- #1
Enzipino
- 13
- 0
Hello,
I've been using Caratheodory's Lemma to prove the Inverse Function Theorem and Fermat's Theorem. I have managed to prove both of them, I would just like someone to look over my proof and tell me if I'm missing anything (i.e. should I clarify any parts of my proof). So here goes:
Inverse Function Theorem - Suppose that $f$ is a one-one continuous function (this means it's bijective right?) on an interval $S$, $f$ is differentiable at a given number ${x}_{0}\in S$, and $f'({x}_{0})\ne0$. Then $f^{-1}$ is differentiable at the number $f({x}_{0}) = {y}_{0}$ and we have $(f^{-1})'({y}_{0})=\frac{1}{f'({x}_{0})}$
For Fermat's Theorem, my book has it in 5 parts but I'm just going to post the 5th part which is the main theorem. I will be using 2 other parts concerning the maximum:
1. If $x$ is not the right endpoint of $S$ and $f'(x)>0$, then $f$ cannot have a max at $x$.
2. If $x$ is not the left endpoint of $S$ and $f'(x)<0$, then $f$ cannot have a max at $x$.
Fermat's Theorem - Suppose that $f$ is a function defined on an interval $S$ and that $f$ is differentiable at a number $x\in S$. If $x$ is not an endpoint of $S$ and $f$ has either a maximum or minimum value at $x$, then we must have $f'(x)=0$.
Any feedback is greatly appreciated. Any tips about how to explain these two things in front of an entire class would be greatly appreciated as well. Thanks!
I've been using Caratheodory's Lemma to prove the Inverse Function Theorem and Fermat's Theorem. I have managed to prove both of them, I would just like someone to look over my proof and tell me if I'm missing anything (i.e. should I clarify any parts of my proof). So here goes:
Inverse Function Theorem - Suppose that $f$ is a one-one continuous function (this means it's bijective right?) on an interval $S$, $f$ is differentiable at a given number ${x}_{0}\in S$, and $f'({x}_{0})\ne0$. Then $f^{-1}$ is differentiable at the number $f({x}_{0}) = {y}_{0}$ and we have $(f^{-1})'({y}_{0})=\frac{1}{f'({x}_{0})}$
Proof: Since $f$ is differentiable at $x_0$, $\exists\phi:S\to\Bbb{R}$ such that $\phi(x_0)=w$ and $f(x) = f(x_0) + (x-x_0)\phi(x)$ [Caratheodory's]
Let $x=f^{-1}(y)$, we have $f(f^{-1}(y)) = f(f^{-1}(y_0)) + (f^{-1}(y)-f^{-1}(y_0))\phi(f^{-1}(y))$.
$\implies f(f^{-1}(y))-f(f^{-1}(y_0)) = (f^{-1}(y)-f^{-1}(y_0))\phi(f^{-1}(y))$
$\implies y-y_0=(f^{-1}(y)-f^{-1}(y_0))\phi(f^{-1}(y))$
$\implies \frac{y-y_0}{(f^{-1}(y)-f^{-1}(y_0))}=\phi(f^{-1}(y))$
Flipping both sides of the equation we have $\frac{(f^{-1}(y)-f^{-1}(y_0))}{y-y_0}=\frac{1}{\phi(f^{-1}(y))}$
$\implies (f^{-1})'({y}_{0})=\frac{1}{\phi(x)}$
$\therefore(f^{-1})'({y}_{0})=\frac{1}{f'(x_0)}$
Q.E.D.
Let $x=f^{-1}(y)$, we have $f(f^{-1}(y)) = f(f^{-1}(y_0)) + (f^{-1}(y)-f^{-1}(y_0))\phi(f^{-1}(y))$.
$\implies f(f^{-1}(y))-f(f^{-1}(y_0)) = (f^{-1}(y)-f^{-1}(y_0))\phi(f^{-1}(y))$
$\implies y-y_0=(f^{-1}(y)-f^{-1}(y_0))\phi(f^{-1}(y))$
$\implies \frac{y-y_0}{(f^{-1}(y)-f^{-1}(y_0))}=\phi(f^{-1}(y))$
Flipping both sides of the equation we have $\frac{(f^{-1}(y)-f^{-1}(y_0))}{y-y_0}=\frac{1}{\phi(f^{-1}(y))}$
$\implies (f^{-1})'({y}_{0})=\frac{1}{\phi(x)}$
$\therefore(f^{-1})'({y}_{0})=\frac{1}{f'(x_0)}$
Q.E.D.
For Fermat's Theorem, my book has it in 5 parts but I'm just going to post the 5th part which is the main theorem. I will be using 2 other parts concerning the maximum:
1. If $x$ is not the right endpoint of $S$ and $f'(x)>0$, then $f$ cannot have a max at $x$.
2. If $x$ is not the left endpoint of $S$ and $f'(x)<0$, then $f$ cannot have a max at $x$.
Fermat's Theorem - Suppose that $f$ is a function defined on an interval $S$ and that $f$ is differentiable at a number $x\in S$. If $x$ is not an endpoint of $S$ and $f$ has either a maximum or minimum value at $x$, then we must have $f'(x)=0$.
Proof: Suppose that $f$ has a maximum value at $x$. Using the converse of 1 and 2 above, we have that $f'(x)\le0$ and $f'(x)\ge0$. Thus, $f'(x)=0$.
Q.E.D.
This proof just seems way too easy which has me second guessing myself but I know that the theorem definitely follows from the other parts so by proving those parts, it should lead me to proving this final part.Q.E.D.
Any feedback is greatly appreciated. Any tips about how to explain these two things in front of an entire class would be greatly appreciated as well. Thanks!