6
$\begingroup$

Let $A$ be an unbounded linear operator of domain $D(A)$ defined on a Banach space $X$. Suppose that $A$ generates a $C_0$-semigroup $T(t)$ which is uniformly bounded. I would like to know if there are additional assumptions under which the following conjecture is true:

For every compact subset $K$ of $X$, there exists a constant $C_K$ such that $$\mid{((T(t)-T(t'))x\mid \leq C_K\mid t-t'\mid},\quad \forall t,t'\in \mathbb{R}_+,\quad \forall x\in K.$$

$\endgroup$

1 Answer 1

6
$\begingroup$

If I understand your question correctly, this would mean that the function $t\mapsto T(t)x$ is Lipschitz continuous, which is equivalent for $x$ to be in the Favard space $\text{Fav}(A)$$. See for example Defintion 8.2. in these lecture notes.

If your space $X$ is reflexive, then the Favard space is exactly $D(A)$, the domain of $A$, see Corollary II.5.21 in

Engel, Klaus-Jochen; Nagel, Rainer, One-parameter semigroups for linear evolution equations, Graduate Texts in Mathematics. 194. Berlin: Springer. xxi, 586 p. (2000). ZBL0952.47036.

Hence if $D(A)\neq X$ (which is the case for unbounded generators), this fails.

Unfortunately, in non-reflexive spaces, where $D(A)\subset \text{Fav}(A)\subset X$ still holds but the Favard space generally fails to be the whole space. Hence, your property cannot be fulfilled even for $K=\{x\}$ singletons if $x\notin \mathrm{Fav}(A)$.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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