A finite-interval method for computing principal-value integrals that is simple, accurate, and fast is described. Rather than using a numerical limit process near the singularity in the integrand, this method uses a power series expansion of the regular part of the integrand around the singularity to approximate the principal value over a finite interval. Outside this interval, where the integrand is well-behaved, conventional methods of integration can be used. © 1998 American Institute of Physics.
REFERENCES
1.
2.
G. C.
John
, J. E.
Hasbun
, and V. A.
Singh
, Comput. Phys.
11
, 293
(1997
).3.
J. R.
Macdonald
and V. I.
Piterbarg
, J. Electroanal. Chem.
428
, 1
(1997
);4.
A. Curnier, Computational Methods in Solid Mechanics (Kluwer, Amsterdam, 1994).
5.
N. I.
Ioakimidis
, Comput. Methods Appl. Mech. Eng.
46
, 1
(1984
); >and references therein.6.
7.
8.
D. B. Hunter and H. V. Smith, BIT 29, 512 (1989).
9.
W. J. Thompson, Atlas for Computing Mathematical Functions (Wiley, New York, 1997).
10.
P. J. Davis and P. Rabinowitz, Methods of Numerical Integration, 2nd ed. (Academic, New York, 1984), Sec. 1.6.1;
This content is only available via PDF.
© 1998 American Institute of Physics.
1998
American Institute of Physics