MathOverflow is a question and answer site for professional mathematicians. 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

Dear community.

I'm looking for survey articles about recent and old results about stability for the elliptic and backward-parabolic Cauchy problem.

For example: Take $\partial_t u + a(t)\partial^2_x u = 0$ with $u(0,x)=u_0(x)$. What is the minimal regularity-assumption on $a$, in terms of a first or second order modulus of continuity to get Hölder continuous dependence of the solution on the initial data. Or: What is the best condition like $|f(t)a(t)|\leq C$ on $a\in C^1(0,T]$ to get the same result. Another question is what happens if we mix this two conditions.

Often is claimed, that Lipschitz is the optimal regularity. But I think this is, like for uniqueness, not true. At least for Zygmund there should be a similar result and also maybe for coefficients with a first order modulus of continuity $\mu$ satisfying the Osgood Condition: $\int\limits_0^1 \frac{ds}{\mu(s)}=+\infty$.

Maybe there is not so much in this direction. For hyperbolic equations there are a lot of results with this kind of conditions.

Best CJ

share|cite|improve this question
Why? What is your background? And what have you searched in? The first hit in Google for "elliptic cauchy stability" already seems relevant: – Willie Wong Apr 9 '11 at 18:33
@Wong: I'm trying something in this direction and I just want an overview about results and methods used to prove such kind of results. Especially I'm interested in the minimal regularity of the coefficients. I searched with google and found this and of course a lot of original papers devoted to the problem. But beside the Alessandrini ed al paper I couldn't really find good reviews on the topic. – CPJ Apr 10 '11 at 11:55
You are aware of course that a backwards parabolic problem can not be solved unless you put extremely strong conditions on the data (i.e. of Gevrey class of order less than 1/2), and when you are in such a situation the regularity of the coefficient does not really matter i.e. a locally integrable $a(t)$ is fine? – Piero D'Ancona Apr 10 '11 at 12:47
@D'Ancona: Yes, I'm aware of the fact that this equation, as the elliptic, is ill-posed even in reasonable situations. I'm just looking for uniqueness and stability (in the sense of John) at the moment. – CPJ Apr 10 '11 at 16:00
@CJ: what is "stability in the sense of John"? Can you give a (hopefully easy to find online) reference? – drbobmeister Apr 11 '11 at 0:48

Your Answer


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

Browse other questions tagged or ask your own question.