1
$\begingroup$

Let $E$ be a Banach space and let $F(y)$ denote the duality map: $$F(y)=\{y^*\in E^*/ \langle y^*,y\rangle =\|y\|^2=\|y^*\|^2\}$$ where $E^*$ is the dual space of $E$.

Are there any sufficient condition for $F$ to be locally Lipschitz apart the case where $E$ is a Hilbert space?

$\endgroup$
1
$\begingroup$

Global Lipschitz continuity

$F$ is globally Lipschitz if and only if the norm of $E$ is uniformly smooth with quadratic modulus of smoothness, meaning that there exists a constant $k>0$ such that the inequality $$ \|x+y\|+\|x-y\|\le 2+k\|y\|^2 $$ holds for all unit vectors $x$ and all $y\in E$. This is an old result which one can usually find in books dealing with moduli of smoothness and convexity. For a proof, see the paper by M. Zemek, Strong monotonicity and Lipschitz-continuity of the duality mapping, Acta Universitatis Carolinae. Vol. 32 (1991), No. 2, 61-64.

For example, the spaces $L^p$ and $\ell^p$ have quadratic modulus of smoothness for $2\le p<\infty$.

Local vs global Lipschitz continuity

If the duality map $F:E\to E^*$ is locally Lipschitz on $E$, then it is globally Lipschitz on $E$. Indeed, being locally Lipschitz implies being Lipschitz in some neighborhood of $0$; in particular, on some sphere $S_r = \{x: \|x\| = 1\}$. By construction, $F$ is degree $1$ homogeneous: $F(tx)=tF(x)$ for $t>0$. It is an elementary exercise to prove that a degree $1$ homogeneous map is globally Lipschitz if and only if it is Lipschitz on some sphere $S_r$.

Local vs global Lipschitz continuity, on a sphere

A more interesting question is to ask when $F$ is locally Lipschitz on the unit sphere $S_1$. Now the preceding argument with a neighborhood of zero doesn't apply, and it a duality map can be locally Lipschitz on $S_1$ without being globally Lipschitz. I suggest looking up the literature on locally uniformly smooth Banach spaces: it should be true that $F$ is locally Lipschitz on $S_1$ iff the norm has a quadratic modulus of smoothness in some neighborhood of every point of $S_1$, with multiplicative constant $k$ depending in the point.

$\endgroup$

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.