Prove that two norms are equivalents.

134 Views Asked by At

Two norms ${∥∙∥}_{1}$ and ${∥∙∥}_{2}$ are equivalents $\iff$ $\exists$ $c_{1}$, $c_{2} > 0$ such that $c_{1}{∥x∥}_{1}\le {∥x∥}_{2} \le c_{2}{∥x∥}_{1}$

Let $X$ - Hilbert space over $\mathbb{R}$, $\dim X=n < \infty$

${∥x∥}_{1}=\sqrt{\left<x\mid x\right>}$

${∥x∥}_{2}=\sqrt{\left<Ax\mid Ax\right>}$

where matrix $A:X \to X$ is nondegenerate on $X$.

I showed that ${∥x∥}_{1}$, ${∥x∥}_{2}$ are norms, but I can't find the $c_{1}$, $c_{2}$ to show that they are equivalent. They must be, since two norms on the same finite dimension space are equivalent. Any ideas how to solve the problem?

1

There are 1 best solutions below

0
On BEST ANSWER

In general, as long as $A:X\to X$ is a bounded linear operator with bounded inverse, the two norms are equivalent. Boundedness means there is $c>0$ with $\|Ax\|\le c\|x\|$. Here, $\|x\|_2=\|Ax\|_1\le c\|x\|$. Also as $A^{-1}$ is bounded, there is a $c'>0$ with $\|A^{-1}x\|\le c'\|x\|$, equivalently $\|x\|\le c'\|Ax\|$. Then $\|x\|_1\le c'\|x\|_2$.

For $X$ finite-dimensional, all operators are bounded. If we use the usual inner product on vectors, and $A$ is a matrix, then the $j$-th entry of $Av$ is $\left<a_j,v\right>$ where $a_j$ is the $j$-th row of $A$ and so is $\le\|a_j\|\|v\|$ by Cauchy-Schwarz. A crude estimate now gives $\|Av\|\le\|v\|\sum_{j=1}^n\|a_j\|$.