5
$\begingroup$

Question:

Consider a semi-linear parabolic equation on a bounded domain $\Omega \subset \mathbb{R}^m$: $$ \frac{\partial f_t}{\partial t} = -\Delta f_t + Q(f_t, df_t), $$ with smooth initial data $f_0 \colon \Omega \to \mathbb{R}^n$. Suppose we know that there exists a smooth solution $f_t$ for all times $t \in [0, \infty)$. Can one find a time $t_0 > 0$ such that the $W^{2,2}$-norm of $f_{t_0}$ (on some smaller domain) is bounded by an expression depending only on $\|f_0\|_{W^{1,2}}$ and on the coefficients of $Q$? Here $t_0$ may depend freely on $f_0$, however.

Motivation:

Due to my complete lack of knowledge on the field of PDE, the question is probably very ill-posed, so here's a bit of background. Consider the Eells-Sampson heat flow equation, which in local coordinates reads: \begin{equation}\tag{1} \frac{\partial f_t^{\alpha}}{\partial{t}} = - \Delta f_t^\alpha + {}^N\Gamma_{\beta\gamma}^\alpha \frac{\partial f^\beta}{\partial x^i}\frac{\partial f^\gamma}{\partial x^j}g^{ij} -{}^M\Gamma_{ij}^k \frac{\partial f^\alpha}{\partial x^k}g^{ij} \end{equation} (here, from the global point of view, we are in Corlette's setting, that is, we are given a representation $\rho \colon \pi_1(M) \to G$ and we are considering $\pi_1(M)$-equivariant maps $f\colon \tilde M \to N = G/K$). We know that for every $\pi_1(M)$-equivariant initial data $f = f_0$ solutions of (1) exist for all time, and that the $L^2$-norm of $d f_t \colon \tilde M \to N$ is decreasing. I would like to use the method above to find another equivariant map $f'= f_{t_0}$ with explicit bounds on the norms of higher derivatives (at least on the $W^{2,2}$ norm) depending only on $E(f) = \|d f_0\|_{L^2}^2$.

Corlette, in his original paper of 1988, in order to construct the solution gives a bound on the $W^{1,2}$ norm of $df$ ($\theta$, in his notations), but this is not what I need, since it also depends on the $L^2$-norm of the tension field $\tau(f)$ ($\Phi$, in his notations); in fact, he considers estimates that hold true for every $t$, while I am really interested in improving the properties of $f_0$.

$\endgroup$

0

Your Answer

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

Browse other questions tagged or ask your own question.