# For what p does this converge?

1. Feb 16, 2008

### fk378

1. The problem statement, all variables and given/known data
Given the limit (B-->inf) [(lnB)^(-p+1)] / (-p+1) - [(ln1)^(-p+1)] / (-p+1)

For what p does this converge?

3. The attempt at a solution

For the left side of the minus sign,
if 1-p<0 --> 0
if 1-p>0 --> inf

but for the ln(1) on the right side of the minus sign, ln(1)=0, so it would be
if 1-p<0 then you would get a zero in the denominator (since ln(1) would be raised to a negative power) and therefore ---> inf
if 1-p>0 then you would have zero as the numerator, but be left with infinity as the other value and hence ---> inf

The book says that for p>1 then it will converge. But how can that be, if you'll end up with a negative exponent for the 0??

Last edited: Feb 16, 2008
2. Feb 16, 2008

### HallsofIvy

Staff Emeritus
Is that really "ln(1)" in the last term? ln(1)= 0 of course. Also parentheses would help.
Finally, you should not use "B" and "b" to mean the same thing but since there is no "b" in function I am going to assume you are. I am also going to assume that is [ln(b^(-p+1))]/[-p+1] - [(ln1)^(-p+1)] / [-p+1] which, since ln(1)= 0 is the same as ln(B-p+1/(-p+1)= (-p+1)ln(B)/(-p+1)= ln(b). So you have [itex]\lim{b\rightarrow \infty} ln(b) which does not converge for any p.

If you meant something else please clarify.

That converges to ln(B

3. Feb 16, 2008

### fk378

Please do not assume that when I stated (lnB)^(-p+1) I meant ln(B^-p+1) because I did not. The exponent is for the entire natural log expression. Therefore your statement regarding the answer to be just lnB is invalid. Thank you though.

Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?