Let $f:[0,1]\rightarrow\mathbb{R}$ be a $\mathcal{C}^{2}([0,1])$ function such that $f(0)=f(1)=1$ and such that the minimum of $f$ is negative. Prove that there must exist a point $\xi\in [0,1]$ such that $f''(\xi)≥8$
Maybe I can use the fact that there exists maximum and minimum in $[0,1]$ so there are two critical points. But I don't know how to proceed. Can someone give me any advice? Thanks before!
I can get to $4$, by mean value theorem.
Proof: Let $x_{\min}$ denote the minimum point. By mean value theorem there is $c_1,c_2$ such that
$$f'(c_1)=\frac{f(x_{\min})-f(0)}{x_{\min}},f'(c_2)=\frac{f(1)-f(x_{\min})}{1-x_{\min}}.$$
And then, by mean value theorem there is $c_3$:
$$f''(c_3)=\frac{f'(c_2)-f'(c_1)}{c_2-c_1}$$
$$=\frac{\frac{f(1)-f(x_{\min})}{1-x_{\min}}-\frac{f(x_{\min})-f(0)}{x_{\min}}}{c_2-c_1}$$ $$>{\frac{1}{1-x_{\min}}+\frac{1}{x_{\min}}}$$ $$=\frac{1}{x_{\min}(1-x_{\min})}\geq 4$$
You can likely get a tighter bound following the comments.
EDIT: The correct way...
By Taylor's theorem, there is $c$ such that
$$f(x)=f(x_{min})+\frac{f''(c)(x-x_{min})^2}{2}.$$
Then
$$f''(c)=2\frac{f(x)-f(x_{min})}{(x-x_{min})^2}$$
Now develop this at $x=0,x=1$ and sum those, then there is $c_1$, $c_2$
$$\frac{f''(c_1)+f''(c_2)}{2}=\frac{1-f(x_{min})}{x_{min}^2}+\frac{1-f(x_{min})}{(1-x_{min})^2}>\frac{1}{x_{min}^2}+\frac{1}{(1-x_{min})^2}$$ $$=\frac{1-2x_{min}+2x_{min}^2}{(1-x_{min})^2x_{min}^2}\geq8$$
which can only hold if $f''(c_1)$ or $f''(c_2)$ is greater than $8$.