Determine the values of "a" and "b" for which the following limit is true: limit as x approaches infinity of ((ax+1)/(bx-1))^x = 4
\[\left(\frac{ax+1}{bx-1}\right)^x=4\]?
yes
find the limit in \(a\) and \(b\)
how do you do that?
oh this is a pain start with by taking the log to get \[x\ln\left(\frac{ax+1}{bx-1}\right)\]
ok, and the other side would equal ln(4), right?
wait, no it wouldn't
yeah now i am confused
we need to compute that limit in \(a\) and \(b\) and it is not of the form to use l'hopital because we don't know if \(a>b\) or if \(a<b\) or what
so now i am thinking this method may not be right
why would you have to know whether a >b or not to use l'hopital's rule?
wouldn't l'hopital's rule work for any values of a and b?
can anyone else help? satellite73 went offline....
I dont know any of this the most i can do is wolfram alpha
\(a=b=-\ln 4\) seems to work
how did you get that?
that doesn't work, try this : \(a=b = \dfrac{1}{\ln 4}\)
1/ln4 gives me 16 http://www.wolframalpha.com/input/?i=lim%28x%5Cto%5Cinfty%29+%28%281%2Fln%284%29*x%2B1%29%2F%281%2Fln%284%29*x-1%29%29%5Ex
so that doesn't work either.. wait il explain once i figure it out
ok
it is \(a=b=\dfrac{1}{\ln 2}\) ! http://www.wolframalpha.com/input/?i=lim%28x%5Cto%5Cinfty%29+%28%281%2Fln%282%29*x%2B1%29%2F%281%2Fln%282%29*x-1%29%29%5Ex
ok, is there any way using calculus to get that answer?
remember how to do long division ?
yes
\[\dfrac{ax+1}{bx-1}=?\]
would that be \[\frac{ a }{ b } + \frac{ 1-\frac{ a }{ b } }{ bx-1 }\] ?
try again
quotient is correct but remainder doesn't look correct..
bx - 1 | ax + 1 | a/b ax - a/b ------------------ 1 + a/b (ax+1)/(bx-1) = a/b + (1 + a/b)/(bx-1) yes ?
true, I messed up while subtracting
\[ \left(\frac{ax+1}{bx-1}\right)^x\\~\\ =\left(\frac{a}{b}+\frac{1+a/b}{bx-1}\right)^x\\~\\ \]
before proceeding further, let me ask you a question
as you make \(n\) large, what number does the below expression approach ? \[(0.9)^n\]
zero
good. how about below expression : \[(1.1)^n\] what number does that approach as you make \(n\) large ?
I presume that should go to infinity, right?
Excellent! as you can see, \((1\pm x)^n\) is either \(0\) or \(\infty\) for nonzero \(x\) as \(n\to\infty\)
Keeping that in mind, lets look at our problem
\[ \left(\frac{ax+1}{bx-1}\right)^x\\~\\ =\left(\frac{a}{b}+\frac{1+a/b}{bx-1}\right)^x\\~\\ \]
As \(x\to \infty\), is there any hope for above expression to be \(4\) if \(\dfrac{a}{b}\) is greater than \(1\) ?
no, because you would be adding a number to 4, so if a/b is greater than 1, than the result will always be greater than 4
In other words, for \(\dfrac{a}{b}\gt 1\), the limit "diverges"
what happens if \(\dfrac{a}{b}\lt 1\) ?
I'm guessing that it would approach 4?
My guess is that it would approach 0
as you can see, i want to argue that \(\dfrac{a}{b}\) has to equal to \(1\)
yes, I understand, if you make b arbitrarily big, then the limit of the whole thing would approach zero, that makes sense
so if the limit does not go to 4 when a/b is greater than 1 or less than 1, that means a/b must equal 1. Which means that a must equal b.
Yes, plug that in \(a=b\)
\[ \left(\frac{ax+1}{bx-1}\right)^x\\~\\ =\left(\frac{a}{b}+\frac{1+a/b}{bx-1}\right)^x\\~\\ \] \(a=b\), so : \[ \left(1+\frac{1+1}{bx-1}\right)^x \]
ok, so that would simplify to \[(1+\frac{ 2 }{ bx-1 })^x\]
would I have to take a common denominator next?
actually, taking a common denominator will bring me back to the original equation
we may use this : \[\lim\limits_{x\to\infty} (1+\dfrac{t}{x})^x = e^t\]
but it seems we need to massage a bit before using that
ok, and how exactly would we denominator to equal to exponent?
\[ \left(\frac{ax+1}{bx-1}\right)^x\\~\\ =\left(\frac{a}{b}+\frac{1+a/b}{bx-1}\right)^x\\~\\ \] \(a=b\), so : \[ =\left(1+\frac{1+1}{bx-1}\right)^x\\~\\ =\left(1+\frac{2/b}{x-1/b}\right)^x\\~\\ \]
fine with above ?
yeah, you just divided both top and bottom with b
way I went about it: \[\lim_{x \rightarrow \infty}(\frac{ax+1}{bx-1})^x=4 \\ \text{ Take }\ln( ) \text{ of both sides: } \\ \ln[\lim_{x \rightarrow \infty}(\frac{ax+1}{bx-1})^x]=\ln(4 ) \\ \text{ you can bring the limit outside } \log \\ \lim_{x \rightarrow \infty} \ln[(\frac{ax+1}{bx-1})^x]=\ln(4) \\ \text{ use power rule for } \log \\ \lim_{x \rightarrow \infty} x \ln (\frac{ax+1}{bx-1})=\ln(4) \\ \lim_{x \rightarrow \infty} \frac{\ln(\frac{ax+1}{bx-1})}{\frac{1}{x}} =\ln(4) \\ \text{ In order for the limit \to actually exist we need } \frac{ax+1}{bx-1} \rightarrow 1 \text{ as } x \rightarrow \infty \\ \text{ so we can have } \frac{0}{0} \\ \text{ and use l'hospital rule } \\ \text{ so this does give you } \frac{a}{b}=1 \implies a=b \\ \text{ So you have } \\ \lim_{x \rightarrow \infty}\frac{\ln(\frac{ax+1}{ax-1})}{\frac{1}{x}}=\ln(4)\] see if you can finish apply l'hospital
\[ \left(\frac{ax+1}{bx-1}\right)^x\\~\\ =\left(\frac{a}{b}+\frac{1+a/b}{bx-1}\right)^x\\~\\ \] \(a=b\), so : \[ =\left(1+\frac{1+1}{bx-1}\right)^x\\~\\ =\left(1+\frac{2/b}{x-1/b}\right)^x\\~\\ =\left(1+\frac{2/b}{x-1/b}\right)^{x-1/b}*\left(1+\frac{2/b}{x-1/b}\right)^{1/b} \\~\\ \] taking limit gives you \[ = e^{2/b}*1 \]
Thank you freckles, I'll try that as soon as I'm done with ganeshie8's way
Ok, so I'm a bit confused as to how the second term equals 1
would it be because as x is approaching infinity (2/b)/(x-1/b) is approaching zero, so 1^(1/b) would just be 1?
Yes. notice that 1/b is a constant with respect to x : \[\lim\limits_{x\to\infty}\left(1+\frac{2/b}{x-1/b}\right)^{1/b} = \left(\lim\limits_{x\to\infty}1+\frac{2/b}{x-1/b}\right)^{1/b}\\~\\~\\ = (1+0)^{1/b}\\~\\ =1 \]
ok, that makes sense
recall, for continuous functions, we have : \[\lim f(g(x)) = f(\lim g(x))\]
\[\lim [f(x)]^n = [\lim f(x)]^n\]
ok so lim[e^(2/b)] equals [lim e]^(2/b) ?
No, we're done. the limit evaluates to \(e^{2/b}\), simply set this equal to \(4\) and solve \(b\)
i got b = 2/ ln(4)
Looks good !
so how does that simplify to 1/ln(2) ?
use below log property : \[\ln(x^a) = a*\ln(x)\]
ok, so it looks like it should be 1/ln(4^(1/2)), but I'm not sure whether that would be the correct way to do it
Easy... ln(4) = ln(2^2) = 2ln(2)
because 4 = 2^2
OHHHH, lol, that was literaly the stupidest question I have asked.....
Thank you so much for your help and your time!
haha thats okay :) my stupidiest question was way stupidier than ur stupidiest q
do scroll up and see freckles method using Lhopital rule...
yes, I'm actually looking at his method right now
Join our real-time social learning platform and learn together with your friends!