$L^q(X) \subset L^p(X)$ if $p\leq q$ and $\mu(X) < \infty$.

1k Views Asked by At

Let $(X,\mathfrak{M},\mu)$ be a measure space. Show that if $\mu(X) < \infty$ then $L^q(X) \subset L^p(X)$ for $1\le p \le q \le \infty$.

Is it enough to define $$E_0 = \{x \in X \, : \, 0 \leq |f(x)| < 1\}$$ and $E_1 = X \setminus E_0$ so that $\mu(E_0), \, \mu(E_1) < \infty$ and $E_0$ is measurable as $E_0 = \{x \in X : |f(x)| \geq 0 \} \cap \{x \in X : |f(x)| \leq 1 \}$ and $f$ is measurable. Therefore if $f \in L^q$,

\begin{eqnarray*} ||f||^p_{L^p} &=& \int_{E_0}|f|^p \, d\mu + \int_{E_1} |f|^p \, d\mu \\ &\leq& \int_{E_0}|f|^p \, d\mu + \int_{E_1} |f|^q \, d\mu \\ &<& \mu(E_0) \,\, + ||f||_{L^q}^q \\ &<& \infty. \end{eqnarray*}

which implies $f \in L^p$.

Doe this suffice? I've seen a proof using Holder's inequality (that has now been given as an answer for those interested), that is much shorter, but does this suffice?

1

There are 1 best solutions below

0
On BEST ANSWER

Your proof is fine. You can also try to show that if $p<q$ then $$\lVert f\rVert_p \leqslant \lVert f\rVert_q \mu(X)^{1/p-1/q}$$

by using Hölder's inequality. This gives a bit more information than your proof. In particular for $X$ a probability space gives $p\mapsto \lVert f\rVert_p$ is increasing.