What is $\left\| f \right\|_{ H_{0}^{k}, H_{0}^{k}}$ norm when $H_{0}^{k}=\left\{u \in H^{k, 2}(M) \mid \int_{M} u \operatorname{vol}_{g}=0\right\}$.

I'm reading a paper Chern-Yamabe flow which said that

Now for $k$ big enough $(k>n)$, the first eigenvalue of the operator $-\Delta$ on the space $H_{0}^{k}=\left\{u \in H^{k, 2}(M) \mid \int_{M} u \mathrm{vol}_{g}=0\right\}$ is strictly negative hence $\left\|e^{-t \Delta}\right\|_{H_{0}^{k}, H_{0}^{k}} \leq$ $C_{0} e^{-c t}$.

I have no idea what is $\left\| - \right\|_{ H_{0}^{k}, H_{0}^{k}}$ norm.

  • $\begingroup$ Taking $K:f\to (\Delta)^{-1}f$ as a (compact) operator between $H^k_0$ and $H^k_0$, it is the usual linear operator norm, perhaps? $\endgroup$
    – username
    Jun 27 at 7:16
  • $\begingroup$ @username given the context, your interpretation is almost certainly the correct one (that $\| e^{-t\Delta}\|_{H_0^k,H_0^k}$ is the operator norm of $e^{-t\Delta}$ as a mapping from $H^k_0$ to itself). Can you post it as an answer? $\endgroup$ Jun 27 at 14:48
  • $\begingroup$ It seems that the $\Delta$ in this paper is $-\Delta$. I add the link in the question. $\endgroup$ Jun 27 at 14:58
  • $\begingroup$ Oof, I don't like this notation. Normally for operator norms I would write $\| - \|_{X\to Y}$ because the one they chose, $\| - \|_{X,Y}$, does not make clear whether they want it to be the operator norm of a mapping from $X\to Y$ or from $Y\to X$. This becomes a bit of an issue later on when they write $\| e^{-t\Delta} \|_{H^k_0, H^0_0}$. At a quick glance it is not obvious to me which direction they meant. $\endgroup$ Jun 27 at 19:13

1 Answer 1


A partial answer to your question. I would think it means: $$ \sup\bigg(\| u(t,\cdot)\|_{H^k_0} : u(t,\cdot)=\exp(-t\Delta) u(0,\cdot),\textrm{ with } \| u(0,\cdot)\|_{H^k_0}=1\bigg). $$


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.