5
$\begingroup$

(Cross-posted from Math.SE)

Suppose I have the Fokker-Planck PDE: $$ \frac{\partial}{\partial t}u(x,t)=-\frac{\partial}{\partial x}(\mu(x,t)u(x,t))+\frac{1}{2}\frac{\partial^2}{\partial x^2}(D(x,t)u(x,t)). $$ Denote the solution $u$ as $u^\mu$ to strength the fact that, in particular, it depends on $\mu$.

Where can I find results that guarantee the continuity of the PDE solution with respect to the coefficients?

That is, results guaranteeing that $$ \mu_n(x,t)\to \mu(x,t) \implies u^{\mu_n}(x,t)\to u^{\mu}(x,t), $$ under some conditions and appropriate definitions for the above convergences. The simpler the result, the better.

$\endgroup$
2
  • $\begingroup$ Do you already have results that guarantee existence-and-uniquness? ​ ​ $\endgroup$
    – user5810
    Aug 7, 2016 at 20:56
  • 2
    $\begingroup$ If you know how to prove existence and uniqueness, you should be able to extend the proof to prove continuous dependence on the coefficients. It's similar to how it's done for ODE's. But you do need to figure out what norm the continuity is with respect to. $\endgroup$
    – Deane Yang
    Aug 7, 2016 at 23:45

2 Answers 2

1
$\begingroup$

Maybe the use of the implicit function theorem could be helpful for this, as it could be used in Banach space and works for some PDE.

See for example section II.3.3 of Hamilton for an application to elliptic PDE, Richard S. Hamilton, MR 656198 The inverse function theorem of Nash and Moser, Bull. Amer. Math. Soc. (N.S.) 7 (1982), no. 1, 65--222.

In your case, I also believe that you have much more than just continuity, provided of course the right assumptions and $D$ and $\mu$ to guarantee the well-posedness of the PDE.

$\endgroup$
0
$\begingroup$

Let me make a suggestion. I will assume $u = 0$ on $\partial \Omega$, since everything can be generalised. The existence of this kind of problem is tricky, and even more so the continuity.

I will suggest you use the weak formulation, you have that for every $\eta \in \mathcal D ([0,T]), \varphi \in H_0^1 (\Omega)$: $$ \int_0^T \int_\Omega u \eta' \varphi + \int_0^T \int_{\Omega} \eta D \nabla u \nabla \varphi = \int_0^T \int_{\Omega} \mu u \nabla \varphi + \int_\Omega u(0) \eta(0) \varphi$$ You should be able to show existence of $u^{\mu_n}$ for every $n$ (maybe going to the mild formulation). Then you want to uniform boundary $u^{\mu_n}$ in some $W^{1,p}$. Then try to pass to the limit in that equation. I would suggest you try $\mu_n \to \mu$ in some $L^p$, since this is most natural setting.

$\endgroup$

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.