. Chakrabarti

A direct function theoretic method is employed to solve certain weakly singular integral equations arising in the study of scattering of surface water waves by vertical barriers with gaps. Such integral equations possess logarithmically singular kernel, and a direct function theoretic method is shown to produce their solutions involving singular integrals of similar types instead of the stronger Cauchy-type singular integrals used by previous workers. Two specific ranges of integration are examined in detail, which involve the following: Case(i) two disjoint finite intervals (0,a)∪(b,c) and (a,b,c being finite ) and Case(ii) a finite union of n disjoint intervals. The connection of such integral equations for Case(i), with a particular water wave scattering problem, is explained clearly, and the important quantities of practical interest (the reflection and transmission coefficients) are determined numerically by using the solution of the associated weakly singular integral equation.


Filomat ◽  
2018 ◽  
Vol 32 (4) ◽  
pp. 1323-1333 ◽  
Author(s):  
Sales Nabavi ◽  
O. Baghani

We deal with some sources of Banach spaces which are closely related to an important issue in applied mathematics i.e. the problem of existence and uniqueness of the solution for the very applicable weakly singular integral equations. In the classical mode, the uniform space (C[a,b], ||.||?) is usually applied to the related discussion. Here, we apply some new types of Banach spaces, in order to extend the area of problems we could discuss. We consider a very general type of singular integral equations involving n weakly singular kernels, for an arbitrary natural number n, without any restrictive assumption of differentiability or even continuity on engaged functions. We show that in appropriate conditions the following multi-singular integral equation of weakly singular type has got exactly a solution in a defined Banach space x(t) = ?p,i=1 ?i/?(^?i) ?t,0 fi(s,x(s)) (tn-tn-1)1-?i,n...(t1-s)1-?i,1 dt + ?(t). In particular we consider the famous fractional Langevin equation and by the method we could extend the region of variations of parameter ?+ ? from interval [0,1) in the earlier works to interval [0,2).


Sign in / Sign up

Export Citation Format

Share Document