firstname.lastname@example.org (John Wojdylo) writes: > >The following definite integral ... >Note that a closed form expression for the other integral that arises in >this problem, ... >is given by integral 3.462 (1.) of Gradshteyn and Ryzhik (1980).
One suggestion, which worked for a nasty one we encountered, is to look at the reference cited in the Russian tables and look up that book, and perhaps its predecessor. A lot of those integrals come from huge collections assembled by various groups and G+R select some of those (as is immediately obvious from the referenced eqn numbers). In turn, the Amsterdam collection I looked up (which had what I needed) mentioned in the preface that this new edition included N new integrals that replaced N others that were not as interesting to them any more!
Many of these integrals are themselves series or other integrals, but in some cases the new integral is really easy to do numerically compared to the original. In other cases, and yours might be one, recasting it so it can be done as a gaussian quadrature might be faster to evaluate than the expression for the integral.
-- James A. Carr <email@example.com> | "Whatever." http://www.scri.fsu.edu/~jac/ | Supercomputer Computations Res. Inst. | George Herbert Walker Bush Florida State, Tallahassee FL 32306 |