legit answer or am i way off base: Show that integral from 1 to infinity: (e^-x) / x dx converges. substitution gets you 0 / infinity L'hospital get you -e^-x/1 or converges to zero???
\[\int\limits_{1}^{\infty}((e^-x)/x) dx\]
Lhopital is for derivatives ...
does the function -1/x^2 converge?
yes by p test 2>1 and p>1 converges
so lets integrate it; and we get the function 1/x does this converge?
the convergence of a derivative is not a gaurentee that the integral will converge ...
now p = 1 and by p test p=1 diverges...sorry slow wrote p test down wrong in class and was doing it backwards for three days
so looks like i should integrate and see what that gets me...
if we integrate this by parts; it might produce a series .....
this will be slow if you want to help elsewhere...
lol, its 130am .... im not really wanting to help anywhere :)
so that looks like (x)(-e^(-x)) - e^(-x)
\[\frac{1}{x}e^{-x}=\frac{1}{x}\sum_{n=0}^{inf}\frac{(-1)^{n+1}}{n!}x^n\] \[\frac{1}{x}e^{-x}=\sum_{n=0}^{inf}\frac{(-1)^{n+1}}{n!}x^{n-1}\]
hmm, -1^n that is ... i seem to have generated a spurious negative :)
if I got that then i would be thinking ratio test? and spurious sign changes is my signature move
this is still not integrated; i know when we derive a summation we add to n each time; so i assume when we integrate we get a n-1 to the summantion since n needs to start at 0; lets pull out the first iteration before we integrate
maybe that is what my key suggests...it just integrates the e^-x and the 1/x disappears in an unexplained poof of smoke
\[\int \frac{1}{x}+\sum_{n=1}^{inf}\frac{(-1)^{n}}{n!}x^{n-1}\ dx\]
\[ln(x) + \sum_{n=0}^{inf}\frac{(-1)^n}{n!(n)}x^n+C\]
uh oh, we got a zero in the denominator with that one :)
it will be ok if we start with 1
yeah, but that ln(x) on the outside doesnt converge ...
so i have a feeling this thing diverges. we could chk it with the wolf
i think they disappeared the 1/x with the idea that if integrate e^-x and converge the multiply 1/x would just make it smaller?
its possible i could have tried to modify it in another manner; i ahvent done enough of these to be too confident
All this converge diverge stuff seems like cheating ....ignore that part just approximate...im just not sure when to do it
well at least in know not to check the limits with l'hospital....seemed like such a good gambit too
:) the summation we got converges by a ratio test .... if we turn the ln(x) into a summation (power series) and add the 2 together I wonder if it pans out
im saving the series expansion for tomorrow...(part three of the final) I'll work the ratio test on that sum/series for practice thanks for the helps
\[ \int_1^c \frac {e^{-x}}x dx \\ u=\frac {1}x, \, dv =e^{-x} dx\\ du =-\frac 1 {x^2},\, v =- e^{-x}\\ \int_1^c \frac {e^{-x}}x dx = \left [- \frac {e^{-x}}x \right]_1^c - \int_1^c \frac {e^{-x}}{x^2}dx=\\ - \frac {e^{-c}}c+ \frac {e^{-1}}1 -\int_1^c \frac {e^{-x}}{x^2}dx \]
Notice that \[ \int_1^c \frac {e^{-x}} {x^2} dx \le \int_1^c \frac {dx} {x^2} = 1 - \frac 1 c\\ 0\le \int_1^\infty \frac {e^{-x}} {x^2} dx \le 1 \] Hence \[ \int_1^\infty \frac {e^{-x}} {x^2} dx \] is convergent. By this and my previous post the original integral is convergent.
so that uses squeeze to prove convergence?
Yes
If c goes to infinity \[ \int_1^\infty \frac {e^{-x}}x dx= \\ -0+ \frac {e^{-1}}1 - k \] Where \[ K = \int_1^\infty \frac {e^{-x}}{x^2}dx \]
K=k
oK
Did you understand it?
i dont think we did anything like that in class - you are subtracting the other integral from the original?
oh ok that is from the integration by parts...now that makes sense excellent thanks
yw
Join our real-time social learning platform and learn together with your friends!