find \lim_{t\rightarrow 0}(ax^t+by^t)^{\frac{1}{t}} when a+b=1
\[\large \lim_{t\rightarrow 0}(ax^t+by^t)^{\frac{1}{t}}\] @UnkleRhaukus ?
When a+b = 1... how did you go about using that substitution?
Series expansion?
@Kainui
1,1/2,inf
Looks like you need to use logarithms.
Yeah
\[\large \lim_{t\rightarrow 0}(ax^t+by^t)^{\frac{1}{t}} = e^{\lim \limits_{t\rightarrow 0} \frac{\ln(ax^t+by^t) }{t}}\] L'hopital next
you should end up with \(x^ay^b\) give it a try i dont want to spoil your fun by posting the solution :P
This was quite a fun limit, and usually I don't really care for limits, but this was quite a surprising result.
the solution without L'hopitals is quite nice too il post it later if somebody is interested..
*Raises hand*
I'm curious @ganeshie8 but I will try to figure it out on my own first give me a few minutes and I'll tell you if I give up.
Mhm sec
I give up, what I tried was making this substitution then using the binomial theorem.\[\Large \lim_{n \to \infty} (ax^{1/n}+by^{1/n})^n = x \sum_{k=0}^\infty \left(\begin{matrix}n \\ k\end{matrix}\right) a^n(\frac{x}{y})^{\frac{k}{n}}(\frac{b}{a})^k\]
Why binomial theorem?
I don't know, I just ran out of ideas. Although I think I messed it up because the n terms inside should still be infinite. It's kind of like how this is false: \[\Large \lim_{n \to \infty}(1+\frac{1}{n})^n \ne \prod_{n=1}^\infty (1+\frac{1}{n})\]
@ganeshie8 what's the answer?
hey Kainui you seem to be having the similar idea as mine
we use below standard limit \[\lim\limits_{n\to 0 } \left(1+nx\right)^{1/n} = e^x\]
I see that was kind of along the lines of what I was thinking.
\[\begin{align} \lim \limits_{t\to 0}(ax^t+by^t)^{1/t} &= \lim \limits_{t\to 0}\left[(1-b)x^t+by^t\right]^{1/t} \\ &= x\lim \limits_{t\to 0}\left[1-b+b\left(\frac{y}{x}\right)^t\right]^{1/t} \\ \end{align}\]
That's fascinating, nice.
say \(y/x\) = u for the time being and lets work that piece \[\begin{align} \lim \limits_{t\to 0}\left[1-b+bu^t\right]^{1/t} &= \lim \limits_{t\to 0}\left[1-b+bu^t\right]^{1/t} \\ &= \lim \limits_{t\to 0}\left[1+b(u^t\color{Red}{-} 1)\right]^{1/t} \\ \end{align}\]
the idea is to use another limit \[\lim_{t\to0}\dfrac{f(x)^t-1}{t}=\ln(f(x))\] this proof is trivial to, we can work it later i hope
\[ \begin{align} \lim \limits_{t\to 0}\left[1-b+bu^t\right]^{1/t} &= \lim \limits_{t\to 0}\left[1-b+bu^t\right]^{1/t} \\ &= \lim \limits_{t\to 0}\left[1+b(u^t\color{Red}{-} 1)\right]^{1/t} \\ &= \lim \limits_{t\to 0}\left[1+bt\frac{(u^t\color{Red}{-} 1)}{t}\right]^{1/t} \\ &= \lim \limits_{t\to 0}\left[1+{b\ln u} t \right]^{1/t} \\ &= e^{{b\ln u}} \\ &= u^b \\ \end{align} \]
*.*
pretty result
That was nice, but where can I find that proof?
that was from a real analysis problem set one sec the proof is easy+pretty, let me pull it up
I'll post it as new question
Join our real-time social learning platform and learn together with your friends!