Integration by parts involving Inx: http://i1220.photobucket.com/albums/dd441/smileesue/math-1.jpg How would you break this integral into two pieces? would it be where the subtraction is? I just don't how to partially integrate it since I don't know which values to assign as u and dv...
separate in to two partial fractions. 1/lnx^2 + 1/lnx
\[\int\limits_{e}^{e ^{2}}(1-Inx)/(Inx)^2 dx = \int\limits_{e}^{e ^{2}}1/(Inx)^2dx - \int\limits_{e}^{e ^{2}}(1/Inx) dx\] That's all I got to.
I tried u = 1/Inx but I can't go on further..
i guess i don't know too ... seems like worth trying.
ugly http://upload.wikimedia.org/wikipedia/en/math/a/3/9/a39117dd6038c21b11c90a777c1f7699.png
Have you tried u = lnx? Then, x = e^u and du = (1/x)dx. The integral becomes: \[\int\limits_{a}^{b} (e^u/u^2 - e^u/u)du\] If I didn't make any careless mistake
And that should be an indefinite integral, a and b are arbitrary limits.
there might be some other elegant method ... changing limits.
http://www.wolframalpha.com/input/?i=Integral+%28%28e%5Ex%29%2F%28x%5E2%29+-+%28e%5Ex%29%2Fx%29 It works :-) haha
In this case, x = lnx, so it becomes -x/lnx
So would you have to partially integrate \[\int\limits_{}^{}(e ^{u}/u ^{2})du - \int\limits_{}^{}(e ^{u}/u)du\] for the respective terms? So for the first term, u=e^u and dv=(1/u^2)du? I am slightly confused
Indeed. But I am unsure that this is the right way. The integrals don't look simplified. I was just amazed that what I did work, haha.
this is getting crazy ... looks something like gamma function
http://en.wikipedia.org/wiki/Gamma_function if the limits were 0 to inf then we'll be able to get the values easily.
Gaah, was typing the reply and the site crashed. Anyway, I think that both integrals are easy. You can do it by definition, i.e., integral of (e^u)/u = exponential integration of u + a constant.
Or maybe not. Yeah, you would have to use the definition if you went through my path. e^u/u is not an elementary function... There has to be another way.
This was what was written below the question. I'm not sure if that'd be a big help though..
What bothers me is that with that substitution (u = 1/lnx), it seems that there is a du factor missing. I am committing a silly mistake somewhere, but my integral keeps getting reduced to u + du times some factor.
Oh I see. I'll put this problem aside for now and try again. Thanks bmp and experimentX for your help though :)
Got it!
Damn, lost my answer.
Anyway, simplify the expression by dividing it by (lnx)^2. Multiply it out by (x/x), then apply the substitution hinted above. We get two integrals, one of them: integral of udx and the other integral of xdu. Then, do the xdu integral by parts, we will have: ux - integral udx, that will cancel out the other integral. So we are left with u*x, that is, x/lnx, QED
-x/lnx*
by dividing it by (lnx)^2, do you mean the original expression? Or do you mean multiplying? Because then wouldn't the denominator become (Inx)^4?
No, I mean just simplifying it to obtain: \[\int\limits_{e}^{e^2} (1/lnx^2 + 1/lnx)dx\]
Hope the font is not too small
Oh I see. Thank you! :)
Join our real-time social learning platform and learn together with your friends!