From: Robert Israel on

> > On Mar 17, 8:40 am, "Robert H. Lewis"
> > <rle...(a)fordham.edu> wrote:
> > > > integral (xe^(2x))/((2x+1)^2)  BY PARTS
> > > Hate to break into all the cute jokes, but the OP's problem
> > > probably has a typo, as it cannot be expressed with elementary
> > > functions.
> >
> > I disagree. I believe that it can be so expressed.
> > Since the OP first wrote the problem a week ago and most teachers
> > don't give more than a week to complete the assignment, I feel
> > that I can post this without spoiling the homework.
> >
> > Let u = xe^(2x) and dv = dx/((2x+1)^2). Then v =
> > -1/(2(2x+1) and
> > du = (x(e^(2x))2 + e^(2x))dx = e^(2x)(2x+1)dx. Hey,
> > that's convenient -- du has a factor of (2x+1) in the
> > numerator and v has a factor of (2x+1) in the denominator!
> >
> > So integrating by parts, we have:...
> > = e^(2x)/((4(2x+1)) + C
> >
> > Thus there does exist an elementary antiderivative.
>
> You're right. A week ago when I posted that I just lazily fed the integral
> to Maple12. It responded:
>
> -(1/2)*ln(xe)*xe^(2*x)/(2*ln(xe)*x+ln(xe))-(1/2)*ln(xe)*Ei(1,
> -2*ln(xe)*x-ln(xe))/xe
>
> with the Ei (exponential integral) function. Knowing a bit about Maple
> (not too much) I applied simplify to the above and got
>
> -(1/2)*(xe^(2*x+1)+2*ln(xe)*Ei(1, -ln(xe)*(2*x+1))*x+ln(xe)*Ei(1,
> -ln(xe)*(2*x+1)))/((2*x+1)*xe)
>
> at which point I trusted Maple and posted the above. Shame on me for
> trusting Maple!

A little knowledge is a dangerous thing. It seems you fed Maple

int(xe^(2*x)/(2*x+1)^2, x);

and Maple interprets xe as a new variable. You should
have tried

int(x*exp(2*x)/(2*x+1)^2, x);

which would return the elementary antiderivative

1/4 * exp(2*x)/(2*x+1)
--
Robert Israel israel(a)math.MyUniversitysInitials.ca
Department of Mathematics http://www.math.ubc.ca/~israel
University of British Columbia Vancouver, BC, Canada