Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm studying the lecture notes by Philippe Clément about Gradient Flows in Metric Spaces.

Now the following problem arises ($X$ is a Hilbert space):

Definition: Let $\phi:X \to (-\infty, \infty]$ be proper (that is it does not only attain $\infty$), lower semicontinuous and $\alpha$-convex for some $\alpha \in \mathbf R$, that is $\phi - \alpha e$ is convex where $e(x) = \frac12 |x|^2$. Define \begin{equation} \psi(y) := \begin{cases} \frac1{2h} |y - x|^2 + \phi(y), &\text{if $y \in D(\phi)$},\\ \infty &\text{otherwise}. \end{cases} \end{equation} this function has a global minimizer (is a lemma) which will be denoted by $J_h x$. Further let $h \in I_\alpha$, that is $I_\alpha = (0, \infty)$ if $\alpha \geq 0$ and $(0, |\alpha|^{-1})$ if $\alpha < 0$. Now we set $$\phi_h(x) = \psi(J_h x).$$

Now the claim is that $\phi_{h_1}(x) \leq \phi_{h_2}(x)$ for $0 < h_2 < h_1 \leq h_\alpha$ and $x \in X$ where $h_\alpha$ is $1$ for nonnegative $\alpha$ and $\frac1{2|\alpha|}$ for negative $\alpha$. Why is this true? I can't seem to figure out, but probably I'm missing something simple. They claim that it should follow from the information which I have given.

share|cite|improve this question
up vote 1 down vote accepted

It's actually not too difficult to see that. Maybe just because it's buried into the notations.

Let $y_1=J_{h_1}x$ and $y_2=J_{h_2}x$. Strictly speaking, $\psi(y)$ also depends on $h$ and you are varying $h$ later. So a better notation would be $\psi_h(y)$, which will be used below to make things clear. Because $y_1$ is a minimizer of $\psi_{h_1}(y)$ \begin{eqnarray} \phi_{h_1}(x)&=\psi_{h_1}(y_1)\le \psi_{h_1}(y_2)=\frac{1}{2h_1}|y_2-x|^2+\phi(y_2)\\ &\le \frac{1}{2h_2}|y_2-x|^2+\phi(y_2)=\psi_{h_2}(y_2)=\phi_{h_2}(x). \end{eqnarray} The last inequality is because, of course, $h_2<h_1$.

I haven't figured out why $h_1$ has to be $\le h_\alpha$, but I guess it has something to do with where the global maximum of $\psi_h$ is obtained. So please check the lemma and you might find the answer.

share|cite|improve this answer
That is indeed not difficult. Thanks. – Jonas Teuwen Apr 7 '11 at 8:29

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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