Suppose we have two equivalent norms, $\|\cdot \| \sim \left\vert\!\left\vert\!\left\vert \cdot\right\vert\!\right\vert\!\right\vert$. We assume that $x_n \rightarrow x $ in $(X,\|\cdot \|)$. Show that $x_n \rightarrow x $ in $(X,\left\vert\!\left\vert\!\left\vert \cdot\right\vert\!\right\vert\!\right\vert)$

We observe since the norms are equivalent, there exits constants $\alpha, \beta$, such that $\alpha \|\cdot \| \leq \left\vert\!\left\vert\!\left\vert \cdot\right\vert\!\right\vert\!\right\vert \leq \beta \|\cdot \|$, with $0 \lt \alpha \leq \beta$

Since, $x_n \rightarrow x$ in $(X,\|\cdot \|$) we know that for all $\epsilon>0, \exists K(\epsilon)$ such that for n>$K(\epsilon),$ we have $\|x_n -x\| \leq \epsilon$

Since, we know $\beta \gt 0$ , then $ \beta \|x_n -x\| \leq \beta \epsilon$, Thus we see that $\left\vert\!\left\vert\!\left\vert x_n-x\right\vert\!\right\vert\!\right\vert\leq \beta \epsilon$.

This is how far I've gotten, our goal to is to find a a $J(\epsilon)$ such that for all $\epsilon>0, \exists J(\epsilon)$ such that for n>$J(\epsilon),$ we have $\left\vert\!\left\vert\!\left\vert x_n-x \right\vert\!\right\vert\!\right\vert \leq \epsilon$

My candidate for $J(\epsilon$) is $J(\epsilon) = \frac{K(\epsilon)}{\beta}$ from my above argument, but I'm not exactly sure how to validate this. If someone could help me on the proof, it would be very appreciated. I believe I'm pretty close, just missing something small.


I think you mean $J(\epsilon) = K(\epsilon/\beta)$.

For all $n \ge J(\epsilon)=K(\epsilon/\beta)$, we have $\|x_n-x\| \le \epsilon/\beta$, so $$||| x_n-x||| \le \beta \|x_n-x\| \le \beta \epsilon/\beta=\epsilon.$$

  • $\begingroup$ Perfect, this is exactly what I was looking for to get me to the end. $\endgroup$ – Vogtster Aug 28 '16 at 22:19

It is far easier to show it this way:

$x_n \rightarrow x $ in $(X,||*||)$, implies that $||x_n - x|| \to 0$ as $n \to \infty $.

Due to norm equivalence, there exist constants $\alpha, \beta$, such that $$\alpha ||x_n - x|| \leq |||x_n - x||| \leq \beta ||x_n - x||,$$

so if both the LHS and RHS goes to zero, then so does the middle term and thus we have shown convergence in the second norm.

  • $\begingroup$ Thanks, this is actually a lot more simple than my argument. $\endgroup$ – Vogtster Aug 28 '16 at 22:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.