Calculus. Finding a power series representation for an integral with natural log
\[f(x)= \int\limits \frac{ \ln(1-t) }{ t }dt\]
\[f ' (x) = \frac{ 1 }{ 1-t }\]=\[\sum_{n=0}^{\infty}t^n\]=\[1+t+t^2...\]
i know thats not the antiderivative... im just feeling like the derivative is where i should start?
Eqautions huh :)
Ok, So lets think this through
Do i take the antiderivative of the sum i found with the derivative?
or does the quotient rule come into play here too somehow?
how did you get that f'?
also is it f(x)=? or f(t)=?
i was thinking ln (1-t) derivative is 1/1-t and then the deriv of t is just 1
they gave me both. f(x)= integral ln(1-t) / t
power series
but you have the integral there you can't differentiate the integral thing like that hmmm... that seems weird because f(x) is a function of x and you have int ln(1-t)/t is a function of t
here is the exact question "Evaluate the indefinite integral as a power series"
\[f(t)=\int\limits t^2 dt \text{ \implies } f'(t)=t^2 \\ \text{ you can take the long way around on this one } \\ f(t)=\frac{t^3}{3}+C \text{ \implies still } f'(t)=t^2 \]
so \[f'(t)=\frac{\ln(1-t)}{t}\] in your case
im not sure why they have f(x) and then use t... that is how they have it...
i can show you the final answer.... i just dont know how they get it
\[C- \sum_{n=1}^{\infty}\frac{ t^n }{ n^2 }\] this is the answer
let's look at ln(1-t)
\[g(t)=\ln(1-t) \\ g'(t)=\frac{-1}{1-t}\]
do you know how to find a power series representation for g'(t)
Does the neg one come from -du=dx?
I think they want you to represent the result of the integral as a power series. one way is to write down the taylor series for ln(1-t) see https://en.wikipedia.org/wiki/Taylor_series#List_of_Maclaurin_series_of_some_common_functions divide by t to get a new series, and integrate term by term to get the answer
chain rule derivative of inside/inside when we have ln( )
yes phi thats what it is
they havent taught us taylor yet..
but you are allow to know: \[g'(t)=- \sum_{n=0}^{\infty} t^n\] right? you can integrate term by term to find g(t)=ln(1-t)
freckles is it \[\sum_{n=0}^{\infty}(x)^n\]
idk if i should use x or t... they got me messed up using both
let's just use t
it doesn't matter you choose your favorite letter and we will use that one
t is fine
ok well \[\sum_{n=0}^{\infty}t^n=\frac{1}{1-t} \\ \text{ so } - \sum_{n=0}^{\infty} =\frac{-1}{1-t}\]
so so far we have ...\[f(t)=\frac{ \ln(1-t) }{ t }, f' (t)= \frac{ -1 }{ 1-t }, \sum_{n=0}^{\infty}t^n\]
\[g(t)=\ln(1-t) \rightarrow g'(t)=\frac{(1-t)'}{1-t}=\frac{0-1}{1-t}=\frac{-1}{1-t} \\ g'(t)=\frac{-1}{1-t}=- \sum_{n=0}^\infty t^n\]
ok no that isn't f' above
to find g(t) integrate g'(t) term by term
and i integrating the derivative of the function f(t)?
do you know how to integrate t^n ?
w.r.t t?
am i
that would be t^n / n
right? integrating t^n= t^n/ n ?
t^(n+1)/(n+1) by power rule
oh yea, add one to the exponent and then drop the new exponent to denom...
\[\text{ so far we have this: } f(t)=\int\limits \frac{\ln(1-t)}{t} dt \\ f'(t)=\frac{\ln(1-t)}{t} \\ \text{ where we wanted \to play with } \ln(1-t) \\ \text{ so I made a new function name } \text{ Let } g(t)=\ln(1-t) \\ \text{ so } g'(t)=\frac{-1}{1-t}=- \sum_{n=0}^\infty t^n \\ \text{ now integrating term by term we can find g from g' } \\ g(t)=- \sum_{n=0}^\infty \frac{t^{n+1}}{n+1} =\ln(1-t) \\ \text{ now } f'(t)=\frac{g(t)}{t}=- \frac{1}{t} \sum _{n=0}^\infty \frac{t^{n+1}}{n+1}\] this is what we have so far
do you have questions with this so far
we still have to do one more thing
i am reading everything, trying to find some sense... the last post you made is making a lot more sense, where you summed everything up..
\[f'(t)=-\sum_{n=0}^\infty \frac{t^{n+1}}{n+1} \cdot t^{-1} =-\sum_{n=0}^\infty \frac{t^n}{n+1}\] now the last step I was talking about is find f from f'
i am trying to understand how n+1 becomes n^2
well you still need to find f given f'
can you do that last step and I will show you the result we get is equivalent to what you have way above
hint: to find f given f' integrate and put your +C I guess we should have put C on that one thing when need from g' to g :p
\[\int\limits \sum_{n=0}^{\infty}\frac{ t^n }{ n+1 }= \sum_{n=0}^{\infty}\frac{ t^{n+1} }{ n+1(n+1) }\]?
yeah sorta you need paranthesis around the other (n+1)
-C
\[-\int\limits \sum_{n=0}^{\infty}\frac{ t^{n+1} }{ (n+1)^2 }= C-\sum_{n=1}^{\infty}\frac{ t^n }{ n^2 }\] because we drop the +1 on the n's when we change the index?
let me give you another summary so far: \[\text{ so far we have this: } f(t)=\int\limits \frac{\ln(1-t)}{t} dt \\ f'(t)=\frac{\ln(1-t)}{t} \\ \text{ where we wanted to play with } \ln(1-t) \\ \text{ so I made a new function name } \text{ Let } g(t)=\ln(1-t) \\ \text{ so } g'(t)=\frac{-1}{1-t}=- \sum_{n=0}^\infty t^n \\ \text{ now integrating term by term we can find g from g' } \\ g(t)=- \sum_{n=0}^\infty \frac{t^{n+1}}{n+1}+C =\ln(1-t) \\ g(0)=0+C=\ln(1-0) \\ \text{ so this integration constant is C=0}\\ \\ \text{ now } f'(t)=\frac{g(t)}{t}=- \frac{1}{t} \sum _{n=0}^\infty \frac{t^{n+1}}{n+1} \] \[f'(t)=\frac{g(t)}{t}=- \sum_{n=0}^\infty \frac{t^n}{n+1} \\ f(t)=- \sum_{n=0}^\infty \frac{t^{n+1}}{(n+1)^2}+C \\ \text{ or } f(t)=C- \sum_{n=0}^\infty \frac{t^{n+1}}{(n+1)^2}\] now our series starts at n=0 and theirs starts at n=1 we can play with what we have to get their exact series thing now we have n=0 on bottom add on one both sides n+1=1 <--we have this ... so this means if we replace old (n+1)'s with (n)'s and start the series at n=1 instead we will get their exact series looking thing though you could leave it like this and it is still correct well you might still need to find the constant of integration so replace (n+1)'s with n's and start series at n=1 we have: \[f(t)=C- \sum_{n=1}^\infty \frac{t^n}{n^2}\]
wow... this one is really hard!
it doesn't matter if you say \[f(t)=C-\sum_{n=0}^\infty \frac{t^{n+1}}{(n+1)^2} \text{ or if you say } f(t)=C-\sum_{n=1}^\infty \frac{t^n}{n^2}\] it is the exact same thing
thank you for the summary! i will need to look over this awhile before i feel i got the hang of this one! i am so grateful tho, thank you!
you can even compare the first terms if you want if you don't see it
so, i take the derivative 1st, and THEN i integrate the power series representation we found by taking the derivative?
here is another example if rewriting the same series with a different starting number: \[\sum_{i=2}^{5} 5^{i+3} \\ \text{ now we have } \\ i=2 \\ \text{ add say } 3 \text{ on both sides } i+3=2+3 \\ \text{ so } i+3 =5 \\ \text{ so we can write } \\ \sum_{i=2}^{5}5^{i+3} \text{ as } \sum_{i=5}^{5+3} 5^i\]
like when dealing with infinities we don't have to worry about changing infinity to infinity+1 because it is still infinity but if you have a ending number that is finite then you would have to change it like start at n=2 and ending at n=5 means if you change it to n+3=2+3=5 and ending at n+3=5+3=8 then you start the series at n=5 and end it at n=8 and change all the (n+3)'s to n's
i don't think we found the derivative to integrate? what do you mean
http://tutorial.math.lamar.edu/Classes/CalcII/PowerSeriesandFunctions.aspx by the way paul is one of my favorite things to look at for notes if you want to look too here is a link --- I took derivative first of the integral function and we got ln(1-t)/t I knew the derivative of ln(1-t) was -1/(1-t) which had a famous power series representation so I played with ln(1-t) first then I came back and divided by t so we could have power series representation for ln(1-t)/t then seen we wanted int (ln(1-t)/t) I integrate the power series of ln(1-t)/t
int means integral
\[-\frac{ 1 }{ t }\int\limits \sum_{n=0}^{\infty}\frac{ t^{n+1} }{ n+1}= -\int\limits \sum_{n=0}^{\infty}\frac{ t^n }{ n+1 }=C-\sum_{n=0}^{\infty}\frac{ t^{n+1} }{ (n+1)^2 }\] This is the part that messed me up the most that you really cleared up for me
i think i meant to ask the question like this--> " If i am to sum up how we performed this in words, I would say we took the derivative of f(t) first to find the power series representation of the function. Then we integrated the power series we found
yea
ok... i think im finally understanding this. and thanks for the pauls note link!
im def gonna re study this a bit... but you helped a whole bunch!
well yeah that -1/t is suppose to be inside the integral sign but coolness :p I'm glad it makes more sense to you
Join our real-time social learning platform and learn together with your friends!