1
$\begingroup$

Let $B \subset \mathbb{R}^N$ be the unit ball and $E = \{u \in H^1_0(B) : u \text{ is radial}\}$. Define the functional $I : E \to \mathbb{R}$ by $$ I(u) = \int_{B} |\nabla u(x)|^2 dx - \frac{1}{p}\int_{B} |x| |u(x)|^{p} dx, $$ for some $p > 2$. Suppose $u \in E$ is a radial critical point of $I$, that is, $$ I'(u) \varphi = \int_{B} \nabla u \nabla \varphi - \int_{B} |x| |u|^{p-2}u \varphi = 0, \quad \forall \varphi \in C^\infty_{0,rad}(B), $$ where $\overline{C^\infty_{0,rad}(B)}^{H^1_0(B)} = E$. How to show $u$ is a weak solution of $$ -\Delta u (x) = |x| u^{p-1}(x), x \in B, \quad u (x) = 0, x \in \partial B, $$ without using the functional $I$ is well defined in $H^1_0(B)$ ? I can't use this fact, because I just know $I$ is well defined in $E$. I know above $PDE$ is equivalent to the $ODE$ $$ -[f''(s) + \frac{N-1}{s} f'(s)] = s |f(s)|^{p-2}f(s), 0 < s < 1, \quad f(0) = \alpha, f'(0) = 0, $$ for some $\alpha \in \mathbb{R}$.

What I tried: Given $\psi \in C^\infty_0(0,1)$ arbitrary, define for $x \neq 0$, $\varphi_0(x) = \psi(|x|)$ and $\varphi_0(0) = \psi(0) = 0$. Notice that $\varphi_0 \in C^\infty_{0,rad}(B)$. So, as $u$ is a critical point of $I$ one has $$ \int_{B} \nabla u \nabla \varphi_0 = \int_{B} |x| |u|^{p-2}u \varphi_0 . $$ If we write $u(x) = f(|x|)$ for some $f \in H^1_0(0,1)$, It is not difficulty to see that $\nabla u(x) \nabla \varphi_0(x) = f'(|x|) \psi'(|x|)$. So, integrating last inequality in polar coordinates one has $$ \sigma_{N}(S^{N-1}) \int_0^1 f'(s) \psi'(s) s^{N-1} ds = \sigma_{N}(S^{N-1}) \int_0^1 s |f(s)|^{p-2} f(s) \psi(s) s^{N-1} ds, $$ that is $$ \int_0^1 f'(s) \psi'(s) s^{N-1} ds = \int_0^1 s |f(s)|^{p-2} f(s) \psi(s) s^{N-1} ds, \quad \forall \psi \in C^\infty_0 (0,1). $$ Is there a way to use my idea to conclude that $$ \int_{B} \nabla u \nabla \varphi - \int_{B} |x| |u|^{p-2}u \varphi = 0, \quad \forall \varphi \in C^\infty_{0}(B) ? $$ Or $$ -\int_0^1 [f''(s) + \frac{N-1}{s} f'(s)] \psi = \int_0^1 s |f(s)|^{p-2}f(s) \psi, \quad \forall \psi \in C^\infty_0(0,1) ? $$

$\endgroup$
2
  • $\begingroup$ Without using that $I$ is well defined on $H^1$, how do you define $I'(u)$ on non-radial functions? If $I:E \to \mathbb{R}$, then $I'(u): E \to \mathbb{R}$ also is at most a linear operator on radial functions. $\endgroup$ Commented Feb 7 at 8:53
  • $\begingroup$ I is defined only on $E$, like in the paper "A Nonlinear Dirichlet Problem on the Unit Ball and Its Applications", from WEI-MING NI. $\endgroup$ Commented Feb 7 at 10:41

0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.