Abstract
In the first part of this work, we derive compact numerical quadrature formulas for finite-range integrals \(I[f]=\int^{b}_{a}f(x)\,dx\), where f(x)=g(x)|x−t|β, β being real. Depending on the value of β, these integrals are defined either in the regular sense or in the sense of Hadamard finite part. Assuming that g∈C ∞[a,b], or g∈C ∞(a,b) but can have arbitrary algebraic singularities at x=a and/or x=b, and letting h=(b−a)/n, n an integer, we derive asymptotic expansions for \({T}^{*}_{n}[f]=h\sum_{1\leq j\leq n-1,\ x_{j}\neq t}f(x_{j})\), where x j =a+jh and t∈{x 1,…,x n−1}. These asymptotic expansions are based on some recent generalizations of the Euler–Maclaurin expansion due to the author (A. Sidi, Euler–Maclaurin expansions for integrals with arbitrary algebraic endpoint singularities, in Math. Comput., 2012), and are used to construct our quadrature formulas, whose accuracies are then increased at will by applying to them the Richardson extrapolation process. We pay particular attention to the case in which β=−2 and f(x) is T-periodic with T=b−a and \(f\in C^{\infty}(-\infty,\infty)\setminus\{t+kT\}^{\infty}_{k=-\infty}\), which arises in the context of periodic hypersingular integral equations. For this case, we propose the remarkably simple and compact quadrature formula \(\widehat{Q}_{n}[f]=h\sum^{n}_{j=1}f(t+jh-h/2)-\pi^{2} g(t)h^{-1}\), and show that \(\widehat{Q}_{n}[f]-I[f]=O(h^{\mu})\) as h→0 ∀μ>0, and that it is exact for a class of singular integrals involving trigonometric polynomials of degree at most n. We show how \(\widehat{Q}_{n}[f]\) can be used for solving hypersingular integral equations in an efficient manner. In the second part of this work, we derive the Euler–Maclaurin expansion for integrals \(I[f]=\int^{b}_{a} f(x)dx\), where f(x)=g(x)(x−t)β, with g(x) as before and β=−1,−3,−5,…, from which suitable quadrature formulas can be obtained. We revisit the case of β=−1, for which the known quadrature formula \(\widetilde{Q}_{n}[f]=h\sum^{n}_{j=1}f(t+jh-h/2)\) satisfies \(\widetilde{Q}_{n}[f]-I[f]=O(h^{\mu})\) as h→0 ∀μ>0, when f(x) is T-periodic with T=b−a and \(f\in C^{\infty}(-\infty,\infty)\setminus\{t+kT\}^{\infty}_{k=-\infty}\). We show that this formula too is exact for a class of singular integrals involving trigonometric polynomials of degree at most n−1. We provide numerical examples involving periodic integrands that confirm the theoretical results.
Similar content being viewed by others
Notes
The usual notation for integrals defined in the sense of the Hadamard finite part (HFP) is \({\mbox{\footnotesize $=$}}\hspace{-1em}\int^{b}_{a} f(x)\,dx\). In this work, we denote them by \(\int^{b}_{a}f(x)\, dx\), as in (1.3), for simplicity. For the definition and properties of Hadamard finite part integrals, see Davis and Rabinowitz [2], Evans [3], or Kythe and Schäferkotter [6], for example. These integrals have most of the properties of regular integrals and some properties that are quite unusual. For example, they are invariant with respect to translation, but they are not necessarily invariant under a scaling of the variable of integration.
The usual notation for integrals defined in the sense of Cauchy principal value (CPV) is \({\mbox{\footnotesize $-$}}\hspace{-1em}\int^{b}_{a} f(x)\,dx\). In this work, we denote them by \(\int^{b}_{a}f(x)\, dx\) for simplicity. For the definition and properties of Cauchy principal value integrals, see [2, 3], or [6], for example.
We can write the expansions in (2.1) in the “simpler” form
$$u(x)\sim\sum^{\infty}_{s=0}c_s \,(x-a)^{\gamma'_s} \quad\mbox{\textit{as}}\ x\to a+, \qquad u(x)\sim\sum ^{\infty}_{s=0}d_s\,(b-x)^{\delta'_s} \quad \mbox{\textit{as}}\ x\to b-, $$ordering the \(\gamma'_{s}\) and the \(\delta'_{s}\) as in (2.2), and allowing now one of the \(\gamma'_{s}\) and/or one of the \(\delta'_{s}\) to be equal to −1. However, this complicates the statements of our results. Therefore, we have chosen to separate these two exponents as in (2.1).
In this case, I[G] is simply the Cauchy principal value of \(\int^{b}_{a} G(x)\,dx\).
References
Capobianco, M.R., Mastroianni, G., Russo, M.G.: Pointwise and uniform approximation of the finite Hilbert transform. In: Approximation and Optimization, vol. I, Cluj-Napoca, 1996, pp. 45–66. Transilvania, Cluj-Napoca (1997)
Davis, P.J., Rabinowitz, P.: Methods of Numerical Integration, 2nd edn. Academic Press, New York (1984)
Evans, G.: Practical Numerical Integration. Wiley, New York (1993)
Gradshteyn, I.S., Ryzhik, I.M.: Table of Integrals, Series, and Products. Academic Press, New York (1980). 4th printing (1983)
Huang, J., Wang, Z., Zhu, R.: Asymptotic error expansions for hypersingular integrals. Adv. Comput. Math. (2011). Published online: September 30. doi:10.1007/s10444-011-9236-x
Kythe, P.K., Schäferkotter, M.R.: Handbook of Computational Methods for Integration. Chapman & Hall/CRC Press, New York (2005)
Ladopoulos, E.G.: Singular Integral Equations: Linear and Non-linear Theory and Its Applications in Science and Engineering. Springer, Berlin (2000)
Lifanov, I.K.: An instance of numerical solution of integral equations with singularities in the periodic case. Differ. Equ. 42(9), 1134–1342 (2006). Translated from Differ. Uravn. 42, 1263–1271 (2006)
Lifanov, I.K., Poltavskii, L.N.: On the numerical solution of hypersingular and singular integral equations on the circle. Differ. Equ. 39(8), 1175–1197 (2003). Translated from Differ. Uravn. 39, 1115–1136 (2003)
Lifanov, I.K., Poltavskii, L.N., Vainikko, G.M.: Hypersingular Integral Equations and Their Applications. CRC Press, New York (2004)
Mastroanni, G., Occorsio, D.: Interlacing properties of the zeros of the orthogonal polynomials and approximation of the Hilbert transform. Comput. Math. Appl. 30, 155–168 (1995)
Mastronardi, N., Occorsio, D.: Some numerical algorithms to evaluate Hadamard finite-part integrals. J. Comput. Appl. Math. 70, 75–93 (1996)
Monegato, G.: Numerical evaluation of hypersingular integrals. J. Comput. Appl. Math. 50, 9–31 (1994)
Navot, I.: An extension of the Euler–Maclaurin summation formula to functions with a branch singularity. J. Math. Phys. 40, 271–276 (1961)
Navot, I.: A further extension of the Euler–Maclaurin summation formula. J. Math. Phys. 41, 155–163 (1962)
Rivlin, T.J.: Chebyshev Polynomials: From Approximation Theory to Algebra and Number Theory, 2nd edn. Wiley, New York (1990)
Sidi, A.: Comparison of some numerical quadrature formulas for weakly singular periodic Fredholm integral equations. Computing 43, 159–170 (1989)
Sidi, A.: Practical Extrapolation Methods: Theory and Applications. Cambridge Monographs on Applied and Computational Mathematics, vol. 10. Cambridge University Press, Cambridge (2003)
Sidi, A.: Euler–Maclaurin expansions for integrals with endpoint singularities: a new perspective. Numer. Math. 98, 371–387 (2004)
Sidi, A.: Euler–Maclaurin expansions for integrals with arbitrary algebraic endpoint singularities. Math. Comput. (2012). Published online: April 10. doi:10.1090/S0025-5718-2012-02597-x
Sidi, A.: Euler–Maclaurin expansions for integrals with arbitrary algebraic-logarithmic endpoint singularities. Constr. Approx. (2011). Published online: September 28. doi:10.1007/s00365-011-9140-0
Sidi, A., Israeli, M.: Quadrature methods for periodic singular and weakly singular Fredholm integral equations. J. Sci. Comput. 3, 201–231 (1988). Originally appeared as Technion Computer Science Dept. Technical Report No. 384 (1985) and also as ICASE Report No. 86-50 (1986)
Stoer, J., Bulirsch, R.: Introduction to Numerical Analysis, 3rd edn. Springer, New York (2002)
Wu, J., Dai, Z., Zhang, X.: The superconvergence of the composite midpoint rule for the finite-part integral. J. Comput. Appl. Math. 233, 1954–1968 (2010)
Wu, J., Sun, W.: The superconvergence of Newton–Cotes rules for the Hadamard finite-part integral on an interval. Numer. Math. 109, 143–165 (2008)
Acknowledgement
This research was supported in part by the United States–Israel Binational Science Foundation grant no. 2008399.
Author information
Authors and Affiliations
Corresponding author
Additional information
This paper is dedicated to the memory of my dear friend Professor Moshe Israeli.
Appendices
Appendix A: Proof of Theorem 5.1
Proof of Part 1
We start by noting that
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equal_HTML.gif)
Making the variable transformation y=2π(x−t)/T in the integral inside the square brackets, and using the fact that the integrand is T-periodic, we obtain
Next,
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equam_HTML.gif)
and since \(\int^{\pi}_{-\pi}\frac{\sin(my)}{\sin^{2}(\frac{1}{2}y)}\,dy=0\) due to its integrand being odd, it follows that
Thus, it suffices to treat only the case m=0,1,…. Now,
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equan_HTML.gif)
which, upon dividing both sides by \(\sin^{2}(\frac{1}{2}y)\), gives the identity
Integrating both sides of this identity over [−π,π], and invoking (A.2), we obtain
Upon setting m=0, and invoking the fact that L −1=L 1, (A.4) gives
while for m≥1, because \(\int^{\pi}_{-\pi}\cos(my)\,dy=0\), (A.4) gives the recurrence relation
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equao_HTML.gif)
whose general solution is of the form L m =Am+B. We can determine A and B by invoking the values of L 0 and L 1. Now, applying to L 0 the known result
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equap_HTML.gif)
we obtain
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equaq_HTML.gif)
As for L 1, by (A.5), we have
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equar_HTML.gif)
Consequently, A=−4π and B=0. Hence, also by the fact that L −m =L m ,
Substituting this in (A.1), we obtain (5.2).
Proof of Part 2
To prove (5.3) and (5.4), we proceed similarly. First, by (4.5),
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equ96_HTML.gif)
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equ97_HTML.gif)
Next,
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equas_HTML.gif)
By the fact that
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equat_HTML.gif)
and since \(\sum^{n}_{j=1}=\sum^{n}_{j=1}w_{n-j+1}\), we have \(\sum^{n}_{j=1} \frac{\sin (my_{j})}{\sin^{2}(\frac{1}{2}y_{j})}=0\). Consequently,
Thus, it suffices to treat only the case m=0,1,….
Next, for every m≥0, there exist unique integers k,r≥0, r≤n−1, such that m=kn+r. (Thus, k=0 and r=0 for m=0, while k=1 and r=0 for m=n.) By the fact that
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equau_HTML.gif)
we realize that
Thus, we need to concern ourselves only with 0≤m≤n−1. (Note that this also implies that |W m,n |≤max0≤i≤n−1|W i,n | for every m, hence \(\{W_{m,n}\}^{\infty}_{m=-\infty}\) is a bounded sequence for fixed n.)
Setting y=y j in (A.3) and summing over j from 1 to n, and invoking (A.9), we obtain
Upon setting m=0 in (A.11), and invoking the fact that W −1,0=W 1,0, we first obtain
Let us now consider 1≤m≤n−1 in (A.11). For this, we need to compute \(\sum^{n}_{j=1}\cos(my_{j})\). Invoking (A.8), we have
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equav_HTML.gif)
since 0<2mπ/n<2π for 1≤m≤n−1. As a result, \(\sum^{n}_{j=1}\cos(my_{j})=0\), and (A.11) becomes
which enables us to compute W m,n for 2≤m≤n when W 0,n and W 1,n are known. The general solution of (A.13) is of the form W m,n =Am+B for some constants A and B, which we determine by invoking two initial values, namely, those of W 0,n and W 1,n . To compute W 0,n , let
and note that ξ 1,…,ξ n are the zeros of T n (ξ), the nth Chebyshev polynomial of the first kind. (For Chebyshev polynomials and their properties, see Rivlin [16], for example.) By (A.9) and (A.14),
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equaw_HTML.gif)
and since ξ n−j+1=−ξ j , j=1,…,n, we have
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equax_HTML.gif)
Now, since ξ j are the zeros of T n (ξ), we have the identity
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equay_HTML.gif)
Letting ξ=1 in this identity, and making use of the facts that T n (1)=1, \(T_{n}'(1)=n^{2}\), we have
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equaz_HTML.gif)
As for W 1,n , using (A.12), we have
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equba_HTML.gif)
Consequently, A=−2n and B=n 2, and hence
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equbb_HTML.gif)
Recalling also the fact that W −m,n =W m,n , we replace m by |m| everywhere, thus obtaining
Finally, substituting (A.15) in (A.7), we obtain (5.3).
The result in (5.4) is obtained by substituting (A.10), with W r,n =n 2−2rn, in (A.8). The result in (5.5) is obtained by subtracting (5.2) from (5.4).
Proof of Part 3
This follows by invoking Part 2 of the theorem in
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equbc_HTML.gif)
□
Appendix B: Proof of Theorem 10.1
Proof of Part 1
We start by noting that
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equbd_HTML.gif)
Making the variable transformation y=2π(x−t)/T in the integral inside the square brackets, and using the fact that the integrand is T-periodic, we obtain
Next,
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Eqube_HTML.gif)
and since \(\int^{\pi}_{-\pi}{\cot(\frac{1}{2}y)}\,{\cos(my)}\,dy=0\) due to its integrand being odd, it follows that
Thus, it suffices to treat only the case m=1,2,…. For m≥1, the integral in (B.2) representing K m is regular. By making the variable transformation \(z=\frac{1}{2}y\) in this integral, and invoking Gradshteyn and Ryzhik [4, p. 366, formula 3.612(7)], we obtain
Substituting this in (B.1), we obtain (10.2), with sgn(0)=0.
Proof of Part 2
By (9.10), we have first
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equ108_HTML.gif)
Next,
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equbf_HTML.gif)
By the fact that
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equbg_HTML.gif)
and since \(\sum^{n}_{j=1}w_{j}=\sum^{n}_{j=1}w_{n-j+1}\), we have \(\sum^{n}_{j=1} \cot(\frac{1}{2}y_{j})\cos (my_{j})=0\). Consequently,
Thus, it suffices to treat only the case m=1,2,….
Next, for every m≥0, there exist unique integers k,r≥0, r≤n−1, such that m=kn+r. (Thus, k=0 and r=0 for m=0, while k=1 and r=0 for m=n, and, therefore, U ±kn,n =0 for k=0,1,….) By the fact that
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equbh_HTML.gif)
we realize that
Thus, we need to concern ourselves only with 1≤m≤n−1. (Note that this also implies that |U m,n |≤max0≤i≤n−1|U i,n | for every m, hence \(\{U_{m,n}\}^{\infty}_{m=-\infty}\) is a bounded sequence for fixed n.)
Now,
Dividing both sides of this identity by \(\sin(\frac{1}{2}y)\), we obtain
Let us now set y=y j in this identity and sum over j from 1 to n. Invoking (B.5), we obtain
Note that the y j in (B.4) are the same as the y j in (A.8), and hence \(\sum^{n}_{j=1}\cos(my_{j})=0\) for 1≤m≤n−1, as shown in the proof of Theorem 5.1 given in Appendix A. Consequently, (B.9) becomes
Therefore, D n =D n−1=⋯=D 2=D 1=n, so that
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equbi_HTML.gif)
Recalling also (B.5), we have
Substituting (B.11) in (B.4), we obtain (10.3). Combining (B.11) with (B.5) and (B.6), we obtain (10.4) with (10.5). The result in (10.6) is obtained by subtracting (10.2) from (10.4).
Proof of Part 3
This follows by invoking Part 2 of the theorem in
![](http://media.springernature.com/full/springer-static/image/art%3A10.1007%2Fs10915-012-9610-y/MediaObjects/10915_2012_9610_Equbj_HTML.gif)
□
Rights and permissions
About this article
Cite this article
Sidi, A. Compact Numerical Quadrature Formulas for Hypersingular Integrals and Integral Equations. J Sci Comput 54, 145–176 (2013). https://doi.org/10.1007/s10915-012-9610-y
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10915-012-9610-y