So I am stuck on this question and I don't really know how to go about it. The question is

Let $X=[0,1]$ and $\mu = \lambda$ be the Lebesgue measure. Consider a measurable function $f:[0,1]\rightarrow \mathbb{R}$ such that $f(x)\geq ||f||_2$ for all $x \in [0,1]$. Prove that there exists an $x \in [0,1]$ such that $f(x_0) =||f||_2$.

So far, I have written out the statements

$||f||_2=\bigg(|f|^2 d\lambda \bigg)$ and $f(x)\geq \bigg( \int |f|^2 d\lambda \bigg)^{\frac{1}{2}}$


Using Holder's inequality $$ \|f\|_2 = \int_0^1 \|f\|_2 \leq \int_0^1 f \leq \left(\int_0^1 |f|^2\right)^{1/2} = \|f\|_2. $$ Hence the non-negative function $g(x) := f(x) - \|f\|_2$ satisfies $\int_0^1 g = 0$, so that $g(x) = 0$ a.e. in $[0,1]$, i.e. $f(x) = \|f\|_2$ a.e. in $[0,1]$.


Since $\{x\in[0,1]:f(x)>\|f\|_{2}\}=\bigcup_{n\in\mathbb{N}}\{x\in[0,1]:f(x)-\frac{1}{n}\geq\|f\|_{2}\}$, if there is not such $x_{0}$, for some $n_0$ we would have for $C_{0}=\{x\in[0,1]:f(x)-\frac{1}{n_{0}}\geq\|f\|_{2}\}$, $\mu(C)>0$. Then,

$$\int_{C}f^{2}\geq \int_{C}\left (\|f\|_{2}+\frac{1}{n_{0}}\right )^{2}= \left (\frac{1}{n_{0}}+\|f\|_{2}\right )^{2}\mu(C) $$ and

$$\int_{[0,1]\backslash C}f^{2}\geq \|f\|_{2}^{2}\mu([0,1]\backslash C) $$

So, $$\int_{[0,1]}f^{2}\geq \|f\|_{2}^{2}+\mu(C)(\frac{1}{n_{0}^{2}}+\frac{2\|f\|_{2}}{n_{0}})>\|f\|_{2}^{2} $$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.