I don't understand WHY the integral of 1/(xlnx) is ln(lnx) + C
Maths and statistics discussion, revision, exam and homework help.
| Announcements | Posted on | |
|---|---|---|
| Important: please read these guidelines before posting about exams on The Student Room | 28-04-2013 | |
-
I don't understand WHY the integral of 1/(xlnx) is ln(lnx) + C
I need to integrate 1/(xlnx)
I'm doing so by integration by parts
1/(xlnx) = 1/x . 1/ lnx
u = 1/lnx dv/dx = 1/x
I integrate and use uv - int(du/dx . v) to get 1 which if definitely wrong
I've looked at how people do it over the net, and they always take u as lnx?
I cannot for the life of me work out why?
When we integrate by parts we: int(u.dv/dx)
why isn't it working for me?
My u value is 1/lnx and my dv/dx is 1/x Becuase when multiply these two together we get 1/(xlnx)Last edited by Sifr; 18-04-2012 at 00:18. -
Re: I don't understand WHY the integral of 1/(xlnx) is ln(lnx) + CYes I know that's why I chose u= 1/ lnx ...(Original post by claret_n_blue)
You can't integrate ln(x), you can only differentiate it to 1/x so therefore you need to set u = ln(x) so you can differentiate it.
Just don't get why everyone gets u = lnx :/ -
Re: I don't understand WHY the integral of 1/(xlnx) is ln(lnx) + CBecause its differential is 1/x, and you already have 1/x in the integrand (for substitution that is).(Original post by Sifr)
Yes I know that's why I chose u= 1/ lnx ...
Just don't get why everyone gets u = lnx :/Last edited by f1mad; 18-04-2012 at 00:41. -
Re: I don't understand WHY the integral of 1/(xlnx) is ln(lnx) + COhh. Well if you differentiate it you get 1/x and this can be substituted for the 1/x. You don't need to use by parts, you use substitution:(Original post by Sifr)
Yes I know that's why I chose u= 1/ lnx ...
Just don't get why everyone gets u = lnx :/

Let u = ln(x)
du = (1/x) dx
-
Re: I don't understand WHY the integral of 1/(xlnx) is ln(lnx) + CNot necessarily, that's why we have various integration methods.(Original post by Sifr)
I can use another method, but it's the principle, every method should work? I'm really pedantic and it bugs me
Consider trying to integrate ln(x).
This isn't possible via substitution, only by parts. -
Re: I don't understand WHY the integral of 1/(xlnx) is ln(lnx) + CSo if you're saying(Original post by Sifr)
I can use another method, but it's the principle, every method should work? I'm really pedantic and it bugs me
then we have
and 

Since the second term of the right hand side is the same as the left hand side, we end up with the wonderous result of -c = 1. This is pretty useless, so since it doesn't get us anywhere we say "ok let's try another method". Use the easier way I suggested earlier and then go to bed -
Re: I don't understand WHY the integral of 1/(xlnx) is ln(lnx) + C
IBP (Integration by Parts) is not necessary in this problem. I encountered this integral in a homework site a long time ago.
Let's use substitution.
is our integral. I see I have a function (the natural logarithm) and its derivative
sitting in the integrand. I'll make the following substitution:

Our integrand now becomes
. This we can do easily, and you should grab
.
-
Re: I don't understand WHY the integral of 1/(xlnx) is ln(lnx) + CThere's no reason to think that integration by parts should work just because you have a product in your integral. Often the best way to deal with a product is integration by substitution, particularly when one of the terms is a function of a function and the other term is the derivative of the 'inside function'.(Original post by Sifr)
I can use another method, but it's the principle, every method should work? I'm really pedantic and it bugs me
Think about it: integration by parts states

Compare this with the product rule (rearranged):

And integration by substitution states

Compare this with the chain rule:

As you can see, integration by parts reverses the product rule (which you can think of as 'differentiation by parts') and integration by substitution reverses the chain rule (which you can think of as 'differentiation by substitution').
Both LHS integrals involve products, but what you get out is a very different type of function. Only in special rare cases are both methods appropriate for calculating the antiderivative.Last edited by nuodai; 18-04-2012 at 08:44. -
Re: I don't understand WHY the integral of 1/(xlnx) is ln(lnx) + C
Not sure if you still need an answer, but I got 1/2 x ln(x)2 + c
used integrating by parts.
u=lnx u'=1/x
v'=1/x v=lnx
integral(1/xlnx)dx = (lnx)2 - integral(1/xlnx)dx
2integral(1/xlnx)dx = (lnx)2
integral (1/xlnx)dx = 1/2 x (lnx)2 + c