The evaluation of trigonometric integrals avoiding spurious discontinuities

Abstract

The tan(x/2) substitution, also called the Weierstrass substitution, is one method currently used by computer-algebra systems for the evaluation of trigonometric integrals. The method needs to be improved, because the expressions obtained using it sometimes contain discontinuities, which unnecessarily limit the domains over which the expressions are correct. We show that the discontinuities are spurious in the following sense: Given an integrand and an expression for its antiderivative that was obtained by the Weierstrass substition, a better expression can be found that is continuous on wider intervals than the first expression and yet is still an antiderivative of the integrand. The origin of the discontinuities is identified, and an algorithm is presented for automatically finding the improved type of antiderivative. The new algorithm also enlarges the set of functions that can be used in the substitution. The algorithm works by first evaluating the given integral using the Weierstrass substitution in the usual way and then removing any spurious discontinuities present in the antiderivative.

Mechanism of drag generation by surface corrugation

“…whereà ¼ B cosð/Þ=2 À A=2,B ¼ ÀB=2 sinð/Þ, c 0 ¼ 12=35M 2 ð1 þÃÞ À2 À a À1 p À1 ð0Þ correspond to pressure normalization p ¼ 0 at n ¼ 0, bc is the floor function which is added to remove spurious discontinuities associated with the inverse tangent function, 28 …”

Challenges of Symbolic Computation: My Favorite Open Problems

“…Progress has also been made on integration, particularly with the papers of Jeffrey (1993Jeffrey ( , 1994Jeffrey ( , 1997 and Jeffrey and Rich (1994). The main contribution of these papers is that it is better to return an integral that is continuous on a domain of maximum extent, rather than trying to fix up spurious singularities and branch cuts later.…”

Effects of wall vibrations on channel flows

“…. stands for the floor function, which has been added to remove spurious discontinuities associated with the arctan function (Jeffrey & Rich 1994), and pressure is normalized with condition p−1 ( 0) = 0. To extract the mean pressure gradient, observe that dp −1 /dξ in (4.6a) is an even function of ξ .…”

“…The resulting sequence of problems begins with the leading-order system being of the form Its solution can be readily determined, i.e. Integration of (A7 a ) results in an expression for pressure in the form of where In the above, symbol stands for the floor function, which has been added to remove spurious discontinuities associated with the arctan function (Jeffrey & Rich 1994), and pressure is normalized with condition . To extract the mean pressure gradient, observe that in (4.6 a ) is an even function of .…”