I am struggling with the following question:

Is it possible to find uncountably many norms on $C[0,1]$ such that no two are Lipschitz equivalent?

I had thought about trying to define norms for each real number r>1:

$$\|f\|_r =\left( \int_0^1 f(x)^r \, dx \right)^{1/r}$$

but I think these are Lipschitz equivalent.

Does anyone have any better ideas?

  • $\begingroup$ "but I don't think these are Lipschitz equivalent" <- But that is exactly what you want, isn't it? That no two of the norms are (Lipschitz) equivalent. [And what you wrote only gives you a norm for $r \geqslant 1$. But there are still uncountably many of those.] $\endgroup$ – Daniel Fischer Nov 6 '15 at 12:52
  • $\begingroup$ Sorry, I meant to say I thought they were Lipschitz equivalent. Maybe they aren't? I can't find counterexamples though. $\endgroup$ – David Smith Nov 6 '15 at 13:54
  • $\begingroup$ No two of these norms are (Lipschitz) equivalent. Hölder's inequality gives $\lVert f\rVert_r \leqslant \lVert f\rVert_s$ for $r < s$, so you need to show that you can't estimate $\lVert f\rVert_s \leqslant C\cdot \lVert f\rVert_r$ then. Play with a modification of $f_s(x) = x^{-1/s}$ (you need to modify that to get a continuous function). $\endgroup$ – Daniel Fischer Nov 6 '15 at 14:24
  • $\begingroup$ Sophisticated answers! A lazier answer that might work (after I get my morning coffee) is $$\|f\|_r= \sup_{0\leq x\leq r}|f(x) | + \int_r^1|f(x)|\,dx$$ for $0<r<1$. It doesn't much to be a norm, although it takes a lot to be a useful norm. $\endgroup$ – B. S. Thomson Nov 6 '15 at 17:25

In addition to the norms suggested in comments, here is one more (inspired by B.S.Thomson): for $t\in [0,1]$, let $$ \|f\|_t = |f(t)| + \int_0^1 |f(x)|\,dx $$ The integral term is only needed to make this a norm rather than a seminorm. The fact that these are mutually nonequivalent follows by considering $f_{a,n}(x)=\max(0,1-n|x-a|)$ which satisfies $$ \|f_{a,n}\|_t \to 0,\quad t\ne a;\qquad \text{yet } \ \|f_{a,n}\|_a = 1 $$


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.