Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

Let $\Omega\subset\mathbb{R}^n$ be a bounded smooth domain. Let $p\in (1,\infty)$ and $K=\{u\in W^{1,p}(\Omega):\ u-g\in W^{1,p}_0(\Omega)\}$, where $g\in W^{1,p}(\Omega)$. Let $\epsilon\geq 0$ and $F:W^{1,p}(\Omega)\rightarrow\mathbb{R}$ defined by $$F(u)=\frac{1}{p}\int_\Omega (\epsilon^2+|\nabla u|^2)^\frac{p}{2}$$

Let $I=\displaystyle\inf_{u\in K} F(u)$ and note that $I\geq 0$, therefore, there exists a sequence $u_n\in K$ such that $F(u_n)\rightarrow I$.

By using the monotonicity of the function $|\ |^\frac{p}{2}$ and the Poincaré inequality, we can show that $\|u_n\|_{1,p}$ is bounded and then extract a subsequence (not relabeled) such that $u_n \rightharpoonup u$ for some $u\in K$. By using the fact that $F$ is convex and lower semi continuous, we can show that $u$ is the "unique" solution of the problem.

Now my problem is: I want to show that in fact $u_n\rightarrow u$ in $W^{1,p}(\Omega)$.

Beyond the facts stated above, we have that $$\int_\Omega (\epsilon^2+|\nabla u|^2)^\frac{p-2}{2}\nabla u\nabla v=0,\ \forall\ v\in W^{1,p}_0(\Omega)$$

where $u$ is the solution of the problem. The last equality appear when we derivate the functional $F$ in the point $u$.

Remark: Suppose $\epsilon=0$, hence in this case, we have that $F(u_n)=\frac{1}{p}\|\nabla u_n\|_p\rightarrow \frac{1}{p}\|\nabla u\|_p=F(u)$. By using the compact immersion $W^{1,p}(\Omega)\hookrightarrow L^p(\Omega)$ we have (up to a seubsequence ) that $\|u_n\|_p\rightarrow \|u\|_p$ and hence $\|u_n\|_{1,p}\rightarrow\|u\|_{1,p}$. As the space $W^{1,p}(\Omega)$ is uniformly convex and $u_n \rightharpoonup u$ we can conclude that $u_n\rightarrow u$ in $W^{1,p}(\Omega)$

In the case $\epsilon>0$ this sill Works: so if we can show that $\|\nabla u_n\|_p\rightarrow \|\nabla u\|_p$, the problem is solved.

Any help is appreciated.


share|cite|improve this question
up vote 0 down vote accepted

I think i have the answer, please verify if it is correct.

First of all, as i have noted in the post, $F$ is convex (in fact it is striclty convex), to see this just derivate it two times and verify that the operator $f''(u)$ is striclty positive for all $u$.

Assumption 1: Suppose $u$ is the strictly local minimum of $F$ in $K$, then for $\delta>0$ there exist a constant $C_\delta>0$ such that $$F(v)>C_\delta+F(u),\ \forall\ v\in K'=K\backslash (B(u,\delta)\cap K)$$ where $B(u,\delta)$ is the ball with center $u$ and ratio $\delta$.

Proof of Assumption 1: Let $I'=\displaystyle\inf_{v\in K'} F(v)$. As the minimum is unique in $K$ we have that $F(u)<I'$ hence, \begin{eqnarray} F(v) &\geq& I' \nonumber \\ &=& (I'-F(u))+F(u), \forall v\in K' \nonumber \end{eqnarray}

So $C_\delta=I'-F(u)$ and this finish the proof os Assumption 1.

Now suppose ad absurdum that $u_n$ does not converge to $u$ in $W^{1,p}(\Omega)$. Therefore we can find some $\delta>0$ and a subsequence of $u_n$ (not relabeled) such that $u_n\in K'$ for all $n$ (I'm using the above notation). By Assumption 1 we have $$F(u_n)\geq C_\delta+F(u)$$

As $C_\delta$ is a positive constant that does not dependes on $n$, we conclude by passing the limit in the laste inequality that $$F(u)\geq C_\delta +F(u)$$

which is an absurd...

share|cite|improve this answer

Your Answer


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

Not the answer you're looking for? Browse other questions tagged or ask your own question.