2
$\begingroup$

Consider the following advection-diffusion equation $$ \begin{cases} u^\epsilon_t + f(u^\epsilon)_x = \epsilon \Delta u^\epsilon\\ u^\epsilon(0,\cdot) = u_0, \end{cases} $$

How can one prove an estimate on the difference $$\Vert u^\epsilon - u^\eta \Vert_{L^2(\mathbb R^N)} $$ where $u^\epsilon$ solves the advection-diffusion equation $$ \begin{cases} u^\epsilon_t + f(u^\epsilon)_x = \epsilon \Delta u^\epsilon\\ u^\epsilon(0,\cdot) = u_0, \end{cases} $$ and $u^\eta$ solves $$ \begin{cases} u^\eta_t + f(u^\eta)_x = \eta \Delta u^\eta\\ u^\eta(0,\cdot) = u_0, \end{cases} $$ for the same initial data $u_0$?

$\endgroup$
1
  • $\begingroup$ Normally you would square the $L^2$ norm, differentiate in $t$, and use the PDEs to get an inequality, to which you apply Gronwall. Have you tried this kind of approach? Where do you get stuck? $\endgroup$
    – Jeff
    Commented Sep 18, 2020 at 16:06

0

You must log in to answer this question.