next up previous
Next: Other Examples of Vector Up: Measuring the Size of Previous: Definition of Vector norm

Proving a Function is a Norm

So we seek to show the following:

Theorem 1.1   The function from $ \mathbb{R}^{n}$ into $ \mathbb{R}$ defined by

$\displaystyle \Vert \mathbf{x}\Vert _{2} = (\sum_{j=1}^{n} x_{j}^{2})^{1/2}
$

is a norm.

Proof. Let $ \mathbf{x}\in \mathbb{R}^{n}$. We prove the proprieties in order.

  1. Since $ x^{2}_{j} \geq 0$ for $ j=1,2,\ldots,n$. Then

    $\displaystyle \sum_{j=1}^{n} x_{j}^{2} \geq 0
$

    The positive branch of the square root function is an increasing function so we have

    $\displaystyle (\sum_{j=1}^{n} x_{j}^{2})^{1/2} \geq 0.
$

  2. Note that $ \mathbf{x}= \mathbf{0}$ is equivalent to $ x_{j} = 0$ for $ j=1,2,\ldots,n$. Of course, this quickly implies

    $\displaystyle (\sum_{j=1}^{n} x_{j}^{2})^{1/2} = 0.
$

    There is a subtle point here. We also need to show

    $\displaystyle (\sum_{j=1}^{n} x_{j}^{2})^{1/2} = 0.
$

    implies that $ x_{j} = 0$ for $ j=1,2,\ldots,n$. There are many reasonable candidates for norms that fail this property. It is often easier to prove the contrapositive. Recall the contrapositive of ($ A$ implies $ B$) is ($ \neg B$ implies $ \neg A$). These expressions are logically equivalent. Prove one and you have proved the other. So we assume $ \mathbf{x}\neq \mathbf{0}$. This means there is an index $ k$ such that $ x_{k} \neq 0$. Since $ f(x) = x^{2}$ has only one root at $ x = 0$, then we know $ x_{k}^{2} > 0$. Furthermore $ x_{j}^{2} \geq 0$ for the remaining components of $ \mathbf{x}$. Hence,

    $\displaystyle (\sum_{j=1}^{n} x_{j}^{2})^{1/2} > 0.
$

    Using the increasing property of the square root function, we get $ \Vert \mathbf{x}\Vert _{2} > 0$. (This may seems like too many words to show something that is obvious, but this obvious property fails enough mathematicians had to create the concept of semi-norm).

  3. Using the fact that $ \sqrt{\alpha^{2}} = \vert\alpha\vert$, quickly get this property.

  4. The triangle inequality is also the problematic part of showing a function is a norm. Let $ \mathbf{x},\mathbf{y}\in \mathbb{R}^{n}$. Then

    \begin{displaymath}
\begin{split}
\Vert \mathbf{x}+ \mathbf{y}\Vert^{2}_{2} &= (...
...thbf{x}\Vert _{2} + \Vert \mathbf{y}\Vert _{2})^{2}
\end{split}\end{displaymath}

$ \qedsymbol$


next up previous
Next: Other Examples of Vector Up: Measuring the Size of Previous: Definition of Vector norm
Michael HIlgers 2002-10-07