2

Discuss the strong and weak convergence of the sequence of functions

$$u_n(x)=\frac{1}{n}\sin nx+2\sqrt{x}$$

in the $W^{1,p}(0,1)$ Sobolev space.

Pointwise limit is $u(x)=2\sqrt{x}$ and can be easily shown that the sequence converge strongly in $L^p(0,1)\ \forall\ 1\le p \le +\infty$.

The derivative of the sequence is $u'_n(x)=\cos nx +\frac{1}{\sqrt{x}}$ and the pointwise limit $u'(x)=\frac{1}{\sqrt{x}}$ (right?).

For $p=1$ I would say that

$$\int_0^1 \cos nx\ dx = \frac{\sin nx}{n}\to 0\text{ as }n\to+\infty$$

so the sequence would converge strongly in $W^{1,1}$, but actually

$$||u'_n(x)-u'(x)||_{L^1} = \int_0^1 |\cos nx|\ dx$$

and $\cos nx$ does not have a fixed sign for $x\in[0,1]$ since $n$ is going to infinity so how to solve it?

For $p=+\infty$

$$||u'_n(x)-u'(x)||_{L^{\infty}} = \max|\cos nx|=1\nrightarrow 0$$

so the sequence doesn't converge strongly in $W^{1,+\infty}$.

For $1<p<+\infty$, since $\max(\cos nx)=1$

$$||u'_n(x)-u'(x)||_{L^p}^p \le 1$$

so we don't have strong convergence, but since $L^p$ is reflexive for $1<p<+\infty$, by Banach-Alaoglu we can extract a subsequence of $u'_n$ converging weakly (but where?). In particular in $L^2$ $\cos nx$ is a subsequence of the trigonometric basis, which converges weakly to $0$ in $L^2$. But for others $p$ where does the subsequence extracted from $u'_n$ weakly converge?

1 Answers1

3

First notice that $u_n$ only belong to some of the $W^{1,p}$ spaces, as we have $u_n'=\cos (nx)+\frac{1}{\sqrt{x}}$ and $$\cos(nx)+\frac{1}{\sqrt{x}}\in L^p(0,1)\iff \frac{1}{\sqrt{x}}\in L^p(0,1)\iff |x|^{-p/2}\in L^1(0,1)\iff p<2 $$ Therefore it is only meaningful to discuss convergence in $W^{1,p}(0,1)$ for $p\in [1,2)$.

There is no pointwise limit for $u_n'(x)=\cos (nx)+\frac{1}{\sqrt{x}}$ because there is no pointwise limit for $\cos(nx)$ except for $x=0$. The sequence keeps on oscillating around $\frac{1}{\sqrt{x}}$.

Your Banach-Alaoglu + reflexivity argument works to prove the existence of a weakly converging subsequence if $p\in (1,\infty)$. The issue is that it says nothing about the weak limit aside from existence.

To understand the weak limit, there is a general result (see my answer here) - if $g:\mathbb{R}\to \mathbb{R}$ is a bounded periodic function such that its mean over a period is $\alpha$, then if $v_n(x):=g(nx)$ we have $v_n\rightharpoonup \alpha$ weakly-star in $L^{\infty}(\mathbb{R})$ for $p<\infty$. In this case, we have $g(x)=\cos x$ whose mean over a period is $\alpha=\int_0^{2\pi}\cos x\, dx = 0$. Therefore, $\cos (nx)\rightharpoonup 0$ weakly-star in $L^{\infty}(\mathbb{R})$, hence $\cos (nx)\rightharpoonup 0$ weakly-star in $L^{\infty}(0,1)$ and thus (since $L^{\infty}(0,1)\hookrightarrow L^p(0,1)$ for $p\in [1,\infty]$) we also have $\cos (nx)\rightharpoonup 0$ weakly-star (and hence weakly, since they are reflexive) in $L^p(0,1)$ for all $p\in (1,\infty)$. Finally, we can include $p=1$ as weak convergence in $L^p(0,1)$ for $p>1$ implies weak convergence in $L^1(0,1)$.

In conclusion, the above argument shows that the sequence $u_n$ is weakly convergent to $2\sqrt{x}$ in $W^{1,p}(0,1)$ for all $p\in [1,2)$.

To study strong convergence, notice that as we have proved, the weak limit of $\cos (nx)$ is $0$. Therefore, since the strong limit must agree with the weak limit when it exists, by contradiction if $u_n'\to u$ strongly in $L^p(0,1)$, then we would have $\cos (nx)\to 0$ strongly in $L^p(0,1)$ and in particular in $L^1(0,1)$, but if $2(k+1)\pi\geq n\geq 2k\pi$, then \begin{align*}\int_0^1|\cos (n x)|\,dx&= \frac{1}{n}\int_0^n|\cos (y)|\,dy\geq \frac{1}{2(k+1)\pi}\int_0^{2k\pi}|\cos y|\,dy\geq \\ &\geq \frac{k}{2(k+1)\pi}\int_0^{2\pi}|\cos y|\,dy=\frac{2}{(1+1/k)\pi}\not \to 0 \end{align*} a contradiction.

Lorenzo Q
  • 5,269
  • How can we say that $\cos nx$ does not admit a.e. convergent subsequence? Is it obvious? Of course $\cos nx$ does not converge a.e. ... – Myunghyun Song Jan 18 '19 at 12:00
  • I agree that this is not entirely obvious. Because for instance, if we fix $x$ then the sequence $\left{\cos (nx)\right}$ is bounded in $\mathbb{R}$ and hence has a converging subsequence. The issue is in finding a converging subsequence which works for a.e. $x\in [a,b]$. With a diagonalization argument you can find a converging subsequence which works for countably many $x$ but a countable set will still have $0$ measure so that doesn't help. – Lorenzo Q Jan 18 '19 at 12:11
  • It is true that there is no such subsequence (even on a set of positive measure). But isn't it what you should show in your answer? – Myunghyun Song Jan 18 '19 at 12:14
  • @Song That was only used to prove that there is no strong convergence. I have edited the answer proving that there is no strong convergence with a completely different approach. – Lorenzo Q Jan 18 '19 at 12:29
  • Thanks for your response. I think it looks good now. – Myunghyun Song Jan 18 '19 at 12:32
  • You're welcome. I wrote without thinking too much the first time. But in general remember that an oscillating sequence weakly converges to its mean on any $L^p$ space for $p\in [1,\infty)$. – Lorenzo Q Jan 18 '19 at 12:42
  • @LorenzoQuarisa thank you for answer. First question: my tutor told me that to find the pointwise limit of the derivative we can derive the pointwise limit of the primitive, that's why I wrote $u'(x)=\frac{1}{\sqrt{x}}$; but from your discussion it seems this not hold, isn't it? Second question: I don't understand your argument in the second line (ie why $p<2$): to check $\frac{1}{\sqrt{x}}\in L^p(0,1)$ I computed $||\frac{1}{\sqrt{x}}||^p_{L^p}=\frac{2}{2-p}<+\infty\ \forall p\in[1,+\infty]\setminus{2}$. So from this it seems that $\frac{1}{\sqrt{x}}\in L^p(0,1)$ for all $p$ except $2$. – sound wave Jan 18 '19 at 13:47
  • 1
  • The issue is that there is no pointwise limit for $\cos (nx)$, the $\frac{1}{\sqrt{x}}$ term is fine. 2. I think your mistake is that you obtain a multiple of $x^{-p/2+1}$ as a primitive and then you plug $0$ into it to obtain $0^{-p/2+1}=0$. Actually, if $p>2$ then $-p/2+1<0$ and so $\lim_{x\to 0^+}x^{-p/2+1}=+\infty$ which shows that the singularity is non-integrable.
  • – Lorenzo Q Jan 18 '19 at 14:06
  • @LorenzoQuarisa Thank you! (1) Ok. (2) Oh yes I made a mistake, $p<2$ is correct. (3) You say that there is no pointwise limit for $u′n$ except for $x=0$, but isn't $0$ a singulariy for $u'_n$? (4) I found this theorem (https://i.imgur.com/WaN7z3u.png) about weak compactness in $W^{1,p}$; it says that for a sequence ${u_n} \in W^{1,p}[a,b]$, if both $u'_n$ and $u_n$ are boundend in the $L^p$-norm by the same positive constant, then $\exists {u{n_k}}$ converging weakly in $W^{1,p}$. For $1<p<2$ the constant exists, so I think is possible to use this result in this case, wht you think? – sound wave Jan 20 '19 at 18:06
  • 1
  • Yes that's a mistake, what is correct is that there is no limit for $\cos(nx)$ except for $x=0$, and so there is no a.e. limit for $u_n'(x)$. 4. This is nothing but the Banach-Alaoglu's theorem since you are deducing weak (sequential) compactness from the boundedness of the sequence $\left{u_n\right}$ in $W^{1,p}$ (the result you linked says more than that though). And, it still says nothing about what the limit is.
  • – Lorenzo Q Jan 20 '19 at 20:10
  • @LorenzoQuarisa (3) Ok. (4) I have to ask to my prof if is important for him to find the limit or it is enough to prove its existence. Thanks for your big help! – sound wave Jan 21 '19 at 08:38
  • @soundwave I think it is required. The weak limit of oscillating sequences like $\cos (nx)$ should be well-known. – Lorenzo Q Jan 22 '19 at 07:56