0

Show that $||x||_2\leq ||x||_1$ on $\Bbb R^d$

My attempt:

First I will prove that $\sum^n_{i=1}x^2_i\leq \sum^n_{i=1}|x_i|$:

Let $x=(x_1,x_2)\in \Bbb R^2$ Then $(x_1+x_2)^2=x^2_1+x^2_2+2x_1x_2\geq x_1^2+x^2_2$

Hence $\sqrt{x_1^2+x_2^2}\leq |x_1|+|x_2|$ . So by induction, $\sum^n_{i=1}x^2_i\leq \sum^n_{i=1}|x_i|$

The above proof I took from this answer

So, $||x||_2=\sqrt{\sum^d_{i=1}x^2_i}\leq \sum^d_{i=1}|x_i|=||x||_1$

Would this be correct?

1 Answers1

2

Let $x= (x_{1},...,x_{n})\in \mathbb{R}^{n}$. Let us define: $$ x^{(1)} = (x_{1},0,...,0)$$ $$x^{(2)} = (0,x_{2},...,0) $$ $$\vdots $$ $$x^{(n)} = (0,0,...,x_{n})$$ Then, we have: $$x = x^{(1)}+\cdots+x^{(n)}$$ We can use the triangle inequality for norms to get: $$||x||_{2} = ||x^{(1)}+\cdots+x^{(n)}||_{2} \le ||x^{(1)}||_{2}+\cdots+||x^{(n)}||_{2} = \sqrt{x_{1}^{2}}+\cdots+\sqrt{x_{n}^{2}} = |x_{1}|+\cdots+|x_{n}| = ||x||_{1}$$

IamWill
  • 4,025