Some kind of inverted Minkowski's inequality?

164 Views Asked by At

Sorry if this is a silly question but I want to confirm a result. I get some kind of inverted Minkowski's inequality for $p\in[2,\infty )$. It is based in Clarkson's inequality what can be written as

$$ 2(\|f\|_p^p+\|g\|_p^p)\leqslant \|f+g\|_p^p+\|f-g\|_p^p\tag1 $$ for $f,g\in L^p(\mu )$. But then we reach the conclusion that $$ \|f\|_p^p+\|g\|_p^p\leqslant \||f|+|g|\|_p^p\tag2 $$ This inequality seems too good to be correct. Can someone check if there is no weird mistake in this result? Thank you.

1

There are 1 best solutions below

0
On BEST ANSWER

No, this is perfectly fine (and rather easy to prove). For $p\ge 1$, the map $x\mapsto x^p$ is convex on $[0,\infty)$. Since it sends $0$ to $0$, it is super-additive on $[0,\infty)$: in fact, by convexity, for all $x\ge 0$ and $y>0$,

$$\frac{(x+y)^p-x^p}{y}\ge \frac{y^p-0^p}y$$

Therefore $$\int(\lvert f\rvert+\lvert g\rvert)^p\ge\int (\lvert f\rvert^p+\lvert g\rvert^p)$$

In fact, when you go on to prove Minkowski, main part of the work is getting rid of those absolute values that just give a weaker reverse inequality.